Predictive Modelling of Linear Growth Faltering Among Pediatric Patients with Diarrhea in Rural Western Kenya: An Explainable Machine Learning Approach

Predictive Modelling of Linear Growth Faltering Among Pediatric Patients with Diarrhea in Rural Western Kenya: An Explainable Machine Learning Approach | 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 Predictive Modelling of Linear Growth Faltering Among Pediatric Patients with Diarrhea in Rural Western Kenya: An Explainable Machine Learning Approach Billy Ogwel, Vincent H. Mzazi, Alex O. Awuor, Caleb Okonji, Raphael O. Anyango, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4047381/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 02 Dec, 2024 Read the published version in BMC Medical Informatics and Decision Making → Version 1 posted 10 You are reading this latest preprint version Abstract Introduction: Stunting affects one-fifth of children globally with diarrhea accounting for an estimated 13.5% of stunting. Identifying risk factors for its precursor, linear growth faltering (LGF), is critical to designing interventions. Moreover, developing new predictive models for LGF using more recent data offers opportunity to improve model performance and capture new insights. We employed machine learning (ML) to derive and validate a predictive model for LGF among children enrolled with diarrhea in the Vaccine Impact on Diarrhea in Africa (VIDA) study and the Enterics for Global Heath (EFGH) ― Shigella study in rural western Kenya. Methods We used 7 ML algorithms to retrospectively build prognostic models for the prediction of LGF (≥ 0.5 decrease in height/length for age z-score [HAZ]) among children 6–35 months. We used de-identified data from the VIDA study (n = 1,473) combined with synthetic data (n = 8,894) in model development, which entailed split-sampling and K-fold cross-validation with over-sampling technique, and data from EFGH-Shigella study (n = 655) for temporal validation. Potential predictors included demographic, household-level characteristics, illness history, anthropometric and clinical data chosen using an explainable model agnostic approach. The champion model was determined based on the area under the curve (AUC) metric. Results The prevalence of LGF in the development and temporal validation cohorts was 187 (16.9%) and 147 (22.4%), respectively. The following variables were associated with LGF in decreasing order: age (16.6%), temperature (6.0%), respiratory rate (4.1%), SAM (3.4%), rotavirus vaccination (3.3%), breastfeeding (3.3%), and skin turgor (2.1%). While all models showed good prediction capability, the gradient boosting model achieved the best performance (AUC% [95% Confidence Interval]: 83.5 [81.6–85.4] and 65.6 [60.8–70.4] on the development and temporal validation datasets, respectively). Conclusion Our findings accentuates the enduring relevance of established predictors of LGF whilst demonstrating the practical utility of ML algorithms for rapid identification of at-risk children. Machine Learning Linear growth faltering Pediatric Diarrhea Prediction Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Diarrhea, a global public health problem with greatest burden in low- and middle-income countries (LMICs) [ 1 ], is a leading etiology of malnutrition among children in LMICs, in part due to anorexia, decreased absorptive function, mucosal damage, catabolism and nutrient exhaustion [ 1 , 2 ]. It has been reported that the cumulative burden of diarrhea days directly correlates with the degree of nutritional failure among children during early childhood and that catch-up growth does not appear to make up for the deficit [ 3 ]. Linear growth faltering (LGF), a precursor to stunting (height-for-age z-score [HAZ] <-2), is one form of malnutrition that results from protracted nutritional deprivation [ 4 ]. Stunting affects one-fifth of children globally and one-third of children in LMIC. In the universal pie of stunting prevalence, diarrhea accounts for an estimated 13.5% [ 5 ]. Additionally, a vicious cycle of diarrhea and malnutrition can occur as malnutrition weakens the body’s defense against future diarrheal episodes resulting in more frequent and longer diarrheal illnesses. The effects of stunting can be severe and protracted, with reduced cognitive development, persistent poor health, and elevated risk of mortality [ 6 ]. Long term complications can include an increased risk of cardiovascular disease, type 2 diabetes, and obesity in adulthood [ 7 , 8 ]. The timely and accurate identification of children at increased risk of LGF is crucial for early nutritional and healthcare interventions as well as efficient allocation of public health resources, efforts that could help to avert the associated negative outcomes. Data-driven predictive models could be leveraged to this end and a number of research efforts exist in the prediction of LGF among children with diarrhea [ 9 , 10 ]. While the existing models provide a valuable starting point, shifts in the study population over time may affect the predictive performance of these models [ 11 , 12 ]. Moreover, development of new models using more recent and pertinent data offers the opportunity to improve model performance and capture new perspectives and insights into this public health problem. We used machine learning (ML), which has been adopted in public health and clinical practice to rapidly develop data-driven clinical prediction models, to develop and temporally validate predictive models for LGF among children aged < 5 years with diarrhea in rural Western Kenya. Methods Data sources This retrospective study used data collected from the Kenyan site (in Siaya County) of two related diarrheal studies: The Vaccine Impact on Diarrhea in Africa (VIDA) study for model development and evaluation; and the Enteric for Global Health (EFGH) Shigella surveillance study for temporal validation. Development cohort VIDA was designed to assess the population-based incidence, etiologies, and adverse clinical consequences of diarrhea following rotavirus vaccine introduction in children aged 0–59 months residing in censused populations in 3 African countries. Moderate-to-severe diarrhea (MSD) cases, defined as children in 3 age strata (0–11, 12–23, and 24–59 months) presenting with diarrhea (defined as ≥ 3 looser-than-normal stools within 24 hours) that began within the past 7 days after ≥ 7 diarrhea-free days and had ≥ 1 of the following: sunken eyes, poor skin turgor, dysentery, intravenous rehydration, or required hospitalization, were enrolled from sentinel health centers (SHCs) serving the health and demographic surveillance systems population at each site. The aim was to enroll 8–9 MSD cases in each age stratum per fortnight. 1–3 diarrhea-free controls matched by age, gender and geographical location were enrolled within 14 days of case enrolment. Follow-ups were conducted between 49–91 days after enrolment. We utilized data collected from cases enrolled at the VIDA Kenya site over a 36 months period from May 2015 and July 2018 restricting to children aged 6–35 months to make the development and temporal validation cohorts comparable. The study design, clinical and epidemiological methods for VIDA have been described elsewhere [ 13 , 14 ]. In addition to the VIDA data (n = 1,106), we generated a synthetic dataset (n = 8,894) based on the VIDA data using the synthpop package [ 15 ] to increase the sample size and to enable the algorithms to generate more stable and reliable predictions that are less sensitive to noise in the data. The variables of the synthetic dataset were compared to the original training dataset with the synthetic dataset demonstrating high similarity to the original dataset (Fig S1 ). The combined dataset (N = 10,000) was used for training and internal validation with a split-sampling conducted in the ratio 3:1 to partition the development data into training and test sets [ 16 ]. Temporal validation cohort The EFGH study set out to establish incidence and consequences of Shigella medically attended diarrhea (MAD) within 7 country sites in Africa, Asia, and Latin America using cross-sectional and longitudinal study designs. MAD cases defined as children aged 6–35 months presenting with diarrhea (defined as ≥ 3 looser-than-normal stools within 24 hours) that began within the past 7 days after ≥ 2 diarrhea-free days were enrolled from SHCs in the study catchment area. Additional eligibility criteria included: residing within the pre-defined study catchment area; primary caregiver and child plan to remain at their current residence for at least the next 4 months; legal guardian consenting to child’s participation in the study as well willingness to be followed-up for 3 months post-enrolment; child is not being referred to a non-EFGH facility at the time of screening; and site enrollment cap has not been met. Follow-ups were conducted at week-4 (24–67 days) and month-3 (84–127 days). Our study utilized data from children enrolled in Kenya between 01 August, 2022 and 31 July, 2023 to temporally validate the champion model. Information on demographic, socio-demographic, epidemiological and clinical characteristics were collected at enrollment by study personnel in both studies. Target variable Consistent with previous studies [ 9 , 10 ], we defined the target variable, LGF, as decrease of 0.5 HAZ or more (Δ HAZ ≥ − 0.5) within 49–91 days of enrollment in VIDA, or within 84–127 days in EFGH. We also computed change in HAZ per month of follow-up and categorized a negative change as LGF in our sensitivity analysis, similar to the definition used by Nasrin et al. [ 17 ]. We excluded children with implausible HAZ values (HAZ > 6 or 3; or length values that were > 1.5 cm lower at follow-up than at enrollment. Predictive variables and feature selection A total of 68 potential candidate predictors collected at enrollment during both studies were considered, including demographic, household-level characteristics, illness history, anthropometric and clinical characteristics collected at enrolment. Missingness patterns were assessed among the features and the missing data points imputed using the Multiple Imputation by Chained Equations (MICE) package [ 18 ]. Furthermore, we conducted feature selection to reduce dimensionality, optimize performance, reduce computational complexity and enhance model interpretability. The feature selection was implemented using the Boruta package [ 19 ] an all relevant feature selection wrapper around the random forest algorithm that selects relevant features by comparing original attributes' importance with importance achievable at random using their permuted copies. We excluded features that were rejected in this process. Moreover, among the confirmed and tentative features, we excluded variables that were not collected in both studies (breastfeeding). Statistical analysis We compared patient characteristics of children with LGF versus those without. Proportions were reported for categorical variables and either chi-square or Fisher`s exact test were performed as appropriate. Wilcoxon rank sum tests were used to compare continuous variables as appropriate. We also compared the prevalence of LGF between the 2 studies. Model development and internal validation To derive the LGF prediction model, we utilized 7 ML algorithms including: Random Forest (RF), Gradient Boosting (GBM), Naive Bayes (NB), Logistic regression (LR), Support vector machine (SVM), K-nearest neighbors (KNN) and Artificial Neural Networks (ANN). The predictive models were developed in the training dataset using 10-fold cross-validation [ 20 ], a valuable step in model development helping to obviate under-fitting or overfitting of the model and ensure robust and well-performing models. Due to the moderate class imbalance in our target variable (LGF), we employed sub-sampling techniques (over-sampling) within the resampling procedure to mitigate the negative impact of class disparity on model fitting [ 21 ]. We then conducted internal validation of the models on the test data evaluating performance using the following metrics: sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and F1-score. Receiver operating characteristic (ROC) curves were constructed and the area under the curve (AUC) and the precision-recall area under the curve (PRAUC) for each model were computed using the precrec package [ 22 ]. We determined the champion model as the model with the best AUC. We also assessed calibration in the developed models using Brier scores (the mean squared error between the actual outcome and the estimated probabilities), Spiegelhalter’s z -test (a formal measurement that serves as a proxy for calibration calculated from the decomposition of Brier score) and its accompanying p-value [ 23 ]. We used Platt scaling approach, in which model estimates are transformed by passing the estimates through a trained sigmoid function, to calibrate the champion model [ 23 ]. To enhance model interpretability, trust and fairness, we conducted explanatory model analysis (EMA) for the top two models using a model agnostic procedure to estimate SHapley Additive exPlanations (SHAPs) attributions drawing on the DALEX package [ 24 ]. Temporal Validation and Business Value Evaluation We further conducted temporal validation on the champion model to assess the robustness and generalizability of the model's performance over time [ 25 ]. To evaluate the business value of the predictive model, modelplotr package [ 26 ] was used to build valuable evaluation plots (cumulative gains, cumulative lift, response and cumulative response plots). Descriptive analysis, predictive modelling for LGF and plotting were all performed in R version 4.2.2 [ 27 ]. Results A total of 1,554 and 706 children were enrolled in the development and temporal validation cohorts, respectively. Among children aged 6–35 months enrolled, 1,106 (71.2%) and 655 (92.7%) had HAZ data that were plausible, respectively. Among those that had plausible HAZ data, 187 (16.9%) and 147 (22.4%) had LGF in the development and temporal validation cohorts, respectively (Fig. 1). Development dataset (VIDA: 2015–2018) Temporal validation dataset (EFGH: 2022–2023) VIDA- Vaccine Impact on Diarrhea in Africa Study EFGH-Enterics for Global Health Shigella Surveillance study MSD-Moderate-to-Severe Diarrhea; MAD-Medically Attended Diarrhea Figure 1. Flowchart of development and temporal validation studies conducted in Siaya County, Kenya This difference in the prevalence of LGF between the development and temporal validation cohorts was statistically significant (p = 0.0042). The median [interquartile range] ΔHAZ between enrollment and follow-up was − 0.21 [-0.42- -0.01] and − 0.24 [-0.48- -0.02] in the development and temporal validation cohorts, respectively. In the sensitivity analysis using the cut-off of negative change in HAZ, the prevalence of LGF was 1,051 (28.7%). Additionally, the constructed synthetic dataset had 8,527 observations and it closely replicated the propensity score distribution of the original development data (VIDA) as evidenced by the comprehensive descriptive analysis that compared each variable (Table S1 ). The characteristics of VIDA participants at enrolment stratified by LGF status are shown in Table 1 . Children who had LGF were younger than those who did not (Median age in months [IQR]: 11 [ 8 – 14 ] vs 17 [ 11 – 24 ], p < 0.001). Furthermore, compared with those who did not have LGF, those with LGF had a higher respiratory rate (Median [IQR]: 38.5 [34.0-42.5] vs 36.0[31.5–39.5], p < 0.001), a higher temperature (Median [IQR]: 37.1 [36.6–37.8] vs 36.8 [36.4–37.5], p < 0.001) and more severe disease (Median Vesikari score [IQR]: 11 [ 9 – 12 ] vs 10 [ 8 – 12 ], p < 0.001). Additionally, caretaker education, breastfeeding, vomiting, wrinkled skin, restless, admission, and intravenous rehydration were significantly associated with LGF (Table 1 ). Table 1 Characteristics of children aged < 5 years seeking care for moderate-to-severe diarrhea in Kenya stratified by Linear Growth Faltering Status, 2015–2018. Linear Growth Faltering Characteristics Yes (n = 187) No (n = 919) p-value* n (%) n (%) Demograhic Median age [IQR] 11 [ 8 – 14 ] 17 [ 11 – 24 ] < 0.001 Age Category 0–11 months 104 (55.6) 259 (28.2) = Secondary ) 78 (41.7) 305 (33.2) 0.026 <= 2 children under 5 yrs 167 (89.3) 839 (91.2) 0.387 <= 4 people sleeping 77 (41.2) 400 (43.6) 0.546 <= 3 Total Assets 158 (84.5) 812 (88.4) 0.142 Refined/Electric Primary Fuel Source 5 (2.7) 39 (4.3) 0.313 Animal ownership 176 (94.1) 836 (91.0) 0.159 Improved water Safely managed 83 (44.3) 431 (46.9) 0.15 Basic 14 (7.5) 112 (12.2) Limited 28 (15.0) 125 (13.6) unimproved/Surface water 62 (33.2) 251 (27.3) Improved Sanitation Safely Managed and Basic 20 (10.7) 106 (11.5) 0.392 Limited 72 (38.5) 306 (33.3) Unimproved/Open Defecation 95 (50.8) 507 (55.2) Clinical characteristics Reported by caretaker Breastfeeding before diarrhea onset None 25 (13.4) 248 (37.9) < 0.001 Exclusive 3 (1.6) 8 (0.9) Partial 159 (85.0) 563 (61.2) Median diarrhea days [IQR] 3 [ 2 – 3 ] 3 [ 2 – 4 ] 0.7196 Stool Type Simple watery 113 (60.4) 532 (57.9) 0.41 Rice watery 5 (2.7) 12 (1.3) Sticky/Mucoid 65 (34.8) 347 (37.8) Bloody 4 (2.1) 28 (3.1) Stool Count 3 27 (14.4) 165 (18.0) 0.489 4–5 101 (54.0) 506 (55.0) 6–10 55 (29.4) 228 (24.8) > 10 4 (2.1) 20 (2.2) Blood in stool 15 (8.0) 108 (11.8) 0.138 Vomiting 127(67.9) 531 (57.8) 0.01 Very Thirsty 156 (83.9) 752 (82.2) 0.582 Drinks poorly 47 (25.1) 232 (25.3) 0.962 Unable to drink 2 (1.1) 27 (2.9) 0.145 Belly Pain 109 (61.2) 508 (57.9) 0.404 Fever 142 (75.9) 709 (77.2) 0.720 Restless 151 (80.8) 710 (77.3) 0.295 Lethargy 123 (65.8) 600 (65.3) 0.898 unconscious 7 (3.7) 32 (3.5) 0.864 Rectal straining 55 (29.4) 211 (23.1) 0.066 Rectal prolapse 2 (1.1) 15 (1.6) 0.565 Cough 103 (55.1) 482 (52.5) 0.511 Difficulty breathing 32 (17.1) 124 (13.5) 0.197 Convulsion 3 (1.6) 17 (1.9) 0.818 Currently Very Thirsty 145 (78.4) 653 (71.7) 0.062 Drinks poorly 40 (21.5) 188 (20.5) 0.747 Sunken Eyes 171 (91.4) 792 (86.3) 0.054 Wrinkled skin 56 (30.6) 211 (23.0) 0.029 Restless 