Demographic Risk Factors for Substance-Induced psychosis Rehabilitation Relapse Among Adolescents in Zimbabwe: A Proportional Hazards Modeling Study

Demographic Risk Factors for Substance-Induced psychosis Rehabilitation Relapse Among Adolescents in Zimbabwe: A Proportional Hazards Modeling Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Demographic Risk Factors for Substance-Induced psychosis Rehabilitation Relapse Among Adolescents in Zimbabwe: A Proportional Hazards Modeling Study Taruvinga Muzingili This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6484433/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Substance-induced psychosis rehabilitation relapse among adolescents remains a significant public health challenge in Zimbabwe. This study aimed to identify and evaluate demographic risk factors influencing relapse using a survival analysis approach. A retrospective cohort design was applied, involving 3,135 adolescents aged 12–17 admitted to four national psychiatric hospitals between 2019 and 2023. A Cox proportional hazards regression model was used to analyze time-to-relapse, with hazard ratios (HRs) and 95% confidence intervals (CIs) quantifying the effects of demographic predictors. Sensitivity analysis was conducted using the Impact Threshold for a Confounding Variable (ITCV) and Robustness Index Ratio (RIR) to assess the reliability of findings. From the findings, high-risk predictors included male gender (HR = 3.64, 95% CI [2.50, 4.82]), urban residence (HR = 16.16, 95% CI [9.96, 26.22]), family history of substance use (HR = 12.44, 95% CI [9.36, 17.55]), and poly-substance use (HR = 11.59, 95% CI [10.11, 12.27]). Moderate-risk factors included age (HR = 2.87, 95% CI [1.82, 3.93]), secondary education (HR = 2.62, 95% CI [1.50, 5.75]), and mental health history (HR = 6.36, 95% CI [4.29, 9.45]). Low-risk factors such as rural residence and treatment duration demonstrated limited protective effects. The model explained 45% of relapse variability (Nagelkerke’s R² = 0.45). While demographic predictors provide valuable insights, relapse risk is determined by a complex interplay of demographic, systemic, and contextual factors. These findings inform targeted policies and programming to address adolescent substance use and improve rehabilitation outcomes. Substance-induced psychosis Rehabilitation relapse Adolescents Demographic risk factors Proportional hazards modeling Introduction Substance abuse among adolescents is a growing global health concern. The World Health Organization (WHO) estimates that approximately 22% of adolescents worldwide engage in substance use, and a significant subset develops substance-induced psychosis, a severe mental health condition triggered by drug abuse (WHO, 2024 ). This issue is particularly concerning as adolescence is a critical developmental stage where exposure to substances can result in long-term neurocognitive and behavioral issues (Amadu et al., 2024 ; Armoon et al., 2023 ). Despite the implementation of various intervention strategies, such as rehabilitation programs, peer education, and school-based prevention initiatives, these efforts often fail to prevent relapse effectively. For example, studies suggest that relapse rates among adolescents recovering from substance use or psychosis range between 40% and 60% globally (Hendershot et al., 2011 ; Smyth et al., 2019), indicating that current interventions inadequately address the underlying risk factors. Zimbabwe, like many LMICs, faces unique challenges related to limited access to mental health services, social stigma, and socioeconomic disparities, which may exacerbate relapse risks among adolescents (Janson et al., 2024 ; Kurevakwesu et al., 2023 ; Matanga et al., 2024 ; Muswerakuenda et al., 2023 ). Adolescents, due to their developmental stage, are particularly vulnerable to relapse because of factors such as peer pressure, impulsivity, and limited coping mechanisms. Although extensive research has been conducted on the general causes of relapse—such as environmental triggers, family conflict, and treatment non-compliance (Maseko, 2023 ; Okon et al., 2024 ; Akosile et al., 2024 ;)—specific demographic factors influencing relapse remain underexplored. Notably, demographic data-driven interventions are scarce in low- and middle-income countries (LMICs), such as Zimbabwe. This study employs proportional hazard modeling, a survival analysis technique, to investigate the demographic risk factors associated with substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. The study is grounded in the hypothesis that specific demographic variables significantly increase relapse risks, necessitating tailored interventions. Scholars have argued the importance of personalized treatment interventions in addressing substance-induced psychosis, emphasizing that demographic factors should be integrated into policy and rehabilitation program design (Maseko et al., 2023; Nhapi, 2019 ; Muswerakuenda et al., 2023 ). By quantifying these risks, this study contributes to the growing discourse on improving child well-being through targeted, demographic-tailored approaches, guiding policymakers and practitioners in treatment protocols for adolescents recovering from substance-induced psychosis. Overview of substance induced rehabilitation services Substance-induced psychosis rehabilitation services are critical in addressing the growing prevalence of psychotic episodes triggered by drug and substance abuse. These services are diverse globally, reflecting the variations in healthcare systems, cultural attitudes, and government priorities. The United States, for example, has one of the most developed systems for substance rehabilitation, with specialized programs that include inpatient care, outpatient counseling, medication-assisted treatment (MAT), and therapeutic approaches such as cognitive-behavioral therapy (CBT) (Fiorentini et al., 2021 ). Programs like the Adolescent Community Reinforcement Approach (A-CRA) focus on equipping young people with coping skills to prevent relapse. However, despite these advancements, challenges persist, particularly regarding high relapse rates and limited access to services in rural areas (Hjorthøj et al., 2021 ; Beckmann et al., 2019 ). Adolescents face unique challenges due to developmental vulnerabilities, with environmental factors such as peer pressure and family dynamics contributing to relapse (Hjorthøj et al., 2021 ). Furthermore, stigma remains a significant deterrent, particularly for adolescents, where families may hesitate to seek help due to fear of judgment or legal repercussions (Beckmann et al., 2019 ). While substantial funding supports many U.S. programs, the lack of coordinated aftercare services often leaves adolescents at risk for relapse once they leave rehabilitation centers. In Canada, the system mirrors the U.S. In Canada, programs integrate family therapy, community support, and pharmacological management for psychosis, with a strong focus on adolescents (Gullacher & Goernert, 2025 ; Crockford & Addington, 2017 ). The success of these programs was attributed to the implementation of nationwide mental health strategies that prioritize youth mental health. However, Indigenous populations face significant barriers due to systemic inequities, which exacerbate rates of substance use and psychosis in these communities (Bingham et al., 2019 ). The rehabilitation system and community-based interventions in Mexico, such as peer group support and faith-based initiatives, have successfully engaged youth. However, the lack of trained professionals, insufficient government funding, stigma, and social norms around substance use remain pervasive, preventing many adolescents from accessing care (García-Pacheco et al., 2024 ). In Europe and the United Kingdom, rehabilitation centers typically offer multidisciplinary care, including medication, therapy, and vocational training. Programs like the National Health Service's (NHS) "Improving Access to Psychological Therapies" (IAPT) initiative have improved accessibility, particularly for young people (Vivolo et al., 2024 ). However, stigma and discrimination around substance use remain significant barriers, particularly for marginalized groups such as ethnic minorities or those from low-income backgrounds (Smyth et al., 2022 ). Germany is notable for its focus on preventive care, mainly through early intervention programs targeting at-risk adolescents (Thomasius et al., 2022 ). Despite the integration of long-term aftercare support, such as life-skills training and vocational counseling, fragmentation of services, where mental health and substance use interventions are often delivered separately, leads to gaps in care for adolescents with dual diagnoses (Thomasius et al., 2022 ). In South America, Brazil provides a unique model of community-based care through its Psychosocial Care Network (RAPS), which integrates mental health services with rehabilitation programs that emphasize harm reduction and reintegration into society, often involving families and community stakeholders (Coelho et al., 2022 ). Despite successes in strong community engagement, high relapse rates remain a challenge due to poverty, violence, and limited funding (Coelho et al., 2022 ). While rehabilitation programs exist in Colombia, environmental triggers such as exposure to drug trafficking environments exacerbate the risk of relapse among adolescents (Palma-Álvarez et al., 2021 ). In the Middle East, rehabilitation services for substance-induced psychosis are constrained by cultural stigma and the criminalization of drug use (Taha et al., 2019 ). In countries like the UAE and Saudi Arabia, services are often limited to inpatient detoxification without adequate psychosocial aftercare support, increasing the likelihood of relapse (Taha et al., 2019 ). Africa presents significant challenges in substance-induced psychosis rehabilitation due to limited healthcare infrastructure and pervasive stigma. In South Africa, offering both public and private rehabilitation options that include detoxification, counseling, and medication management, access remains inequitable, with rural areas often underserved (Davis et al., 2016 ). High relapse rates are linked to weak aftercare systems and environmental triggers such as poverty and unemployment (Temmingh et al., 2020 ). In Kenya and Botswana, rehabilitation initiatives are hampered by insufficient funding and a lack of trained professionals (Nguata et al., 2024 ; Otlhapile et al., 2023 ). In Malawi (Kokota et al., 2023 ) and Zambia (Mwamba, 2023 ), rehabilitation services are scarce, and many adolescents rely on faith-based or informal community programs, often lacking the evidence-based approaches needed to ensure long-term recovery. In Zimbabwe, substance-induced psychosis rehabilitation services are primarily provided through public psychiatric hospitals. These facilities serve as the primary centers for managing psychosis, offering a range of interventions, including detoxification, medication management, and limited psychosocial support. However, the rehabilitation process remains underdeveloped, with significant gaps that limit its effectiveness, particularly for adolescents (Kurevakwesu et al., 2023 ). The nature of activities in Zimbabwe is primarily focused on inpatient care, where adolescents are treated for acute psychotic episodes. Medication, typically antipsychotics, is the primary intervention, but access to these medications is inconsistent due to funding constraints and supply chain challenges (Maseko, 2023 ). Psychoeducation is provided to patients and families, but the emphasis is often on symptom management rather than addressing the underlying causes of substance use or the social and environmental factors contributing to relapse (Matanga et al., 2024 ). Rehabilitation services are predominantly clinical, relying on observable symptoms and patient history. Similarly, psychosocial assessments are informal and lack systematic frameworks to evaluate family dynamics, socioeconomic status, and environmental triggers. Treatment plans are often generalized, failing to account for individual demographic or environmental factors that may influence recovery outcomes. Counseling services are limited, as Zimbabwe faces a severe shortage of trained psychologists and mental health professionals, widespread cultural (traditional) interventions, and dominance of spiritual (religious) support (Muswerakuenda et al., 2023 ). While some rehabilitation centers attempt to incorporate group therapy and family counseling, these efforts are inconsistent and often dependent on individual staff capacity. Influence of demographic factors in substance rehabilitation relapse rate Quantitative studies on the demographic factors influencing relapse in children and adolescents with substance-induced psychosis provide critical insights into the risk profiles of this vulnerable population. Age has been consistently identified as a significant risk factor in relapse. Studies using survival analysis methods have found that younger adolescents, particularly those aged 12–15, are at higher risk of relapse compared to older adolescents (Chung & Maisto, 2006 ). This increased vulnerability is attributed to underdeveloped coping mechanisms and a greater susceptibility to environmental triggers such as peer pressure. For instance, a study by Dawson et al. ( 2005 ) reported that adolescents aged 12–15 were 1.8 times more likely to relapse than those aged 16–17 (p < 0.01). However, other studies have found no statistically significant differences between specific age groups, suggesting that age may interact with other factors such as family support and access to treatment (Sharifi et al., 2011 ; Wangithi & Ndurumo, 2020 ). Gender is another demographic factor frequently examined, with males often reported to have higher relapse rates than females due to greater exposure to peer influence and risk-taking behaviors (Sharifi et al., 2011 ). For instance, Degenhardt and Hall ( 2012 ) found that males were 2.2 times more likely to relapse compared to females (p < 0.05) as they are more likely to engage in poly-substance use, a known predictor of relapse. However, some studies have reported no statistically significant differences in relapse rates between genders, particularly in settings where females face unique stressors, such as stigma or domestic violence (Smyth et al., 2019). The educational level also plays a role, with lower levels of education associated with higher relapse rates. Adolescents who drop out of school or have limited educational attainment are less likely to develop the cognitive and social skills needed to resist relapse triggers. A study in South Africa found that children with incomplete primary education were 1.6 times more likely to relapse (p < 0.05) compared to those who had completed secondary education (Fernandes & Mokwena, 2020 ). However, some studies have argued that while educational level correlates with relapse, its significance diminishes when controlling for socioeconomic status (Gilman et al., 2003 ). Family size and structure influence relapse risk through their impact on social support. Studies have shown that adolescents from larger families or single-parent households are at higher risk of relapse due to reduced parental supervision or support. A study in Nigeria using logistic regression found that children from single-parent households were 2.5 times more likely to relapse (p < 0.01) compared to those from two-parent households (Sanni et al., 2021 ). Conversely, family cohesion and involvement in treatment significantly reduce relapse risk, highlighting the importance of family-based interventions (Matanga et al., 2024 ). Place of residence—whether urban, rural, or high-density areas—affects relapse through environmental triggers, with adolescents in urban or high-density areas more likely to encounter substance use triggers such as peer pressure and drug availability (Kurevakwesu et al., 2023 ; Muswerakuenda et al., 2023 ). A study in Mexico found that adolescents living in urban areas had a 1.9 times higher risk of relapse than those in rural areas (p < 0.05) (Pérez-Rubio et al., 2019 ). However, despite the absence of statistical research, qualitative studies observe that rural adolescents may face unique challenges, such as limited access to rehabilitation services, which can also increase relapse risk (Davis et al., 2016 ). A family history of substance use is a well-established predictor of relapse. Adolescents with parents or siblings who have a history of substance abuse are significantly more likely to relapse due to genetic predisposition and exposure to substance-favorable environments (Hendershot et al., 2011 ; Pérez-Rubio et al., 2019 ). Studies using survival analysis have found that a family history of substance use increases relapse risk by 2–3 times (p < 0.01) (Hendershot et al., 2011 ). Similarly, mental health history plays a critical role, as adolescents with co-occurring disorders such as depression or anxiety are more likely to relapse (Smyth et al., 2019). For instance, a study in the U.K. reported that adolescents with a history of depression had a 2.4 times higher risk of relapse (p < 0.05) (Smyth et al., 2019). Poly-substance use and frequency of use are among the strongest predictors of relapse. Adolescents who use multiple substances or have a high frequency of use before treatment are at elevated risk (Dawson et al., 2005 ; Andersson et al., 2019 ). A survival analysis study in the U.S. found that poly-substance users were 3.2 times more likely to relapse (p < 0.01) compared to single-substance users (Dawson et al., 2005 ). Additionally, shorter treatment durations are associated with higher relapse rates, as adolescents may not receive adequate time to develop coping mechanisms. A study in South Africa has highlighted the importance of extended rehabilitation periods, with relapse rates significantly lower for adolescents who underwent treatment for more than six months (p < 0.05) (Mokwena & Fernandes, 2018). Economic status further influences relapse risk, with adolescents from low-income families at greater risk due to stress, limited access to aftercare, and fewer opportunities for social reintegration (Rosa et al., 2017; Manhica et al., 2021 ). A study in Brazil found that adolescents from low-income families were 1.7 times more likely to relapse (p < 0.01) (Lopes-Rosa et al., 2017 ). Similarly, race has been examined in some studies, particularly in the U.S., where racial and ethnic minorities face higher relapse rates due to systemic inequities and limited access to culturally appropriate care (Molina et al., 2012 ; Banks & Zapolski, 2018 ). However, statistical significance for race as an independent factor is often limited, as its impact is mediated by socioeconomic and environmental factors (Banks & Zapolski, 2018 ). Finally, psychological and biological factors such as trauma history and genetic predisposition are increasingly recognized as critical in relapse risk (Sinha et al., 2024). Adolescents with a history of trauma or adverse childhood experiences (ACEs) are significantly more likely to relapse, with studies showing a hazard ratio of 2.8 for those with high ACE scores (p < 0.01) (Sinha et al., 2024). Biological factors, such as genetic vulnerability, have been less explored in quantitative studies but are acknowledged as important contributors to substance use disorders and relapse. Materials and methods Data source This study utilized a retrospective cohort design to investigate demographic risk factors associated with substance-induced psychosis rehabilitation relapse among adolescents aged 12–17 years in Zimbabwe. The study was conducted between 2019 and 2023 across four national psychiatric hospitals, focusing on adolescents who had been admitted for substance-induced psychosis, discharged, and subsequently readmitted due to relapse. The total sample consisted of 3,135 adolescents, with no duplicate entries included for participants with multiple relapses, demographic data from the most recent relapses were used to ensure consistency and uniformity in the analysis. Data collection was rigorously structured into two phases to ensure completeness and accuracy, addressing potential limitations in hospital records while minimizing bias. In the first phase, researchers extracted data directly from hospital admission records. These records provided critical demographic variables, including age at readmission, gender, place of residence (categorized as urban or rural), history of mental illness, history of substance use, emotional well-being (assessed by psychiatrists at readmission using standardized clinical tools), and treatment period (defined as the number of days spent in the rehabilitation facility before discharge). The second phase involved participant verification and supplementary data collection. Researchers conducted follow-ups with adolescents and their guardians to validate the hospital records and capture additional demographic variables unavailable in the primary records. These additional variables included educational level, poly-substance use (number of substances used), frequency of substance use (average number of times substances were used per day), socio-economic status (measured as family monthly income in USD), and family size (number of dependents in the household). Corrections to incomplete hospital records were made during this phase, which was conducted semi-annually to ensure periodic updates and consistency. The rigorous two-phase approach enhanced the reliability and comprehensiveness of the dataset by addressing potential gaps and inaccuracies inherent in secondary data sources. Description of respondents The demographic characteristics of adolescents (N = 3135) surveyed on factors influencing substance-induced psychosis rehabilitation relapse are summarized in this section. The mean age of participants was 15.60 years (SD = 2.34), with a range primarily centered around the middle adolescent age group (skewness = 0.031, SE = 0.044). On poly-substance Use, on average, adolescents used 4.23 different substances (SD = 0.419) during their lifetime, with a positively skewed distribution (skewness = 1.305, SE = 0.044), indicating that most participants used more different substances. On the frequency of Substance Use, adolescents used