An Actuarial Perspective on the Discrepancies in Health Insurance Pricing: Evidence from India

An Actuarial Perspective on the Discrepancies in Health Insurance Pricing: Evidence from India | 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 An Actuarial Perspective on the Discrepancies in Health Insurance Pricing: Evidence from India Nikhil Rathee, Rupel Nargunam This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6613076/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study investigates the actuarial fairness of pricing health insurance policies by examining discrepancies within the Indian health insurance and their impact on medical costs. Health insurance helps an individual in financing their medical expenses, for a fee called premium which is typically the reflection of expected cost of medical services. However, discrepancies such as misinformation, accessibility to health care services, hospital quality and Inconsistencies in claims processing, increase costs associated with health care of individuals participating in the health insurance, thereby challenging the fairness of these policies. Structural Equation Modelling (SEM) has been employed to develop latent variables representing these discrepancies and Hierarchical Linear Modelling (HLM) has been used to assess their effect on the cost of medical care. The findings of this study suggest the presence of discrepancies in the Indian health insurance sector and the results support the significant impact on the increase in medical expenses. The study concludes with policy recommendations aimed at enhancing the efficiency, effectiveness and fairness of the health insurance policies in India Pricing Health Insurance Actuarial Fairness Discrepancies Misinformation Claim process inconsistencies Access to Medical Services Figures Figure 1 Figure 2 Figure 3 Figure 4 1. INTRODUCTION Health insurance is a form of collectivism, by construct it is designed to enable individuals to pool their risk of incurring medical expenses. Since the health insurance acts as a binding agreement between the insurer and insured, its realised value is influenced by the multiple factors such as unequal distribution of medical resources, lack of human resources, poor infrastructure, and low-quality medical services. These factors contribute to the disparities in delivery of health care under health insurance and overthrow the underlying principle of designing health insurance contracts. Data from the Indian Human Development Survey-II (IHDS-II) reveals that approximately 40.8 percent of households with health insurance have incurred additional medical expenses that exceed the risk cover. In 2011, India had a doctor-to-population ratio of 1:1457 which is lower than the WHO’s recommended ratio of 1:1000 (Deo, 2013 ). This is emphasizing the critical shortage of human resources in the health sector. Furthermore, India faces numerous obstacles in its health insurance sector, such as resource distribution, infrastructure problems and administrative inefficiencies (Mavalankar et al., 2000) which disrupt the delivery of quality care under health insurance schemes. These unaddressed challenges and associated costs distort the principle of indemnification by which insurance contracts are designed, resulting in health insurance contracts where the expected cost of premiums does not match the true expected cost of medical care. The objective of this article is to investigate the aforementioned discrepancies present in the Indian health insurance network by examining the nature of these factors individually and highlighting their impact on the cost of medical care for insured households. A comprehensive analysis of these discrepancies is expected to help the pricing actuary to determine the expected costs more efficiently and determine premium rates which are in line with the fundamental principles of insurance contracts. This study focuses on examining the individual impact of each of these factors on expected cost of health care, to account for the variations in costs incurred by individuals from different socio-economic backgrounds. It is crucial as the presence discrepancies cause fluctuations in medical costs, which would in turn indicate that health insurance contracts are not priced accurately and may tend to deviate from the principle of actuarial fairness. 1.1 Pricing Health Insurance and Actuarial Fairness To comprehend actuarial fairness in health insurance, it is essential to understand the significance of the definition of health insurance Pitacco, E. (2014). Health insurance is described as a tool for transferring the risk of healthcare costs to another party in exchange for single or series of payments, called insurance premium. The determinisation of premium is of critical importance and should be determined with fairness. Donahue and Barocas (2021, p. 186) define the cost of insurance as "pricing in which each participant in the insurance pool pays their expected costs." In other words, for health insurance to be actuarially fair, the premium paid in advance should be equal to the expected cost of healthcare. 