134 (71.7) 557 (60.6) 0.004 Lethargy/unconscious 23 (12.3) 151 (16.4) 0.157 Dry mouth 142 (75.9) 658 (71.7) 0.235 Fast breathing 24 (12.8) 100 (10.9) 0.44 Home ORS use 21 (11.2) 86 (9.4) 0.43 Home Zinc use 8 (4.3) 34 (3.7) 0.706 Assessed by Clinician Temperature [IQR] 37.1 [36.6–37.9] 36.8 [36.4–37.5] < 0.001 Measured Fever (≥ 37.5 o C) 99 (52.9) 342 (37.2) < 0.001 Median Respiratory rate [IQR] 38.5 [34.0-42.5] 36.0 [31.5–39.5] < 0.001 Chest indrawing 4 (2.1) 9 (1.0) 0.180 Sunken eyes 177 (94.7) 848 (92.3) 0.255 Dry mouth 183 (97.9) 903 (98.3) 0.71 Skin turgor (slow/very slow) 78 (41.7) 391 (42.6) 0.833 Mental Status Normal 73 (39.0) 380 (41.4) 0.052 Restless/Irritable 108 (57.8) 530 (57.7) Lethargic/Unconscious 6 (3.2) 9 (0.9) Rectal prolapse 0 (0) 3 (0.3) 0.434 Bipedal edema 2 (1.1) 5 (0.5) 0.337 Abnormal hair 9 (4.8) 43 (4.7) 0.937 Under Nutrition 21 (11.2) 109 (11.9) 0.807 Flaky Skin 2 (1.1) 5 (0.5) 0.409 Severe Acute Malnutrition (SAM) 24 (12.8) 78 (8.5) 0.061 Wasting 13 (7.0) 39 (4.2) 0.111 Admission 27 (14.4) 87 (9.5) 0.043 Diarrhea Duration (≥ 7 days) 70 (37.4) 327 (35.6) 0.631 any_antibiotic 78 (41.7) 402 (43.7) 0.609 Rotavirus vaccination doses 0 2 (1.1) 19 (2.4) 0.385 1 8 (4.6) 25 (3.1) 2 166 (94.3) 764 (94.5) ORS at facility 186 (99.5) 914 (99.9) 0.311 Zinc at facility 183 (97.9) 887 (96.9) 0.494 IV rehydration 31 (16.6) 92 (10.1) 0.01 Dehydration None 8 (4.3) 35 (3.8) 0.747 Some 126 (67.4) 645 (70.2) Severe 53 (28.3) 239 (26.0) Vesikari Score Mild 13 (7.0) 71 (7.7) 0.088 Moderate 75 (40.1) 442 (48.1) Severe 99 (52.9) 406 (44.2) Median Vesikari score [IQR] 11 [ 9 – 12 ] 10 [ 8 – 12 ] 0.0003 Diagnosis Dysentery 10 (5.4) 58 (6.4) 0.605 Malaria 85 (45.5) 361 (39.5) 0.131 Pneumonia 12 (6.4) 37 (4.1) 0.152 Bacterial Infection 14 (7.5) 93 (10.2) 0.258 Malnutrition 15 (8.0) 68 (7.4) 0.784 β− Includes electricity, propane, butane, natural gas; SAM defined as WHZ < − 3 or MUAC < 115 millimeters, or the presence of bilateral pitting edema; ORS-Oral rehydration solution *P-value computed using either chi-square or Fisher`s exact test were performed as appropriate for categorical variables and Wilcoxon rank sum tests were used to compare continuous variables From the feature selection analysis, the confirmed variables in order of importance were age (16.6%), temperature (6.0%), respiratory rate (4.1%) and breastfeeding (3.3%). SAM (3.4%), rotavirus vaccination (3.3%), and skin turgor (2.1%) were tentative features (Fig. 2). Green, yellow, red and blue boxplots represent the Z scores of confirmed, tentative, rejected and shadow features, respectively. Confirmed and tentative features: Age; temperature; respiratory rate; severe acute malnutrition (SAM); rotavirus vaccination; breastfeeding; skin turgor Figure 2. Feature selection for linear growth faltering among children aged < 5 years presenting with moderate to severe diarrhea in rural western Kenya, 2015–2018 In addition to age, respiratory rate, temperature and breastfeeding, the following features were selected: confirmed (stunting at baseline [5.2%], vomit [4.0%], Vesikari score (3.7%) and sunken eyes [3.6%]) and tentative (bacterial infection diagnosis [2.5%]) in the sensitivity analysis using a cut-off of negative change in HAZ (Figure S2). Model Performance We evaluated seven ML algorithms in the prediction of LGF. From the developed models, sensitivity was highest in the RF model (80.7%), followed by the ANN (79.5%), SVM (77.3%), NB (76.5%), GBM (75.6%), LR (75.4%) and lowest in the KNN model (72.4%). The specificity ranged from 58.2–71.8%. Specifically, the specificity of the GBM model was the highest (71.8%), followed by RF (70.1%), LR (61.9%), NB and SVM (61.6%), KNN (61.4%) and lowest in the ANN model (58.2%). The PPV ranged between 27.4% − 34.9% while the NPV ranged between 92.3% − 94.8%. The AUC of the models ranged from 73.4–83.5% with the GBM model having the highest AUC (83.5%, 95% Confidence Interval [95% CI]: 81.6–85.4) (Table 2 ). Table 2 Model performance of linear growth faltering prediction β models using combined data (Original and synthetic data) Algorithm Sensitivity % [95% CI] Specificity % [95% CI] PPV % [95% CI] NPV % [95% CI] F1-Score [95% CI] AUC % [95% CI] PRAUC % [95% CI] RF 80.7 [76.5–84.4] 70.1 [68.1–72.1] 34.9 [31.9–38.0] 94.8 [93.6–95.9] 48.7 [16.8–59.7] 82.8 [80.8–84.8] 96.0 [93.8–96.2] GBM 75.6 [71.2–79.7] 71.8 [69.8–73.7] 34.7 [31.6–37.9] 93.7 [92.4–94.8] 47.6 [13.9–74.5] 83.5 [81.6–85.4] 96.2 [94.9–96.5] NB 76.1 [71.7–80.1] 61.6 [59.5–63.7] 28.2 [25.6–31.0] 92.8 [91.4–94.1] 40.2 [12.0-42.4] 75.6 [73.3–77.9] 94.0 [92.1–95.0] LR 75.4 [70.9–79.4] 61.9 [59.7–64.0] 28.2 [25.5–30.9] 92.7 [91.2–94.0] 38.2 [3.1–64.2] 73.7 [71.3–76.1] 93.0 [91.1–94.0] SVM 77.3 [73.0-81.2] 61.6 [59.5–63.7] 28.6 [25.9–31.3] 93.2 [91.7–94.5] 41.7 [9.3–56.8] 73.4 [71.0-75.8] 93.0 [91.6–94.1] KNN 72.4 [69.7–75.0] 61.4 [59.3–63.5] 27.6 [25.0-30.3] 92.3 [90.8–93.6] 40.2 [6.7–67.0] 74.8 [72.3–77.2] 93.0 [90.8–93.6] ANN 79.5 [75.3–83.3] 58.2 [56.1–60.4] 27.4 [24.9–30.0] 93.5 [92.0-94.7] 40.8 [9.8–58.5] 73.6 [71.3–76.0] 93.0 [90.9–94.1] β− Linear growth faltering defined as Δ HAZ ≥ − 0.5 *RF-Random Forest; GBM-Gradient Boosting; NB- Naïve Bayes; LR-Logistic Regression; SVM- Support vector machine; KNN-K-nearest neighbors; ANN-Artificial Neural Networks; 95% CI- 95% Confidence Interval; PPV- Positive Predictive Value; NPV- Negative Predictive Value; AUC- Area under the Curve; PRAUC- Precision Recall Area under the Curve The GBM model was chosen as the champion model. The receiver operating characteristic (ROC) curves for LGF prediction models are shown in Figure S3. Moreover, in the sensitivity analysis using only the VIDA data in development, the model performance ranged between 63.0%-82.6%, 55.9%-78.6%, 27.3%-33.7%, 91.0%-94.2%, 40.3%-44.3%, 68.0%-75.5%, and 90.6%-94.4% for sensitivity, specificity, PPV, NPV, F1-score, AUC and PRAUC, respectively (Table 3 ). All models showed a decline in predictive performance during sensitivity analysis except for the SVM model, which had a marginal increase. Table 3 Model performance of linear growth faltering prediction β models using original training data only Algorithm Sensitivity % [95% CI] Specificity % [95% CI] PPV % [95% CI] NPV % [95% CI] F1-Score [95% CI] AUC % [95% CI] PRAUC % [95% CI] RF 52.2 [36.9–67.1] 78.6 [72.7–83.7] 32.9 [22.3–44.9] 89.1 [84.0–93.0] 40.3 [10.9–51.4] 70.3 [61.8–78.7] 90.6 [88.4–90.8] GBM 80.4 [66.1–90.6] 63.3 [56.7–69.6] 30.6 [22.5–39.6] 94.2 [89.2–97.3] 44.3 [12.9–55.8] 75.5 [68.2–82.8] 93.6 [92.3–93.9] NB 63.0 [47.5–76.8] 75.1 [69.0-80.6] 33.7 [23.9–44.7] 91.0 [86.0-94.7] 43.9 [5.6–60.3] 73.6 [66.1–81.2] 93.0 [91.1–94.0] LR 73.9 [58.9–85.7] 63.3 [56.7–69.6] 28.8 [20.8–37.9] 92.4 [87.0–96.0] 41.5 [7.8–52.1] 73.8 [67.0-80.5] 93.9 [92.0-94.9] SVM 71.7 [56.5–84.0] 65.9 [59.4–72.1] 29.7 [21.4–39.1] 92.1 [86.8–95.7] 42.0 [7.2–51.9] 75.2 [68.9–81.5] 94.4 [93.0-95.5] KNN 82.6 [68.6–92.2] 56.8 [50.1–63.3] 27.7 [20.4–36.0] 94.2 [88.9–97.5] 41.5 [12.2–51.8] 73.1 [66.3–79.9] 93.6 [91.4–94.2] ANN 82.6 [68.6–92.2] 55.9 [49.2–62.4] 27.3 [20.1–35.5] 94.1 [88.7–97.4] 41.1 [12.0-57.9] 68.0 [60.5–75.6] 91.4 [89.3–92.5] β− Linear growth faltering defined as Δ HAZ ≥ − 0.5 *RF-Random Forest; GBM-Gradient Boosting; NB- Naïve Bayes; LR-Logistic Regression; SVM- Support vector machine; KNN-K-nearest neighbors; ANN-Artificial Neural Networks; 95% CI- 95% Confidence Interval; PPV- Positive Predictive Value; NPV- Negative Predictive Value; AUC- Area under the Curve; PRAUC- Precision Recall Area under the Curve In the sensitivity analysis using the second definition of LGF (negative change in HAZ), the model performance ranged between 45.8%-73.1%, 53.2%-76.6%, 79.0%-90.5%, 28.6%-48.5%, 58.3%-80.9%, 58.0%-82.4%, and 29.0%-62.6% for sensitivity, specificity, PPV, NPV, F1-score, AUC and PRAUC, respectively (Table S2). In this scenario, all models exhibited a drop in predictive performance except for the SVM model, which had a marginal increase and the RF model which registered same performance as in the primary analysis. Overall the Brier scores were relatively high and ranged between 0.19–2.50 (Table 4 ).The Spiegelhalter’s p-value showed that all the models were not properly calibrated (p < 0.05). The performance of the calibrated GBM model was largely similar to its uncalibrated form with the model having an AUC of 83.7%. Table 4 Calibration results of linear growth faltering prediction models. Algorithm Brier Score Spiegelhalter Z-score Spiegelhalter p-value RF 0.19 16.83 < 0.0001 GBM 2.50 208.10 < 0.0001 NB 2.18 101.02 < 0.0001 LR 2.16 85.02 < 0.0001 SVM 2.16 85.88 < 0.0001 KNN 2.21 109.88 < 0.0001 ANN 2.17 84.07 < 0.0001 *RF-Random Forest; GBM-Gradient Boosting; NB- Naïve Bayes; LR-Logistic Regression; SVM- Support vector machine; KNN-K-nearest neighbors; ANN-Artificial Neural Networks; Explanatory Model Analysis The EMA results for the top 2 models in the primary analysis were similar though the degree of importance varied across models with no SAM, no skin turgor, no rotavirus vaccine, age, elevated temperature and respiratory rate being predictive of LGF (Fig. 3 ). Similarly, in the sensitivity analysis using the second definition of LGF, the direction of association was similar between the two models although the magnitude of importance varied. In addition to age, respiratory rate and temperature, the following factors were also identified to be predictive of LGF: severity of disease, no vomiting, stunting at baseline, bacterial infection and lack of sunken eyes (Fig. 3 ). Business Value Evaluation of Champion Model From the business value evaluation of our champion model (GBM), the cumulative gains plot shows that the model is able to select ~ 60% of the target class (LGF) if we select the top-20% cases based on our model. Additionally, from the cumulative lift plot, our champion model is able to identify ~ 3 times higher number of the target class compared to a random selection if we pick the top-20% observations based on model probability. Lastly, from the cumulative response plot, 48% of observations in the top-20% cases based on model probability belong to the target class (Fig. 4). *Scenario 1- Predicting linear growth faltering using a cut-off of Δ HAZ ≥ − 0.5 * age = 9: 9 months; Rotavirus_vacc = 2:2 doses of rotavirus vaccine; cur_wrinkledskin = 0: normal skin; SAM = 0: No severe acute malnutrition (SAM) *Scenario 2- Predicting linear growth faltering using change in haz/month (negative change in linear growth is deemed growth faltering) *age = 9: 9 months; vesikari_cat = 3: Severe disease based on Vesikari score; vomit = 1: Vomitting; Stunting_base = 0: No stunting at baseline; bacterial_infec = 0: No bacterial infection; sunken_eyes = 1: sunken eyes. Figure 4. Business value plots for the Gradient Boosting (GBM) Model for linear growth faltering Temporal Validation We observed a decline in model performance on the temporal validation dataset with the AUC dropping by ~ 18%. Additionally, all metrics dropped in temporal validation with the GBM model achieving 53.7%, 67.7%, 32.5%, 83.5%, 40.5%, 65.6% and 86.4% for sensitivity, specificity, PPV, NPV, F1-score, AUC and PRAUC, respectively (Fig. 5). PPV- Positive Predictive Value; NPV- Negative Predictive Value; AUC- Area under the Curve; PRAUC- Precision Recall Area under the Curve Figure 5. Performance of champion model in development (2015–2018) and temporal validation (2022–2023) datasets. Discussion The study findings illuminate a comprehensive exploration into the prediction of LGF among pediatric patients presenting with diarrhea, employing a robust ML framework. The study involved the development and temporal validation of predictive models using diverse cohorts, revealing distinct prevalence rates and influencing factors associated with LGF. Key features linked to this outcome, such as age, breastfeeding, rotavirus vaccination, respiratory rate, temperature and SAM, were identified through extensive feature selection and their impact on risk prediction was estimated using SHAP attribution. The ML algorithms exhibited varying performance with GBM model emerging as the champion model, demonstrating promising business value. However, the temporal validation uncovered a notable decline in model performance, emphasizing the dynamic nature of health data and the need for ongoing model evaluation and adaptation. This discussion delves into the nuanced interpretation of these results, shedding light on the implications for predictive modeling in the context of pediatric diarrheal outcomes and the broader landscape of healthcare. Despite the impact of rotavirus vaccine introduction on the epidemiology of diarrhea and pathogen landscape, we identified similar predictors, in addition to rotavirus vaccination, to previous modelling efforts [ 9 , 10 ] that used data collected pre-vaccine introduction─ age, breastfeeding, respiratory rate, temperature, absence SAM and stunting at baseline. This finding underscores the enduring importance of these risk factors and the need for comprehensive, sustained, and adaptable public health strategies to combat LGF. Furthermore, we observed that rotavirus vaccination was inversely associated with LGF a finding that is consistent with those of Loli and Carcamo who studied the impact of vaccination on HAZ in Peruvian children aged 6–60 months [ 28 ]. This finding could be due to rotavirus vaccination substantially reducing the incidence and severity of rotavirus infections, curbing the immediate impact of diarrheal diseases on nutrient absorption and consequently diarrhea-mediated growth faltering [ 28 ]. Bolstering rotavirus vaccination is a possible strategy that could be leveraged by policy makers and public health experts to reduce stunting in such settings. Moreover, from a modelling perspective, this finding on predictors generates confidence in the relevance and stability of these variables in different contexts and epidemiological periods, enhancing model transferability and generalizability. These variables have been documented as risk factors for LGF in previous studies. Specifically, age is a significant determinant of LGF among pediatric populations following a diarrheal episode [ 9 , 10 , 29 ]. Infants and very young children face heightened vulnerability to nutritional and health challenges due to their ongoing physiological maturation, which is exacerbated during diarrheal illness leading to pronounced impacts on nutrient absorption and utilization cumulatively contributing to a heightened prevalence of LGF among younger children. Stunting has been shown to be irreversible to a large extent after reaching 24 months of age [ 30 ]. Therefore, the timely identification of at-risk children (infants and toddlers) facilitates the implementation of effective preventive strategies during this critical window of opportunity in early childhood. Moreover, breastfeeding has been shown to be a protective factor against LGF in infants by providing essential nutrients, antibodies, fostering a healthy gut microbiome, and reducing exposure to environmental and food contamination [ 31 , 32 ]. Its nutrient-dense composition supports optimal growth and development, while immune protection reduces the risk of infections that can impede growth. Contrary to existing evidence [ 33 , 34 ], we observed children without SAM to be at increased risk of LGF. Despite majority of factors predisposing children to SAM and stunting being similar, we observed a discordant relationship between the two and this may require further investigation to gain insights into this finding. Elevated baseline temperature and respiratory rate signal are markers of disease severity, and particularly those affecting the gastrointestinal tract, may lead to nutritional deficiencies and hinder linear growth [ 9 , 10 ]. Additionally, elevated respiratory rate and temperature may indicate increased energy expenditure, potentially due to the body's efforts to combat infections or inflammation. This increased energy demand can divert resources away from growth-related processes, impacting linear growth. Tree-based ensembles showed good predictive performance with the GBM model narrowly outperforming the RF model in the prediction of LGF. Our champion model outperformed existing models by Brander et al. (AUC = 67.0%) [ 9 ] and (AUC = 75.0%) Ahmed et al. [ 10 ]. The improvement in model performance could be attributed to the robust modelling approach employed. Moreover, the predictive prowess of tree-based ensembles may have also contributed to this improvement. This strong discriminatory ability of the champion model has significant public health implications as it reinforces the feasibility and efficacy of ML algorithms in timely identification of children, at increased risk of LGF, for early nutritional and healthcare interventions. The model can enhance the efficiency of resource allocation by facilitating targeted screening as well as providing healthcare providers with a valuable tool for informed decision-making, enabling tailored interventions based on individual children risk profiles. However, the decline in model performance during temporal validation while consistent with findings from Ahmed et al. [ 10 ] raises important considerations. Spectral differences in the severity of diarrhea among children in the development and validation cohorts, coupled with potential shifts in the study population over time, highlight challenges in maintaining consistent predictive accuracy. This finding highlights the need for monitoring and periodic retraining of the model in order to maintain its predictive performance. Our primary analysis that used combined data (VIDA and synthetic data) in model development had better performance than the sensitivity analysis that only used VIDA data. This result emphasizes the importance of synthetic data in addressing challenges associated with imbalanced, limited, or privacy-sensitive real-world datasets, providing a means to augment and diversify the data pool [ 35 , 36 ]. This approach overcomes issues of data scarcity, facilitates more comprehensive model training, and enhances generalization. It