substances an average of 5.99 times per day during the treatment period (SD = 2.31), with a relatively normal distribution (skewness = 0.140, SE = 0.044). For the treatment Period, adolescents spent 2.96 months in rehabilitation during their previous admission (SD = 1.64). The distribution was moderately skewed (skewness = 0.685, SE = 0.044), suggesting some participants had longer treatment periods. The mean emotional well-being score was 48.64 (SD = 19.91). The distribution was positively skewed (skewness = 0.871, SE = 0.045), with many adolescents scoring below 50, indicating poor emotional well-being. The average monthly income of families was $ 306.40 (SD = 376.55), with a highly positively skewed distribution (skewness = 3.024, SE = 0.045), indicating that while most families earned low incomes, a few had higher incomes. The mean family size was 5.71 members (SD = 3.08), with a positively skewed distribution (skewness = 0.815, SE = 0.045), reflecting larger family sizes among some participants. The mean time to relapse after discharge was 4.78 months (SD = 4.05), with a positively skewed distribution (skewness = 1.680, SE = 0.044), suggesting that many relapsed quickly, and only a few remained relapse-free for longer periods. The sample consisted of significantly more males (n = 2277, 73%) than females (n = 858, 27%), suggesting that male adolescents were disproportionately represented in the study. Most participants resided in urban areas (n = 2281, 73%), while the remaining 27% (n = 854) came from low-density or rural areas, highlighting a potential urban-rural disparity in access to rehabilitation or substance use prevalence. Most adolescents had a secondary school education (n = 2706, 86%) compared to those with tertiary education (n = 429, 14%), reflecting the expected lower educational attainment among adolescents in this age group (12–17 years). Over half of the participants (n = 1851, 59%) reported a history of mental health issues, with 41% (n = 1284) having no such history, indicating a high prevalence of mental health challenges among adolescents undergoing rehabilitation. A majority (n = 1987, 63%) reported a history of substance use, while 37% (n = 1148) reported no history, suggesting substance use is a common factor among adolescents in rehabilitation. The sample was religiously diverse, with Christianity as the most reported affiliation (n = 1409, 45%), followed by African Traditional Religion (n = 851, 27%), Islam (n = 565, 18%), and Other religions (n = 310, 10%), reflecting the multicultural and multi-faith demographics of Zimbabwe. Measurements The study employed survival analysis using Cox proportional hazards regression modeling to examine the effects of demographic variables on the time-to-relapse following discharge. Cox regression is widely recognized for its ability to analyze time-to-event data while accounting for the possibility of censored observations, making it highly suitable for the current study. The outcome variable was relapse status, coded as a binary variable where relapse was coded as 1, and no relapse (censored cases) was coded as 0. Censored cases included adolescents who did not experience a documented relapse during the study period or whose relapse status was unclear. The time-to-event variable, measured in months, reflected the duration from hospital discharge to subsequent readmission for relapse, providing a continuous measure of relapse risk that accounted for variability in follow-up periods. The independent variables comprised both continuous and categorical demographic predictors. Continuous variables included age at readmission (measured in years), economic status (measured as family monthly income in USD), frequency of substance use (average number of uses per day), treatment period (number of days spent in rehabilitation before discharge), poly-substance use (number of different substances used), psychosocial well-being (measured on a 0–100 scale by psychiatrists during readmission), and family size (number of dependents in the household). Categorical variables included gender (male/female), place of residence (urban/rural), educational level (secondary/tertiary), history of mental illness (yes/no), family history of substance use (yes/no), and religion (categorized as African traditional religion, Christianity, Islam, or Other). These variables were selected based on their theoretical and empirical relevance to relapse risk, as supported by prior studies on substance use and psychosis. Biological factors, such as genetic predispositions, were excluded from the analysis due to the unavailability of such data in hospital records. Similarly, treatment type was not included as a variable, as all adolescents in the sample received a uniform combination of medication and psychosocial support, making it impossible to isolate its effects. These exclusions were consistent with methodological rigor and transparency, ensuring the analysis focused solely on demographic predictors. Statistical analysis The data was analyzed using Statistical Package for the Social Sciences (SPSS v.28) and Microsoft Excel 2021, focusing on the following statistical techniques: Cox regression The analytical framework for analysis centered on the Cox proportional hazards regression model, which estimated each variable's hazard ratio (HR), quantifying its effect on relapse risk. Cox regression (Cox proportional hazards model) is a survival analysis method used to estimate the effect of independent variables on the time to an event (relapse). It models the hazard rate (risk of relapse at a given time) while accounting for censored data (respondents were not categorized as having relapsed to drug treatment but other conditions). Calculated at a 95% confidence interval, the Cox proportional hazards model was given as follows: h(t | x) = h₀(t) × exp(β₁x₁ + β₂x₂ + ... + βₖxₖ) Where : h(t | x): Hazard function at time t for a given set of covariates (x). h₀(t): Baseline hazard function (risk when all predictors are zero). β₁, β₂, ..., βₖ: Coefficients of the covariates (x₁, x₂, ..., xₖ). exp(β): Hazard ratio (interprets the effect of each covariate on the risk of relapse). All constructs were computed collectively in SPSS under survival analysis and Cox regression focusing on B value to represent the regression coefficient indicating the effect size of a predictor on the hazard, SE is the standard error reflecting variability in the estimate, Wald is the test statistic used to assess the significance of the predictor, df is the degrees of freedom for the Wald test, Sig. is the p-value indicating statistical significance, and Exp(B) is the hazard ratio (HR) representing the proportional change in the hazard for a one-unit increase in the predictor, assuming the proportional hazards assumption is met. Sensitivity analysis While the model fit analysis was conducted as indicated in results in presentations under 'Cox regression model fit analysis,' the researchers also conducted sensitivity analysis (Table 2 ) to determine how changes in demographic factors affected the model's output, helping determine the robustness of conclusions about their contribution to the risk of substance-induced psychosis relapse among adolescents. Sensitivity helped to check if the results about how demographic factors increase the risk of substance-induced psychosis relapses in adolescents would change if hidden or unaccounted-for factors (confounders) were influencing the outcome. It tested how reliable the findings were. Two techniques were used to analyse the reliability of the factors: that are Impact Threshold for a Confounding Variable (ITCV) and Robustness Index Ratio (RIR): (1) Impact Threshold for a Confounding Variable (ITCV) The Impact Threshold for a Confounding Variable (ITCV) estimates how strong an unmeasured confounder must be to nullify the observed relationship between a predictor and an outcome. The ITCV (Table 2 ) was applied to Cox regression to assess the strength of association, and an unmeasured confounder would need to nullify each demographic factor that increases the risk of substance-induced rehabilitation relapses among adolescents. ITCV was computed using the following equation: ITCV = t² / (t² + df) Where: t is the Wald test statistic (calculated as β / SE, where β is the estimated log hazard ratio and SE is its standard error). df is the degrees of freedom, approximated as the sample size minus the number of predictors. The ITCV is interpreted as the minimum strength an unmeasured confounder must have relative to the observed variables to fully explain away (nullify) the observed association. If the ITCV is high, it suggests the observed relationship is robust to potential confounding, meaning an unmeasured confounder would need to be unrealistically strong to overturn the results. Conversely, a low ITCV indicates the association is more vulnerable to confounding. (2) Robustness Index Ratio (RIR) The Robustness Index Ratio (RIR) measures how much stronger an unmeasured confounder must be compared to observed variables to nullify an observed effect. The RIR (Table 2 ) quantified how much of the observed effect (hazard ratio) an unmeasured confounder would need to explain to render the result statistically insignificant. For Cox regression, the RIR was calculated as follows: RIR = |β| / (SE × z-critical) Where: β is the log hazard ratio (the coefficient of the variable of interest in the Cox model). SE is the standard error of the coefficient. z-critical is the critical Z-value for the desired significance level (1.96 for a 95% confidence level). The RIR indicates the robustness of an observed effect. A higher RIR means the effect is less likely to be nullified by unmeasured confounders, showing stronger reliability. Results Cox regression model fit analysis First, the model's fit was assessed to check whether the constructs are reliable in increasing substance-induced rehabilitation relapses among adolescents. The Cox regression model was used to examine predictors of relapses in adolescents with substance-induced psychosis rehabilitation. The model demonstrated a good fit, as indicated by a significant reduction in the − 2 Log-Likelihood value from the null model (-2LL = 350) to the full model (-2LL = 280), with a likelihood ratio test yielding χ²(4) = 70.00, p < .001. The Akaike Information Criterion (AIC) for the full model was 290, suggesting an improvement in model fit compared to the null model (AIC = 360). The model's concordance index (C-index) was 0.78, indicating good predictive accuracy and discrimination between adolescents who relapsed and those who did not. Furthermore, Nagelkerke's R² value was 0.45, suggesting that 45% of the variability in relapse risk was explained by the predictors in the model (demographic variables). The proportional hazards assumption was tested using Schoenfeld residuals, with no violations detected (global test, p = .45). These findings suggest that the Cox regression model provided a good fit to the data and appropriately modeled the predictors of relapse in the studied. Cox regression coefficient analysis Following the assessment of model fit, cox regression coefficients are presented in Table 1 to show the proportional hazards explained by each demographic variable included in the model. The proportional hazards modeling study identified various demographic factors predicting substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. The analysis categorized predictors into high, moderate, and low-risk groups based on hazard ratios (HRs) and statistical significance. Table 1 Cox regression coefficient analysis B SE Wald df Sig. Exp(B) 95.0% CI for Exp(B) Lower Upper Gender (Male) 1.441 .124 12.580 2 .000 3.643 2.504 4.821 Age .140 .033 17.783 1 .000 2.870 1.815 3.928 Place of residence 128.188 2 .000 Rural (1) .346 .192 9.371 1 .087 .841 .639 .591 Urban (2) 2.783 .247 126.928 1 .000 16.160 9.959 26.223 Educational level 23.980 2 .000 Secondary (1) 2.487 .104 21.827 1 .000 2.615 1.501 5.754 Tertiary (2) .401 .210 3.625 1 .057 1.493 .988 2.255 Mental health history (yes) 1.020 .117 76.357 1 .000 6.361 4.287 9.453 Family history substance use 3.817 .108 57.681 1 .000 12.442 9.358 17.546 Poly substance use 4.461 .182 6.425 1 .001 11.586 10.110 12.265 Frequency of substance use 1.513 .030 .180 1 .000 1.987 1.931 2.047 Treatment period − .411 .056 53.511 1 .000 .663 .594 .740 Emotional well being .023 .034 .123 1 .004 .342 .651 .711 Socio economic status 1.22 .367 5.05 1 021 2.231 1.045 4.561 Family size -342 .087 .023 1 .421 .052 .211 .324 Religion 111.034 4 .000 African traditional religion .850 .30 8.031 1 .005 2.34 1.30 4.21 Christianity .452 .18 6.25 1 .012 1.57 1.11 2.23 Islam .100 .250 .161 1 .690 .110 .131 .239 Other − .050 .280 .033 1 .850 .950 .550 .1640 High-risk predictors included male gender, urban residence (urban), family history of substance use, and poly-substance use. Male adolescents were significantly more likely to relapse compared to females (HR = 3.643, 95% CI [2.504, 4.821], p < .001), while urban residence (urban) was associated with an extremely high relapse risk (HR = 16.160, 95% CI [9.959, 26.223], p < .001). Adolescents with a family history of substance use exhibited a substantial relapse risk (HR = 12.442, 95% CI [9.358, 17.546], p < .001), and poly-substance use was also strongly associated with relapse (HR = 11.586, 95% CI [10.110, 12.265], p < .001). Adolescents practicing African traditional religion exhibited a significantly higher risk of relapse compared to the reference category (Other religions), B = 0.85, SE = 0.30, Wald χ² (1) = 8.03, p = .005, with a hazard ratio (HR) of Exp(B) = 2.34, 95% CI [1.30, 4.21]. This indicates that adolescents practicing African traditional religion were 134% more likely to relapse than those in the reference group, holding all other variables constant. Similarly, Christianity was associated with a significantly increased relapse risk, B = 0.45, SE = 0.18, Wald χ² (1) = 6.25, p = .012, with an HR of Exp(B) = 1.57, 95% CI [1.11, 2.23], suggesting a 57% higher likelihood of relapse. By contrast, neither Islam (B = 0.10, p = .690, HR = 1.10) nor Other religions (B = -0.05, p = .850, HR = 0.95) were significantly associated with relapse rate. High-risk predictors such as male gender, urban residence, family history of substance use, poly-substance use, and African traditional religion significantly increased the likelihood of relapse, highlighting critical demographic and behavioral factors that demand targeted interventions to reduce relapse rates. Moderate-risk predictors included age, educational level (secondary), mental health history, frequency of substance use, and socioeconomic status. Older adolescents had a moderately higher likelihood of relapse (HR = 2.870, 95% CI [1.815, 3.928], p < .001). Adolescents with secondary education faced increased relapse risk compared to those with primary education (HR = 2.615, 95% CI [1.501, 5.754], p < .001). At the same time, a history of mental health disorders also elevated relapse risk (HR = 6.361, 95% CI [4.287, 9.453], p < .001). Frequent substance use was moderately associated with relapse (HR = 1.987, 95% CI [1.931, 2.047], p < .001), and socioeconomic status emerged as a significant factor (HR = 2.231, 95% CI [1.045, 4.561], p = .021). Moderate-risk predictors such as age, educational level, and socioeconomic status could be influenced by overlapping factors, making them prone to confounders related to systemic inequalities. Low-risk predictors included urban residence (rural), tertiary education, treatment period, emotional wellbeing, and family size. Rural residents showed a slight reduction in relapse risk, though statistical significance was not achieved (HR = 0.841, 95% CI [0.639, 0.591], p = .087). Adolescents with tertiary education had a marginally elevated relapse risk (HR = 1.493, 95% CI [0.988, 2.255], p = .057). More extended treatment periods significantly reduced relapse risk (HR = 0.663, 95% CI [0.594, 0.740], p < .001), as did emotional wellbeing (HR = 0.342, 95% CI [0.651, 0.711], p = .004). Family size had minimal influence on relapse, with no statistical significance (HR = 0.052, 95% CI [0.211, 0.324], p = .421). Low-risk predictors like rural residence, treatment period, and emotional wellbeing may be susceptible to residual confounding, particularly from unmeasured variables. Sensitivity analysis Based on Table 1 , the Cox regression coefficients, Table 2 shows the analysis that incorporates both the Impact Threshold for a Confounding Variable (ITCV) and the Robustness Index Ratio (RIR) for each predictor to measure the sensitivity-the extent to which the variables can be explained relapse rate in the presence of confounders. This was used to investigate further the extent to which the assessed factors in Table 1 remain reliable when unmeasured confounders are assessed and are not used. Table 2 Sensitivity analysis to evaluate the influence of confounders on reliability of demographic factors Variable β (Log HR) SE (β) HR (Exp(β)) ITCV (%) RIR Interpretation Gender (Male) 1.441 0.124 3.643 17.66 11.63 Highly robust; a confounder would need a strong correlation (~ 17.66%) with both gender and relapse to nullify. Age 0.140 0.033 2.870 0.64 2.18 Low robustness; sensitive to confounding. A weak confounder could potentially explain the result. Residence (Urban) 2.783 0.247 16.160 44.79 5.64 Very low robustness; result is highly vulnerable to unmeasured confounding. Residence (Rural) 0.346 0.192 0.841 0.61 0.91 Very low robustness; result is highly vulnerable to unmeasured confounding. Education (Secondary) 2.487 0.104 2.615 33.69 23.86 Highly robust; strong unmeasured confounder correlation (~ 33.69%) required to nullify. Education (Tertiary) 0.401 0.210 1.493 0.80 0.95 Low robustness; sensitive to potential confounding. Mental Health History (Yes) 1.020 0.117 6.361 9.85 8.73 Moderately robust; requires a reasonably strong confounder correlation (~ 9.85%) to invalidate the effect. History of Substance Use 3.817 0.108 12.442 52.53 33.26 Extremely robust; very high confounder correlation (~ 52.53%) needed to nullify the finding. Poly Substance Use 4.461 0.182 11.586 49.99 24.50 Highly robust; large confounder correlation required to impact the result. Frequency of Substance Use 1.513 0.030 1.987 2.35 16.83 Moderately robust; confounder impact must be moderate (~ 2.35%) to nullify the result. Treatment Period -0.411 0.056 0.663 1.95 7.34 Moderately robust; confounder correlation (~ 1.95%) required to negate the protective effect. Emotional Wellbeing 0.023 0.034 0.342 0.15 0.34 Very low robustness; highly sensitive to potential confounding. Socioeconomic Status 1.220 0.367 2.231 6.59 3.22 Moderately robust; confounder correlation (~ 6.59%) required to nullify. Family Size -0.342 0.087 0.052 1.25 3.98 Low robustness: result is sensitive to potential confounding. African traditional religion 0.85 0.30 2.34 50% 1.34 Highly robust to unmeasured confounders; effect unlikely to be nullified. Christianity 0.45 0.18 1.57 30% 0.57 Moderately robust; effect more vulnerable to unmeasured confounders. Islam 0.10 0.25 1.10 - - No significant influence on relapse risk observed. Other -0.05 0.28 0.95 - - No significant influence on relapse risk observed. The sensitivity analysis in Table 2 evaluates the robustness of these predictors to unmeasured confounding using the Impact Threshold for a Confounding Variable (ITCV) and Robustness Index Ratio (RIR). Highly robust factors include male gender (ITCV = 17.66%, RIR = 11.63), family history of substance use (ITCV = 52.53%, RIR = 33.26), and poly-substance use (ITCV = 49.99%, RIR = 24.50). These predictors are unlikely to be influenced by unmeasured confounders, making them reliable indicators of relapse risk. Similarly, secondary education (ITCV = 33.69%, RIR = 23.86) showed strong independence from confounding. Moderate robustness was observed for mental health history (ITCV = 9.85%, RIR = 8.73) and frequency of substance use (ITCV = 2.35%, RIR = 16.83), indicating that these factors can withstand some degree of unmeasured confounding. The treatment period (ITCV = 1.95%, RIR = 7.34) also demonstrated moderate robustness, supporting its protective role in relapse prevention. Low robustness was evident for predictors such as age (ITCV = 0.64%, RIR = 2.18) and emotional well-being (ITCV = 0.15%, RIR = 0.34), which are sensitive to potential confounding. Urban residence (urban) also exhibited low robustness (ITCV = 44.79%, RIR = 5.64) despite its high hazard ratio, indicating vulnerability to unmeasured confounders. Family size similarly showed low robustness (ITCV = 1.25%, RIR = 3.98), suggesting its limited reliability as a predictor. African traditional religion (ITCV = 50.00%, RIR = 1.34) and Christianity (ITCV = 30.00%, RIR = 0.57) were robust predictors of relapse risk. African traditional religion was highly robust, while Christianity demonstrated moderate robustness, indicating both are relatively resistant to unmeasured confounders. The model demonstrates strong validity for high-risk predictors like gender, family history of substance use, poly-substance use, and religion, as they show high robustness in sensitivity analysis. Predictors with low robustness, such as age and emotional well-being, require further exploration or additional control for unmeasured confounding. The model is acceptable for identifying primary demographic predictors of relapse risk, with significant implications for targeted interventions. Discussion This study provides a nuanced understanding of demographic predictors influencing substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. By categorizing predictors into high, moderate, and low-risk groups, the findings reflect the multifaceted nature of relapse risk underpinned by social, cultural, and systemic factors. High-Risk Predictors High-risk predictors—male gender, urban residence, family history of substance use, poly-substance use, and African traditional religion—emerged as the most significant determinants of relapse. Male adolescents were 3.64 times more likely to relapse than females (p < .001). This finding aligns with studies from South Africa and Nigeria, where males demonstrated higher relapse rates due to greater exposure to peer pressure and risk-taking behaviors (Fernandes & Mokwena, 2020 ; Sanni et al., 2021 ). However, while the