1.2 Discrepancies Discrepancies are unexpected differences that suggest something is wrong between two sets of conditions and needs to be explained. A study at the time when foreign direct investment in the Indian health insurance industry was severely restricted, highlights challenges such as the distribution of resources, improper infrastructure, poor healthcare quality, administrative inefficiencies, lack of awareness, and adverse selection (Mavalankar and Bhat 2000). In this study, we focus on factors that deviate the cost of health insurance from being actuarially fair. More specifically, we identify qualitative variables that are expected to have an impact on the pricing of health care costs. In that regard, this study highlights four discrepancies in the delivery of health care under health insurance schemes in the Indian context, namely: Inconsistencies in claims processing, misinformation, quality of medical care, and access to medical services. The following subsections give detailed descriptions of the discrepancies addressed by this study: 1.2.1 Misinformation Inconsistencies in claims processing and misinformation are not uncommon in the health insurance environment. They are observed in the issuance of insurance and during claim submissions. Misinformation arises from differences in knowledge between individuals and insurers, leading to misunderstandings about coverage and benefits. During claim submission, insured individuals experience inconvenience and dissatisfaction, which result in the ongoing insurance policy being overpriced or risk cover going underutilized regarding their actual risk experience. Leinsdorf et al. (1980) suggest that misinformation affects the equilibrium in insurance markets due to lack of access to accurate information that explains the full disclosures of insurance contracts, leading to market inefficiencies such as mispriced policies, adverse selection, and moral hazard. Further, the delay in timely updation of health insurance provider directories can cause insured individuals to inadvertently select out-of-network providers (Kleban, 2019 ). 1.2.2 Accessibility and Quality of Health care Accessibility and quality of health care discrepancies related to the additional expenditure incurred by individuals. These costs are associated with obtaining medical care services in the same vicinity or for access better quality medical services provided by hospitals within the insurance network. When individuals incur significantly higher expenses to access these services, it indicates that the insurance premium does not accurately represent the true medical costs borne by the insured. Consequently, these unaccounted inconsistencies lead to the deviation from actuarially fair health insurances policies, as the premium does not reflect the actual expected cost of medical care. Assessing patient satisfaction from the quality of health care services provided, Raposo et al. ( 2009 ) finds factors such as cleanliness, temperature, comfort, effective administration, and good nursing behaviour significantly influence patients' positive perceptions of facilities, thereby enhancing overall satisfaction. Xiong et al. ( 2018 ) examine the role of health insurance within the context of the Universal Medical Insurance System (UMIS) in China, in enhancing both the accessibility and affordability of medical services across diverse populations and regions, suggest that a unified system helps to create a more inclusive and sustainable health insurance system. 2. Data and Modelling Methodology 2.1 Data The primary focus of this study is to quantify the discrepancies in health care the delivery of health care under insurance schemes. In this regard, a set of variables that constitute the factors mentioned in section 1.2 are chosen. The data sources include the India Human Development Survey II (IHDS-II), which covers over 42,000 households across India and provides data on health insurance coverage, health utilization, and economic status. The Indian Insurance Statistics (IIS) Handbook 2011-12, published by the IRDAI, offers detailed statistics on the insurance market, including policy numbers, consumer complaints, and insurer performance metrics. The National Health Profile (NHP) 2011, compiled by the Central Bureau of Health Intelligence, provides information on health infrastructure, healthcare professionals, disease prevalence, and public health finances. Additionally, data from Tata AIG Ltd. on hospitals registered under ROHINI (Registry of Hospitals in Network of Insurance) offers insights into the accessibility and quality of services from managed hospitals. Table 1: Table describing the variables used for research. Variable Description Summary Statistic Data Source STATEID Identifier for the state of each surveyed household. Numeric value from 1 to 34 IHDS-II CO34 Medical cost incurred by a household (in-patient) Mean − 6414.4 Standard deviation − 34604.72 Max – 4000000 Min − 0 IHDS-II QC8 Medical Treatment wait time Mean − 28.10795 minutes. Standard deviation – 33.78 Max – 760 minutes. Min – 0 minutes. IHDS-II OG5 Respondent knowledgeable about health and education expense Very little knowledge (1) Somewhat (2) Very knowledgeable (3) IHDS-II QC3 Medical treatment location Same Village/Town (1) Another Village (2) Other Town (3) District Town (4) IHDS-II HHEDUC Highest adult education none (0) 1st class (1) – 12 t h class (12) 1-year post-secondary 2-years post-secondary Undergraduate level of educations (15) Above Undergraduate level of educations (16) IHDS-II NOC Number of complaints filed against insurance providers. Mean – 627 Standard deviation – 839.67 Max – 3516 Min – 5 Indian Insurance Statistics Handbook Nogovdr Number Allopathic doctors in a state Mean – 3494 Standard deviation – 3836.59 Max – 14509 Min – 19 NHP AcceNABH status of