contributes to overcoming biases, ensuring model fairness, and accommodating the complexity of risk factors influencing a health outcome. Ultimately, the strategic use of synthetic data strengthens the reliability, generalizability, and ethical integrity of predictive models, offering a pathway for more effective and personalized healthcare interventions. However, synthetic data may advance bias propagation since any biases in the primary data will be reflected in the generated data and this may perpetuate and even exacerbate healthcare disparities if they exist [ 37 ]. Furthermore, in the second sensitivity analysis using a cutoff of any negative change in HAZ, we observed a substantial decline in model performance compared to using a cutoff of a decrease of 0.5 HAZ or more. These results imply that using a specific cutoff criteria for defining LGF can significantly impact the performance of the predictive model. Different cutoff criteria may be more appropriate in different contexts, and the choice should be informed by clinical expertise and relevance considering the specific context of the healthcare setting, study population (varying age categories), and the clinical significance of HAZ changes. It also underscores the dynamic nature of model performance, necessitating ongoing evaluation and adaptation to maintain optimal cutoff criteria. Our study, while commendable, has limitations, notably the exclusion of pathogen data during model development to maintain practical applicability, despite its influence on LGF. Future research should address this gap, as well as focus on the acceptability and impact of ML models on clinical practice and patient outcomes. The cost-effectiveness of deploying these models is also crucial for practical implementation in diverse healthcare settings. Exploring these facets will contribute significantly to enhancing understanding and ensuring the effective use of ML models in healthcare. Conclusion The study's findings emphasize the enduring relevance of established predictors of LGF. Addressing multifaceted challenges in pediatric LGF requires sustained efforts with adaptive interventions for these risk factors. The study demonstrates the practical use of ML algorithms for rapid identification of at-risk children. A decline in model performance during temporal validation highlights the dynamic nature of health data, necessitating continuous evaluation and adaptation. Additionally, the study shows the viability of integrating synthetic data to enhance model robustness, providing a pathway for more comprehensive and ethical predictive modeling in healthcare. Declarations Ethics approval and consent to participate The VIDA protocol was approved by the Institutional Review Board of the University of Maryland School of Medicine, Baltimore, MD, USA (UMB Protocol #: HM-HP-00062472) and the Kenya Medical Research Institute (KEMRI) Scientific and Ethical Review Unit (SERU) (SERU#2996). The EFGH protocol was approved by the KEMRI SERU (SERU#4362). Written informed consent was sought from caregivers in both studies before initiation of study procedures. Additionally, ethical approval for undertaking the current study was sought from the health research ethics committee of the University of South Africa, College of Agricultural Sciences (2023/CAES_HREC/2192). Consent for publication Not applicable. Competing Interest Authors declare no conflict of interest. Disclosure The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Kenya Medical Research Institute or any collaborating institutions. Funding This work was supported by the Bill & Melinda Gates Foundation (grant INV-045988). The funders did not play any role in the study and interpretation of its outcome. Author Contribution BO, VHM, KDT, PBP and RO conceived the study and contributed to study design and implementation. BO, VHM and KDT analyzed and interpreted the data. BO drafted the manuscript and all authors critically reviewed the manuscript for intellectual content and approved the final manuscript. All authors read and approved the final manuscript. Acknowledgements We appreciate the contributions and efforts of KEMRI-CGHR staff involved in the data collection, data management, and laboratory testing of samples in the two studies. We also wish to thank the study participants and the ministry of health staff for supporting both studies. Moreover, we would like to acknowledge the use of artificial intelligence (AI) technology for grammar checking and proofreading of this manuscript. Availability of data and materials The data used for the modelling in this study belongs to KEMRI and restrictions apply to the availability of these data. Data cleaning, pre-processing and model development were done in R version 4.1.2. The programming code for R is available upon request addressed to the corresponding author: Billy Ogwel ( [email protected] ). References World Health Organization. Diarrhoeal disease. 2017. Available at: https://www.who.int/news-room/fact-sheets/detail/diarrhoeal-disease . Accessed 19 February 2022. Ferdous F, Das SK, Ahmed S, et al. Severity of Diarrhea and Malnutrition among Under Five-Year-Old Children in Rural Bangladesh. Am J Trop Med Hyg. 2013;89:223–8. Checkley W, Buckley G, Gilman RH, et al. Multi-country analysis of the effects of diarrhoea on childhood stunting. Int J Epidemiol. 2008;37:816. Lenters L, Wazny K, Bhutta ZA. Management of Severe and Moderate Acute Malnutrition in Children. In: Black RE, Laxminarayan R, Temmerman M, Walker N, eds. Reproductive, Maternal, Newborn, and Child Health: Disease Control Priorities, Third Edition (Volume 2). Washington (DC): The International Bank for Reconstruction and Development / The World Bank, 2016. Available at: http://www.ncbi.nlm.nih.gov/books/NBK361900/ . Accessed 27 November 2020. Danaei G, Andrews KG, Sudfeld CR, et al. Risk Factors for Childhood Stunting in 137 Developing Countries: A Comparative Risk Assessment Analysis at Global, Regional, and Country Levels. PLoS Med. 2016;13:e1002164. Wierzba TF, Muhib F. Exploring the broader consequences of diarrhoeal diseases on child health. Lancet Global Health. 2018;6:e230–1. Guerrant RL, DeBoer MD, Moore SR, Scharf RJ, Lima AAM. The impoverished gut–a triple burden of diarrhoea, stunting and chronic disease. Nat Rev Gastroenterol Hepatol. 2013;10:220–9. ROTA Council. The broader impact of early childhood diarrhea. 2019. Available at: https://preventrotavirus.org/wp-content/uploads/2019/05/ROTA-Brief6-LastingImpact-SP-1-3.pdf . Accessed 1 November 2022. Brander RL, Pavlinac PB, Walson JL, et al. Determinants of linear growth faltering among children with moderate-to-severe diarrhea in the Global Enteric Multicenter Study. BMC Med. 2019;17:214. Ahmed SM, Brintz BJ, Pavlinac PB, et al. Derivation and external validation of clinical prediction rules identifying children at risk of linear growth faltering. Elife. 2023;12:e78491. Rahmani K, Thapa R, Tsou P, et al. Assessing the effects of data drift on the performance of machine learning models used in clinical sepsis prediction. Int J Med Informatics. 2023;173:104930. Sahiner B, Chen W, Samala RK, Petrick N. Data drift in medical machine learning: implications and potential remedies. Br J Radiol 2023;:20220878. Powell H, Liang Y, Neuzil KM, et al. A Description of the Statistical Methods for the Vaccine Impact on Diarrhea in Africa (VIDA) Study. Clin Infect Dis. 2023;76:S5–11. Nasrin D, Liang Y, Powell H, et al. Moderate-to-Severe Diarrhea and Stunting Among Children Younger Than 5 Years: Findings From the Vaccine Impact on Diarrhea in Africa (VIDA) Study. Clin Infect Dis. 2023;76:S41–8. Nowok B, Raab GM, Dibben C. synthpop: Bespoke Creation of Synthetic Data in R . J Stat Soft. 2016; 74. Available at: http://www.jstatsoft.org/v74/i11/ . Accessed 23 September 2023. Joseph VR. Sci J. 2022;15:531–8. Optimal ratio for data splitting. Statistical Analysis and Data Mining: The ASA Data. Nasrin D, Blackwelder WC, Sommerfelt H, et al. Pathogens Associated With Linear Growth Faltering in Children With Diarrhea and Impact of Antibiotic Treatment: The Global Enteric Multicenter Study. J Infect Dis. 2021;224:S848–55. van Buuren S, Groothuis-Oudshoorn K, Vink G et al. Package ‘mice’. 2021. Available at: https://cran.r-project.org/web/packages/mice/mice.pdf . Accessed 31 May 2021. Kursa MB, Rudnicki WR. Package ‘Boruta’. 2020. Available at: https://cran.r-project.org/web/packages/Boruta/Boruta.pdf . Accessed 31 May 2021. Refaeilzadeh P, Tang L, Liu H. Cross-Validation. In: LIU L, ÖZSU MT,Encyclopedia of Database Systems. Boston, Springer MA. US, 2009: 532–538. Available at: https://doi.org/10.1007/978-0-387-39940-9_565 . Accessed 17 October 2023. Bach M, Werner A, Żywiec J, Pluskiewicz W. The study of under- and over-sampling methods’ utility in analysis of highly imbalanced data on osteoporosis. Inf Sci. 2017;384:174–90. Saito T, Rehmsmeier M. precrec: Calculate Accurate Precision-Recall and ROC (Receiver Operator Characteristics) Curves. 2023; Available at: https://CRAN.R-project.org/package=precrec . Accessed 10 February 2023. Huang Y, Li W, Macheret F, Gabriel RA, Ohno-Machado L. A tutorial on calibration measurements and calibration models for clinical prediction models. J Am Med Inform Assoc. 2020;27:621–33. Biecek P, Maksymiuk S, Baniecki H, DALEX. : moDel Agnostic Language for Exploration and eXplanation. 2023; Available at: https://CRAN.R-project.org/package=DALEX . Accessed 10 February 2023. Cowley LE, Farewell DM, Maguire S, Kemp AM. Methodological standards for the development and evaluation of clinical prediction rules: a review of the literature. Diagn Prognostic Res. 2019;3:16. Nagelkerke J. modelplotr: Plots to evaluate the business value of predictive models. 2020; Available at: https://cran.r-project.org/web/packages/modelplotr/vignettes/modelplotr.html . Accessed 19 November 2022. R Core Team. R: The R Project for Statistical Computing. 2021. Available at: https://www.r-project.org/ . Accessed 3 December 2021. Loli S, Carcamo CP. Rotavirus vaccination and stunting: Secondary Data Analysis from the Peruvian Demographic and Health Survey. Vaccine. 2020;38:8010–5. Benjamin-Chung J, Mertens A, Colford JM, et al. Early-childhood linear growth faltering in low- and middle-income countries. Nature. 2023;621:550–7. Victora CG, de Onis M, Hallal PC, Blössner M, Shrimpton R. Worldwide timing of growth faltering: revisiting implications for interventions. Pediatrics. 2010;125:e473–480. Silverberg SL, Qamar H, Keya FK, et al. Do Early Infant Feeding Practices and Modifiable Household Behaviors Contribute to Age-Specific Interindividual Variations in Infant Linear Growth? Evidence from a Birth Cohort in Dhaka, Bangladesh. Curr Developments Nutr. 2021;5:nzab077. Kramer MS, Kakuma R. Optimal duration of exclusive breastfeeding. Cochrane Database Syst Rev. 2012; 2012:CD003517. Ngari MM, Iversen PO, Thitiri J, et al. Linear growth following complicated severe malnutrition: 1-year follow-up cohort of Kenyan children. Arch Dis Child. 2019;104:229–35. Garenne M, Myatt M, Khara T, Dolan C, Briend A. Concurrent wasting and stunting among under-five children in Niakhar, Senegal. Matern Child Nutr. 2018;15:e12736. Giuffrè M, Shung DL. Harnessing the power of synthetic data in healthcare: innovation, application, and privacy. npj Digit Med. 2023;6:1–8. Gonzales A, Guruswamy G, Smith SR. Synthetic data in health care: A narrative review. PLOS Digit Health. 2023;2:e0000082. Marwala T, Fournier-Tombs E, Stinckwich S. The Use of Synthetic Data to Train AI Models: Opportunities and Risks for Sustainable Development. 2023. Additional Declarations No competing interests reported. Supplementary Files SupplementalMaterials12Mar2024.pdf Cite Share Download PDF Status: Published Journal Publication published 02 Dec, 2024 Read the published version in BMC Medical Informatics and Decision Making → Version 1 posted Editorial decision: Revision requested 22 Oct, 2024 Reviews received at journal 02 Oct, 2024 Reviewers agreed at journal 30 Aug, 2024 Reviews received at journal 20 Jun, 2024 Reviewers agreed at journal 13 Jun, 2024 Reviewers invited by journal 31 May, 2024 Editor invited by journal 13 Mar, 2024 Submission checks completed at journal 13 Mar, 2024 Editor assigned by journal 13 Mar, 2024 First submitted to journal 08 Mar, 2024 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-4047381","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":279489787,"identity":"51389f0d-12f2-4e01-bab1-16562b1ab9df","order_by":0,"name":"Billy Ogwel","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA10lEQVRIiWNgGAWjYLCChw0SMnzsDUCWgQWRWhIbJHjYeA6AtEgQrYWBh00iAcQkQgs//xrDB4k7LHjYJJ9f3fCjQIKBv707Aa8WyRlvjA0SzwAdJp1TdrMH6DCJM2c34NVicOPsNonENrCWtBs8QC0GErkEtWz/AdYieSbt5h+itJzv3cYA1iLBfuw2UbZIzuD/LAH2C08O220ZAwkegn7h5z+W+OHjjjo5fvbjz26++WMjx9/ei18LAyQ6QIDHAEziVw625gCMxf6AsOpRMApGwSgYkQAA7S1DqH4PKsMAAAAASUVORK5CYII=","orcid":"","institution":"Kenya Medical Research Institute- Center for Global Health Research (KEMRI-CGHR)","correspondingAuthor":true,"prefix":"","firstName":"Billy","middleName":"","lastName":"Ogwel","suffix":""},{"id":279489788,"identity":"49ae2172-f2a2-426e-af19-361af3293262","order_by":1,"name":"Vincent H. Mzazi","email":"","orcid":"","institution":"University of South Africa","correspondingAuthor":false,"prefix":"","firstName":"Vincent","middleName":"H.","lastName":"Mzazi","suffix":""},{"id":279489789,"identity":"321d5337-8061-4466-8862-2f569b23e0c3","order_by":2,"name":"Alex O. Awuor","email":"","orcid":"","institution":"Kenya Medical Research Institute- Center for Global Health Research (KEMRI-CGHR)","correspondingAuthor":false,"prefix":"","firstName":"Alex","middleName":"O.","lastName":"Awuor","suffix":""},{"id":279489790,"identity":"e7c37be3-3134-49a4-8f62-34648f40b176","order_by":3,"name":"Caleb Okonji","email":"","orcid":"","institution":"Kenya Medical Research Institute- Center for Global Health Research (KEMRI-CGHR)","correspondingAuthor":false,"prefix":"","firstName":"Caleb","middleName":"","lastName":"Okonji","suffix":""},{"id":279489791,"identity":"180c053e-a3f3-42bc-927b-dea513905901","order_by":4,"name":"Raphael O. Anyango","email":"","orcid":"","institution":"Kenya Medical Research Institute- Center for Global Health Research (KEMRI-CGHR)","correspondingAuthor":false,"prefix":"","firstName":"Raphael","middleName":"O.","lastName":"Anyango","suffix":""},{"id":279489792,"identity":"0d525179-894d-40a7-934a-9fe5c67a72f8","order_by":5,"name":"Caren Oreso","email":"","orcid":"","institution":"Kenya Medical Research Institute- Center for Global Health Research (KEMRI-CGHR)","correspondingAuthor":false,"prefix":"","firstName":"Caren","middleName":"","lastName":"Oreso","suffix":""},{"id":279489793,"identity":"b48cbccf-d0b7-420d-bd6d-bdf6eb053cf1","order_by":6,"name":"John B. Ochieng","email":"","orcid":"","institution":"Kenya Medical Research Institute- Center for Global Health Research (KEMRI-CGHR)","correspondingAuthor":false,"prefix":"","firstName":"John","middleName":"B.","lastName":"Ochieng","suffix":""},{"id":279489794,"identity":"1f75c483-5585-4802-b963-c2e6b58ebaec","order_by":7,"name":"Stephen Munga","email":"","orcid":"","institution":"Kenya Medical Research Institute- Center for Global Health Research (KEMRI-CGHR)","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"","lastName":"Munga","suffix":""},{"id":279489795,"identity":"4c83bb67-0983-44b3-a7fd-47e3510ec219","order_by":8,"name":"Dilruba Nasrin","email":"","orcid":"","institution":"University of Maryland School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Dilruba","middleName":"","lastName":"Nasrin","suffix":""},{"id":279489796,"identity":"cfc0bb47-c879-4fd6-ad6d-2136dedaf781","order_by":9,"name":"Kirkby D. Tickell","email":"","orcid":"","institution":"University of Washington","correspondingAuthor":false,"prefix":"","firstName":"Kirkby","middleName":"D.","lastName":"Tickell","suffix":""},{"id":279489797,"identity":"e66778d8-60a7-47ef-9f20-b96819e6a97e","order_by":10,"name":"Patricia B. Pavlinac","email":"","orcid":"","institution":"University of Washington","correspondingAuthor":false,"prefix":"","firstName":"Patricia","middleName":"B.","lastName":"Pavlinac","suffix":""},{"id":279489798,"identity":"b5a17e0a-c433-432b-aea5-2e4984cccca9","order_by":11,"name":"Karen L. Kotloff","email":"","orcid":"","institution":"University of Maryland School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Karen","middleName":"L.","lastName":"Kotloff","suffix":""},{"id":279489800,"identity":"873d0f61-e672-451c-84c2-70cf6ac04ed0","order_by":12,"name":"Richard Omore","email":"","orcid":"","institution":"Kenya Medical Research Institute- Center for Global Health Research (KEMRI-CGHR)","correspondingAuthor":false,"prefix":"","firstName":"Richard","middleName":"","lastName":"Omore","suffix":""}],"badges":[],"createdAt":"2024-03-08 17:00:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4047381/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4047381/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12911-024-02779-7","type":"published","date":"2024-12-02T15:57:18+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":52789252,"identity":"d5ba116c-dfb0-4f19-b6f3-45073832e894","added_by":"auto","created_at":"2024-03-15 19:44:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":22641,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of development and temporal validation studies conducted in Siaya County, Kenya\u003c/p\u003e\n\u003cp\u003eVIDA- Vaccine Impact on Diarrhea in Africa Study\u003c/p\u003e\n\u003cp\u003eEFGH-Enterics for Global Health Shigella Surveillance study\u003c/p\u003e\n\u003cp\u003eMSD-Moderate-to-Severe Diarrhea; MAD-Medically Attended Diarrhea\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4047381/v1/c248d342b1b7fbcd0c42e433.png"},{"id":52789256,"identity":"98d254e6-fb33-4149-adde-f682539f9cdb","added_by":"auto","created_at":"2024-03-15 19:45:00","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":308847,"visible":true,"origin":"","legend":"\u003cp\u003eFeature selection for linear growth faltering among children aged \u0026lt; 5 years presenting with moderate to severe diarrhea in rural western Kenya, 2015-2018\u003c/p\u003e\n\u003cp\u003eGreen, yellow, red and blue boxplots represent the Z scores of confirmed, tentative, rejected and shadow features, respectively.