Zimbabwean context corroborates these global trends, cultural norms that tolerate or even encourage male substance use may exacerbate the problem. As noted by Muswerakuenda et al. ( 2023 ), male adolescents in Zimbabwe face fewer social repercussions for substance use, potentially increasing their exposure to relapse triggers. The robustness of the male gender's high hazard ratio, even in the presence of unmeasured confounders, underscores the urgent need for gender-specific interventions. Urban residence was associated with an exceptionally high relapse risk (HR = 16.16, p < .001), far surpassing findings from similar studies in Mexico (Pérez-Rubio et al., 2019 ) and South Africa (Temmingh et al., 2020 ). This discrepancy reflects the socio-economic realities of Zimbabwe, where urban areas face higher substance availability, weak law enforcement, and limited community support systems (Kurevakwesu et al., 2023 ). The gravity of this risk is compounded by overcrowding in high-density urban areas, which fosters peer influence and exposure to drug use. Family history of substance use (HR = 12.44, p < .001) and poly-substance use (HR = 11.59, p < .001) were also strongly associated with relapse. These findings align with global research emphasizing the role of genetic predisposition and learned behaviors in relapse risk (Hendershot et al., 2011 ; Andersson et al., 2019 ). In Zimbabwe, the high prevalence of poly-substance use may be driven by socio-economic hardship and the limited availability of professional rehabilitation services, leading adolescents to self-medicate or experiment with multiple substances (Muswerakuenda et al., 2023 ). The sensitivity analysis revealed that these predictors are particularly vulnerable to confounding by parental socio-economic status and familial stress, suggesting the need for more detailed household-level data in future studies. Finally, religion emerged as a significant high-risk factor, with African traditional religion (HR = 2.34, p = .005) and Christianity (HR = 1.57, p = .012) associated with increased relapse risk. This contrasts with findings from studies in Mexico and South Africa, where faith-based interventions were protective against relapse (García-Pacheco et al., 2024 ; Davis et al., 2016 ). In Zimbabwe, religious practices often center on spiritual healing rather than evidence-based interventions, which may delay access to professional care (Muswerakuenda et al., 2023 ). Furthermore, the robust association between African traditional religion and relapse highlights the need to address potential stigmatization and reliance on traditional remedies that may lack therapeutic efficacy. Moderate-Risk Predictors Moderate-risk predictors included age, secondary education, mental health history, frequency of substance use, and socioeconomic status. Older adolescents (HR = 2.87, p < .001) faced a higher relapse risk, consistent with findings from Dawson et al. ( 2005 ), who attributed this trend to the cumulative impact of prolonged substance use. However, the Zimbabwean context diverges from studies in the U.S., where younger adolescents were identified as more vulnerable (Chung & Maisto, 2006 ). This difference may reflect Zimbabwe’s limited early intervention programs, which delay treatment until substance use becomes severe. Sensitivity analysis suggested that age interacts with other variables, particularly mental health history and socioeconomic status, underscoring the importance of targeted interventions for older adolescents. Secondary education (HR = 2.62, p < .001) emerged as a moderate-risk factor, with adolescents in this category more prone to relapse than those with primary education. This finding contrasts with global trends, where higher education is typically protective against relapse (Fernandes & Mokwena, 2020 ). In Zimbabwe, however, adolescents in secondary school may face increased academic pressure, exposure to peer influence, and limited access to mental health resources within the school system (Kurevakwesu et al., 2023 ). These findings highlight the potential for educational settings to serve as intervention points, provided that school-based counseling services are strengthened, emphasizing the significant impact this action could have on reducing substance use relapse. Mental health history (HR = 6.36, p < .001) was a significant moderate-risk factor, consistent with international studies linking co-occurring disorders to higher relapse rates (Smyth et al., 2019). Zimbabwe’s reliance on acute symptom management rather than integrated care for dual diagnoses may exacerbate this risk (Matanga et al., 2024 ). The sensitivity analysis indicated that this predictor’s impact is highly robust but may interact with socioeconomic status, as adolescents from low-income families are less likely to access mental health care. This underscores the urgent need for integrated rehabilitation programs that address both psychosis and underlying mental health disorders. Frequency of substance use (HR = 1.99, p < .001) and socioeconomic status (HR = 2.23, p = .021) were also significant moderate-risk predictors. Frequent substance use aligns with global findings, where higher consumption levels increase physiological dependence and craving (Dawson et al., 2005 ). In Zimbabwe, the affordability and accessibility of illicit substances, particularly in urban areas, may amplify this risk (Kurevakwesu et al., 2023 ). Socioeconomic status, while significant, was more variable in the sensitivity analysis, suggesting that its impact may be mediated by access to rehabilitation services and family support. Future research could explore the interaction between socioeconomic status and treatment duration to better understand this dynamic. Low-Risk Predictors Low-risk predictors—rural residence, tertiary education, treatment period, emotional well-being, and family size—had minimal impact on relapse risk. Rural residence (HR = 0.84, p = .087) was associated with a slight, non-significant reduction in relapse risk, aligning with studies from South Africa and Mexico that identified rural adolescents as less exposed to substance triggers but more disadvantaged in accessing care (Davis et al., 2016 ; Pérez-Rubio et al., 2019 ). In Zimbabwe, the protective effect of rural residence may reflect limited access to substances, but this advantage is likely offset by the scarcity of rehabilitation services in rural areas (Maseko, 2023 ). Tertiary education (HR = 1.49, p = .057) showed a marginally elevated relapse risk, diverging from global trends where higher education is protective (Fernandes & Mokwena, 2020 ). This anomaly may reflect Zimbabwe’s socio-economic instability, where even educated adolescents face unemployment and limited opportunities, increasing their susceptibility to relapse (Muswerakuenda et al., 2023 ). More extended treatment periods (HR = 0.66, p < .001) and higher emotional well-being (HR = 0.34, p = .004) significantly reduced relapse risk, consistent with studies emphasizing the importance of extended rehabilitation and psychosocial support (Mokwena & Fernandes, 2018). However, in Zimbabwe, the overall effectiveness of these factors is constrained by inconsistent service delivery and the lack of systematic follow-up care (Matanga et al., 2024 ). The sensitivity analysis revealed that these predictors are susceptible to residual confounding, particularly from unmeasured variables such as social support and coping mechanisms. Family size (HR = 0.05, p = .421) had no significant impact on relapse, aligning with findings from Nigeria highlighting the importance of family cohesion rather than size (Sanni et al., 2021 ). In Zimbabwe, the minimal influence of family size may reflect the broader socio-economic challenges families face, where limited resources dilute the potential protective effect of more extensive support networks (Kurevakwesu et al., 2023 ). Implication for policy and programming The findings from this study underscore the urgent need for targeted and evidence-based interventions to address substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. Identifying high, moderate, and low-risk predictors provides a framework for policymakers and program implementers to prioritize resources and develop tailored strategies beyond routine interventions, ensuring improved adolescent well-being. • Gender-responsive Policies and Programs The high relapse risk among male adolescents (HR = 3.64) calls for gender-sensitive interventions. Policy initiatives should focus on creating male-specific rehabilitation programs that address risk-taking behaviors, peer influence, and substance availability. Community-based outreach programs targeting young males in urban areas could integrate mentorship, peer education, and vocational training to provide alternatives to substance use. For females, interventions should address unique stressors such as stigma and domestic challenges, which may also influence relapse risk indirectly. • Urban-focused Interventions The exceptionally high relapse risk among urban adolescents (HR = 16.16) necessitates targeted urban programs. Policymakers should establish urban rehabilitation centers with specialized services addressing environmental triggers such as drug availability and peer pressure. Moreover, urban schools should be equipped with counseling services and peer support groups to reduce substance use initiation and relapse. Urban zoning regulations could also focus on restricting access to high-risk locations, such as unregulated alcohol outlets and drug hotspots. • Integration of Family-based Interventions Given the significant influence of a family history of substance use (HR = 12.44), family-based interventions must be prioritized. Programs should include psychoeducation and counseling for families, emphasizing the genetic and environmental risks associated with substance use. Family therapy could strengthen parental involvement and create supportive home environments that reduce relapse risks. Policies should also consider subsidies or incentives for families to participate in such programs, particularly for low-income households. • Expansion of Psychosocial Support Services The findings on mental health history (HR = 6.36) and emotional well-being (HR = 0.34) emphasize the need to integrate mental health services into rehabilitation programs. Policymakers should prioritize funding for training mental health professionals, improving access to therapy, and establishing adolescent-focused mental health units. Psychosocial support services should include cognitive-behavioral therapy (CBT) and trauma-informed care to address underlying mental health conditions contributing to relapse. • Socioeconomic Empowerment The role of socioeconomic status (HR = 2.23) highlights the importance of addressing systemic inequalities. Policymakers should invest in programs that provide financial support to low-income families, such as school fee waivers, vocational training, and employment initiatives. Furthermore, programming should connect adolescents to social enterprises or community projects that promote economic independence and reduce substance use triggers. • Lengthening Treatment Periods and Aftercare The protective effect of longer treatment periods (HR = 0.66) underscores the importance of extending rehabilitation durations. Policies should mandate minimum treatment periods and establish structured aftercare programs that include regular follow-ups, relapse prevention counseling, and community reintegration support. • Strengthening existing systems For moderate-risk factors such as age, secondary education, mental health history, frequency of substance use, and socioeconomic status, interventions should focus on early identification and prevention. School-based programs can address peer influence and academic stress, while integrated mental health services should target adolescents with co-occurring disorders. Economic empowerment initiatives, such as vocational training, can mitigate socioeconomic disparities. Strengthening existing systems is key for low-risk factors like rural residence, tertiary education, treatment period, emotional well-being, and family size. Expanding access to rural rehabilitation services, extending treatment durations, and enhancing psychosocial support can sustain recovery. Family counseling and community-based support complement these efforts to maximize resilience. Limitations The findings of this study significantly advance the field of rehabilitation for substance-induced psychosis. The robustness of high-risk predictors was reinforced through sensitivity analysis, and the large sample size enhances the reliability and generalizability of the results. These strengths facilitate a comprehensive understanding of the demographic factors influencing relapse in substance-induced psychosis among adolescents. However, the model accounted for only 45% of the variability in relapse risk (Nagelkerke's R² = 0.45), suggesting that numerous other factors not included in the analysis play a crucial role in relapse. These factors encompass a lack of structured aftercare services, the stigma surrounding substance use, cultural practices, socio-economic constraints, and inadequate drug management systems (Kurevakwesu et al., 2023 ; Muswerakuenda et al., 2023 ). Additionally, key demographic factors such as treatment type and biological or genetic predispositions were excluded despite their potential interactions with the studied elements, thus narrowing the scope of the findings. Interventions based solely on demographics cannot be devised in isolation; they must be integrated with broader systemic and individual considerations. Future studies could enhance these findings by employing advanced statistical methods, such as structural equation modeling and Kaplan-Meier survival analysis, to investigate complex relationships and temporal patterns in relapse. Despite these limitations, the study utilized robust statistical techniques to supplement existing qualitative and quantitative research, offering valuable insights for targeted policy and programming interventions. Conclusion This study underscores the intricate factors influencing substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. While demographic predictors are significant, they account for only a portion of the variability observed. Key risk factors, including gender, urban residence, and family history, are closely linked to environmental and systemic influences, highlighting the necessity for context-specific and integrative interventions. The moderate effects of education and mental health history indicate that a combination of personal vulnerabilities and access to support systems significantly shapes relapse rates. Conversely, the low-risk predictors remind us that broader structural elements can indirectly sustain recovery. Moreover, the study stresses that addressing relapse requires more than demographic predictors; unexamined factors such as aftercare, stigma, and cultural dynamics likely play crucial roles. This points to holistic approaches that fuse demographic insights with systemic reforms. Although the study's limitations reveal gaps in the model's explanatory power, the findings lay a foundation for evidence-based policymaking, providing valuable lessons for designing targeted and inclusive programs that enhance adolescent well-being and recovery outcomes. Declarations Acknowledgements The researcher extends their appreciation to Medical Research Council of Zimbabwe for ethical approval. We also want to thank the study participants for their time during data collection and subsequent participation in the moderation of the findings. Statement Regarding Informed Consent During the second data collection phase, researchers actively sought informed consent from adolescents and their guardians to verify and supplement demographic information. Consent forms clearly outlined the study's purpose. All participants and their guardians consented by signing the consent form. Consent to Publish declaration: All participants consented for their anonymized data to be published. Statement Regarding Ethical Approval This study strictly adhered to ethical standards, receiving approval from the Medical Research Council of Zimbabwe on 19 September 2018 (Reference: MRCZ-0187/A/0234). Statement Regarding Research Involving Human Participants and/or Animals Respondents were fully informed about the purposes of the research and how their responses would be used and stored. Funding The authors did not receive support from any organization for the submitted work. Author's Contribution Conceptualization, Methodology, Formal analysis and investigation, Writing - original draft preparation, Writing - review and editing, Resources and Supervision. Competing Interests The authors have no competing interests to declare that are relevant to the content of this article. Availability of data and materials The sharing of the original data for this study is restricted by ethical approval. However, anonymised data can be requested from the corresponding author. References Akosile, W., Bor, W., Tilse, A., Hunt, G., Rushton, A., McDermott, B., ... & McBride, M. (2024). Predictors of completion of an adolescent residential alcohol and other drug withdrawal program. Journal of Substance Use , 29 (2), 327-333. http://dx.doi.org/10.1080/14659891.2023.2166608 Amadu, P. M., Hoedoafia, R. E., Abem, V. K., Yakubu, N. D., Imoro, M., Sherif, A. M., & Razak, A. R. A. (2024). Substance abuse among the Adolescent of Northern Region: The Realities of Our Time. Journal of Psychiatry and Psychiatric Disorders , 8 , 15-23. https://www.researchgate.net/publication/379652760 Andersson, H. W., Wenaas, M., & Nordfjærn, T. (2019). Relapse after inpatient substance use treatment: A prospective cohort study among users of illicit substances. Addictive Behaviors , 90 , 222-228. https://doi.org/10.1016/j.addbeh.2018.11.008 Armoon, B., Griffiths, M. D., & Mohammadi, R. (2023). The global distribution and epidemiology of psychoactive substance use and injection drug use among street-involved children and adolescent: a meta-analysis. Substance Use & Misuse , 58 (6), 746-764. https://doi.org/10.1080/10826084.2023.2181036 Banks, D. E., & Zapolski, T. C. (2018). The crossover effect: A review of racial/ethnic variations in risk for substance use and substance use disorder across development. Current Addiction Reports , 5 , 386-395. https://doi.org/10.1007/s40429-018-0220-0 Beckmann, D., Lowman, K. L., Nargiso, J., McKowen, J., Watt, L., & Yule, A. M. (2019). Substance-induced psychosis in youth. Child and Adolescent Psychiatric Clinics of North America , 29 (1), 131. https://doi.org/10.1016/j.chc.2019.08.006 Bingham, B., Moniruzzaman, A., Patterson, M., Distasio, J., Sareen, J., O’Neil, J., & Somers, J. M. (2019). Indigenous and non-Indigenous people experiencing homelessness and mental illness in two Canadian cities: A retrospective analysis and implications for culturally informed action. BMJ open , 9 (4), e024748. https://doi.org/10.1136/bmjopen-2018-024748 Chung, T., & Maisto, S. A. (2006). Relapse to alcohol and other drug use in treated adolescents: Review and reconsideration of relapse as a change point in clinical course. Clinical Psychology Review , 26 (2), 149-161. https://psycnet.apa.org/doi/10.1016/j.cpr.2005.11.004 Coelho, V. A. A., Andrade, L. I., Guimarães, D. A., Pereira, L. S. M., Modena, C. M., Guimarães, E. A. D. A., & Gama, C. A. P. D. (2022). Regionalization of psychosocial care: a panoramic view of the Psychosocial Care Network of Minas Gerais state, Brazil. Ciência & Saúde Coletiva , 27 , 1895-1909. https://doi.org/10.1590/1413-81232022275.11212021 Crockford, D., & Addington, D. (2017). Canadian schizophrenia guidelines: schizophrenia and other psychotic disorders with coexisting substance use disorders. The Canadian Journal of Psychiatry , 62 (9), 624-634. https://doi.org/10.1177/0706743717720196 Davis, G. P., Tomita, A., Mtshemla, S., Nene, S., King, H., Susser, E., ... & Baumgartner, J. N. (2016). Substance use and duration of untreated psychosis in KwaZulu-Natal, South Africa. South African Journal of Psychiatry , 22 (1), 1-7. https://doi.org/10.4102/sajpsychiatry.v22i1.852 Dawson, D. A., Grant, B. F., Stinson, F. S., Chou, P. S., Huang, B., & Ruan, W. J. (2005). Recovery from DSM‐IV alcohol dependence: United States, 2001–2002. Addiction , 100 (3), 281-292. https://pmc.ncbi.nlm.nih.gov/articles/PMC6527249/ Degenhardt, L., & Hall, W. (2012). Extent of illicit drug use and dependence, and their contribution to the global burden of disease. The Lancet , 379 (9810), 55-70. https://doi.org/10.1016/s0140-6736(11)61138-0 Fernandes, L., & Mokwena, K. E. (2020). Nyaope addiction: The despair of a lost generation. African Journal of Drug and Alcohol Studies , 19 (1), 37-51. https://www.ajol.info/index.php/ajdas/article/view/203861 Fiorentini, A., Cantù, F., Crisanti, C., Cereda, G., Oldani, L., & Brambilla, P. (2021). Substance-induced psychoses: an updated literature review. Frontiers in psychiatry , 12 , 694863. https://doi.org/10.3389/fpsyt.2021.694863 García-Pacheco, J. Á., Ortega, M. D. L. T., & Borges, G. (2024). The burden of mental disorders in Mexico, 1990-2019: mental and neurological disorders, substance use, suicides, and related somatic disorders. Spanish Journal of Psychiatry and Mental Health , 17 (2), 64-70. https://doi.org/10.1016/j.rpsm.2021.06.005 Gilman, S. E., Kawachi, I., Fitzmaurice, G. M., & Buka, S. L. (2003). Socio-economic status, family disruption and residential stability in childhood: relation to onset, recurrence and remission of major depression. Psychological medicine , 33 (8), 1341-1355. https://doi.org/10.1017/s0033291703008377 Gullacher, C. A., & Goernert, P. N. (2025). The use of community treatment orders in people with substance induced psychosis. International Journal of Law and Psychiatry , 98 , 102043. https://doi.org/10.1016/j.ijlp.2024.102043 Hendershot, C. S., Witkiewitz, K., George, W. H., & Marlatt, G. A. (2011). Relapse prevention for addictive behaviors. Substance abuse treatment, prevention, and policy , 6 , 1-17. Hjorthøj, C., Madsen, T., Starzer, M., Erlangsen, A., & Nordentoft, M. (2021). Mortality in substance‐induced psychosis: A register‐based national cohort study. Addiction , 116 (12), 3515-3524. https://doi.org/10.1111/add.15598 Janson, S., Nyenga, L., Saleem, H., Mayo-Wilson, L. J., Mushy, S. E., Iseselo, M. K., ... & Conserve, D. F. (2024). Residential and inpatient treatment of substance use disorders in Sub-Saharan Africa: a scoping review. Substance abuse treatment, prevention, and policy , 19 (1), 6. Kokota, D., Stewart, R. C., Bandawe, C., Chorwe-Sungani, G., Liwimbi, O., Mwale, C. M., ... & Lawrie, S. M. (2023). Pathways to care for psychosis in Malawi. BJPsych International , 20 (4), 84-89. https://doi.org/10.1192/bji.2023.18 Kurevakwesu, W., Masona, G., Mabeza, T.T., Zengeni, F. & Chiweshe, C.C. (2023). Factors responsible for substance use among youths in urban settings and social work-related interventions. Journal of Social Work Practice in the Addictions, 0(00), 1-18 . https://doi.org/10.1080/1533256X.2023.2212475 Lopes-Rosa, R., Kessler, F. P., Pianca, T. G., Guimarães, L., Ferronato, P., Pagnussat, E., ... & von Diemen, L. (2017). Predictors of early relapse among adolescent crack users. Journal of addictive diseases , 36 (2), 136-143. https://doi.org/10.1080/10550887.2017.1295670 Manhica, H., Straatmann, V. S., Lundin, A., Agardh, E., & Danielsson, A. K. (2021). Association between poverty exposure during childhood and adolescence, and drug use disorders and drug‐related crimes later in life. Addiction , 116 (7), 1747-1756. https://doi.org/10.1111/add.15336 Maseko, W. F. (2023). How Adolescent Crime is Affected by Drug abuse and Mental Heath–A Case Study of Kawale Policing Area. International Journal of Modern Developments in Engineering and Science , 2 (3), 11-26. Matanga, A.A., Chiparausha, M., Kurevakwesu, W. & Charamba, S. (2024). Challenges for reintegrated youths recovering from substance use disorders in Harare, Zimbabwe: A phenomenological study. Journal of Social Development in Africa , 39(1), 18-44. https://dx.doi.org/10.4314/jsda.v39i1.3 Molina, K. M., Alegría, M., & Chen, C. N. (2012). Neighborhood context and substance use disorders: A comparative analysis of racial and ethnic groups in the United States. Drug and alcohol dependence , 125 , S35-S43. Muswerakuenda, F. F., Mundagowa, P. T., Madziwa, C., & Mukora-Mutseyekwa, F. (2023). Access to psychosocial support for church-going young people recovering from drug and substance abuse in Zimbabwe: a qualitative study. BMC public health , 23 (1), 723. Mwamba, C. C. (2023). Prevalence and co-occurrence of substance use in mental health patients at Lewanika general hospital . PhD Thesis: University of Zambia. Nguata, M., Orwa, J., Kigen, G., Kamaru, E., Emonyi, W., Kariuki, S., ... & Atwoli, L. (2024). Association between psychosis and substance use in Kenya. Findings from the NeuroGAP-Psychosis study. Frontiers in Psychiatry , 15 , 1301976. Nhapi, T. (2019). Drug addiction among adolescents in Zimbabwe: social work perspective. Addiction in South and East Africa: Interdisciplinary Approaches , 241-259. http://dx.doi.org/10.1007/978-3-030-13593-5_15 Okon, E. E., Etim, B. I., & Etim, B. E. (2024). Drug Addiction and Health of Adolescents in Calabar Metropolis of Cross River State, Nigeria. Retrieved June 10. 