hospitals, which reflects their compliance with the specific healthcare quality and the safety standards set by accreditation bodies such as NABH. NABH - Pre-entry Level NABH - Higher level Pre-accredited Tata AIG general insurance Source: Author’s contribution using IHDS, IIS, NHP, TATA AIG Tables 1, 2, and 3 provide an overview of the variables used in this study. Table 1 presents the variables collected for the research, while Tables 2 and 3 display the transformed quantitative and qualitative variables, respectively. In Table 1, STATEID, CO34, OG5, QC8, QC3, and HHEDUC are variables sourced from IHDS-II. Among these, STATEID, OG5, QC3, and HHEDUC are qualitative variables, while CO34 and QC8 are quantitative. The NOC variable, collected from the IIS Handbook 2011-12, is quantitative. The qualitative variables used in this study are Nogovdr, sourced from NHP 2011, and AcceNABH, obtained from Tata AIG Ltd. As these variables are collated from different sources and represent varying levels of aggregation. IHDS-II provides household-level data, IIS Handbook 2011-12 offers data from complaint addressal centres, NHP supplies state-level aggregate data, and Tata AIG’s data lists individual hospitals with NABH accreditation in each Indian state. Suitable transformation was applied to normalise the varying levels. The NOC variable from the IIS Handbook was proportionally distributed across states based on population coverage by complaint addressal centres. The "AcceNABH" variable from Tata AIG was aggregated at the state level to represent the number of hospitals with specific accreditation making them quantitative variable from qualitative. Following these transformations, the variables were combined with IHDS-II data, resulting in a dataset with both household-level and state-level aggregates. From Table 1, it is evident that the standard deviation of some quantitative variables is 100 times greater than that of other variables, implying that the variance of these variables would be 10,000 times higher, which could impact the modelling process. To achieve a better fit, natural logarithmic transformations of the variables were applied to all the quantitative variables. The transformed variables are given in Table 2. Table 2: Table describing transformation of the quantitative variables. Variable Name Description Summary Statistics log_CO34 log_CO34 = log (CO34 + 1) Mean – 3.57 Std dev – 4.41 Max – 14.22 Min – 0 log_QC8 Log_QC8 = log (QC8 + 1) Mean − 2.85 Std dev – 1.09 Max – 6.58 Min – 0 log_NABH_P log_NABH_P = log (NABH_pre_entry + 1), where NABH_pre_entry is the total number of hospitals with NABH accreditation entry level. Mean – 2.14 Std dev – 1.67 Max – 5.15 Min – 0 log_NABH_H log_NABH_H = log (NABH_Higher + 1), where NABH_Higher is the total number of hospitals with NABH accreditation level above entry and below full accreditation. Mean – 2.51 Std dev – 1.70 Max – 4.80 Min – 0 log_NABH_A log_NABH_A = log (NABH_accrediated + 1), where NABH_accrediated is the total number of hospitals with NABH full accreditation. Mean – 2.31 Std dev – 1.63 Max – 4.90 Min – 0 log_NOC log_NOC = log (noc + 1) Mean − 6.40 Std dev – 1.20 Max – 8.16 Min – 2.30 log_num_doc log_num_doc = log (Nogovdr + 1) Mean − 8.33 Std dev – 0.832 Max – 9.58 Min – 4.04 Source: Author’s Contribution using IHDS, IIS, NHP, TATA AIG From Table 1, the qualitative variable has been also transformed into dummy variables such as: OG5 has been divided into OG5_1, OG5_2 and OG5_ 3; QC3 has been divided into QC3_1, QC3_2, QC3_3 and QC3_4; HHEDUC variable has been grouped into 2 categories, above undergraduate level of education and below undergraduate level of education, captured with Education_Group. The transformed dummy variables are given in Table 3. Table 3: Table describing transformation of the qualitative variables. Variables Description Counts OG5_1 OG5_1 = 1, if household claim to have very little knowledge of medical expenses. OG5_1 = 0, ow Count of 1s – 403 Count of 0s – 4171 OG5_2 OG5_2 = 1, if household claim to have some knowledge of medical expenses. OG5_2 = 0, ow Count of 1s – 1144 Count of 0s – 3430 OG5_3 OG5_3 = 1, if household claim to have very knowledge of medical expenses. OG5_2 = 0, ow Count of 1s – 3027 Count of 0s – 1547 Education_Group Education_Group = 1, if the highest adult education is undergraduate level of education and above or else = 0 Count of 1s – 1094 Count of 0s – 3480 QC3_1 𝑄𝐶3_1 = 1, if the medical treatment location is Same Village/Town or QC3_1 = 0 Count of 1s – 2863 Count of 0s – 1171 QC3_2 𝑄𝐶3_2 = 1, if the medical treatment location is Another Village. QC3_2 = 0, ow Count of 1s – 731 Count of 0s – 3843 QC3_3 𝑄𝐶3_3 = 1, if the medical treatment location is Other Town QC3_3 = 0, ow Count of 1s – 723 Count of 0s – 3851 QC3_4 𝑄𝐶3_4 = 1, if the medical treatment location is District Town. QC3_4 = 0, ow Count of 1s – 257 Count of 0s – 4317 Source: Author’s Contribution using IHDS, IIS, NHP, TATA AIG. 2.2 Empirical Analysis As the discrepancies are qualitative in nature and cannot be measured directly from the data, they are modelled as latent variables using modelling techniques. Structural Equation Modelling (SEM) is used to define the latent variables and Hierarchical Linear Modelling (HLM) is to control for state-specific constant. As the latent variables are constructed from data sources of which some data represent state-level aggregates, the latent variable developed would also comprise of similar attribute. In such circumstances there arises the need to use models which control for state-specific constant, this motivates the use of HLM in this study. 