\u003c/p\u003e\n\u003cp\u003eConfirmed and tentative features: \u003cem\u003eAge; temperature; respiratory rate; severe acute malnutrition (SAM); rotavirus vaccination; breastfeeding; skin turgor\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4047381/v1/5a1ddd3fb6e3006892d625ff.jpeg"},{"id":52789245,"identity":"0c2f495c-b39d-419a-b2c7-2cc7b4fa0672","added_by":"auto","created_at":"2024-03-15 19:44:54","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":435363,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP attributions for the Top 2 linear growth faltering models\u003c/p\u003e\n\u003cp\u003e*Scenario 1- Predicting linear growth faltering using a cut-off of Δ HAZ ≥ − 0.5\u003c/p\u003e\n\u003cp\u003e* age=9: 9 months; Rotavirus_vacc =2:2 doses of rotavirus vaccine; cur_wrinkledskin=0: normal skin; SAM=0: No severe acute malnutrition (SAM)\u003c/p\u003e\n\u003cp\u003e*Scenario 2- Predicting linear growth faltering using change in haz/month (negative change in linear growth is deemed growth faltering)\u003c/p\u003e\n\u003cp\u003e*age=9: 9 months; vesikari_cat=3: Severe disease based on Vesikari score; vomit=1: Vomitting; Stunting_base=0: No stunting at baseline; bacterial_infec=0: No bacterial infection; sunken_eyes =1: sunken eyes.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4047381/v1/2b1d54a818a1c506ff1e80fd.jpeg"},{"id":52789241,"identity":"e1090f14-6473-4730-8d82-570b3989d473","added_by":"auto","created_at":"2024-03-15 19:44:53","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":45676,"visible":true,"origin":"","legend":"\u003cp\u003eBusiness value plots for the Gradient Boosting (GBM) Model for linear growth faltering\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-4047381/v1/84f48c566dd5b50e11d61819.png"},{"id":52789239,"identity":"9fe08550-2163-4f7d-88bb-afdadb0c6805","added_by":"auto","created_at":"2024-03-15 19:44:51","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":36262,"visible":true,"origin":"","legend":"\u003cp\u003ePerformance of champion model in development (2015-2018) and temporal validation (2022-2023) datasets.\u003c/p\u003e\n\u003cp\u003ePPV- Positive Predictive Value; NPV- Negative Predictive Value; AUC- Area under the Curve; PRAUC- Precision Recall Area under the Curve\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-4047381/v1/4f3c2182ef27e8b7091b274a.png"},{"id":70964658,"identity":"e526e552-e2a9-4fb9-a04c-96072b737bfd","added_by":"auto","created_at":"2024-12-09 16:13:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2001382,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4047381/v1/b059ba14-686e-417b-97f8-c50dad9cffa3.pdf"},{"id":52789240,"identity":"240417f8-8c7a-4dda-a119-def4dde6b0fb","added_by":"auto","created_at":"2024-03-15 19:44:53","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":824800,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementalMaterials12Mar2024.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4047381/v1/05ebd541d8e190b5b52f8c86.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Predictive Modelling of Linear Growth Faltering Among Pediatric Patients with Diarrhea in Rural Western Kenya: An Explainable Machine Learning Approach","fulltext":[{"header":"Introduction","content":"\u003cp\u003e \u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eDiarrhea, a global public health problem with greatest burden in low- and middle-income countries (LMICs) [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e], is a leading etiology of malnutrition among children in LMICs, in part due to anorexia, decreased absorptive function, mucosal damage, catabolism and nutrient exhaustion [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. It has been reported that the cumulative burden of diarrhea days directly correlates with the degree of nutritional failure among children during early childhood and that catch-up growth does not appear to make up for the deficit [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Linear growth faltering (LGF), a precursor to stunting (height-for-age z-score [HAZ] \u0026lt;-2), is one form of malnutrition that results from protracted nutritional deprivation [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Stunting affects one-fifth of children globally and one-third of children in LMIC. In the universal pie of stunting prevalence, diarrhea accounts for an estimated 13.5% [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Additionally, a vicious cycle of diarrhea and malnutrition can occur as malnutrition weakens the body’s defense against future diarrheal episodes resulting in more frequent and longer diarrheal illnesses. The effects of stunting can be severe and protracted, with reduced cognitive development, persistent poor health, and elevated risk of mortality [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Long term complications can include an increased risk of cardiovascular disease, type 2 diabetes, and obesity in adulthood [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003eThe timely and accurate identification of children at increased risk of LGF is crucial for early nutritional and healthcare interventions as well as efficient allocation of public health resources, efforts that could help to avert the associated negative outcomes. Data-driven predictive models could be leveraged to this end and a number of research efforts exist in the prediction of LGF among children with diarrhea [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. While the existing models provide a valuable starting point, shifts in the study population over time may affect the predictive performance of these models [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Moreover, development of new models using more recent and pertinent data offers the opportunity to improve model performance and capture new perspectives and insights into this public health problem. We used machine learning (ML), which has been adopted in public health and clinical practice to rapidly develop data-driven clinical prediction models, to develop and temporally validate predictive models for LGF among children aged \u0026lt; 5 years with diarrhea in rural Western Kenya.\u003c/p\u003e "},{"header":"Methods","content":"\u003cp\u003eData sources\u003c/p\u003e\u003cp\u003eThis retrospective study used data collected from the Kenyan site (in Siaya County) of two related diarrheal studies: The Vaccine Impact on Diarrhea in Africa (VIDA) study for model development and evaluation; and the Enteric for Global Health (EFGH) \u003cem\u003eShigella\u003c/em\u003e surveillance study for temporal validation.\u003c/p\u003e\u003cp\u003eDevelopment cohort\u003c/p\u003e\u003cp\u003eVIDA was designed to assess the population-based incidence, etiologies, and adverse clinical consequences of diarrhea following rotavirus vaccine introduction in children aged 0–59 months residing in censused populations in 3 African countries. Moderate-to-severe diarrhea (MSD) cases, defined as children in 3 age strata (0–11, 12–23, and 24–59 months) presenting with diarrhea (defined as ≥ 3 looser-than-normal stools within 24 hours) that began within the past 7 days after ≥ 7 diarrhea-free days and had ≥ 1 of the following: sunken eyes, poor skin turgor, dysentery, intravenous rehydration, or required hospitalization, were enrolled from sentinel health centers (SHCs) serving the health and demographic surveillance systems population at each site. The aim was to enroll 8–9 MSD cases in each age stratum per fortnight. 1–3 diarrhea-free controls matched by age, gender and geographical location were enrolled within 14 days of case enrolment. Follow-ups were conducted between 49–91 days after enrolment. We utilized data collected from cases enrolled at the VIDA Kenya site over a 36 months period from May 2015 and July 2018 restricting to children aged 6–35 months to make the development and temporal validation cohorts comparable. The study design, clinical and epidemiological methods for VIDA have been described elsewhere [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn addition to the VIDA data (n = 1,106), we generated a synthetic dataset (n = 8,894) based on the VIDA data using the synthpop package [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e] to increase the sample size and to enable the algorithms to generate more stable and reliable predictions that are less sensitive to noise in the data. The variables of the synthetic dataset were compared to the original training dataset with the synthetic dataset demonstrating high similarity to the original dataset (Fig \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). The combined dataset (N = 10,000) was used for training and internal validation with a split-sampling conducted in the ratio 3:1 to partition the development data into training and test sets [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eTemporal validation cohort\u003c/p\u003e\u003cp\u003eThe EFGH study set out to establish incidence and consequences of \u003cem\u003eShigella\u003c/em\u003e medically attended diarrhea (MAD) within 7 country sites in Africa, Asia, and Latin America using cross-sectional and longitudinal study designs. MAD cases defined as children aged 6–35 months presenting with diarrhea (defined as ≥ 3 looser-than-normal stools within 24 hours) that began within the past 7 days after ≥ 2 diarrhea-free days were enrolled from SHCs in the study catchment area. Additional eligibility criteria included: residing within the pre-defined study catchment area; primary caregiver and child plan to remain at their current residence for at least the next 4 months; legal guardian consenting to child’s participation in the study as well willingness to be followed-up for 3 months post-enrolment; child is not being referred to a non-EFGH facility at the time of screening; and site enrollment cap has not been met. Follow-ups were conducted at week-4 (24–67 days) and month-3 (84–127 days). Our study utilized data from children enrolled in Kenya between 01 August, 2022 and 31 July, 2023 to temporally validate the champion model.\u003c/p\u003e\u003cp\u003eInformation on demographic, socio-demographic, epidemiological and clinical characteristics were collected at enrollment by study personnel in both studies.\u003c/p\u003e\u003cp\u003eTarget variable\u003c/p\u003e\u003cp\u003eConsistent with previous studies [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], we defined the target variable, LGF, as decrease of 0.5 HAZ or more (Δ HAZ ≥ − 0.5) within 49–91 days of enrollment in VIDA, or within 84–127 days in EFGH. We also computed change in HAZ per month of follow-up and categorized a negative change as LGF in our sensitivity analysis, similar to the definition used by Nasrin et al. [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. We excluded children with implausible HAZ values (HAZ \u0026gt; 6 or \u0026lt; − 6 and change in (Δ) HAZ \u0026gt; 3; or length values that were \u0026gt; 1.5 cm lower at follow-up than at enrollment.\u003c/p\u003e\u003cp\u003ePredictive variables and feature selection\u003c/p\u003e\u003cp\u003eA total of 68 potential candidate predictors collected at enrollment during both studies were considered, including demographic, household-level characteristics, illness history, anthropometric and clinical characteristics collected at enrolment. Missingness patterns were assessed among the features and the missing data points imputed using the Multiple Imputation by Chained Equations (MICE) package [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Furthermore, we conducted feature selection to reduce dimensionality, optimize performance, reduce computational complexity and enhance model interpretability. The feature selection was implemented using the Boruta package [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e] an all relevant feature selection wrapper around the random forest algorithm that selects relevant features by comparing original attributes' importance with importance achievable at random using their permuted copies. We excluded features that were rejected in this process. Moreover, among the confirmed and tentative features, we excluded variables that were not collected in both studies (breastfeeding).\u003c/p\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cp\u003eWe compared patient characteristics of children with LGF versus those without. Proportions were reported for categorical variables and either chi-square or Fisher`s exact test were performed as appropriate. Wilcoxon rank sum tests were used to compare continuous variables as appropriate. We also compared the prevalence of LGF between the 2 studies.\u003c/p\u003e\u003cp\u003eModel development and internal validation\u003c/p\u003e\u003cp\u003eTo derive the LGF prediction model, we utilized 7 ML algorithms including: Random Forest (RF), Gradient Boosting (GBM), Naive Bayes (NB), Logistic regression (LR), Support vector machine (SVM), K-nearest neighbors (KNN) and Artificial Neural Networks (ANN). The predictive models were developed in the training dataset using 10-fold cross-validation [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e], a valuable step in model development helping to obviate under-fitting or overfitting of the model and ensure robust and well-performing models. Due to the moderate class imbalance in our target variable (LGF), we employed sub-sampling techniques (over-sampling) within the resampling procedure to mitigate the negative impact of class disparity on model fitting [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. We then conducted internal validation of the models on the test data evaluating performance using the following metrics: sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and F1-score. Receiver operating characteristic (ROC) curves were constructed and the area under the curve (AUC) and the precision-recall area under the curve (PRAUC) for each model were computed using the precrec package [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. We determined the champion model as the model with the best AUC. We also assessed calibration in the developed models using Brier scores (the mean squared error between the actual outcome and the estimated probabilities), Spiegelhalter’s \u003cem\u003ez\u003c/em\u003e-test (a formal measurement that serves as a proxy for calibration calculated from the decomposition of Brier score) and its accompanying p-value [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. We used Platt scaling approach, in which model estimates are transformed by passing the estimates through a trained sigmoid function, to calibrate the champion model [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. To enhance model interpretability, trust and fairness, we conducted explanatory model analysis (EMA) for the top two models using a model agnostic procedure to estimate SHapley Additive exPlanations (SHAPs) attributions drawing on the DALEX package [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eTemporal Validation and Business Value Evaluation\u003c/p\u003e\u003cp\u003eWe further conducted temporal validation on the champion model to assess the robustness and generalizability of the model's performance over time [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. To evaluate the business value of the predictive model, modelplotr package [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e] was used to build valuable evaluation plots (cumulative gains, cumulative lift, response and cumulative response plots). Descriptive analysis, predictive modelling for LGF and plotting were all performed in R version 4.2.2 [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e].\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eA total of 1,554 and 706 children were enrolled in the development and temporal validation cohorts, respectively. Among children aged 6–35 months enrolled, 1,106 (71.2%) and 655 (92.7%) had HAZ data that were plausible, respectively. Among those that had plausible HAZ data, 187 (16.9%) and 147 (22.4%) had LGF in the development and temporal validation cohorts, respectively (Fig.\u0026nbsp;1).\u003c/p\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eDevelopment dataset (VIDA: 2015–2018) Temporal validation dataset (EFGH: 2022–2023)\u003c/h2\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eVIDA- Vaccine Impact on Diarrhea in Africa Study\u003c/p\u003e \u003cp\u003eEFGH-Enterics for Global Health Shigella Surveillance study\u003c/p\u003e \u003cp\u003eMSD-Moderate-to-Severe Diarrhea; MAD-Medically Attended Diarrhea\u003c/p\u003e \u003cp\u003eFigure 1. Flowchart of development and temporal validation studies conducted in Siaya County, Kenya\u003c/p\u003e \u003cp\u003eThis difference in the prevalence of LGF between the development and temporal validation cohorts was statistically significant (p = 0.0042). The median [interquartile range] ΔHAZ between enrollment and follow-up was − 0.21 [-0.42- -0.01] and − 0.24 [-0.48- -0.02] in the development and temporal validation cohorts, respectively. In the sensitivity analysis using the cut-off of negative change in HAZ, the prevalence of LGF was 1,051 (28.7%). Additionally, the constructed synthetic dataset had 8,527 observations and it closely replicated the propensity score distribution of the original development data (VIDA) as evidenced by the comprehensive descriptive analysis that compared each variable (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe characteristics of VIDA participants at enrolment stratified by LGF status are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Children who had LGF were younger than those who did not (Median age in months [IQR]: 11 [\u003cspan additionalcitationids=\"CR9 CR10 CR11 CR12 CR13\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e–\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] vs 17 [\u003cspan additionalcitationids=\"CR12 CR13 CR14 CR15 CR16 CR17 CR18 CR19 CR20 CR21 CR22 CR23\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e–\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e], p \u003cb\u003e\u0026lt;\u003c/b\u003e 0.001). Furthermore, compared with those who did not have LGF, those with LGF had a higher respiratory rate (Median [IQR]: 38.5 [34.0-42.5] vs 36.0[31.5–39.5], p \u0026lt; 0.001), a higher temperature (Median [IQR]: 37.1 [36.6–37.8] vs 36.8 [36.4–37.5], p \u0026lt; 0.001) and more severe disease (Median Vesikari score [IQR]: 11 [\u003cspan additionalcitationids=\"CR10 CR11\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e–\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] vs 10 [\u003cspan additionalcitationids=\"CR9 CR10 CR11\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e–\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e], p \u0026lt; 0.001). Additionally, caretaker education, breastfeeding, vomiting, wrinkled skin, restless, admission, and intravenous rehydration were significantly associated with LGF (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\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\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\u003eCharacteristics of children aged \u0026lt; 5 years seeking care for moderate-to-severe diarrhea in Kenya stratified by Linear Growth Faltering Status, 2015–2018.