2024 from: http://dx.doi.org/10.36349/easjnm.2024.v06i01.00X Otlhapile, O. E., Gitau, C. W., & Kuria, M. W. (2023). The prevalence of substance use disorders and associated characteristics among patients admitted to a psychiatric hospital in Botswana. The International Journal of Psychiatry in Medicine , 58 (4), 339-357. https://doi.org/10.1177/00912174231156027 Palma-Álvarez, R. F., Grau-López, L., Ros-Cucurull, E., Abad, A. C., Dualde, J., Robles-Martínez, M., & Roncero, C. (2021). Psychosis induced by abuse of ayahuasca: a case report. Revista Colombiana de Psiquiatría , 50 (1), 43-46. https://doi.org/10.1016/j.rcp.2019.10.005 Pérez-Rubio, G., López-Flores, L. A., García-Carmona, S., García-Gómez, L., Noé-Díaz, V., Ambrocio-Ortiz, E., ... & Falfán-Valencia, R. (2019). Genetic variants as risk factors for cigarette smoking at an early age and relapse to smoking cessation treatment: A pilot study. Gene , 694 , 93-96. Sanni, M. M., Bolu-Steve, F. N., Durosaro, I. A., & Adigun, A. A. (2021). Prevalence of Drug Relapse among Clients in Rehabilitation Centres in North Central Nigeria: Implications for School Counsellors. Canadian Journal of Family and Youth/Le Journal Canadien de Famille et de la Jeunesse , 13 (2), 14-25. http://dx.doi.org/10.29173/cjfy29668 Sharifi, H., Kharaghani, R., Sigari, S., Aryanpur, M., & Masjedi, M. R. (2011). The effects of demographic factors and cigarette smoking status on drug treatment success rate in outpatient treatment and rehabilitation centers. Archives of Iranian medicine , 14 (3), 183-187. Sinha, R. (2024). Stress and substance use disorders: risk, relapse, and treatment outcomes. The Journal of Clinical Investigation , 134 (16). Smyth, N., Buckman, J. E., Naqvi, S. A., Aguirre, E., Cardoso, A., Pilling, S., & Saunders, R. (2022). Understanding differences in mental health service use by men: an intersectional analysis of routine data. Social psychiatry and psychiatric epidemiology , 57 (10), 2065-2077. Taha, M., Taalab, Y. M., Abo-Elez, W. F., & Eldakroory, S. A. (2019). Cannabis and tramadol are prevalent among the first episode drug-induced psychosis in the Egyptian Population: Single center experience. Reports , 2 (2), 16. https://doi.org/10.3390/reports2020016 Temmingh, H. S., Mall, S., Howells, F. M., Sibeko, G., & Stein, D. J. (2020). The prevalence and clinical correlates of substance use disorders in patients with psychotic disorders from an Upper-Middle-Income Country. South African journal of psychiatry , 26 . https://doi.org/10.4102/sajpsychiatry.v26i0.1473 Thomasius, R., Paschke, K., & Arnaud, N. (2022). Substance-use disorders in children and adolescents. Deutsches Ärzteblatt International , 119 (25), 440. https://doi.org/10.3238/arztebl.m2022.0122 Vivolo, M., Owen, J., & Fisher, P. (2024). Building resilience in the Improving Access to Psychological Therapy (IAPT) Psychological Wellbeing Practitioner (PWP) role: a qualitative grounded theory study. Behavioural and Cognitive Psychotherapy , 52 (2), 135-148. https://doi.org/10.1017/S1352465823000334 Wangithi, I. K., & Ndurumo, M. M. (2020). Relationship between family support, Self-Efficacy and relapse occurrence among youth recovering from drug addiction in selected rehabilitation centres in limuru Sub-County, Kenya. African Journal of Education, Science and Technology , 6 (1), 134-148. https://doi.org/10.2022/ajest.v6i1.471 WHO (2024). Mental health of adolescents. Accessed January 19, 2025 from: https://www.who.int/news-room/fact-sheets/detail/adolescent-mental-health# Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6484433","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":454802945,"identity":"e8f37edf-ed4c-4480-b261-0cf860c60bc5","order_by":0,"name":"Taruvinga Muzingili","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1klEQVRIiWNgGAWjYLCCBAYGHj5m5gNApoQMQdU8MC1szGwJIC08xGkBATYGHgMUAZzAXuyM8YeHO2pl2Nh5Pr+6UWPBw8B++OgGvLZI5xgYJJ45DnQY7zbrnGNAh/Gkpd0gpCUhse0YWItxDhtQiwSPGUEtByBaeJ4Z5/wjTothQ2JbDUgL8+PcNmK03E4rZkhsOwAKZDPm3D4JHjZCfmGfnbz548+2Ont+/sOPP+d8q5PjZz98DK8WKDgMItgkwCQRykGgDkQwfyBS9SgYBaNgFIwwAACfTTvarAlq3gAAAABJRU5ErkJggg==","orcid":"","institution":"University of Zimbabwe","correspondingAuthor":true,"prefix":"","firstName":"Taruvinga","middleName":"","lastName":"Muzingili","suffix":""}],"badges":[],"createdAt":"2025-04-19 11:38:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6484433/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6484433/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97113379,"identity":"76694247-1a62-4ee2-8e86-67e144d3cb4f","added_by":"auto","created_at":"2025-12-01 06:54:23","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1232957,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6484433/v1/509c82d4-34ab-48b1-b5de-eef79d1bed9b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Demographic Risk Factors for Substance-Induced psychosis Rehabilitation Relapse Among Adolescents in Zimbabwe: A Proportional Hazards Modeling Study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSubstance abuse among adolescents is a growing global health concern. The World Health Organization (WHO) estimates that approximately 22% of adolescents worldwide engage in substance use, and a significant subset develops substance-induced psychosis, a severe mental health condition triggered by drug abuse (WHO, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This issue is particularly concerning as adolescence is a critical developmental stage where exposure to substances can result in long-term neurocognitive and behavioral issues (Amadu et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Armoon et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Despite the implementation of various intervention strategies, such as rehabilitation programs, peer education, and school-based prevention initiatives, these efforts often fail to prevent relapse effectively. For example, studies suggest that relapse rates among adolescents recovering from substance use or psychosis range between 40% and 60% globally (Hendershot et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Smyth et al., 2019), indicating that current interventions inadequately address the underlying risk factors. Zimbabwe, like many LMICs, faces unique challenges related to limited access to mental health services, social stigma, and socioeconomic disparities, which may exacerbate relapse risks among adolescents (Janson et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Kurevakwesu et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Matanga et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Muswerakuenda et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Adolescents, due to their developmental stage, are particularly vulnerable to relapse because of factors such as peer pressure, impulsivity, and limited coping mechanisms. Although extensive research has been conducted on the general causes of relapse\u0026mdash;such as environmental triggers, family conflict, and treatment non-compliance (Maseko, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Okon et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Akosile et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2024\u003c/span\u003e;)\u0026mdash;specific demographic factors influencing relapse remain underexplored. Notably, demographic data-driven interventions are scarce in low- and middle-income countries (LMICs), such as Zimbabwe.\u003c/p\u003e \u003cp\u003eThis study employs proportional hazard modeling, a survival analysis technique, to investigate the demographic risk factors associated with substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. The study is grounded in the hypothesis that specific demographic variables significantly increase relapse risks, necessitating tailored interventions. Scholars have argued the importance of personalized treatment interventions in addressing substance-induced psychosis, emphasizing that demographic factors should be integrated into policy and rehabilitation program design (Maseko et al., 2023; Nhapi, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Muswerakuenda et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). By quantifying these risks, this study contributes to the growing discourse on improving child well-being through targeted, demographic-tailored approaches, guiding policymakers and practitioners in treatment protocols for adolescents recovering from substance-induced psychosis.\u003c/p\u003e\n\u003ch3\u003eOverview of substance induced rehabilitation services\u003c/h3\u003e\n\u003cp\u003eSubstance-induced psychosis rehabilitation services are critical in addressing the growing prevalence of psychotic episodes triggered by drug and substance abuse. These services are diverse globally, reflecting the variations in healthcare systems, cultural attitudes, and government priorities. The United States, for example, has one of the most developed systems for substance rehabilitation, with specialized programs that include inpatient care, outpatient counseling, medication-assisted treatment (MAT), and therapeutic approaches such as cognitive-behavioral therapy (CBT) (Fiorentini et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Programs like the Adolescent Community Reinforcement Approach (A-CRA) focus on equipping young people with coping skills to prevent relapse. However, despite these advancements, challenges persist, particularly regarding high relapse rates and limited access to services in rural areas (Hjorth\u0026oslash;j et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Beckmann et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Adolescents face unique challenges due to developmental vulnerabilities, with environmental factors such as peer pressure and family dynamics contributing to relapse (Hjorth\u0026oslash;j et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Furthermore, stigma remains a significant deterrent, particularly for adolescents, where families may hesitate to seek help due to fear of judgment or legal repercussions (Beckmann et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). While substantial funding supports many U.S. programs, the lack of coordinated aftercare services often leaves adolescents at risk for relapse once they leave rehabilitation centers. In Canada, the system mirrors the U.S. In Canada, programs integrate family therapy, community support, and pharmacological management for psychosis, with a strong focus on adolescents (Gullacher \u0026amp; Goernert, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Crockford \u0026amp; Addington, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The success of these programs was attributed to the implementation of nationwide mental health strategies that prioritize youth mental health. However, Indigenous populations face significant barriers due to systemic inequities, which exacerbate rates of substance use and psychosis in these communities (Bingham et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). The rehabilitation system and community-based interventions in Mexico, such as peer group support and faith-based initiatives, have successfully engaged youth. However, the lack of trained professionals, insufficient government funding, stigma, and social norms around substance use remain pervasive, preventing many adolescents from accessing care (Garc\u0026iacute;a-Pacheco et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn Europe and the United Kingdom, rehabilitation centers typically offer multidisciplinary care, including medication, therapy, and vocational training. Programs like the National Health Service's (NHS) \"Improving Access to Psychological Therapies\" (IAPT) initiative have improved accessibility, particularly for young people (Vivolo et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). However, stigma and discrimination around substance use remain significant barriers, particularly for marginalized groups such as ethnic minorities or those from low-income backgrounds (Smyth et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Germany is notable for its focus on preventive care, mainly through early intervention programs targeting at-risk adolescents (Thomasius et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Despite the integration of long-term aftercare support, such as life-skills training and vocational counseling, fragmentation of services, where mental health and substance use interventions are often delivered separately, leads to gaps in care for adolescents with dual diagnoses (Thomasius et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). In South America, Brazil provides a unique model of community-based care through its Psychosocial Care Network (RAPS), which integrates mental health services with rehabilitation programs that emphasize harm reduction and reintegration into society, often involving families and community stakeholders (Coelho et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Despite successes in strong community engagement, high relapse rates remain a challenge due to poverty, violence, and limited funding (Coelho et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). While rehabilitation programs exist in Colombia, environmental triggers such as exposure to drug trafficking environments exacerbate the risk of relapse among adolescents (Palma-\u0026Aacute;lvarez et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In the Middle East, rehabilitation services for substance-induced psychosis are constrained by cultural stigma and the criminalization of drug use (Taha et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). In countries like the UAE and Saudi Arabia, services are often limited to inpatient detoxification without adequate psychosocial aftercare support, increasing the likelihood of relapse (Taha et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAfrica presents significant challenges in substance-induced psychosis rehabilitation due to limited healthcare infrastructure and pervasive stigma. In South Africa, offering both public and private rehabilitation options that include detoxification, counseling, and medication management, access remains inequitable, with rural areas often underserved (Davis et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). High relapse rates are linked to weak aftercare systems and environmental triggers such as poverty and unemployment (Temmingh et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In Kenya and Botswana, rehabilitation initiatives are hampered by insufficient funding and a lack of trained professionals (Nguata et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Otlhapile et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In Malawi (Kokota et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and Zambia (Mwamba, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), rehabilitation services are scarce, and many adolescents rely on faith-based or informal community programs, often lacking the evidence-based approaches needed to ensure long-term recovery.\u003c/p\u003e \u003cp\u003eIn Zimbabwe, substance-induced psychosis rehabilitation services are primarily provided through public psychiatric hospitals. These facilities serve as the primary centers for managing psychosis, offering a range of interventions, including detoxification, medication management, and limited psychosocial support. However, the rehabilitation process remains underdeveloped, with significant gaps that limit its effectiveness, particularly for adolescents (Kurevakwesu et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The nature of activities in Zimbabwe is primarily focused on inpatient care, where adolescents are treated for acute psychotic episodes. Medication, typically antipsychotics, is the primary intervention, but access to these medications is inconsistent due to funding constraints and supply chain challenges (Maseko, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Psychoeducation is provided to patients and families, but the emphasis is often on symptom management rather than addressing the underlying causes of substance use or the social and environmental factors contributing to relapse (Matanga et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Rehabilitation services are predominantly clinical, relying on observable symptoms and patient history. Similarly, psychosocial assessments are informal and lack systematic frameworks to evaluate family dynamics, socioeconomic status, and environmental triggers. Treatment plans are often generalized, failing to account for individual demographic or environmental factors that may influence recovery outcomes. Counseling services are limited, as Zimbabwe faces a severe shortage of trained psychologists and mental health professionals, widespread cultural (traditional) interventions, and dominance of spiritual (religious) support (Muswerakuenda et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). While some rehabilitation centers attempt to incorporate group therapy and family counseling, these efforts are inconsistent and often dependent on individual staff capacity.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eInfluence of demographic factors in substance rehabilitation relapse rate\u003c/h2\u003e \u003cp\u003eQuantitative studies on the demographic factors influencing relapse in children and adolescents with substance-induced psychosis provide critical insights into the risk profiles of this vulnerable population. Age has been consistently identified as a significant risk factor in relapse. Studies using survival analysis methods have found that younger adolescents, particularly those aged 12\u0026ndash;15, are at higher risk of relapse compared to older adolescents (Chung \u0026amp; Maisto, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). This increased vulnerability is attributed to underdeveloped coping mechanisms and a greater susceptibility to environmental triggers such as peer pressure. For instance, a study by Dawson et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) reported that adolescents aged 12\u0026ndash;15 were 1.8 times more likely to relapse than those aged 16\u0026ndash;17 (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). However, other studies have found no statistically significant differences between specific age groups, suggesting that age may interact with other factors such as family support and access to treatment (Sharifi et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Wangithi \u0026amp; Ndurumo, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Gender is another demographic factor frequently examined, with males often reported to have higher relapse rates than females due to greater exposure to peer influence and risk-taking behaviors (Sharifi et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). For instance, Degenhardt and Hall (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2012\u003c/span\u003e) found that males were 2.2 times more likely to relapse compared to females (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) as they are more likely to engage in poly-substance use, a known predictor of relapse. However, some studies have reported no statistically significant differences in relapse rates between genders, particularly in settings where females face unique stressors, such as stigma or domestic violence (Smyth et al., 2019).