2.3.1 Structural Equation Modelling (SEM) SEM, also known as path analysis, is a statistical technique used to understand complex relationships between observed and latent variables (Muthén, 1983). This method is useful in testing theoretical models to examine the direct and indirect interaction of variables. Latent variables, representing abstract concepts, are inferred from measurable indicators. SEM uses factor analysis to validate these relationships. The general form of the measurement model which relates the observed variables (X) to latent variables (𝜉) is given as follows: Latent variables, representing abstract concepts, are inferred from measurable indicators. SEM uses factor analysis to validate these relationships. The measurement model links observed variables (X) to latent variables (𝜉) with the equation X = Λ_X ξ + 𝛿 (1) where, Λ_X represents factor loadings and 𝛿 represents measurement errors. The fitted model used for construction of the latent variables is given as follows: M \(\:I\:={\:\alpha\:}_{1}\:OG5\_2\:+\:{\alpha\:}_{2}\:OG5\_3\:+\:{\alpha\:}_{3}\:Education\_Group\:+\:{\epsilon\:}_{MI}\) (2) $$\:NetMedQ\:=\:{\beta\:}_{1}\:log\_NABH\_H\:+\:{\beta\:}_{2}\:log\_NABH\_P\:+\:{\beta\:}_{3}\:log\_NABH\_A\:+\:{\epsilon\:}_{NetMedQ}$$ 3 $$\:MSA\:=\:{\gamma\:}_{1}\:QC3\_2\:+\:{\gamma\:}_{2}\:QC3\_3\:+\:{\gamma\:}_{3}\:log\_QC8\:+\:{\gamma\:}_{4}\:log\_num\_doc\:+\:{\epsilon\:}_{MSA}$$ 4 where, MI, NetMedQ, and MSA are the latent variables constructed from the observed variables in Tables 2 and 3; OG5_2, OG5_3, and Education_Group for MI, log_NABH_H, log_NABH_P, log_NABH_A for NetMedQ and QC3_2, QC3_3, log_QC8 and log_num_doc for MSA. The coefficients α, β, and γ represent the values associated with each variable, and ε represents the error terms. The values for all the coefficients α i , β i , and γ i , i = 1,2,3, of the latent variables MI, NetMedQ, and MSA used in this study are estimated. 2.3.2 Hierarchical Linear Modelling (HLM) HLM also referred to as multilevel modelling is ideal for analysing clustered data. Unlike ordinary least squares (OLS) regression, HLM accounts for the correlation of data within clusters, which violates the independence of errors assumption (Santos, 2023). HLM effectively handles data structured at multiple levels, allowing for regression equations at each level and accommodating both fixed and random effects. Given that this research involves both aggregate data and household-level data, it suggests that HLM is the most suited and effective model for the analysis. Level 1 Model (Within-Group Relationships) The general form of the model which describes the relationship between a dependent variable and one or more independent variables within a group is given below: \(\:{Y}_{ij}={\beta\:}_{0j}+{\beta\:}_{1j}{X}_{ij}+{r}_{ij}\) ​ (5) where, 𝑌 𝑖𝑗 is the outcome variable for individual i in group j, 𝑋 ij is the predictor variable for individual i in group j, 𝛽 ij and 𝛽 ij are the coefficients for group j, which can vary from group to group, 𝑟​ ij represents the residuals at the individual level, assumed to be normally distributed. Level 2 Model (Between-Group Relationships) : The general form of the model which explains how the coefficients 𝛽 0𝑗 and 𝛽 1𝑗 ​ from the Level 1 model vary across groups is given below: $$\:{\beta\:}_{0j}={\gamma\:}_{00}+{\gamma\:}_{01}{W}_{j}+{u}_{0j}$$ 6 $$\:{\beta\:}_{01}={\gamma\:}_{10}+{\gamma\:}_{11}{W}_{j}+{u}_{1j}$$ 7 where, 𝛾 00 and 𝛾 10 are the intercepts for 𝛽 0j​ and 𝛽 ij ​, respectively, 𝛾 01 ​ and 𝛾 11 ​ describe how 𝛽 0j ​ and 𝛽 1j change with a group-level predictor 𝑊 j ​, 𝑢 0j and 𝑢 1j ​ are the random effects for group j, indicating variation across groups not explained by 𝑊 j . The fitted model used for regressing the natural logarithm of medical cost of the insured on the above specified latent variables is given below: $$\:{log}_{e}{\left(CO34\right)}_{ij}={\lambda\:}_{0}+{\lambda\:}_{1}\:{MI\_scores}_{ij}+{\lambda\:}_{2}\:{NetMedQ\_scores}_{ij}+{\lambda\:}_{3}{\:MSA\_scores}_{ij}\:+{{\lambda\:}_{4}\:log\_NOC}_{ij}+{u}_{0j}+{ϵ}_{ij}$$ 8 where, λ 0 is the intercept term and λ i’s are the predictor variables’ coefficients, namely MI_scores, NetMedQ_scores, MSA_scores and log_NOC, log e (CO34) is the predicted variable; u is random effect error term; ε is residual error term. 3. Results The empirical results of the study are presented in this section. 3.1 Descriptive Analysis Plots of the latent variables, namely average medical cost insured by households, average waiting period for medical treatment, distribution of ROHINI with NABH accreditations, and Number of complaints respectively in India are given in Figs. 1, 2, 3 and 4. respectively In Fig. 1, the average medical cost incurred by the household participating varies from INR500 to INR24000, with maximum in Delhi INR23,968.30 and lowest by Chandigarh and Manipur. In Fig. 2 the average waiting period before receiving medical treatment varies from 13 to 68 minutes, with Jammu and Kashmir (including the territory of Ladakh) showing the maximum average waiting time of 67 minutes and lowest in Punjab 13 minutes. In Fig. 3, the state-wise distribution of hospitals with NABH accreditation is displayed. Tripura has no hospitals with NABH accreditation while Uttar Pradesh shows the maximum number of hospitals with NABH accreditation