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eLinear Growth Faltering\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes (n = 187)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo (n = 919)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ep-value*\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003en (%)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003en (%)\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDemograhic\u003c/b\u003e\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian age [IQR]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11 [\u003cspan additionalcitationids=\"CR9 CR10 CR11 CR12 CR13\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e–\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17 [\u003cspan additionalcitationids=\"CR12 CR13 CR14 CR15 CR16 CR17 CR18 CR19 CR20 CR21 CR22 CR23\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e–\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt; 0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge Category\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0–11 months\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e104 (55.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e259 (28.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt; 0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12–23 months\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e74 (39.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e428 (46.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e24–59 months\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9 (4.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e232 (25.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender: Female\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e83 (44.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e428 (46.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.584\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHousehold Details\u003c/b\u003e\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCaretaker education ( \u0026gt; = Secondary )\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e78 (41.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e305 (33.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.026\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;= 2 children under 5 yrs\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e167 (89.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e839 (91.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.387\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;= 4 people sleeping\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e77 (41.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e400 (43.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.546\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;= 3 Total Assets\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e158 (84.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e812 (88.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.142\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRefined/Electric Primary Fuel Source\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5 (2.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39 (4.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.313\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnimal ownership\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e176 (94.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e836 (91.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.159\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eImproved water\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSafely managed\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e83 (44.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e431 (46.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBasic\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14 (7.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e112 (12.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLimited\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e28 (15.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e125 (13.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eunimproved/Surface water\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e62 (33.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e251 (27.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eImproved Sanitation\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSafely Managed and Basic\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20 (10.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e106 (11.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.392\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLimited\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e72 (38.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e306 (33.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnimproved/Open Defecation\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e95 (50.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e507 (55.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eClinical characteristics\u003c/b\u003e\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eReported by caretaker\u003c/b\u003e\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBreastfeeding before diarrhea onset\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25 (13.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e248 (37.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt; 0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExclusive\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 (1.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8 (0.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePartial\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e159 (85.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e563 (61.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian diarrhea days [IQR]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e–\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 [\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e–\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.7196\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStool Type\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSimple watery\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e113 (60.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e532 (57.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.41\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRice watery\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5 (2.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12 (1.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSticky/Mucoid\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e65 (34.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e347 (37.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBloody\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (2.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e28 (3.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStool Count\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e27 (14.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e165 (18.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.489\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4–5\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e101 (54.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e506 (55.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6–10\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e55 (29.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e228 (24.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt; 10\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (2.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20 (2.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBlood in stool\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15 (8.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e108 (11.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.138\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVomiting\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e127(67.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e531 (57.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.01\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVery Thirsty\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e156 (83.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e752 (82.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.582\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDrinks poorly\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e47 (25.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e232 (25.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.962\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnable to drink\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (1.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27 (2.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.145\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBelly Pain\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e109 (61.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e508 (57.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.404\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFever\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e142 (75.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e709 (77.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.720\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRestless\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e151 (80.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e710 (77.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.295\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLethargy\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e123 (65.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e600 (65.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.898\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eunconscious\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7 (3.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e32 (3.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.864\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRectal straining\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e55 (29.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e211 (23.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.066\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRectal prolapse\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (1.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15 (1.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.565\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCough\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e103 (55.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e482 (52.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.511\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDifficulty breathing\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e32 (17.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e124 (13.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.197\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConvulsion\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 (1.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17 (1.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.818\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCurrently\u003c/b\u003e\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVery Thirsty\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e145 (78.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e653 (71.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.062\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDrinks poorly\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e40 (21.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e188 (20.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.747\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSunken Eyes\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e171 (91.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e792 (86.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.054\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWrinkled skin\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e56 (30.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e211 (23.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.029\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRestless\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e134 (71.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e557 (60.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.004\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLethargy/unconscious\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23 (12.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e151 (16.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.157\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDry mouth\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e142 (75.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e658 (71.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.235\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFast breathing\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24 (12.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e100 (10.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.44\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHome ORS use\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21 (11.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e86 (9.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.43\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHome Zinc use\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (4.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34 (3.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.706\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAssessed by Clinician\u003c/b\u003e\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTemperature [IQR]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e37.1 [36.6–37.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36.8 [36.4–37.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt; 0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMeasured Fever (≥ 37.5\u003csup\u003eo\u003c/sup\u003eC)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e99 (52.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e342 (37.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt; 0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian Respiratory rate [IQR]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38.5 [34.0-42.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36.0 [31.5–39.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e\u0026lt; 0.001\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChest indrawing\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (2.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9 (1.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.180\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSunken eyes\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e177 (94.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e848 (92.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.255\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDry mouth\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e183 (97.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e903 (98.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.71\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSkin turgor (slow/very slow)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e78 (41.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e391 (42.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMental Status\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNormal\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e73 (39.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e380 (41.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.052\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRestless/Irritable\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e108 (57.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e530 (57.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLethargic/Unconscious\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (3.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9 (0.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRectal prolapse\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 (0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (0.