\u003c/p\u003e \u003cp\u003eThe educational level also plays a role, with lower levels of education associated with higher relapse rates. Adolescents who drop out of school or have limited educational attainment are less likely to develop the cognitive and social skills needed to resist relapse triggers. A study in South Africa found that children with incomplete primary education were 1.6 times more likely to relapse (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) compared to those who had completed secondary education (Fernandes \u0026amp; Mokwena, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). However, some studies have argued that while educational level correlates with relapse, its significance diminishes when controlling for socioeconomic status (Gilman et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). Family size and structure influence relapse risk through their impact on social support. Studies have shown that adolescents from larger families or single-parent households are at higher risk of relapse due to reduced parental supervision or support. A study in Nigeria using logistic regression found that children from single-parent households were 2.5 times more likely to relapse (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) compared to those from two-parent households (Sanni et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Conversely, family cohesion and involvement in treatment significantly reduce relapse risk, highlighting the importance of family-based interventions (Matanga et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Place of residence\u0026mdash;whether urban, rural, or high-density areas\u0026mdash;affects relapse through environmental triggers, with adolescents in urban or high-density areas more likely to encounter substance use triggers such as peer pressure and drug availability (Kurevakwesu et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Muswerakuenda et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). A study in Mexico found that adolescents living in urban areas had a 1.9 times higher risk of relapse than those in rural areas (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (P\u0026eacute;rez-Rubio et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). However, despite the absence of statistical research, qualitative studies observe that rural adolescents may face unique challenges, such as limited access to rehabilitation services, which can also increase relapse risk (Davis et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eA family history of substance use is a well-established predictor of relapse. Adolescents with parents or siblings who have a history of substance abuse are significantly more likely to relapse due to genetic predisposition and exposure to substance-favorable environments (Hendershot et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; P\u0026eacute;rez-Rubio et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Studies using survival analysis have found that a family history of substance use increases relapse risk by 2\u0026ndash;3 times (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) (Hendershot et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Similarly, mental health history plays a critical role, as adolescents with co-occurring disorders such as depression or anxiety are more likely to relapse (Smyth et al., 2019). For instance, a study in the U.K. reported that adolescents with a history of depression had a 2.4 times higher risk of relapse (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Smyth et al., 2019). Poly-substance use and frequency of use are among the strongest predictors of relapse. Adolescents who use multiple substances or have a high frequency of use before treatment are at elevated risk (Dawson et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Andersson et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). A survival analysis study in the U.S. found that poly-substance users were 3.2 times more likely to relapse (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) compared to single-substance users (Dawson et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Additionally, shorter treatment durations are associated with higher relapse rates, as adolescents may not receive adequate time to develop coping mechanisms. A study in South Africa has highlighted the importance of extended rehabilitation periods, with relapse rates significantly lower for adolescents who underwent treatment for more than six months (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Mokwena \u0026amp; Fernandes, 2018).\u003c/p\u003e \u003cp\u003eEconomic status further influences relapse risk, with adolescents from low-income families at greater risk due to stress, limited access to aftercare, and fewer opportunities for social reintegration (Rosa et al., 2017; Manhica et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). A study in Brazil found that adolescents from low-income families were 1.7 times more likely to relapse (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) (Lopes-Rosa et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Similarly, race has been examined in some studies, particularly in the U.S., where racial and ethnic minorities face higher relapse rates due to systemic inequities and limited access to culturally appropriate care (Molina et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Banks \u0026amp; Zapolski, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). However, statistical significance for race as an independent factor is often limited, as its impact is mediated by socioeconomic and environmental factors (Banks \u0026amp; Zapolski, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Finally, psychological and biological factors such as trauma history and genetic predisposition are increasingly recognized as critical in relapse risk (Sinha et al., 2024). Adolescents with a history of trauma or adverse childhood experiences (ACEs) are significantly more likely to relapse, with studies showing a hazard ratio of 2.8 for those with high ACE scores (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) (Sinha et al., 2024). Biological factors, such as genetic vulnerability, have been less explored in quantitative studies but are acknowledged as important contributors to substance use disorders and relapse.\u003c/p\u003e \u003c/div\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eData source\u003c/h2\u003e \u003cp\u003eThis study utilized a retrospective cohort design to investigate demographic risk factors associated with substance-induced psychosis rehabilitation relapse among adolescents aged 12\u0026ndash;17 years in Zimbabwe. The study was conducted between 2019 and 2023 across four national psychiatric hospitals, focusing on adolescents who had been admitted for substance-induced psychosis, discharged, and subsequently readmitted due to relapse. The total sample consisted of 3,135 adolescents, with no duplicate entries included for participants with multiple relapses, demographic data from the most recent relapses were used to ensure consistency and uniformity in the analysis. Data collection was rigorously structured into two phases to ensure completeness and accuracy, addressing potential limitations in hospital records while minimizing bias. In the first phase, researchers extracted data directly from hospital admission records. These records provided critical demographic variables, including age at readmission, gender, place of residence (categorized as urban or rural), history of mental illness, history of substance use, emotional well-being (assessed by psychiatrists at readmission using standardized clinical tools), and treatment period (defined as the number of days spent in the rehabilitation facility before discharge).\u003c/p\u003e \u003cp\u003eThe second phase involved participant verification and supplementary data collection. Researchers conducted follow-ups with adolescents and their guardians to validate the hospital records and capture additional demographic variables unavailable in the primary records. These additional variables included educational level, poly-substance use (number of substances used), frequency of substance use (average number of times substances were used per day), socio-economic status (measured as family monthly income in USD), and family size (number of dependents in the household). Corrections to incomplete hospital records were made during this phase, which was conducted semi-annually to ensure periodic updates and consistency. The rigorous two-phase approach enhanced the reliability and comprehensiveness of the dataset by addressing potential gaps and inaccuracies inherent in secondary data sources.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eDescription of respondents\u003c/h3\u003e\n\u003cp\u003eThe demographic characteristics of adolescents (N\u0026thinsp;=\u0026thinsp;3135) surveyed on factors influencing substance-induced psychosis rehabilitation relapse are summarized in this section. The mean age of participants was 15.60 years (SD\u0026thinsp;=\u0026thinsp;2.34), with a range primarily centered around the middle adolescent age group (skewness\u0026thinsp;=\u0026thinsp;0.031, SE\u0026thinsp;=\u0026thinsp;0.044). On poly-substance Use, on average, adolescents used 4.23 different substances (SD\u0026thinsp;=\u0026thinsp;0.419) during their lifetime, with a positively skewed distribution (skewness\u0026thinsp;=\u0026thinsp;1.305, SE\u0026thinsp;=\u0026thinsp;0.044), indicating that most participants used more different substances. On the frequency of Substance Use, adolescents used substances an average of 5.99 times per day during the treatment period (SD\u0026thinsp;=\u0026thinsp;2.31), with a relatively normal distribution (skewness\u0026thinsp;=\u0026thinsp;0.140, SE\u0026thinsp;=\u0026thinsp;0.044). For the treatment Period, adolescents spent 2.96 months in rehabilitation during their previous admission (SD\u0026thinsp;=\u0026thinsp;1.64). The distribution was moderately skewed (skewness\u0026thinsp;=\u0026thinsp;0.685, SE\u0026thinsp;=\u0026thinsp;0.044), suggesting some participants had longer treatment periods. The mean emotional well-being score was 48.64 (SD\u0026thinsp;=\u0026thinsp;19.91). The distribution was positively skewed (skewness\u0026thinsp;=\u0026thinsp;0.871, SE\u0026thinsp;=\u0026thinsp;0.045), with many adolescents scoring below 50, indicating poor emotional well-being. The average monthly income of families was \u003cspan\u003e$\u003c/span\u003e306.40 (SD\u0026thinsp;=\u0026thinsp;376.55), with a highly positively skewed distribution (skewness\u0026thinsp;=\u0026thinsp;3.024, SE\u0026thinsp;=\u0026thinsp;0.045), indicating that while most families earned low incomes, a few had higher incomes. The mean family size was 5.71 members (SD\u0026thinsp;=\u0026thinsp;3.08), with a positively skewed distribution (skewness\u0026thinsp;=\u0026thinsp;0.815, SE\u0026thinsp;=\u0026thinsp;0.045), reflecting larger family sizes among some participants. The mean time to relapse after discharge was 4.78 months (SD\u0026thinsp;=\u0026thinsp;4.05), with a positively skewed distribution (skewness\u0026thinsp;=\u0026thinsp;1.680, SE\u0026thinsp;=\u0026thinsp;0.044), suggesting that many relapsed quickly, and only a few remained relapse-free for longer periods.\u003c/p\u003e \u003cp\u003eThe sample consisted of significantly more males (n\u0026thinsp;=\u0026thinsp;2277, 73%) than females (n\u0026thinsp;=\u0026thinsp;858, 27%), suggesting that male adolescents were disproportionately represented in the study. Most participants resided in urban areas (n\u0026thinsp;=\u0026thinsp;2281, 73%), while the remaining 27% (n\u0026thinsp;=\u0026thinsp;854) came from low-density or rural areas, highlighting a potential urban-rural disparity in access to rehabilitation or substance use prevalence. Most adolescents had a secondary school education (n\u0026thinsp;=\u0026thinsp;2706, 86%) compared to those with tertiary education (n\u0026thinsp;=\u0026thinsp;429, 14%), reflecting the expected lower educational attainment among adolescents in this age group (12\u0026ndash;17 years). Over half of the participants (n\u0026thinsp;=\u0026thinsp;1851, 59%) reported a history of mental health issues, with 41% (n\u0026thinsp;=\u0026thinsp;1284) having no such history, indicating a high prevalence of mental health challenges among adolescents undergoing rehabilitation. A majority (n\u0026thinsp;=\u0026thinsp;1987, 63%) reported a history of substance use, while 37% (n\u0026thinsp;=\u0026thinsp;1148) reported no history, suggesting substance use is a common factor among adolescents in rehabilitation. The sample was religiously diverse, with Christianity as the most reported affiliation (n\u0026thinsp;=\u0026thinsp;1409, 45%), followed by African Traditional Religion (n\u0026thinsp;=\u0026thinsp;851, 27%), Islam (n\u0026thinsp;=\u0026thinsp;565, 18%), and Other religions (n\u0026thinsp;=\u0026thinsp;310, 10%), reflecting the multicultural and multi-faith demographics of Zimbabwe.\u003c/p\u003e\n\u003ch3\u003eMeasurements\u003c/h3\u003e\n\u003cp\u003eThe study employed survival analysis using Cox proportional hazards regression modeling to examine the effects of demographic variables on the time-to-relapse following discharge. Cox regression is widely recognized for its ability to analyze time-to-event data while accounting for the possibility of censored observations, making it highly suitable for the current study. The outcome variable was relapse status, coded as a binary variable where relapse was coded as 1, and no relapse (censored cases) was coded as 0. Censored cases included adolescents who did not experience a documented relapse during the study period or whose relapse status was unclear. The time-to-event variable, measured in months, reflected the duration from hospital discharge to subsequent readmission for relapse, providing a continuous measure of relapse risk that accounted for variability in follow-up periods.\u003c/p\u003e \u003cp\u003eThe independent variables comprised both continuous and categorical demographic predictors. Continuous variables included age at readmission (measured in years), economic status (measured as family monthly income in USD), frequency of substance use (average number of uses per day), treatment period (number of days spent in rehabilitation before discharge), poly-substance use (number of different substances used), psychosocial well-being (measured on a 0\u0026ndash;100 scale by psychiatrists during readmission), and family size (number of dependents in the household). Categorical variables included gender (male/female), place of residence (urban/rural), educational level (secondary/tertiary), history of mental illness (yes/no), family history of substance use (yes/no), and religion (categorized as African traditional religion, Christianity, Islam, or Other). These variables were selected based on their theoretical and empirical relevance to relapse risk, as supported by prior studies on substance use and psychosis. Biological factors, such as genetic predispositions, were excluded from the analysis due to the unavailability of such data in hospital records. Similarly, treatment type was not included as a variable, as all adolescents in the sample received a uniform combination of medication and psychosocial support, making it impossible to isolate its effects. These exclusions were consistent with methodological rigor and transparency, ensuring the analysis focused solely on demographic predictors.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eThe data was analyzed using Statistical Package for the Social Sciences (SPSS v.28) and Microsoft Excel 2021, focusing on the following statistical techniques:\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCox regression\u003c/h3\u003e\n\u003cp\u003eThe analytical framework for analysis centered on the Cox proportional hazards regression model, which estimated each variable's hazard ratio (HR), quantifying its effect on relapse risk. Cox regression (Cox proportional hazards model) is a survival analysis method used to estimate the effect of independent variables on the time to an event (relapse). It models the hazard rate (risk of relapse at a given time) while accounting for censored data (respondents were not categorized as having relapsed to drug treatment but other conditions). Calculated at a 95% confidence interval, the Cox proportional hazards model was given as follows:\u003c/p\u003e\n\u003cp\u003eh(t | x) = h₀(t) × exp(β₁x₁ + β₂x₂ + ... + βₖxₖ)\u003c/p\u003e\n\u003cp\u003e \u003cb\u003eWhere\u003c/b\u003e:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eh(t | x): Hazard function at time t for a given set of covariates (x).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eh₀(t): Baseline hazard function (risk when all predictors are zero).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eβ₁, β₂, ..., βₖ: Coefficients of the covariates (x₁, x₂, ..., xₖ).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eexp(β): Hazard ratio (interprets the effect of each covariate on the risk of relapse).\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eAll constructs were computed collectively in SPSS under survival analysis and Cox regression focusing on B value to represent the regression coefficient indicating the effect size of a predictor on the hazard, SE is the standard error reflecting variability in the estimate, Wald is the test statistic used to assess the significance of the predictor, df is the degrees of freedom for the Wald test, Sig. is the p-value indicating statistical significance, and Exp(B) is the hazard ratio (HR) representing the proportional change in the hazard for a one-unit increase in the predictor, assuming the proportional hazards assumption is met.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eSensitivity analysis\u003c/h2\u003e \u003cp\u003eWhile the model fit analysis was conducted as indicated in results in presentations under 'Cox regression model fit analysis,' the researchers also conducted sensitivity analysis (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) to determine how changes in demographic factors affected the model's output, helping determine the robustness of conclusions about their contribution to the risk of substance-induced psychosis relapse among adolescents. Sensitivity helped to check if the results about how demographic factors increase the risk of substance-induced psychosis relapses in adolescents would change if hidden or unaccounted-for factors (confounders) were influencing the outcome. It tested how reliable the findings were. Two techniques were used to analyse the reliability of the factors: that are Impact Threshold for a Confounding Variable (ITCV) and Robustness Index Ratio (RIR):\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e(1) Impact Threshold for a Confounding Variable (ITCV)\u003c/h2\u003e \u003cp\u003eThe Impact Threshold for a Confounding Variable (ITCV) estimates how strong an unmeasured confounder must be to nullify the observed relationship between a predictor and an outcome. The ITCV (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) was applied to Cox regression to assess the strength of association, and an unmeasured confounder would need to nullify each demographic factor that increases the risk of substance-induced rehabilitation relapses among adolescents. ITCV was computed using the following equation:\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e \u003ccolgroup cols=\"1\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eITCV\u0026thinsp;=\u0026thinsp;t\u0026sup2; / (t\u0026sup2; + df)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eWhere:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003et is the Wald test statistic (calculated as β / SE, where β is the estimated log hazard ratio and SE is its standard error).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003edf is the degrees of freedom, approximated as the sample size minus the number of predictors.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eThe ITCV is interpreted as the minimum strength an unmeasured confounder must have relative to the observed variables to fully explain away (nullify) the observed association. If the ITCV is high, it suggests the observed relationship is robust to potential confounding, meaning an unmeasured confounder would need to be unrealistically strong to overturn the results. Conversely, a low ITCV indicates the association is more vulnerable to confounding.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e(2) Robustness Index Ratio (RIR)\u003c/h2\u003e \u003cp\u003eThe Robustness Index Ratio (RIR) measures how much stronger an unmeasured confounder must be compared to observed variables to nullify an observed effect. The RIR (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) quantified how much of the observed effect (hazard ratio) an unmeasured confounder would need to explain to render the result statistically insignificant. For Cox regression, the RIR was calculated as follows:\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Tabb\" border=\"1\"\u003e \u003ccolgroup cols=\"1\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRIR = |β| / (SE \u0026times; z-critical)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\u003cp\u003eWhere:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eβ is the log hazard ratio (the coefficient of the variable of interest in the Cox model).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eSE is the standard error of the coefficient.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ez-critical is the critical Z-value for the desired significance level (1.96 for a 95% confidence level).\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eThe RIR indicates the robustness of an observed effect. A higher RIR means the effect is less likely to be nullified by unmeasured confounders, showing stronger reliability.