totalling to 93,884. In Fig. 4, the number of complaints filed state-wise against non-life insurance ranging from 5 to 3,600 is displayed. The union territory Daman and Diu show has the least number of complaints equal to 9, while Maharashtra has the maximum number of complaints filed totalling to 3,516. Comparing Figs. 1 and 2, we see that Jammu and Kashmir, Kerala, Tamil Nadu, West Bengal and Karnataka show darker colouring which indicates that the average waiting time is more than the average medical cost, whereas the colouring is lighter in Madhya Pradesh. This suggests that medical care cost and average waiting time are positively related that may be attributed to high demand of the services which impacts the duration of waiting time to receive medical attention. Similarly, Figs. 3 and 4, suggest that Rajasthan, Gujarat, Maharashtra, Karnataka, Bihar showing contrast in the colouring shades which suggest that the number of hospitals with NABH accreditation and number of complaints filed against non-life insurance schemes have a negative relationship. Figures 2 and 3 suggest that the average waiting period for medical treatment and the distribution of hospital in the insurance network are negatively correlated for the states Uttar Pradesh, Bihar, Madhya Pradesh, Jammu and Kashmir. Kerala, West Bengal Andhra Pradesh, Assam, Bihar, Karnataka show contrast in the colouring shades which can be inferred as the number of hospitals in an insurance network is high than the average waiting time to receive medical care is less. 3.2 Latent Variable Development – Structural Equation Modelling The SEM is fitted on the following three latent variables: Misinformation (MI), Network Medical Quality (NetMedQ), and Medical Service Availability (MSA). The estimates of the fitted model are given in Table 4. Table 4. Estimators of parameters from SEM Latent Variable Indicator Estimate Std. Err z-value P-value MI OG5_2 1.000 - - - OG5_3 -1.325 *** 0.179 -7.414 0.000 Education_Group -0.091 *** 0.017 -5.238 0.000 NetMedQ log_NABH_H 1.000 - - - log_NABH_P 0.981 *** 0.004 247.339 0.000 log_NABH_A 0.959 *** 0.004 244.920 0.000 MSA QC3_2 1.000 - - - QC3_3 -0.917 * 0.486 -1.884 0.060 log_QC8 8.170 *** 2.570 3.179 0.001 log_num_doc -21.462 *** 6.691 -3.208 0.001 Source: Author’s Contribution using IHDS, IIS, NHP, TATA AIG From the estimated values of the fitted model given in Table 4, we see that the coefficient α 2 of OG5_3 and α 3 of Education_Group are negative and significant at the 1% level. OG5_2 at 1.000 serves as a scaling factor for the latent variable MI. The interpretation for α 2 and α 3 is as follows: if the insured individual claims to be very knowledgeable about medical and education expenditure (OG5_3 = 1), the MI variable decreases by 1.325 units compared to those with no knowledge; however, if the highest adult education in the household is a undergraduate level of education's degree or above (Education_Group = 1), the MI variable decreases by 0.091 units compared to households with lower education levels, repectively. For NetMedQ, the estimated coefficients β 2 of NABH_P and β 3 of NABH_A are positive and significant at the 1% level. Log_NABH_H at 1.000 serves as a scaling factor for the latent variable NetMedQ. The interpretation of β 2 and β 3 as follows: if the number of hospitals with pre-entry level accreditation increases by 1%, the NetMedQ variable increases by 0.981 units. Similarly, a 1% increase in hospitals with pre-accreditation leads to a 0.959 unit increase in Network Hospital Quality, respectively. For MSA, the estimated coefficient γ 2 of QC3_3 is significant at the 10% level, while the coefficients γ 1 , γ 3 , and γ 4 are significant at the 1% level, with positive and negative signs respectively. QC3_2 at 1.000 serves as a scaling factor for the latent variable MSA. The interpretation of γ 2 , γ 3 and γ 4 as follows: if persons get treatment in other town the MSA will decrease by 0.917 units, a percentage change in the waiting time of medical treatment (log_QC8) leads to an 8.170 unit increase in MSA. However, a 1% increase in the number of doctors (log_num_doc) results in a 21.462 unit decrease in MSA. After estimating these parameters, the scores for the latent variables MI, MSA, and NetMedQ were computed as MI_scores, MSA_scores, and NetMedQ_scores. These scores are used in the hierarchical linear model (HLM) for further analysis. 3.3 Regression – Hierarchical Linear Modelling Since HLM can accommodate for both fixed and random effects in a model, it is used in the furtherance of this analysis. The level 1 model which accounts for within-group relationship are the part of fixed effects and level 2 model explain how the coefficients (intercept and slopes) from the Level 1 model may vary across groups, which are capture by random effects. Random Effects The estimates of the fitted level 2 model from HLM are given result in Table 5 Table 5: Random Effects estimates from HLM. Group Component Estimate Std. Dev. STATEID (Intercept) 2.187 1.479 Residual 17.098 4.135 Source: Author’s Contribution using IHDS, IIS, NHP, TATA AIG The STATEID group accounts for the variation in the baseline level of log_CO34 across different Indian states. The estimated variance of 2.187 signifies a significant difference in the starting point of log_CO34. The standard deviation of 1.479 indicates that, on average, states deviate from the mean intercept by about 1.479 units of log_CO34. This substantial heterogeneity suggests