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.434\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBipedal edema\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (1.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5 (0.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.337\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbnormal hair\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9 (4.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43 (4.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.937\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnder Nutrition\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21 (11.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e109 (11.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.807\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFlaky Skin\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (1.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5 (0.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.409\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSevere Acute Malnutrition (SAM)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24 (12.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e78 (8.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.061\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWasting\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13 (7.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39 (4.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.111\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAdmission\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e27 (14.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e87 (9.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.043\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiarrhea Duration (≥ 7 days)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e70 (37.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e327 (35.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.631\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eany_antibiotic\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e78 (41.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e402 (43.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.609\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRotavirus vaccination doses\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (1.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19 (2.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.385\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (4.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25 (3.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e166 (94.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e764 (94.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eORS at facility\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e186 (99.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e914 (99.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.311\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZinc at facility\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e183 (97.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e887 (96.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.494\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIV rehydration\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e31 (16.6)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e92 (10.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.01\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDehydration\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (4.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e35 (3.8)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.747\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSome\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e126 (67.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e645 (70.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSevere\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e53 (28.3)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e239 (26.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVesikari Score\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMild\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13 (7.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e71 (7.7)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.088\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModerate\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e75 (40.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e442 (48.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSevere\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e99 (52.9)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e406 (44.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian Vesikari score [IQR]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11 [\u003cspan additionalcitationids=\"CR10 CR11\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e–\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10 [\u003cspan additionalcitationids=\"CR9 CR10 CR11\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e–\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.0003\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDiagnosis\u003c/b\u003e\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=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDysentery\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10 (5.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e58 (6.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.605\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMalaria\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e85 (45.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e361 (39.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.131\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePneumonia\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12 (6.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e37 (4.1)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.152\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBacterial Infection\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14 (7.5)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e93 (10.2)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.258\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMalnutrition\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15 (8.0)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e68 (7.4)\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.784\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003e \u003csup\u003eβ−\u003c/sup\u003e Includes electricity, propane, butane, natural gas; SAM defined as WHZ \u0026lt; − 3 or MUAC \u0026lt; 115 millimeters, or the presence of bilateral pitting edema; ORS-Oral rehydration solution\u003c/p\u003e \u003cp\u003e*P-value computed using either chi-square or Fisher`s exact test were performed as appropriate for categorical variables and Wilcoxon rank sum tests were used to compare continuous variables\u003c/p\u003e \u003cp\u003eFrom the feature selection analysis, the confirmed variables in order of importance were age (16.6%), temperature (6.0%), respiratory rate (4.1%) and breastfeeding (3.3%). SAM (3.4%), rotavirus vaccination (3.3%), and skin turgor (2.1%) were tentative features (Fig.\u0026nbsp;2).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eGreen, yellow, red and blue boxplots represent the Z scores of confirmed, tentative, rejected and shadow features, respectively.\u003c/p\u003e \u003cp\u003eConfirmed and tentative features: \u003cem\u003eAge; temperature; respiratory rate; severe acute malnutrition (SAM); rotavirus vaccination; breastfeeding; skin turgor\u003c/em\u003e\u003c/p\u003e \u003cp\u003eFigure 2. Feature selection for linear growth faltering among children aged \u0026lt; 5 years presenting with moderate to severe diarrhea in rural western Kenya, 2015–2018\u003c/p\u003e \u003cp\u003eIn addition to age, respiratory rate, temperature and breastfeeding, the following features were selected: confirmed (stunting at baseline [5.2%], vomit [4.0%], Vesikari score (3.7%) and sunken eyes [3.6%]) and tentative (bacterial infection diagnosis [2.5%]) in the sensitivity analysis using a cut-off of negative change in HAZ (Figure S2).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eModel Performance\u003c/h2\u003e \u003cp\u003eWe evaluated seven ML algorithms in the prediction of LGF. From the developed models, sensitivity was highest in the RF model (80.7%), followed by the ANN (79.5%), SVM (77.3%), NB (76.5%), GBM (75.6%), LR (75.4%) and lowest in the KNN model (72.4%). The specificity ranged from 58.2–71.8%. Specifically, the specificity of the GBM model was the highest (71.8%), followed by RF (70.1%), LR (61.9%), NB and SVM (61.6%), KNN (61.4%) and lowest\u003c/p\u003e \u003cp\u003ein the ANN model (58.2%). The PPV ranged between 27.4% − 34.9% while the NPV ranged between 92.3% − 94.8%. The AUC of the models ranged from 73.4–83.5% with the GBM model having the highest AUC (83.5%, 95% Confidence Interval [95% CI]: 81.6–85.4) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\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=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" 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\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\u003eModel performance of linear growth faltering prediction\u003csup\u003eβ\u003c/sup\u003e models using combined data (Original and synthetic data)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlgorithm\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSensitivity % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSpecificity % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePPV % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNPV % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eF1-Score [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAUC % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePRAUC % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e80.7 [76.5–84.4]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e70.1 [68.1–72.1]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e34.9 [31.9–38.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e94.8 [93.6–95.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e48.7 [16.8–59.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e82.8 [80.8–84.8]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e96.0 [93.8–96.2]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGBM\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e75.6 [71.2–79.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e71.8 [69.8–73.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e34.7 [31.6–37.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e93.7 [92.4–94.8]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e47.6 [13.9–74.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e83.5 [81.6–85.4]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e96.2 [94.9–96.5]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNB\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e76.1 [71.7–80.1]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e61.6 [59.5–63.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e28.2 [25.6–31.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e92.8 [91.4–94.1]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e40.2 [12.0-42.4]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e75.6 [73.3–77.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e94.0 [92.1–95.0]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e75.4 [70.9–79.4]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e61.9 [59.7–64.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e28.2 [25.5–30.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e92.7 [91.2–94.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e38.2 [3.1–64.2]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e73.7 [71.3–76.1]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e93.0 [91.1–94.0]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSVM\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e77.3 [73.0-81.2]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e61.6 [59.5–63.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e28.6 [25.9–31.3]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e93.2 [91.7–94.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e41.7 [9.3–56.8]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e73.4 [71.0-75.8]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e93.0 [91.6–94.1]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKNN\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e72.4 [69.7–75.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e61.4 [59.3–63.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e27.6 [25.0-30.3]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e92.3 [90.8–93.6]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e40.2 [6.7–67.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e74.8 [72.3–77.2]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e93.0 [90.8–93.6]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eANN\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e79.5 [75.3–83.3]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e58.2 [56.1–60.4]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e27.4 [24.9–30.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e93.5 [92.0-94.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e40.8 [9.8–58.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e73.6 [71.3–76.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e93.0 [90.9–94.1]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"8\"\u003e\u003csup\u003eβ−\u003c/sup\u003e Linear growth faltering defined as Δ HAZ ≥ − 0.5\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd colspan=\"8\"\u003e*RF-Random Forest; GBM-Gradient Boosting; NB- Naïve Bayes; LR-Logistic Regression; SVM- Support vector machine; KNN-K-nearest neighbors; ANN-Artificial Neural Networks;\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003e95% CI- 95% Confidence Interval; PPV- Positive Predictive Value; NPV- Negative Predictive Value; AUC- Area under the Curve; PRAUC- Precision Recall Area under the Curve\u003c/p\u003e \u003cp\u003eThe GBM model was chosen as the champion model. The receiver operating characteristic (ROC) curves for LGF prediction models are shown in Figure S3. Moreover, in the sensitivity analysis using only the VIDA data in development, the model performance ranged between 63.0%-82.6%, 55.9%-78.6%, 27.3%-33.7%, 91.0%-94.2%, 40.3%-44.3%, 68.0%-75.5%, and 90.6%-94.4% for sensitivity, specificity, PPV, NPV, F1-score, AUC and PRAUC, respectively (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). All models showed a decline in predictive performance during sensitivity analysis except for the SVM model, which had a marginal increase.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\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=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" 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\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eModel performance of linear growth faltering prediction \u003csup\u003eβ\u003c/sup\u003e models using original training data only\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlgorithm\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSensitivity % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSpecificity % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePPV % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNPV % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eF1-Score [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAUC % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePRAUC % [95% CI]\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e52.2 [36.9–67.1]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e78.6 [72.7–83.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e32.9 [22.3–44.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e89.1 [84.0–93.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e40.3 [10.9–51.4]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e70.3 [61.8–78.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e90.6 [88.4–90.8]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGBM\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e80.4 [66.1–90.6]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e63.3 [56.7–69.6]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e30.6 [22.5–39.6]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e94.2 [89.2–97.3]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e44.3 [12.9–55.8]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e75.5 [68.2–82.8]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e93.6 [92.3–93.9]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNB\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e63.0 [47.5–76.8]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e75.1 [69.0-80.6]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e33.7 [23.9–44.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e91.0 [86.0-94.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e43.9 [5.6–60.3]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e73.6 [66.1–81.2]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e93.0 [91.1–94.0]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e73.9 [58.9–85.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e63.3 [56.7–69.6]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e28.8 [20.8–37.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e92.4 [87.0–96.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e41.5 [7.8–52.1]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e73.8 [67.0-80.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e93.9 [92.0-94.9]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSVM\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e71.7 [56.5–84.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e65.9 [59.4–72.1]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e29.7 [21.4–39.1]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e92.1 [86.8–95.7]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e42.0 [7.2–51.