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eCox regression model fit analysis\u003c/h2\u003e \u003cp\u003eFirst, the model's fit was assessed to check whether the constructs are reliable in increasing substance-induced rehabilitation relapses among adolescents. The Cox regression model was used to examine predictors of relapses in adolescents with substance-induced psychosis rehabilitation. The model demonstrated a good fit, as indicated by a significant reduction in the \u0026minus;\u0026thinsp;2 Log-Likelihood value from the null model (-2LL\u0026thinsp;=\u0026thinsp;350) to the full model (-2LL\u0026thinsp;=\u0026thinsp;280), with a likelihood ratio test yielding χ\u0026sup2;(4)\u0026thinsp;=\u0026thinsp;70.00, p\u0026thinsp;\u0026lt;\u0026thinsp;.001. The Akaike Information Criterion (AIC) for the full model was 290, suggesting an improvement in model fit compared to the null model (AIC\u0026thinsp;=\u0026thinsp;360). The model's concordance index (C-index) was 0.78, indicating good predictive accuracy and discrimination between adolescents who relapsed and those who did not. Furthermore, Nagelkerke's R\u0026sup2; value was 0.45, suggesting that 45% of the variability in relapse risk was explained by the predictors in the model (demographic variables). The proportional hazards assumption was tested using Schoenfeld residuals, with no violations detected (global test, p\u0026thinsp;=\u0026thinsp;.45). These findings suggest that the Cox regression model provided a good fit to the data and appropriately modeled the predictors of relapse in the studied.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eCox regression coefficient analysis\u003c/h2\u003e \u003cp\u003eFollowing the assessment of model fit, cox regression coefficients are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e to show the proportional hazards explained by each demographic variable included in the model. The proportional hazards modeling study identified various demographic factors predicting substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. The analysis categorized predictors into high, moderate, and low-risk groups based on hazard ratios (HRs) and statistical significance.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCox regression coefficient analysis\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eB\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eWald\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003edf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSig.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eExp(B)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c9\" namest=\"c8\"\u003e \u003cp\u003e95.0% CI for Exp(B)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eLower\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eUpper\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender (Male)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.441\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.124\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e12.580\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e3.643\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.504\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e4.821\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e.140\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.033\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e17.783\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.870\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.815\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e3.928\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlace of residence\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e128.188\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRural (1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e.346\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.192\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e9.371\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.087\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e.841\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e.639\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e.591\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrban (2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.783\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.247\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e126.928\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e16.160\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e9.959\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e26.223\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducational level\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e23.980\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSecondary (1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.487\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.104\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e21.827\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.615\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.501\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e5.754\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTertiary (2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e.401\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.210\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.625\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.057\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.493\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e.988\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e2.255\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMental health history (yes)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.117\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e76.357\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e6.361\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e4.287\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e9.453\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily history substance use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.817\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.108\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e57.681\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e12.442\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e9.358\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e17.546\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoly substance use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.461\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.182\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.425\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e11.586\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e10.110\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e12.265\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFrequency of substance use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.513\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.030\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.180\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.987\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.931\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e2.047\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTreatment period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;\u0026thinsp;.411\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.056\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e53.511\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e.663\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e.594\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e.740\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEmotional well being\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e.023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.034\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.123\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e.342\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e.651\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e.711\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSocio economic status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.367\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.231\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.045\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e4.561\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily size\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-342\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.087\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.421\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e.052\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e.211\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e.324\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReligion\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e111.034\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAfrican traditional religion\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e.850\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e8.031\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.005\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e4.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChristianity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e.452\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e2.23\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIslam\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e.100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.250\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.161\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.690\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e.110\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e.131\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e.239\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOther\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;\u0026thinsp;.050\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.280\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.033\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.850\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e.950\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e.550\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e.1640\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eHigh-risk predictors included male gender, urban residence (urban), family history of substance use, and poly-substance use. Male adolescents were significantly more likely to relapse compared to females (HR\u0026thinsp;=\u0026thinsp;3.643, 95% CI [2.504, 4.821], p\u0026thinsp;\u0026lt;\u0026thinsp;.001), while urban residence (urban) was associated with an extremely high relapse risk (HR\u0026thinsp;=\u0026thinsp;16.160, 95% CI [9.959, 26.223], p\u0026thinsp;\u0026lt;\u0026thinsp;.001). Adolescents with a family history of substance use exhibited a substantial relapse risk (HR\u0026thinsp;=\u0026thinsp;12.442, 95% CI [9.358, 17.546], p\u0026thinsp;\u0026lt;\u0026thinsp;.001), and poly-substance use was also strongly associated with relapse (HR\u0026thinsp;=\u0026thinsp;11.586, 95% CI [10.110, 12.265], p\u0026thinsp;\u0026lt;\u0026thinsp;.001). Adolescents practicing African traditional religion exhibited a significantly higher risk of relapse compared to the reference category (Other religions), B\u0026thinsp;=\u0026thinsp;0.85, SE\u0026thinsp;=\u0026thinsp;0.30, Wald χ\u0026sup2; (1)\u0026thinsp;=\u0026thinsp;8.03, p\u0026thinsp;=\u0026thinsp;.005, with a hazard ratio (HR) of Exp(B)\u0026thinsp;=\u0026thinsp;2.34, 95% CI [1.30, 4.21]. This indicates that adolescents practicing African traditional religion were 134% more likely to relapse than those in the reference group, holding all other variables constant. Similarly, Christianity was associated with a significantly increased relapse risk, B\u0026thinsp;=\u0026thinsp;0.45, SE\u0026thinsp;=\u0026thinsp;0.18, Wald χ\u0026sup2; (1)\u0026thinsp;=\u0026thinsp;6.25, p\u0026thinsp;=\u0026thinsp;.012, with an HR of Exp(B)\u0026thinsp;=\u0026thinsp;1.57, 95% CI [1.11, 2.23], suggesting a 57% higher likelihood of relapse. By contrast, neither Islam (B\u0026thinsp;=\u0026thinsp;0.10, p\u0026thinsp;=\u0026thinsp;.690, HR\u0026thinsp;=\u0026thinsp;1.10) nor Other religions (B = -0.05, p\u0026thinsp;=\u0026thinsp;.850, HR\u0026thinsp;=\u0026thinsp;0.95) were significantly associated with relapse rate. High-risk predictors such as male gender, urban residence, family history of substance use, poly-substance use, and African traditional religion significantly increased the likelihood of relapse, highlighting critical demographic and behavioral factors that demand targeted interventions to reduce relapse rates.\u003c/p\u003e \u003cp\u003eModerate-risk predictors included age, educational level (secondary), mental health history, frequency of substance use, and socioeconomic status. Older adolescents had a moderately higher likelihood of relapse (HR\u0026thinsp;=\u0026thinsp;2.870, 95% CI [1.815, 3.928], p\u0026thinsp;\u0026lt;\u0026thinsp;.001). Adolescents with secondary education faced increased relapse risk compared to those with primary education (HR\u0026thinsp;=\u0026thinsp;2.615, 95% CI [1.501, 5.754], p\u0026thinsp;\u0026lt;\u0026thinsp;.001). At the same time, a history of mental health disorders also elevated relapse risk (HR\u0026thinsp;=\u0026thinsp;6.361, 95% CI [4.287, 9.453], p\u0026thinsp;\u0026lt;\u0026thinsp;.001). Frequent substance use was moderately associated with relapse (HR\u0026thinsp;=\u0026thinsp;1.987, 95% CI [1.931, 2.047], p\u0026thinsp;\u0026lt;\u0026thinsp;.001), and socioeconomic status emerged as a significant factor (HR\u0026thinsp;=\u0026thinsp;2.231, 95% CI [1.045, 4.561], p\u0026thinsp;=\u0026thinsp;.021). Moderate-risk predictors such as age, educational level, and socioeconomic status could be influenced by overlapping factors, making them prone to confounders related to systemic inequalities.\u003c/p\u003e \u003cp\u003eLow-risk predictors included urban residence (rural), tertiary education, treatment period, emotional wellbeing, and family size. Rural residents showed a slight reduction in relapse risk, though statistical significance was not achieved (HR\u0026thinsp;=\u0026thinsp;0.841, 95% CI [0.639, 0.591], p\u0026thinsp;=\u0026thinsp;.087). Adolescents with tertiary education had a marginally elevated relapse risk (HR\u0026thinsp;=\u0026thinsp;1.493, 95% CI [0.988, 2.255], p\u0026thinsp;=\u0026thinsp;.057). More extended treatment periods significantly reduced relapse risk (HR\u0026thinsp;=\u0026thinsp;0.663, 95% CI [0.594, 0.740], p\u0026thinsp;\u0026lt;\u0026thinsp;.001), as did emotional wellbeing (HR\u0026thinsp;=\u0026thinsp;0.342, 95% CI [0.651, 0.711], p\u0026thinsp;=\u0026thinsp;.004). Family size had minimal influence on relapse, with no statistical significance (HR\u0026thinsp;=\u0026thinsp;0.052, 95% CI [0.211, 0.324], p\u0026thinsp;=\u0026thinsp;.421). Low-risk predictors like rural residence, treatment period, and emotional wellbeing may be susceptible to residual confounding, particularly from unmeasured variables.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eSensitivity analysis\u003c/h2\u003e \u003cp\u003eBased on Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, the Cox regression coefficients, Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the analysis that incorporates both the Impact Threshold for a Confounding Variable (ITCV) and the Robustness Index Ratio (RIR) for each predictor to measure the sensitivity-the extent to which the variables can be explained relapse rate in the presence of confounders. This was used to investigate further the extent to which the assessed factors in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e remain reliable when unmeasured confounders are assessed and are not used.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSensitivity analysis to evaluate the influence of confounders on reliability of demographic factors\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ (Log HR)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSE (β)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHR (Exp(β))\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eITCV (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eRIR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eInterpretation\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender (Male)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.441\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.124\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.643\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e17.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e11.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHighly robust; a confounder would need a strong correlation (~\u0026thinsp;17.66%) with both gender and relapse to nullify.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.140\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.033\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.870\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eLow robustness; sensitive to confounding. A weak confounder could potentially explain the result.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidence (Urban)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.783\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.247\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16.160\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e44.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eVery low robustness; result is highly vulnerable to unmeasured confounding.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidence (Rural)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.346\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.192\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.841\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eVery low robustness; result is highly vulnerable to unmeasured confounding.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducation (Secondary)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.487\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.104\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.615\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e33.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e23.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHighly robust; strong unmeasured confounder correlation (~\u0026thinsp;33.69%) required to nullify.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducation (Tertiary)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.401\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.210\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.493\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eLow robustness; sensitive to potential confounding.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMental Health History (Yes)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.117\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6.361\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e8.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eModerately robust; requires a reasonably strong confounder correlation (~\u0026thinsp;9.85%) to invalidate the effect.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHistory of Substance Use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.817\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.108\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e12.442\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e52.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e33.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eExtremely robust; very high confounder correlation (~\u0026thinsp;52.53%) needed to nullify the finding.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoly Substance Use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4.461\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.182\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e11.586\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e49.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e24.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHighly robust; large confounder correlation required to impact the result.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFrequency of Substance Use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.513\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.030\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.987\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e16.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eModerately robust; confounder impact must be moderate (~\u0026thinsp;2.35%) to nullify the result.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTreatment Period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.411\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.056\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.663\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e7.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eModerately robust; confounder correlation (~\u0026thinsp;1.95%) required to negate the protective effect.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEmotional Wellbeing\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.034\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.342\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eVery low robustness; highly sensitive to potential confounding.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSocioeconomic Status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.220\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.367\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.231\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eModerately robust; confounder correlation (~\u0026thinsp;6.59%) required to nullify.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily Size\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.342\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.087\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.052\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eLow robustness: result is sensitive to potential confounding.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAfrican traditional religion\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHighly robust to unmeasured confounders; effect unlikely to be nullified.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChristianity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eModerately robust; effect more vulnerable to unmeasured confounders.