that there might be state-specific factors influencing the baseline levels of log_CO34. These factors could potentially include variations in policies, economic disparities, or differences in healthcare infrastructure across states. The Residual group captures the remaining unexplained variation in log_CO34 that the model could not account for through the fixed effects or the random intercepts by statewise categorization. Fixed Effects The estimates of the fitted level 1 model from HLM are given in Table 6 Table 6: Fixed Effects estimates from HLM. Predictor Estimate Std. Error t value MI_scores 0.5969 ** 0.2344 2.546 log_NOC 0.8305 *** 0.2315 3.588 MSA_scores -30.8388 ** 12.1783 -2.532 NetMedQ_scores -0.4908 ** 0.2440 -2.012 Source: Author’s Contribution using IHDS, IIS, NHP, TATA AIG The estimated coefficient λ 1 of MI_scores is significant at the 5% level, indicating a positive relationship between MI and log_CO34. This implies that a one unit increase in misinformation leads to a 0.59% increase in the cost of medical care by the insured. The estimated coefficient λ 4 of log_NOC is significant at the 1% level, showing that a 1% increase in the number of complaints results in a 0.83% increase in medical care costs. The estimated coefficient λ 3 of MSA_scores is significant at the 5% level. This shows a negative relationship between medical service availability and medical care costs, where a one unit increase in MSA_scores leads to a 30.83% decrease in medical care costs. Lastly, the estimated λ 2 of NetMedQ_scores is significant at the 5% level. This suggests that a one unit increase in the number of hospitals with good health care quality standards leads to a 0.49% decrease in medical care costs. 4. Discussion The results of model fitting given in Tables 4, 5 and 6 indicate the presence of discrepancies health care provided by insurance schemes. This significantly influences the medical costs incurred by households participating in health insurance policies. The following discussion delves into the specific impact of each discrepancy on actuarial fairness within health insurance. 4.1 Impact of discrepancies in medical care services under health insurance schemes in India 4.1.1 Misinformation (MI) In this study, the differences in knowledge between insured and insurer is defined as misinformation. This latent variable comprises of the variables which correspond to level of knowledge of medical information and education background. This is quantiatively represented in Table 4 as: OG5_1, OG5_2, OG5_3, and Education_Group. The estimated results indicate that the presence of misconceptions and lack of knowledge on health conditions and the related medical treatments. Households with very little knowledge represented by OG5_1, spend less on medical care, often choosing cheaper, lower quality services. Those with some knowledge represented by OG5_2, have the capacity to spend more on higher quality, more expensive health care services. Highly knowledgeable households represented by OG5_3 make cost-effective choices for quality care. The categorization across levels of education shows that households with an undergraduate level of education or higher incur lower costs. The results of our analysis confirm the presence of information gap which leads to inaccurate risk assessments, causing health insurance mispricing and making it actuarially unfair (Kleindorfer & Kunreuther, 1980). Hence, awareness programs on diseases, health care and access to basic education can lead to a better management of medical expenses, leading to fairer health insurance costs which aligning premiums with actual healthcare expenses and risks. 4.1.2 Network Medical Care Quality (NETMEDQ) In this study, the quality of health care provided by the empanelled hospitals within the insurance framework are examined. The latent variable assesses the quality of medical care within an insurance network is constructed taking into account its crucial role in measuring healthcare costs and insurance fairness. It includes three indicators, namely, log_NABH_H, log_NABH_P, and log_NABH_A, which are tabulated in Table 4. The estimated results show a positive relationship with NetMedQ which suggest that increase in the number of NABH accredited hospitals improve medical care provided by the network of empanelled hospitals. We see that a higher quality network tends to reduce healthcare costs by providing effective and efficient care, reducing the need for repeated treatments. Failure to maintain high-quality care in the network leads to overpriced insurance policies, as households might seek more expensive out-of-network services (Raposo, Alves, & Duarte, 2009), resulting in actuarial unfairness. To ensure fairness, insurers must maintain rigorous quality standards and increase the number of accredited hospitals, aligning premiums with the true value of services provided. 