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e75.2 [68.9–81.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e94.4 [93.0-95.5]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKNN\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e82.6 [68.6–92.2]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e56.8 [50.1–63.3]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e27.7 [20.4–36.0]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e94.2 [88.9–97.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e41.5 [12.2–51.8]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e73.1 [66.3–79.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e93.6 [91.4–94.2]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eANN\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e82.6 [68.6–92.2]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e55.9 [49.2–62.4]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e27.3 [20.1–35.5]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e94.1 [88.7–97.4]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e41.1 [12.0-57.9]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e68.0 [60.5–75.6]\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e91.4 [89.3–92.5]\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"8\"\u003e\u003csup\u003eβ−\u003c/sup\u003e Linear growth faltering defined as Δ HAZ ≥ − 0.5\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd colspan=\"8\"\u003e*RF-Random Forest; GBM-Gradient Boosting; NB- Naïve Bayes; LR-Logistic Regression; SVM- Support vector machine; KNN-K-nearest neighbors; ANN-Artificial Neural Networks;\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003e95% CI- 95% Confidence Interval; PPV- Positive Predictive Value; NPV- Negative Predictive Value; AUC- Area under the Curve; PRAUC- Precision Recall Area under the Curve\u003c/p\u003e \u003cp\u003eIn the sensitivity analysis using the second definition of LGF (negative change in HAZ), the model performance ranged between 45.8%-73.1%, 53.2%-76.6%, 79.0%-90.5%, 28.6%-48.5%, 58.3%-80.9%, 58.0%-82.4%, and 29.0%-62.6% for sensitivity, specificity, PPV, NPV, F1-score, AUC and PRAUC, respectively (Table S2). In this scenario, all models exhibited a drop in predictive performance except for the SVM model, which had a marginal increase and the RF model which registered same performance as in the primary analysis.\u003c/p\u003e \u003cp\u003eOverall the Brier scores were relatively high and ranged between 0.19–2.50 (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).The Spiegelhalter’s p-value showed that all the models were not properly calibrated (p \u0026lt; 0.05). The performance of the calibrated GBM model was largely similar to its uncalibrated form with the model having an AUC of 83.7%.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\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\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCalibration results of linear growth faltering prediction models.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlgorithm\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBrier Score\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSpiegelhalter Z-score\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSpiegelhalter p-value\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.19\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e16.83\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt; 0.0001\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGBM\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.50\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e208.10\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt; 0.0001\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNB\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.18\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e101.02\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt; 0.0001\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.16\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e85.02\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt; 0.0001\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSVM\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.16\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e85.88\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt; 0.0001\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKNN\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.21\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e109.88\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt; 0.0001\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eANN\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.17\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e84.07\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt; 0.0001\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"4\"\u003e*RF-Random Forest; GBM-Gradient Boosting; NB- Naïve Bayes; LR-Logistic Regression; SVM- Support vector machine; KNN-K-nearest neighbors; ANN-Artificial Neural Networks;\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eExplanatory Model Analysis\u003c/h2\u003e \u003cp\u003eThe EMA results for the top 2 models in the primary analysis were similar though the degree of importance varied across models with no SAM, no skin turgor, no rotavirus vaccine, age, elevated temperature and respiratory rate being predictive of LGF (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Similarly, in the sensitivity analysis using the second definition of LGF, the direction of association was similar between the two models although the magnitude of importance varied. In addition to age, respiratory rate and temperature, the following factors were also identified to be predictive of LGF: severity of disease, no vomiting, stunting at baseline, bacterial infection and lack of sunken eyes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eBusiness Value Evaluation of Champion Model\u003c/h2\u003e \u003cp\u003eFrom the business value evaluation of our champion model (GBM), the cumulative gains plot shows that the model is able to select ~ 60% of the target class (LGF) if we select the top-20% cases based on our model. Additionally, from the cumulative lift plot, our champion model is able to identify ~ 3 times higher number of the target class compared to a random selection if we pick the top-20% observations based on model probability. Lastly, from the cumulative response plot, 48% of observations in the top-20% cases based on model probability belong to the target class (Fig.\u0026nbsp;4).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e*Scenario 1- Predicting linear growth faltering using a cut-off of Δ HAZ ≥ − 0.5\u003c/p\u003e \u003cp\u003e* age = 9: 9 months; Rotavirus_vacc = 2:2 doses of rotavirus vaccine; cur_wrinkledskin = 0: normal skin; SAM = 0: No severe acute malnutrition (SAM)\u003c/p\u003e \u003cp\u003e*Scenario 2- Predicting linear growth faltering using change in haz/month (negative change in linear growth is deemed growth faltering)\u003c/p\u003e \u003cp\u003e*age = 9: 9 months; vesikari_cat = 3: Severe disease based on Vesikari score; vomit = 1: Vomitting; Stunting_base = 0: No stunting at baseline; bacterial_infec = 0: No bacterial infection; sunken_eyes = 1: sunken eyes.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure 4. Business value plots for the Gradient Boosting (GBM) Model for linear growth faltering\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eTemporal Validation\u003c/h2\u003e \u003cp\u003eWe observed a decline in model performance on the temporal validation dataset with the AUC dropping by ~ 18%. Additionally, all metrics dropped in temporal validation with the GBM model achieving 53.7%, 67.7%, 32.5%, 83.5%, 40.5%, 65.6% and 86.4% for sensitivity, specificity, PPV, NPV, F1-score, AUC and PRAUC, respectively (Fig.\u0026nbsp;5).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003ePPV- Positive Predictive Value; NPV- Negative Predictive Value; AUC- Area under the Curve; PRAUC- Precision Recall Area under the Curve\u003c/p\u003e \u003cp\u003eFigure 5. Performance of champion model in development (2015–2018) and temporal validation (2022–2023) datasets.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe study findings illuminate a comprehensive exploration into the prediction of LGF among pediatric patients presenting with diarrhea, employing a robust ML framework. The study involved the development and temporal validation of predictive models using diverse cohorts, revealing distinct prevalence rates and influencing factors associated with LGF. Key features linked to this outcome, such as age, breastfeeding, rotavirus vaccination, respiratory rate, temperature and SAM, were identified through extensive feature selection and their impact on risk prediction was estimated using SHAP attribution. The ML algorithms exhibited varying performance with GBM model emerging as the champion model, demonstrating promising business value. However, the temporal validation uncovered a notable decline in model performance, emphasizing the dynamic nature of health data and the need for ongoing model evaluation and adaptation. This discussion delves into the nuanced interpretation of these results, shedding light on the implications for predictive modeling in the context of pediatric diarrheal outcomes and the broader landscape of healthcare.\u003c/p\u003e \u003cp\u003eDespite the impact of rotavirus vaccine introduction on the epidemiology of diarrhea and pathogen landscape, we identified similar predictors, in addition to rotavirus vaccination, to previous modelling efforts [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] that used data collected pre-vaccine introduction─ age, breastfeeding, respiratory rate, temperature, absence SAM and stunting at baseline. This finding underscores the enduring importance of these risk factors and the need for comprehensive, sustained, and adaptable public health strategies to combat LGF. Furthermore, we observed that rotavirus vaccination was inversely associated with LGF a finding that is consistent with those of Loli and Carcamo who studied the impact of vaccination on HAZ in Peruvian children aged 6\u0026ndash;60 months [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. This finding could be due to rotavirus vaccination substantially reducing the incidence and severity of rotavirus infections, curbing the immediate impact of diarrheal diseases on nutrient absorption and consequently diarrhea-mediated growth faltering [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Bolstering rotavirus vaccination is a possible strategy that could be leveraged by policy makers and public health experts to reduce stunting in such settings. Moreover, from a modelling perspective, this finding on predictors generates confidence in the relevance and stability of these variables in different contexts and epidemiological periods, enhancing model transferability and generalizability.\u003c/p\u003e \u003cp\u003eThese variables have been documented as risk factors for LGF in previous studies. Specifically, age is a significant determinant of LGF among pediatric populations following a diarrheal episode [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Infants and very young children face heightened vulnerability to nutritional and health challenges due to their ongoing physiological maturation, which is exacerbated during diarrheal illness leading to pronounced impacts on nutrient absorption and utilization cumulatively contributing to a heightened prevalence of LGF among younger children. Stunting has been shown to be irreversible to a large extent after reaching 24 months of age [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Therefore, the timely identification of at-risk children (infants and toddlers) facilitates the implementation of effective preventive strategies during this critical window of opportunity in early childhood. Moreover, breastfeeding has been shown to be a protective factor against LGF in infants by providing essential nutrients, antibodies, fostering a healthy gut microbiome, and reducing exposure to environmental and food contamination [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Its nutrient-dense composition supports optimal growth and development, while immune protection reduces the risk of infections that can impede growth.\u003c/p\u003e \u003cp\u003eContrary to existing evidence [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e], we observed children without SAM to be at increased risk of LGF. Despite majority of factors predisposing children to SAM and stunting being similar, we observed a discordant relationship between the two and this may require further investigation to gain insights into this finding. Elevated baseline temperature and respiratory rate signal are markers of disease severity, and particularly those affecting the gastrointestinal tract, may lead to nutritional deficiencies and hinder linear growth [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Additionally, elevated respiratory rate and temperature may indicate increased energy expenditure, potentially due to the body's efforts to combat infections or inflammation. This increased energy demand can divert resources away from growth-related processes, impacting linear growth.\u003c/p\u003e \u003cp\u003eTree-based ensembles showed good predictive performance with the GBM model narrowly outperforming the RF model in the prediction of LGF. Our champion model outperformed existing models by Brander et al. (AUC\u0026thinsp;=\u0026thinsp;67.0%) [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e] and (AUC\u0026thinsp;=\u0026thinsp;75.0%) Ahmed et al. [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. The improvement in model performance could be attributed to the robust modelling approach employed. Moreover, the predictive prowess of tree-based ensembles may have also contributed to this improvement. This strong discriminatory ability of the champion model has significant public health implications as it reinforces the feasibility and efficacy of ML algorithms in timely identification of children, at increased risk of LGF, for early nutritional and healthcare interventions. The model can enhance the efficiency of resource allocation by facilitating targeted screening as well as providing healthcare providers with a valuable tool for informed decision-making, enabling tailored interventions based on individual children risk profiles. However, the decline in model performance during temporal validation while consistent with findings from Ahmed et al. [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] raises important considerations. Spectral differences in the severity of diarrhea among children in the development and validation cohorts, coupled with potential shifts in the study population over time, highlight challenges in maintaining consistent predictive accuracy. This finding highlights the need for monitoring and periodic retraining of the model in order to maintain its predictive performance.\u003c/p\u003e \u003cp\u003eOur primary analysis that used combined data (VIDA and synthetic data) in model development had better performance than the sensitivity analysis that only used VIDA data. This result emphasizes the importance of synthetic data in addressing challenges associated with imbalanced, limited, or privacy-sensitive real-world datasets, providing a means to augment and diversify the data pool [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. This approach overcomes issues of data scarcity, facilitates more comprehensive model training, and enhances generalization. It contributes to overcoming biases, ensuring model fairness, and accommodating the complexity of risk factors influencing a health outcome. Ultimately, the strategic use of synthetic data strengthens the reliability, generalizability, and ethical integrity of predictive models, offering a pathway for more effective and personalized healthcare interventions. However, synthetic data may advance bias propagation since any biases in the primary data will be reflected in the generated data and this may perpetuate and even exacerbate healthcare disparities if they exist [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. Furthermore, in the second sensitivity analysis using a cutoff of any negative change in HAZ, we observed a substantial decline in model performance compared to using a cutoff of a decrease of 0.5 HAZ or more. These results imply that using a specific cutoff criteria for defining LGF can significantly impact the performance of the predictive model. Different cutoff criteria may be more appropriate in different contexts, and the choice should be informed by clinical expertise and relevance considering the specific context of the healthcare setting, study population (varying age categories), and the clinical significance of HAZ changes. It also underscores the dynamic nature of model performance, necessitating ongoing evaluation and adaptation to maintain optimal cutoff criteria.\u003c/p\u003e \u003cp\u003eOur study, while commendable, has limitations, notably the exclusion of pathogen data during model development to maintain practical applicability, despite its influence on LGF. Future research should address this gap, as well as focus on the acceptability and impact of ML models on clinical practice and patient outcomes. The cost-effectiveness of deploying these models is also crucial for practical implementation in diverse healthcare settings. Exploring these facets will contribute significantly to enhancing understanding and ensuring the effective use of ML models in healthcare.