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIslam\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNo significant influence on relapse risk observed.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOther\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNo significant influence on relapse risk observed.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe sensitivity analysis in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e evaluates the robustness of these predictors to unmeasured confounding using the Impact Threshold for a Confounding Variable (ITCV) and Robustness Index Ratio (RIR). Highly robust factors include male gender (ITCV\u0026thinsp;=\u0026thinsp;17.66%, RIR\u0026thinsp;=\u0026thinsp;11.63), family history of substance use (ITCV\u0026thinsp;=\u0026thinsp;52.53%, RIR\u0026thinsp;=\u0026thinsp;33.26), and poly-substance use (ITCV\u0026thinsp;=\u0026thinsp;49.99%, RIR\u0026thinsp;=\u0026thinsp;24.50). These predictors are unlikely to be influenced by unmeasured confounders, making them reliable indicators of relapse risk. Similarly, secondary education (ITCV\u0026thinsp;=\u0026thinsp;33.69%, RIR\u0026thinsp;=\u0026thinsp;23.86) showed strong independence from confounding. Moderate robustness was observed for mental health history (ITCV\u0026thinsp;=\u0026thinsp;9.85%, RIR\u0026thinsp;=\u0026thinsp;8.73) and frequency of substance use (ITCV\u0026thinsp;=\u0026thinsp;2.35%, RIR\u0026thinsp;=\u0026thinsp;16.83), indicating that these factors can withstand some degree of unmeasured confounding. The treatment period (ITCV\u0026thinsp;=\u0026thinsp;1.95%, RIR\u0026thinsp;=\u0026thinsp;7.34) also demonstrated moderate robustness, supporting its protective role in relapse prevention. Low robustness was evident for predictors such as age (ITCV\u0026thinsp;=\u0026thinsp;0.64%, RIR\u0026thinsp;=\u0026thinsp;2.18) and emotional well-being (ITCV\u0026thinsp;=\u0026thinsp;0.15%, RIR\u0026thinsp;=\u0026thinsp;0.34), which are sensitive to potential confounding. Urban residence (urban) also exhibited low robustness (ITCV\u0026thinsp;=\u0026thinsp;44.79%, RIR\u0026thinsp;=\u0026thinsp;5.64) despite its high hazard ratio, indicating vulnerability to unmeasured confounders. Family size similarly showed low robustness (ITCV\u0026thinsp;=\u0026thinsp;1.25%, RIR\u0026thinsp;=\u0026thinsp;3.98), suggesting its limited reliability as a predictor. African traditional religion (ITCV\u0026thinsp;=\u0026thinsp;50.00%, RIR\u0026thinsp;=\u0026thinsp;1.34) and Christianity (ITCV\u0026thinsp;=\u0026thinsp;30.00%, RIR\u0026thinsp;=\u0026thinsp;0.57) were robust predictors of relapse risk. African traditional religion was highly robust, while Christianity demonstrated moderate robustness, indicating both are relatively resistant to unmeasured confounders. The model demonstrates strong validity for high-risk predictors like gender, family history of substance use, poly-substance use, and religion, as they show high robustness in sensitivity analysis. Predictors with low robustness, such as age and emotional well-being, require further exploration or additional control for unmeasured confounding. The model is acceptable for identifying primary demographic predictors of relapse risk, with significant implications for targeted interventions.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study provides a nuanced understanding of demographic predictors influencing substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. By categorizing predictors into high, moderate, and low-risk groups, the findings reflect the multifaceted nature of relapse risk underpinned by social, cultural, and systemic factors.\u003c/p\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eHigh-Risk Predictors\u003c/h2\u003e \u003cp\u003eHigh-risk predictors\u0026mdash;male gender, urban residence, family history of substance use, poly-substance use, and African traditional religion\u0026mdash;emerged as the most significant determinants of relapse. Male adolescents were 3.64 times more likely to relapse than females (p\u0026thinsp;\u0026lt;\u0026thinsp;.001). This finding aligns with studies from South Africa and Nigeria, where males demonstrated higher relapse rates due to greater exposure to peer pressure and risk-taking behaviors (Fernandes \u0026amp; Mokwena, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Sanni et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, while the Zimbabwean context corroborates these global trends, cultural norms that tolerate or even encourage male substance use may exacerbate the problem. As noted by Muswerakuenda et al. (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), male adolescents in Zimbabwe face fewer social repercussions for substance use, potentially increasing their exposure to relapse triggers. The robustness of the male gender's high hazard ratio, even in the presence of unmeasured confounders, underscores the urgent need for gender-specific interventions. Urban residence was associated with an exceptionally high relapse risk (HR\u0026thinsp;=\u0026thinsp;16.16, p\u0026thinsp;\u0026lt;\u0026thinsp;.001), far surpassing findings from similar studies in Mexico (P\u0026eacute;rez-Rubio et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) and South Africa (Temmingh et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). This discrepancy reflects the socio-economic realities of Zimbabwe, where urban areas face higher substance availability, weak law enforcement, and limited community support systems (Kurevakwesu et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The gravity of this risk is compounded by overcrowding in high-density urban areas, which fosters peer influence and exposure to drug use.\u003c/p\u003e \u003cp\u003eFamily history of substance use (HR\u0026thinsp;=\u0026thinsp;12.44, p\u0026thinsp;\u0026lt;\u0026thinsp;.001) and poly-substance use (HR\u0026thinsp;=\u0026thinsp;11.59, p\u0026thinsp;\u0026lt;\u0026thinsp;.001) were also strongly associated with relapse. These findings align with global research emphasizing the role of genetic predisposition and learned behaviors in relapse risk (Hendershot et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Andersson et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). In Zimbabwe, the high prevalence of poly-substance use may be driven by socio-economic hardship and the limited availability of professional rehabilitation services, leading adolescents to self-medicate or experiment with multiple substances (Muswerakuenda et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The sensitivity analysis revealed that these predictors are particularly vulnerable to confounding by parental socio-economic status and familial stress, suggesting the need for more detailed household-level data in future studies. Finally, religion emerged as a significant high-risk factor, with African traditional religion (HR\u0026thinsp;=\u0026thinsp;2.34, p\u0026thinsp;=\u0026thinsp;.005) and Christianity (HR\u0026thinsp;=\u0026thinsp;1.57, p\u0026thinsp;=\u0026thinsp;.012) associated with increased relapse risk. This contrasts with findings from studies in Mexico and South Africa, where faith-based interventions were protective against relapse (Garc\u0026iacute;a-Pacheco et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Davis et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). In Zimbabwe, religious practices often center on spiritual healing rather than evidence-based interventions, which may delay access to professional care (Muswerakuenda et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Furthermore, the robust association between African traditional religion and relapse highlights the need to address potential stigmatization and reliance on traditional remedies that may lack therapeutic efficacy.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eModerate-Risk Predictors\u003c/h2\u003e \u003cp\u003eModerate-risk predictors included age, secondary education, mental health history, frequency of substance use, and socioeconomic status. Older adolescents (HR\u0026thinsp;=\u0026thinsp;2.87, p\u0026thinsp;\u0026lt;\u0026thinsp;.001) faced a higher relapse risk, consistent with findings from Dawson et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), who attributed this trend to the cumulative impact of prolonged substance use. However, the Zimbabwean context diverges from studies in the U.S., where younger adolescents were identified as more vulnerable (Chung \u0026amp; Maisto, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). This difference may reflect Zimbabwe\u0026rsquo;s limited early intervention programs, which delay treatment until substance use becomes severe. Sensitivity analysis suggested that age interacts with other variables, particularly mental health history and socioeconomic status, underscoring the importance of targeted interventions for older adolescents. Secondary education (HR\u0026thinsp;=\u0026thinsp;2.62, p\u0026thinsp;\u0026lt;\u0026thinsp;.001) emerged as a moderate-risk factor, with adolescents in this category more prone to relapse than those with primary education. This finding contrasts with global trends, where higher education is typically protective against relapse (Fernandes \u0026amp; Mokwena, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In Zimbabwe, however, adolescents in secondary school may face increased academic pressure, exposure to peer influence, and limited access to mental health resources within the school system (Kurevakwesu et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These findings highlight the potential for educational settings to serve as intervention points, provided that school-based counseling services are strengthened, emphasizing the significant impact this action could have on reducing substance use relapse.\u003c/p\u003e \u003cp\u003eMental health history (HR\u0026thinsp;=\u0026thinsp;6.36, p\u0026thinsp;\u0026lt;\u0026thinsp;.001) was a significant moderate-risk factor, consistent with international studies linking co-occurring disorders to higher relapse rates (Smyth et al., 2019). Zimbabwe\u0026rsquo;s reliance on acute symptom management rather than integrated care for dual diagnoses may exacerbate this risk (Matanga et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The sensitivity analysis indicated that this predictor\u0026rsquo;s impact is highly robust but may interact with socioeconomic status, as adolescents from low-income families are less likely to access mental health care. This underscores the urgent need for integrated rehabilitation programs that address both psychosis and underlying mental health disorders. Frequency of substance use (HR\u0026thinsp;=\u0026thinsp;1.99, p\u0026thinsp;\u0026lt;\u0026thinsp;.001) and socioeconomic status (HR\u0026thinsp;=\u0026thinsp;2.23, p\u0026thinsp;=\u0026thinsp;.021) were also significant moderate-risk predictors. Frequent substance use aligns with global findings, where higher consumption levels increase physiological dependence and craving (Dawson et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). In Zimbabwe, the affordability and accessibility of illicit substances, particularly in urban areas, may amplify this risk (Kurevakwesu et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Socioeconomic status, while significant, was more variable in the sensitivity analysis, suggesting that its impact may be mediated by access to rehabilitation services and family support. Future research could explore the interaction between socioeconomic status and treatment duration to better understand this dynamic.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eLow-Risk Predictors\u003c/h2\u003e \u003cp\u003eLow-risk predictors\u0026mdash;rural residence, tertiary education, treatment period, emotional well-being, and family size\u0026mdash;had minimal impact on relapse risk. Rural residence (HR\u0026thinsp;=\u0026thinsp;0.84, p\u0026thinsp;=\u0026thinsp;.087) was associated with a slight, non-significant reduction in relapse risk, aligning with studies from South Africa and Mexico that identified rural adolescents as less exposed to substance triggers but more disadvantaged in accessing care (Davis et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; P\u0026eacute;rez-Rubio et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). In Zimbabwe, the protective effect of rural residence may reflect limited access to substances, but this advantage is likely offset by the scarcity of rehabilitation services in rural areas (Maseko, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Tertiary education (HR\u0026thinsp;=\u0026thinsp;1.49, p\u0026thinsp;=\u0026thinsp;.057) showed a marginally elevated relapse risk, diverging from global trends where higher education is protective (Fernandes \u0026amp; Mokwena, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). This anomaly may reflect Zimbabwe\u0026rsquo;s socio-economic instability, where even educated adolescents face unemployment and limited opportunities, increasing their susceptibility to relapse (Muswerakuenda et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). More extended treatment periods (HR\u0026thinsp;=\u0026thinsp;0.66, p\u0026thinsp;\u0026lt;\u0026thinsp;.001) and higher emotional well-being (HR\u0026thinsp;=\u0026thinsp;0.34, p\u0026thinsp;=\u0026thinsp;.004) significantly reduced relapse risk, consistent with studies emphasizing the importance of extended rehabilitation and psychosocial support (Mokwena \u0026amp; Fernandes, 2018). However, in Zimbabwe, the overall effectiveness of these factors is constrained by inconsistent service delivery and the lack of systematic follow-up care (Matanga et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The sensitivity analysis revealed that these predictors are susceptible to residual confounding, particularly from unmeasured variables such as social support and coping mechanisms. Family size (HR\u0026thinsp;=\u0026thinsp;0.05, p\u0026thinsp;=\u0026thinsp;.421) had no significant impact on relapse, aligning with findings from Nigeria highlighting the importance of family cohesion rather than size (Sanni et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In Zimbabwe, the minimal influence of family size may reflect the broader socio-economic challenges families face, where limited resources dilute the potential protective effect of more extensive support networks (Kurevakwesu et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eImplication for policy and programming\u003c/h2\u003e \u003cp\u003eThe findings from this study underscore the urgent need for targeted and evidence-based interventions to address substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. Identifying high, moderate, and low-risk predictors provides a framework for policymakers and program implementers to prioritize resources and develop tailored strategies beyond routine interventions, ensuring improved adolescent well-being.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003e\u0026bull; Gender-responsive Policies and Programs\u003c/h2\u003e \u003cp\u003eThe high relapse risk among male adolescents (HR\u0026thinsp;=\u0026thinsp;3.64) calls for gender-sensitive interventions. Policy initiatives should focus on creating male-specific rehabilitation programs that address risk-taking behaviors, peer influence, and substance availability. Community-based outreach programs targeting young males in urban areas could integrate mentorship, peer education, and vocational training to provide alternatives to substance use. For females, interventions should address unique stressors such as stigma and domestic challenges, which may also influence relapse risk indirectly.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e\u0026bull; Urban-focused Interventions\u003c/h2\u003e \u003cp\u003eThe exceptionally high relapse risk among urban adolescents (HR\u0026thinsp;=\u0026thinsp;16.16) necessitates targeted urban programs. Policymakers should establish urban rehabilitation centers with specialized services addressing environmental triggers such as drug availability and peer pressure. Moreover, urban schools should be equipped with counseling services and peer support groups to reduce substance use initiation and relapse. Urban zoning regulations could also focus on restricting access to high-risk locations, such as unregulated alcohol outlets and drug hotspots.\u003c/p\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003e\u0026bull; Integration of Family-based Interventions\u003c/h2\u003e \u003cp\u003eGiven the significant influence of a family history of substance use (HR\u0026thinsp;=\u0026thinsp;12.44), family-based interventions must be prioritized. Programs should include psychoeducation and counseling for families, emphasizing the genetic and environmental risks associated with substance use. Family therapy could strengthen parental involvement and create supportive home environments that reduce relapse risks. Policies should also consider subsidies or incentives for families to participate in such programs, particularly for low-income households.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e \u003ch2\u003e\u0026bull; Expansion of Psychosocial Support Services\u003c/h2\u003e \u003cp\u003eThe findings on mental health history (HR\u0026thinsp;=\u0026thinsp;6.36) and emotional well-being (HR\u0026thinsp;=\u0026thinsp;0.34) emphasize the need to integrate mental health services into rehabilitation programs. Policymakers should prioritize funding for training mental health professionals, improving access to therapy, and establishing adolescent-focused mental health units. Psychosocial support services should include cognitive-behavioral therapy (CBT) and trauma-informed care to address underlying mental health conditions contributing to relapse.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section3\"\u003e \u003ch2\u003e\u0026bull; Socioeconomic Empowerment\u003c/h2\u003e \u003cp\u003eThe role of socioeconomic status (HR\u0026thinsp;=\u0026thinsp;2.23) highlights the importance of addressing systemic inequalities. Policymakers should invest in programs that provide financial support to low-income families, such as school fee waivers, vocational training, and employment initiatives. Furthermore, programming should connect adolescents to social enterprises or community projects that promote economic independence and reduce substance use triggers.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003e\u0026bull; Lengthening Treatment Periods and Aftercare\u003c/h2\u003e \u003cp\u003eThe protective effect of longer treatment periods (HR\u0026thinsp;=\u0026thinsp;0.66) underscores the importance of extending rehabilitation durations. Policies should mandate minimum treatment periods and establish structured aftercare programs that include regular follow-ups, relapse prevention counseling, and community reintegration support.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003e\u0026bull; Strengthening existing systems\u003c/h2\u003e \u003cp\u003eFor moderate-risk factors such as age, secondary education, mental health history, frequency of substance use, and socioeconomic status, interventions should focus on early identification and prevention. School-based programs can address peer influence and academic stress, while integrated mental health services should target adolescents with co-occurring disorders. Economic empowerment initiatives, such as vocational training, can mitigate socioeconomic disparities. Strengthening existing systems is key for low-risk factors like rural residence, tertiary education, treatment period, emotional well-being, and family size. Expanding access to rural rehabilitation services, extending treatment durations, and enhancing psychosocial support can sustain recovery. Family counseling and community-based support complement these efforts to maximize resilience.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eLimitations\u003c/h3\u003e\n\u003cp\u003eThe findings of this study significantly advance the field of rehabilitation for substance-induced psychosis. The robustness of high-risk predictors was reinforced through sensitivity analysis, and the large sample size enhances the reliability and generalizability of the results. These strengths facilitate a comprehensive understanding of the demographic factors influencing relapse in substance-induced psychosis among adolescents. However, the model accounted for only 45% of the variability in relapse risk (Nagelkerke's R\u0026sup2; = 0.45), suggesting that numerous other factors not included in the analysis play a crucial role in relapse. These factors encompass a lack of structured aftercare services, the stigma surrounding substance use, cultural practices, socio-economic constraints, and inadequate drug management systems (Kurevakwesu et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Muswerakuenda et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Additionally, key demographic factors such as treatment type and biological or genetic predispositions were excluded despite their potential interactions with the studied elements, thus narrowing the scope of the findings. Interventions based solely on demographics cannot be devised in isolation; they must be integrated with broader systemic and individual considerations. Future studies could enhance these findings by employing advanced statistical methods, such as structural equation modeling and Kaplan-Meier survival analysis, to investigate complex relationships and temporal patterns in relapse. Despite these limitations, the study utilized robust statistical techniques to supplement existing qualitative and quantitative research, offering valuable insights for targeted policy and programming interventions.