4.1.3 Medical Service Availability (MSA) The accessibility to quality health care is a crucial determinant in determining the costs associated with medical services. The latent variable used in this study measures the accessibility to medical services in an Indian state. This latent variable comprises of the variables which correspond to the geographical location of the health care provider and number of allopathy doctors available in the respective Indian state. This is quantiatively respresented in Table 4 as: QC3_2, QC3_3, log_QC8 and log_num_doc. The estimated results show that QC3_2 and log_QC8 show a positive relationship with MSA, suggesting that when medical services are more accessible and when more healthcare providers such as hospitals or clinics are available, there is a positive impact on the perception towards medical care. However, we see that the estimates of QC3_3 and log_num_doc show a negative relationship with MSA, this indicates the presence of challenges in accessibility to quality health care. This is most common in rural parts of India where medical services are predominantly available in distant locations and it reduces overall accessibility. Hence, enhancing service availability through expanded infrastructure and clearer communication about coverage limitations could help mitigate these issues (Xiong, et al., 2018 ). 4.1.4 Inconsistencies in claims processing (log_NOC) In this study, the inconveniences caused by the insurer and the insured at the time of submission of health insurance claims is called as inconsistencies in claims processing. This variable is pivotal in determining the fairness of health insurance pricing as indemnification is the underlying principle of insurance. It is quantitively measured by the logarithmic transformation of the number of complaints from an Indian state and is given in Table 6. As shown in Table 6, a proportional increase in medical costs is associated with a corresponding rise in the number of complaints. An increase in complaints also indicates which can result in the mispricing of health insurance, either making it overpriced or underpriced. Consequently, as inconsistencies in the claims process increase, so do medical costs, leading to actuarial unfairness in health insurance. Therefore, it is essential to streamline the claims processing by including measures that allow for transparency, active grievance redressal which will minimize inconsistencies and lead to fairness in health insurance. 5. Conclusion This study begins by highlighting critical findings from the Indian Human Development Survey-II (IHDS-II), which reveal that approximately 40.8 percent of households with health insurance have incurred additional medical expenses exceeding their risk cover. With a doctor-to-population ratio of 1:1457 in 2011, far below the WHO’s recommended 1:1000, the shortage of healthcare resources in India is evident (Deo, 2013 ). Furthermore, systemic issues such as uneven resource distribution, poor infrastructure, and administrative inefficiencies (Mavalankar et al., 2000) continue to disrupt the delivery of quality care under health insurance schemes. These unaddressed challenges lead to a mismatch between the expected cost of premiums and the actual cost of medical care, raising concerns about the fairness of actuarial principles underlying these insurance policies. Building on these foundational observations, the study employed SEM and HLM to examine how discrepancies like misinformation, accessibility, hospital quality, and claims process inconsistencies impact medical costs and the fairness of health insurance pricing. The results of the study show statistically significant estimates of these discrepancies which lead to the mismatch of experience adjustments used in actuarial valuation in comparison to the realised costs incurred by insured households. Misinformation, driven by knowledge gaps, leads to differences in medical spending. Households with less skills obtained through education tend to incur lower initial costs, possibly sacrificing quality, while more informed households make cost-effective decisions. Reducing misinformation through education can align consumer perceptions with reality, supporting actuarial fairness. The quality of care within insurance networks, as indicated by NABH accredited hospitals, enhances health outcomes and reduces costs. Improved facilities in hospitals and accountability in the role of medical practitioners could help insurers maintain fair premium structures, ensuring customers receive the true value for their premiums. Medical service availability significantly impacts costs associated with good quality healthcare. Limited access to services often leads to higher expenditures due to travel or higher local prices, resulting in unfair high premiums. Expanding infrastructure and clear communication about coverage could mitigate these issues. Inconsistencies in the claim process, reflected by rising complaints, suggest inefficiencies that increase medical costs. Including customer friendly policies such as transparency, grievance redressal within the scope of processing health insurance claims can reduce these inconsistencies. In summary, this study highlights the importance of experience studies in actuarial science and the need for updating experience assumptions to maintain actuarial fairness in pricing health insurance. The discrepancies present in health insurance costs, including misinformation, inadequate healthcare resources, and systemic inefficiencies, significantly impact the alignment between insurance premiums and actual medical costs. Addressing these issues by improving public knowledge on healthcare schemes and practices, enhancing the quality of healthcare provided, and expanding accessibility to medical services, can lead to more equitable insurance pricing that truly reflects the risks and costs faced by households. By adopting these strategies, policy makers and insurers can foster greater trust and satisfaction among policyholders which could lead to client retention and at