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThe study's findings emphasize the enduring relevance of established predictors of LGF. Addressing multifaceted challenges in pediatric LGF requires sustained efforts with adaptive interventions for these risk factors. The study demonstrates the practical use of ML algorithms for rapid identification of at-risk children. A decline in model performance during temporal validation highlights the dynamic nature of health data, necessitating continuous evaluation and adaptation. Additionally, the study shows the viability of integrating synthetic data to enhance model robustness, providing a pathway for more comprehensive and ethical predictive modeling in healthcare.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e \u003cp\u003e The VIDA protocol was approved by the Institutional Review Board of the University of Maryland School of Medicine, Baltimore, MD, USA (UMB Protocol #: HM-HP-00062472) and the Kenya Medical Research Institute (KEMRI) Scientific and Ethical Review Unit (SERU) (SERU#2996). The EFGH protocol was approved by the KEMRI SERU (SERU#4362). Written informed consent was sought from caregivers in both studies before initiation of study procedures. Additionally, ethical approval for undertaking the current study was sought from the health research ethics committee of the University of South Africa, College of Agricultural Sciences (2023/CAES_HREC/2192).\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConsent for publication\u003c/strong\u003e \u003cp\u003eNot applicable.\u003c/p\u003e \u003cp\u003e \u003ch2\u003eCompeting Interest\u003c/h2\u003e \u003cp\u003eAuthors declare no conflict of interest.\u003c/p\u003e \u003ch2\u003eDisclosure\u003c/h2\u003e \u003cp\u003eThe findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Kenya Medical Research Institute or any collaborating institutions.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis work was supported by the Bill \u0026amp; Melinda Gates Foundation (grant INV-045988). The funders did not play any role in the study and interpretation of its outcome.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eBO, VHM, KDT, PBP and RO conceived the study and contributed to study design and implementation. BO, VHM and KDT analyzed and interpreted the data. BO drafted the manuscript and all authors critically reviewed the manuscript for intellectual content and approved the final manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003eWe appreciate the contributions and efforts of KEMRI-CGHR staff involved in the data collection, data management, and laboratory testing of samples in the two studies. We also wish to thank the study participants and the ministry of health staff for supporting both studies. Moreover, we would like to acknowledge the use of artificial intelligence (AI) technology for grammar checking and proofreading of this manuscript.\u003c/p\u003e\u003ch2\u003eAvailability of data and materials\u003c/h2\u003e \u003cp\u003eThe data used for the modelling in this study belongs to KEMRI and restrictions apply to the availability of these data. Data cleaning, pre-processing and model development were done in R version 4.1.2. The programming code for R is available upon request addressed to the corresponding author: Billy Ogwel ([email protected]).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eWorld Health Organization. Diarrhoeal disease. 2017. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/news-room/fact-sheets/detail/diarrhoeal-disease\u003c/span\u003e\u003cspan address=\"https://www.who.int/news-room/fact-sheets/detail/diarrhoeal-disease\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 19 February 2022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFerdous F, Das SK, Ahmed S, et al. Severity of Diarrhea and Malnutrition among Under Five-Year-Old Children in Rural Bangladesh. Am J Trop Med Hyg. 2013;89:223\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheckley W, Buckley G, Gilman RH, et al. Multi-country analysis of the effects of diarrhoea on childhood stunting. Int J Epidemiol. 2008;37:816.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLenters L, Wazny K, Bhutta ZA. Management of Severe and Moderate Acute Malnutrition in Children. In: Black RE, Laxminarayan R, Temmerman M, Walker N, eds. Reproductive, Maternal, Newborn, and Child Health: Disease Control Priorities, Third Edition (Volume 2). Washington (DC): The International Bank for Reconstruction and Development / The World Bank, 2016. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.ncbi.nlm.nih.gov/books/NBK361900/\u003c/span\u003e\u003cspan address=\"http://www.ncbi.nlm.nih.gov/books/NBK361900/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 27 November 2020.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDanaei G, Andrews KG, Sudfeld CR, et al. Risk Factors for Childhood Stunting in 137 Developing Countries: A Comparative Risk Assessment Analysis at Global, Regional, and Country Levels. PLoS Med. 2016;13:e1002164.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWierzba TF, Muhib F. Exploring the broader consequences of diarrhoeal diseases on child health. Lancet Global Health. 2018;6:e230\u0026ndash;1.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuerrant RL, DeBoer MD, Moore SR, Scharf RJ, Lima AAM. The impoverished gut\u0026ndash;a triple burden of diarrhoea, stunting and chronic disease. Nat Rev Gastroenterol Hepatol. 2013;10:220\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eROTA Council. The broader impact of early childhood diarrhea. 2019. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://preventrotavirus.org/wp-content/uploads/2019/05/ROTA-Brief6-LastingImpact-SP-1-3.pdf\u003c/span\u003e\u003cspan address=\"https://preventrotavirus.org/wp-content/uploads/2019/05/ROTA-Brief6-LastingImpact-SP-1-3.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 1 November 2022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrander RL, Pavlinac PB, Walson JL, et al. Determinants of linear growth faltering among children with moderate-to-severe diarrhea in the Global Enteric Multicenter Study. BMC Med. 2019;17:214.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAhmed SM, Brintz BJ, Pavlinac PB, et al. Derivation and external validation of clinical prediction rules identifying children at risk of linear growth faltering. Elife. 2023;12:e78491.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRahmani K, Thapa R, Tsou P, et al. Assessing the effects of data drift on the performance of machine learning models used in clinical sepsis prediction. Int J Med Informatics. 2023;173:104930.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSahiner B, Chen W, Samala RK, Petrick N. Data drift in medical machine learning: implications and potential remedies. Br J Radiol 2023;:20220878.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePowell H, Liang Y, Neuzil KM, et al. A Description of the Statistical Methods for the Vaccine Impact on Diarrhea in Africa (VIDA) Study. Clin Infect Dis. 2023;76:S5\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNasrin D, Liang Y, Powell H, et al. Moderate-to-Severe Diarrhea and Stunting Among Children Younger Than 5 Years: Findings From the Vaccine Impact on Diarrhea in Africa (VIDA) Study. Clin Infect Dis. 2023;76:S41\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNowok B, Raab GM, Dibben C. synthpop: Bespoke Creation of Synthetic Data in \u003cem\u003eR\u003c/em\u003e. J Stat Soft. 2016; 74. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.jstatsoft.org/v74/i11/\u003c/span\u003e\u003cspan address=\"http://www.jstatsoft.org/v74/i11/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 23 September 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJoseph VR. Sci J. 2022;15:531\u0026ndash;8. Optimal ratio for data splitting. Statistical Analysis and Data Mining: The ASA Data.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNasrin D, Blackwelder WC, Sommerfelt H, et al. Pathogens Associated With Linear Growth Faltering in Children With Diarrhea and Impact of Antibiotic Treatment: The Global Enteric Multicenter Study. J Infect Dis. 2021;224:S848\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Buuren S, Groothuis-Oudshoorn K, Vink G et al. Package \u0026lsquo;mice\u0026rsquo;. 2021. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cran.r-project.org/web/packages/mice/mice.pdf\u003c/span\u003e\u003cspan address=\"https://cran.r-project.org/web/packages/mice/mice.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 31 May 2021.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKursa MB, Rudnicki WR. Package \u0026lsquo;Boruta\u0026rsquo;. 2020. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cran.r-project.org/web/packages/Boruta/Boruta.pdf\u003c/span\u003e\u003cspan address=\"https://cran.r-project.org/web/packages/Boruta/Boruta.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 31 May 2021.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRefaeilzadeh P, Tang L, Liu H. Cross-Validation. In: LIU L, \u0026Ouml;ZSU MT,Encyclopedia of Database Systems. Boston, Springer MA. US, 2009: 532\u0026ndash;538. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-0-387-39940-9_565\u003c/span\u003e\u003cspan address=\"10.1007/978-0-387-39940-9_565\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 17 October 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBach M, Werner A, Żywiec J, Pluskiewicz W. The study of under- and over-sampling methods\u0026rsquo; utility in analysis of highly imbalanced data on osteoporosis. Inf Sci. 2017;384:174\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaito T, Rehmsmeier M. precrec: Calculate Accurate Precision-Recall and ROC (Receiver Operator Characteristics) Curves. 2023; Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://CRAN.R-project.org/package=precrec\u003c/span\u003e\u003cspan address=\"https://CRAN.R-project.org/package=precrec\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 10 February 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang Y, Li W, Macheret F, Gabriel RA, Ohno-Machado L. A tutorial on calibration measurements and calibration models for clinical prediction models. J Am Med Inform Assoc. 2020;27:621\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBiecek P, Maksymiuk S, Baniecki H, DALEX. : moDel Agnostic Language for Exploration and eXplanation. 2023; Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://CRAN.R-project.org/package=DALEX\u003c/span\u003e\u003cspan address=\"https://CRAN.R-project.org/package=DALEX\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 10 February 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCowley LE, Farewell DM, Maguire S, Kemp AM. Methodological standards for the development and evaluation of clinical prediction rules: a review of the literature. Diagn Prognostic Res. 2019;3:16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNagelkerke J. modelplotr: Plots to evaluate the business value of predictive models. 2020; Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cran.r-project.org/web/packages/modelplotr/vignettes/modelplotr.html\u003c/span\u003e\u003cspan address=\"https://cran.r-project.org/web/packages/modelplotr/vignettes/modelplotr.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 19 November 2022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR Core Team. R: The R Project for Statistical Computing. 2021. Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.r-project.org/\u003c/span\u003e\u003cspan address=\"https://www.r-project.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 3 December 2021.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLoli S, Carcamo CP. Rotavirus vaccination and stunting: Secondary Data Analysis from the Peruvian Demographic and Health Survey. Vaccine. 2020;38:8010\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBenjamin-Chung J, Mertens A, Colford JM, et al. Early-childhood linear growth faltering in low- and middle-income countries. Nature. 2023;621:550\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVictora CG, de Onis M, Hallal PC, Bl\u0026ouml;ssner M, Shrimpton R. Worldwide timing of growth faltering: revisiting implications for interventions. Pediatrics. 2010;125:e473\u0026ndash;480.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSilverberg SL, Qamar H, Keya FK, et al. Do Early Infant Feeding Practices and Modifiable Household Behaviors Contribute to Age-Specific Interindividual Variations in Infant Linear Growth? Evidence from a Birth Cohort in Dhaka, Bangladesh. Curr Developments Nutr. 2021;5:nzab077.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKramer MS, Kakuma R. Optimal duration of exclusive breastfeeding. Cochrane Database Syst Rev. 2012; 2012:CD003517.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNgari MM, Iversen PO, Thitiri J, et al. Linear growth following complicated severe malnutrition: 1-year follow-up cohort of Kenyan children. Arch Dis Child. 2019;104:229\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGarenne M, Myatt M, Khara T, Dolan C, Briend A. Concurrent wasting and stunting among under-five children in Niakhar, Senegal. Matern Child Nutr. 2018;15:e12736.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGiuffr\u0026egrave; M, Shung DL. Harnessing the power of synthetic data in healthcare: innovation, application, and privacy. npj Digit Med. 2023;6:1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGonzales A, Guruswamy G, Smith SR. Synthetic data in health care: A narrative review. PLOS Digit Health. 2023;2:e0000082.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarwala T, Fournier-Tombs E, Stinckwich S. The Use of Synthetic Data to Train AI Models: Opportunities and Risks for Sustainable Development. 2023.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Machine Learning, Linear growth faltering, Pediatric, Diarrhea, Prediction","lastPublishedDoi":"10.21203/rs.3.rs-4047381/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4047381/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eIntroduction:\u003c/h2\u003e \u003cp\u003eStunting affects one-fifth of children globally with diarrhea accounting for an estimated 13.5% of stunting. Identifying risk factors for its precursor, linear growth faltering (LGF), is critical to designing interventions. Moreover, developing new predictive models for LGF using more recent data offers opportunity to improve model performance and capture new insights. We employed machine learning (ML) to derive and validate a predictive model for LGF among children enrolled with diarrhea in the Vaccine Impact on Diarrhea in Africa (VIDA) study and the Enterics for Global Heath (EFGH) ― Shigella study in rural western Kenya.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe used 7 ML algorithms to retrospectively build prognostic models for the prediction of LGF (\u0026ge;\u0026thinsp;0.5 decrease in height/length for age z-score [HAZ]) among children 6\u0026ndash;35 months. We used de-identified data from the VIDA study (n\u0026thinsp;=\u0026thinsp;1,473) combined with synthetic data (n\u0026thinsp;=\u0026thinsp;8,894) in model development, which entailed split-sampling and K-fold cross-validation with over-sampling technique, and data from EFGH-Shigella study (n\u0026thinsp;=\u0026thinsp;655) for temporal validation. Potential predictors included demographic, household-level characteristics, illness history, anthropometric and clinical data chosen using an explainable model agnostic approach. The champion model was determined based on the area under the curve (AUC) metric.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe prevalence of LGF in the development and temporal validation cohorts was 187 (16.9%) and 147 (22.4%), respectively. The following variables were associated with LGF in decreasing order: age (16.6%), temperature (6.0%), respiratory rate (4.1%), SAM (3.4%), rotavirus vaccination (3.3%), breastfeeding (3.3%), and skin turgor (2.1%). While all models showed good prediction capability, the gradient boosting model achieved the best performance (AUC% [95% Confidence Interval]: 83.5 [81.6\u0026ndash;85.4] and 65.6 [60.8\u0026ndash;70.4] on the development and temporal validation datasets, respectively).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eOur findings accentuates the enduring relevance of established predictors of LGF whilst demonstrating the practical utility of ML algorithms for rapid identification of at-risk children.\u003c/p\u003e","manuscriptTitle":"Predictive Modelling of Linear Growth Faltering Among Pediatric Patients with Diarrhea in Rural Western Kenya: An Explainable Machine Learning Approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-03-15 19:44:29","doi":"10.21203/rs.3.rs-4047381/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-10-22T10:26:28+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-02T08:45:17+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"180587170332839588641944006782516036976","date":"2024-08-30T12:20:02+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-06-20T15:01:02+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"151803615612659092374797845696463664326","date":"2024-06-13T10:53:54+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-05-31T07:35:57+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-03-13T08:07:16+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-03-13T08:03:07+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-03-13T08:03:07+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Informatics and Decision Making","date":"2024-03-08T16:58:07+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e58322ec-5cbe-4afc-88d4-8476210dde94","owner":[],"postedDate":"March 15th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-12-09T16:00:48+00:00","versionOfRecord":{"articleIdentity":"rs-4047381","link":"https://doi.org/10.1186/s12911-024-02779-7","journal":{"identity":"bmc-medical-informatics-and-decision-making","isVorOnly":false,"title":"BMC Medical Informatics and Decision Making"},"publishedOn":"2024-12-02 15:57:18","publishedOnDateReadable":"December 2nd, 2024"},"versionCreatedAt":"2024-03-15 19:44:29","video":"","vorDoi":"10.1186/s12911-024-02779-7","vorDoiUrl":"https://doi.org/10.1186/s12911-024-02779-7","workflowStages":[]},"version":"v1","identity":"rs-4047381","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4047381","identity":"rs-4047381","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}