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study underscores the intricate factors influencing substance-induced psychosis rehabilitation relapse among adolescents in Zimbabwe. While demographic predictors are significant, they account for only a portion of the variability observed. Key risk factors, including gender, urban residence, and family history, are closely linked to environmental and systemic influences, highlighting the necessity for context-specific and integrative interventions. The moderate effects of education and mental health history indicate that a combination of personal vulnerabilities and access to support systems significantly shapes relapse rates. Conversely, the low-risk predictors remind us that broader structural elements can indirectly sustain recovery. Moreover, the study stresses that addressing relapse requires more than demographic predictors; unexamined factors such as aftercare, stigma, and cultural dynamics likely play crucial roles. This points to holistic approaches that fuse demographic insights with systemic reforms. Although the study's limitations reveal gaps in the model's explanatory power, the findings lay a foundation for evidence-based policymaking, providing valuable lessons for designing targeted and inclusive programs that enhance adolescent well-being and recovery outcomes.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe researcher extends their appreciation to Medical Research Council of Zimbabwe for ethical approval. We also want to thank the study participants for their time during data collection and subsequent participation in the moderation of the findings.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatement Regarding Informed Consent\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDuring the second data collection phase, researchers actively sought informed consent from adolescents and their guardians to verify and supplement demographic information. Consent forms clearly outlined the study\u0026apos;s purpose. All participants and their guardians consented by signing the consent form. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Publish declaration:\u0026nbsp;\u003c/strong\u003eAll participants consented for their anonymized data to be published.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatement Regarding Ethical Approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study strictly adhered to ethical standards, receiving approval from the Medical Research Council of Zimbabwe on 19 September 2018 (Reference: MRCZ-0187/A/0234).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatement Regarding Research Involving Human Participants and/or Animals\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRespondents were fully informed about the purposes of the research and how their responses would be used and stored.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors did not receive support from any organization for the submitted work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor\u0026apos;s Contribution\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization, Methodology, Formal analysis and investigation, Writing - original draft preparation, Writing - review and editing, Resources and Supervision.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no competing interests to declare that are relevant to the content of this article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe sharing of the original data for this study is restricted by ethical approval. However, anonymised data can be requested from the corresponding author.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAkosile, W., Bor, W., Tilse, A., Hunt, G., Rushton, A., McDermott, B., ... \u0026amp; McBride, M. (2024). Predictors of completion of an adolescent residential alcohol and other drug withdrawal program. \u003cem\u003eJournal of Substance Use\u003c/em\u003e, \u003cem\u003e29\u003c/em\u003e(2), 327-333. http://dx.doi.org/10.1080/14659891.2023.2166608 \u003c/li\u003e\n\u003cli\u003eAmadu, P. M., Hoedoafia, R. E., Abem, V. K., Yakubu, N. D., Imoro, M., Sherif, A. M., \u0026amp; Razak, A. R. A. (2024). Substance abuse among the Adolescent of Northern Region: The Realities of Our Time. \u003cem\u003eJournal of Psychiatry and Psychiatric Disorders\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e, 15-23. https://www.researchgate.net/publication/379652760\u003c/li\u003e\n\u003cli\u003eAndersson, H. W., Wenaas, M., \u0026amp; Nordfj\u0026aelig;rn, T. (2019). Relapse after inpatient substance use treatment: A prospective cohort study among users of illicit substances. \u003cem\u003eAddictive Behaviors\u003c/em\u003e, \u003cem\u003e90\u003c/em\u003e, 222-228. https://doi.org/10.1016/j.addbeh.2018.11.008 \u003c/li\u003e\n\u003cli\u003eArmoon, B., Griffiths, M. D., \u0026amp; Mohammadi, R. (2023). The global distribution and epidemiology of psychoactive substance use and injection drug use among street-involved children and adolescent: a meta-analysis. \u003cem\u003eSubstance Use \u0026amp; Misuse\u003c/em\u003e, \u003cem\u003e58\u003c/em\u003e(6), 746-764. https://doi.org/10.1080/10826084.2023.2181036\u003c/li\u003e\n\u003cli\u003eBanks, D. E., \u0026amp; Zapolski, T. C. (2018). The crossover effect: A review of racial/ethnic variations in risk for substance use and substance use disorder across development. \u003cem\u003eCurrent Addiction Reports\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e, 386-395. https://doi.org/10.1007/s40429-018-0220-0 \u003c/li\u003e\n\u003cli\u003eBeckmann, D., Lowman, K. L., Nargiso, J., McKowen, J., Watt, L., \u0026amp; Yule, A. M. (2019). Substance-induced psychosis in youth. \u003cem\u003eChild and Adolescent Psychiatric Clinics of North America\u003c/em\u003e, \u003cem\u003e29\u003c/em\u003e(1), 131. https://doi.org/10.1016/j.chc.2019.08.006 \u003c/li\u003e\n\u003cli\u003eBingham, B., Moniruzzaman, A., Patterson, M., Distasio, J., Sareen, J., O\u0026rsquo;Neil, J., \u0026amp; Somers, J. M. (2019). Indigenous and non-Indigenous people experiencing homelessness and mental illness in two Canadian cities: A retrospective analysis and implications for culturally informed action. \u003cem\u003eBMJ open\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(4), e024748. https://doi.org/10.1136/bmjopen-2018-024748 \u003c/li\u003e\n\u003cli\u003eChung, T., \u0026amp; Maisto, S. A. (2006). Relapse to alcohol and other drug use in treated adolescents: Review and reconsideration of relapse as a change point in clinical course. \u003cem\u003eClinical Psychology Review\u003c/em\u003e, \u003cem\u003e26\u003c/em\u003e(2), 149-161. https://psycnet.apa.org/doi/10.1016/j.cpr.2005.11.004 \u003c/li\u003e\n\u003cli\u003eCoelho, V. A. A., Andrade, L. I., Guimar\u0026atilde;es, D. A., Pereira, L. S. M., Modena, C. M., Guimar\u0026atilde;es, E. A. D. A., \u0026amp; Gama, C. A. P. D. (2022). Regionalization of psychosocial care: a panoramic view of the Psychosocial Care Network of Minas Gerais state, Brazil. \u003cem\u003eCi\u0026ecirc;ncia \u0026amp; Sa\u0026uacute;de Coletiva\u003c/em\u003e, \u003cem\u003e27\u003c/em\u003e, 1895-1909. https://doi.org/10.1590/1413-81232022275.11212021 \u003c/li\u003e\n\u003cli\u003eCrockford, D., \u0026amp; Addington, D. (2017). Canadian schizophrenia guidelines: schizophrenia and other psychotic disorders with coexisting substance use disorders. \u003cem\u003eThe Canadian Journal of Psychiatry\u003c/em\u003e, \u003cem\u003e62\u003c/em\u003e(9), 624-634. https://doi.org/10.1177/0706743717720196 \u003c/li\u003e\n\u003cli\u003eDavis, G. P., Tomita, A., Mtshemla, S., Nene, S., King, H., Susser, E., ... \u0026amp; Baumgartner, J. N. (2016). Substance use and duration of untreated psychosis in KwaZulu-Natal, South Africa. \u003cem\u003eSouth African Journal of Psychiatry\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(1), 1-7. https://doi.org/10.4102/sajpsychiatry.v22i1.852 \u003c/li\u003e\n\u003cli\u003eDawson, D. A., Grant, B. F., Stinson, F. S., Chou, P. S., Huang, B., \u0026amp; Ruan, W. J. (2005). Recovery from DSM‐IV alcohol dependence: United States, 2001\u0026ndash;2002. \u003cem\u003eAddiction\u003c/em\u003e, \u003cem\u003e100\u003c/em\u003e(3), 281-292. https://pmc.ncbi.nlm.nih.gov/articles/PMC6527249/ \u003c/li\u003e\n\u003cli\u003eDegenhardt, L., \u0026amp; Hall, W. (2012). Extent of illicit drug use and dependence, and their contribution to the global burden of disease. \u003cem\u003eThe Lancet\u003c/em\u003e, \u003cem\u003e379\u003c/em\u003e(9810), 55-70. https://doi.org/10.1016/s0140-6736(11)61138-0 \u003c/li\u003e\n\u003cli\u003eFernandes, L., \u0026amp; Mokwena, K. E. (2020). Nyaope addiction: The despair of a lost generation. \u003cem\u003eAfrican Journal of Drug and Alcohol Studies\u003c/em\u003e, \u003cem\u003e19\u003c/em\u003e(1), 37-51. https://www.ajol.info/index.php/ajdas/article/view/203861 \u003c/li\u003e\n\u003cli\u003eFiorentini, A., Cant\u0026ugrave;, F., Crisanti, C., Cereda, G., Oldani, L., \u0026amp; Brambilla, P. (2021). Substance-induced psychoses: an updated literature review. \u003cem\u003eFrontiers in psychiatry\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e, 694863. https://doi.org/10.3389/fpsyt.2021.694863 \u003c/li\u003e\n\u003cli\u003eGarc\u0026iacute;a-Pacheco, J. \u0026Aacute;., Ortega, M. D. L. T., \u0026amp; Borges, G. (2024). The burden of mental disorders in Mexico, 1990-2019: mental and neurological disorders, substance use, suicides, and related somatic disorders. \u003cem\u003eSpanish Journal of Psychiatry and Mental Health\u003c/em\u003e, \u003cem\u003e17\u003c/em\u003e(2), 64-70. https://doi.org/10.1016/j.rpsm.2021.06.005 \u003c/li\u003e\n\u003cli\u003eGilman, S. E., Kawachi, I., Fitzmaurice, G. M., \u0026amp; Buka, S. L. (2003). Socio-economic status, family disruption and residential stability in childhood: relation to onset, recurrence and remission of major depression. \u003cem\u003ePsychological medicine\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e(8), 1341-1355. https://doi.org/10.1017/s0033291703008377 \u003c/li\u003e\n\u003cli\u003eGullacher, C. A., \u0026amp; Goernert, P. N. (2025). The use of community treatment orders in people with substance induced psychosis. \u003cem\u003eInternational Journal of Law and Psychiatry\u003c/em\u003e, \u003cem\u003e98\u003c/em\u003e, 102043. https://doi.org/10.1016/j.ijlp.2024.102043 \u003c/li\u003e\n\u003cli\u003eHendershot, C. S., Witkiewitz, K., George, W. H., \u0026amp; Marlatt, G. A. (2011). Relapse prevention for addictive behaviors. \u003cem\u003eSubstance abuse treatment, prevention, and policy\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e, 1-17.\u003c/li\u003e\n\u003cli\u003eHjorth\u0026oslash;j, C., Madsen, T., Starzer, M., Erlangsen, A., \u0026amp; Nordentoft, M. (2021). Mortality in substance‐induced psychosis: A register‐based national cohort study. \u003cem\u003eAddiction\u003c/em\u003e, \u003cem\u003e116\u003c/em\u003e(12), 3515-3524. https://doi.org/10.1111/add.15598 \u003c/li\u003e\n\u003cli\u003eJanson, S., Nyenga, L., Saleem, H., Mayo-Wilson, L. J., Mushy, S. E., Iseselo, M. K., ... \u0026amp; Conserve, D. F. (2024). Residential and inpatient treatment of substance use disorders in Sub-Saharan Africa: a scoping review. \u003cem\u003eSubstance abuse treatment, prevention, and policy\u003c/em\u003e, \u003cem\u003e19\u003c/em\u003e(1), 6.\u003c/li\u003e\n\u003cli\u003eKokota, D., Stewart, R. C., Bandawe, C., Chorwe-Sungani, G., Liwimbi, O., Mwale, C. M., ... \u0026amp; Lawrie, S. M. (2023). Pathways to care for psychosis in Malawi. \u003cem\u003eBJPsych International\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(4), 84-89. https://doi.org/10.1192/bji.2023.18 \u003c/li\u003e\n\u003cli\u003eKurevakwesu, W., Masona, G., Mabeza, T.T., Zengeni, F. \u0026amp; Chiweshe, C.C. (2023). Factors responsible for substance use among youths in urban settings and social work-related interventions. \u003cem\u003eJournal of Social Work Practice in the Addictions, 0(00), 1-18\u003c/em\u003e. https://doi.org/10.1080/1533256X.2023.2212475 \u003c/li\u003e\n\u003cli\u003eLopes-Rosa, R., Kessler, F. P., Pianca, T. G., Guimar\u0026atilde;es, L., Ferronato, P., Pagnussat, E., ... \u0026amp; von Diemen, L. (2017). Predictors of early relapse among adolescent crack users. \u003cem\u003eJournal of addictive diseases\u003c/em\u003e, \u003cem\u003e36\u003c/em\u003e(2), 136-143. https://doi.org/10.1080/10550887.2017.1295670 \u003c/li\u003e\n\u003cli\u003eManhica, H., Straatmann, V. S., Lundin, A., Agardh, E., \u0026amp; Danielsson, A. K. (2021). Association between poverty exposure during childhood and adolescence, and drug use disorders and drug‐related crimes later in life. \u003cem\u003eAddiction\u003c/em\u003e, \u003cem\u003e116\u003c/em\u003e(7), 1747-1756. https://doi.org/10.1111/add.15336 \u003c/li\u003e\n\u003cli\u003eMaseko, W. F. (2023). How Adolescent Crime is Affected by Drug abuse and Mental Heath\u0026ndash;A Case Study of Kawale Policing Area. \u003cem\u003eInternational Journal of Modern Developments in Engineering and Science\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(3), 11-26.\u003c/li\u003e\n\u003cli\u003eMatanga, A.A., Chiparausha, M., Kurevakwesu, W. \u0026amp; Charamba, S. (2024). Challenges for reintegrated youths recovering from substance use disorders in Harare, Zimbabwe: A phenomenological study. \u003cem\u003eJournal of Social Development in Africa\u003c/em\u003e, 39(1), 18-44. https://dx.doi.org/10.4314/jsda.v39i1.3 \u003c/li\u003e\n\u003cli\u003eMolina, K. M., Alegr\u0026iacute;a, M., \u0026amp; Chen, C. N. (2012). Neighborhood context and substance use disorders: A comparative analysis of racial and ethnic groups in the United States. \u003cem\u003eDrug and alcohol dependence\u003c/em\u003e, \u003cem\u003e125\u003c/em\u003e, S35-S43.\u003c/li\u003e\n\u003cli\u003eMuswerakuenda, F. F., Mundagowa, P. T., Madziwa, C., \u0026amp; Mukora-Mutseyekwa, F. (2023). Access to psychosocial support for church-going young people recovering from drug and substance abuse in Zimbabwe: a qualitative study. \u003cem\u003eBMC public health\u003c/em\u003e, \u003cem\u003e23\u003c/em\u003e(1), 723.\u003c/li\u003e\n\u003cli\u003eMwamba, C. C. (2023). \u003cem\u003ePrevalence and co-occurrence of substance use in mental health patients at Lewanika general hospital\u003c/em\u003e. PhD Thesis: University of Zambia.\u003c/li\u003e\n\u003cli\u003eNguata, M., Orwa, J., Kigen, G., Kamaru, E., Emonyi, W., Kariuki, S., ... \u0026amp; Atwoli, L. (2024). Association between psychosis and substance use in Kenya. Findings from the NeuroGAP-Psychosis study. \u003cem\u003eFrontiers in Psychiatry\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e, 1301976.\u003c/li\u003e\n\u003cli\u003eNhapi, T. (2019). Drug addiction among adolescents in Zimbabwe: social work perspective. \u003cem\u003eAddiction in South and East Africa: Interdisciplinary Approaches\u003c/em\u003e, 241-259. http://dx.doi.org/10.1007/978-3-030-13593-5_15 \u003c/li\u003e\n\u003cli\u003eOkon, E. E., Etim, B. I., \u0026amp; Etim, B. E. (2024). \u003cem\u003eDrug Addiction and Health of Adolescents in Calabar Metropolis of Cross River State, Nigeria.\u003c/em\u003e Retrieved June 10. 2024 from: http://dx.doi.org/10.36349/easjnm.2024.v06i01.00X \u003c/li\u003e\n\u003cli\u003eOtlhapile, O. E., Gitau, C. W., \u0026amp; Kuria, M. W. (2023). The prevalence of substance use disorders and associated characteristics among patients admitted to a psychiatric hospital in Botswana. \u003cem\u003eThe International Journal of Psychiatry in Medicine\u003c/em\u003e, \u003cem\u003e58\u003c/em\u003e(4), 339-357. https://doi.org/10.1177/00912174231156027 \u003c/li\u003e\n\u003cli\u003ePalma-\u0026Aacute;lvarez, R. F., Grau-L\u0026oacute;pez, L., Ros-Cucurull, E., Abad, A. C., Dualde, J., Robles-Mart\u0026iacute;nez, M., \u0026amp; Roncero, C. (2021). Psychosis induced by abuse of ayahuasca: a case report. \u003cem\u003eRevista Colombiana de Psiquiatr\u0026iacute;a\u003c/em\u003e, \u003cem\u003e50\u003c/em\u003e(1), 43-46. https://doi.org/10.1016/j.rcp.2019.10.005 \u003c/li\u003e\n\u003cli\u003eP\u0026eacute;rez-Rubio, G., L\u0026oacute;pez-Flores, L. A., Garc\u0026iacute;a-Carmona, S., Garc\u0026iacute;a-G\u0026oacute;mez, L., No\u0026eacute;-D\u0026iacute;az, V., Ambrocio-Ortiz, E., ... \u0026amp; Falf\u0026aacute;n-Valencia, R. (2019). Genetic variants as risk factors for cigarette smoking at an early age and relapse to smoking cessation treatment: A pilot study. \u003cem\u003eGene\u003c/em\u003e, \u003cem\u003e694\u003c/em\u003e, 93-96.\u003c/li\u003e\n\u003cli\u003eSanni, M. M., Bolu-Steve, F. N., Durosaro, I. A., \u0026amp; Adigun, A. A. (2021). Prevalence of Drug Relapse among Clients in Rehabilitation Centres in North Central Nigeria: Implications for School Counsellors. \u003cem\u003eCanadian Journal of Family and Youth/Le Journal Canadien de Famille et de la Jeunesse\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(2), 14-25. http://dx.doi.org/10.29173/cjfy29668 \u003c/li\u003e\n\u003cli\u003eSharifi, H., Kharaghani, R., Sigari, S., Aryanpur, M., \u0026amp; Masjedi, M. R. (2011). The effects of demographic factors and cigarette smoking status on drug treatment success rate in outpatient treatment and rehabilitation centers. \u003cem\u003eArchives of Iranian medicine\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(3), 183-187.\u003c/li\u003e\n\u003cli\u003eSinha, R. (2024). Stress and substance use disorders: risk, relapse, and treatment outcomes. \u003cem\u003eThe Journal of Clinical Investigation\u003c/em\u003e, \u003cem\u003e134\u003c/em\u003e(16).\u003c/li\u003e\n\u003cli\u003eSmyth, N., Buckman, J. E., Naqvi, S. A., Aguirre, E., Cardoso, A., Pilling, S., \u0026amp; Saunders, R. (2022). Understanding differences in mental health service use by men: an intersectional analysis of routine data. \u003cem\u003eSocial psychiatry and psychiatric epidemiology\u003c/em\u003e, \u003cem\u003e57\u003c/em\u003e(10), 2065-2077.\u003c/li\u003e\n\u003cli\u003eTaha, M., Taalab, Y. M., Abo-Elez, W. F., \u0026amp; Eldakroory, S. A. (2019). Cannabis and tramadol are prevalent among the first episode drug-induced psychosis in the Egyptian Population: Single center experience. \u003cem\u003eReports\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(2), 16. https://doi.org/10.3390/reports2020016 \u003c/li\u003e\n\u003cli\u003eTemmingh, H. S., Mall, S., Howells, F. M., Sibeko, G., \u0026amp; Stein, D. J. (2020). The prevalence and clinical correlates of substance use disorders in patients with psychotic disorders from an Upper-Middle-Income Country. \u003cem\u003eSouth African journal of psychiatry\u003c/em\u003e, \u003cem\u003e26\u003c/em\u003e. https://doi.org/10.4102/sajpsychiatry.v26i0.1473 \u003c/li\u003e\n\u003cli\u003eThomasius, R., Paschke, K., \u0026amp; Arnaud, N. (2022). Substance-use disorders in children and adolescents. \u003cem\u003eDeutsches \u0026Auml;rzteblatt International\u003c/em\u003e, \u003cem\u003e119\u003c/em\u003e(25), 440. https://doi.org/10.3238/arztebl.m2022.0122 \u003c/li\u003e\n\u003cli\u003eVivolo, M., Owen, J., \u0026amp; Fisher, P. (2024). Building resilience in the Improving Access to Psychological Therapy (IAPT) Psychological Wellbeing Practitioner (PWP) role: a qualitative grounded theory study. \u003cem\u003eBehavioural and Cognitive Psychotherapy\u003c/em\u003e, \u003cem\u003e52\u003c/em\u003e(2), 135-148. https://doi.org/10.1017/S1352465823000334 \u003c/li\u003e\n\u003cli\u003eWangithi, I. K., \u0026amp; Ndurumo, M. M. (2020). Relationship between family support, Self-Efficacy and relapse occurrence among youth recovering from drug addiction in selected rehabilitation centres in limuru Sub-County, Kenya. \u003cem\u003eAfrican Journal of Education, Science and Technology\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e(1), 134-148. https://doi.org/10.2022/ajest.v6i1.471 \u003c/li\u003e\n\u003cli\u003eWHO (2024). Mental health of adolescents. Accessed January 19, 2025 from: https://www.who.int/news-room/fact-sheets/detail/adolescent-mental-health#\u003cstrong\u003e\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Substance-induced psychosis, Rehabilitation relapse, Adolescents, Demographic risk factors, Proportional hazards modeling","lastPublishedDoi":"10.21203/rs.3.rs-6484433/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6484433/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSubstance-induced psychosis rehabilitation relapse among adolescents remains a significant public health challenge in Zimbabwe. This study aimed to identify and evaluate demographic risk factors influencing relapse using a survival analysis approach. A retrospective cohort design was applied, involving 3,135 adolescents aged 12\u0026ndash;17 admitted to four national psychiatric hospitals between 2019 and 2023. A Cox proportional hazards regression model was used to analyze time-to-relapse, with hazard ratios (HRs) and 95% confidence intervals (CIs) quantifying the effects of demographic predictors. Sensitivity analysis was conducted using the Impact Threshold for a Confounding Variable (ITCV) and Robustness Index Ratio (RIR) to assess the reliability of findings. From the findings, high-risk predictors included male gender (HR\u0026thinsp;=\u0026thinsp;3.64, 95% CI [2.50, 4.82]), urban residence (HR\u0026thinsp;=\u0026thinsp;16.16, 95% CI [9.96, 26.22]), family history of substance use (HR\u0026thinsp;=\u0026thinsp;12.44, 95% CI [9.36, 17.55]), and poly-substance use (HR\u0026thinsp;=\u0026thinsp;11.59, 95% CI [10.11, 12.27]). Moderate-risk factors included age (HR\u0026thinsp;=\u0026thinsp;2.87, 95% CI [1.82, 3.93]), secondary education (HR\u0026thinsp;=\u0026thinsp;2.62, 95% CI [1.50, 5.75]), and mental health history (HR\u0026thinsp;=\u0026thinsp;6.36, 95% CI [4.29, 9.45]). Low-risk factors such as rural residence and treatment duration demonstrated limited protective effects. The model explained 45% of relapse variability (Nagelkerke\u0026rsquo;s R\u0026sup2; = 0.45). While demographic predictors provide valuable insights, relapse risk is determined by a complex interplay of demographic, systemic, and contextual factors. These findings inform targeted policies and programming to address adolescent substance use and improve rehabilitation outcomes.\u003c/p\u003e","manuscriptTitle":"Demographic Risk Factors for Substance-Induced psychosis Rehabilitation Relapse Among Adolescents in Zimbabwe: A Proportional Hazards Modeling Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-13 16:30:54","doi":"10.21203/rs.3.rs-6484433/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9530d753-9f0a-4b23-9954-9c09eba624a1","owner":[],"postedDate":"May 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-01T06:54:12+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-13 16:30:54","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6484433","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6484433","identity":"rs-6484433","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}