large a healthier society. The scope of this analysis is restricted to the IHDS-II data. Future research should continue to identify and mitigate additional discrepancies within the health insurance sector, while also exploring the ethical and legal implications of these disparities, especially in relation to vulnerable populations, to promote equity and efficiency in the health insurance sector. Declarations COMPETING INTERESTS The authors do not have any financial competing interest in the publication of this paper. This research was done solely for academic purposes alone. ETHICS DECLARATION Not applicable. FUNDING There was no funding involved in the design of study, data collection, analysis and interpretation as this research was a part of the dissertation report carried out as part of the curriculum of the M.A. Actuarial Economics program at the Madras School of Economics, Chennai, India. Author Contribution Nikhil Rathee: Writing – original draft, Conceptualization, Data curation, Formal analysis, Methodology, Investigation. Rupel Nargunam: Writing – review & editing, Supervision, Methodology, Formal analysis, Data curation, Conceptualization, Visualization, Validation.All authors reviewed the manuscript ACKNOWLEDGEMENTS Only the authors worked upon this paper and there were no other individuals who have contributed towards this article. The usual disclaimers apply. Data Availability The data sources include the India Human Development Survey II (IHDS-II), which covers over 42,000 households across India and provides data on health insurance coverage, health utilization, and economic status. The Indian Insurance Statistics (IIS) Handbook 2011-12, published by the IRDAI, offers detailed statistics on the insurance market, including policy numbers, consumer complaints, and insurer performance metrics. The National Health Profile (NHP) 2011, compiled by the Central Bureau of Health Intelligence, provides information on health infrastructure, healthcare professionals, disease prevalence, and public health finances. Additionally, data from Tata AIG Ltd. on hospitals registered under ROHINI (Registry of Hospitals in Network of Insurance) offers insights into the accessibility and quality of services from managed hospitals. References ABEL-SMITH, BRIAN. “Health Insurance in Developing Countries: Lessons from Experience.” Health Policy and Planning 7, no. 3 (1992): 215–26. http://www.jstor.org/stable/45090068. Al-Rawashdeh, Firas. 2016. Determinates of demand on purchasing insurance policies: Case of Jordan. International Review of Management and Marketing, 6, 1094-1102. Arnold, Jens Matthias, Beata Javorcik, Molly Lipscomb, and Aaditya Mattoo. "Services reform and manufacturing performance: Evidence from India." The Economic Journal 126, no. 590 (2016): 1-39. Deo, Madhav G. "Doctor population ratio for India-The reality." Indian Journal of Medical Research 137, no. 4 (2013): 632-635. Desai, Sonalde, Amaresh Dubey, B. L. Joshi, Mitali Sen, Abusaleh Shariff, and Reeve Vanneman. "India human development survey." College Park, Maryland: University of Maryland (2008). Donahue, Kate, and Solon Barocas. "Better together? how externalities of size complicate notions of solidarity and actuarial fairness." In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, pp. 185-195. 2021. Eckstein, David, Vera Künzel, and Laura Schäfer. The global climate risk index 2021. Bonn: Germanwatch, 2021. https://bvearmb.do/handle/123456789/1306 Gollier, Christian. "The determinants of the insurance demand by firms." Toulouse School of Economics (LERNA and IDEI). Toulouse: Toulouse School of Economics, 2007. http://tse-fr.eu/pub/7278 Heras, Antonio J., Pierre-Charles Pradier, and David Teira. "What was fair in actuarial fairness?." History of the Human Sciences 33, no. 2 (2020): 91-114. Hoyt, Robert E., and Ho Khang. "On the demand for corporate property insurance." Journal of Risk and Insurance (2000): 91-107. Huta, Veronika. "When to use hierarchical linear modeling." The quantitative methods for psychology 10, no. 1 (2014): 13-28. Kajwang, Ben. "Contribution of Industrial All Risks Insurance to the Growth of Manufacturing Firms in Kenya." Journal of Developing Country Studies 5, no. 1 (2020): 36-48. Kleban, Jaclyn. "Countering Misinformation in the Health Care System: The Case for Stricter Regulations Within Health Insurance Provider Directories." CARDOzO L. REv. 41 (2019): 1185. Kleindorfer, Paul R., and Howard Kunreuther. "Misinformation and equilibrium in insurance markets." In Economic Analysis of Regulated Markets, pp. 67-90. London: Palgrave Macmillan UK, 1980. Krummaker, Simone. "Corporate Demand for Insurance: Empirical Evidence from Germany." (2016). https://openaccess.city.ac.uk/id/eprint/24066/ Levy, Helen, and David Meltzer. "The impact of health insurance on health." Annu. Rev. Public Health 29, no. 1 (2008): 399-409. Lieberthal, Robert D., and Robert D. Lieberthal. "Defining Health Insurance." What Is Health Insurance (Good) For? An Examination of Who Gets It, Who Pays for It, and How to Improve It (2016): 3-32. Linnerooth-Bayer, Joanne, and Reinhard Mechler. "Insurance against losses from natural disasters in developing countries." IIASA, Background paper for UN WESS, July (2008). Mavalankar, Dileep, and Ramesh Bhat. "Health insurance in India: opportunities, challenges and concerns." Ahmedabad: Indian Institute of Management (2000): 1-16. Mayers, David, and Clifford W. Smith Jr. "On the corporate demand for insurance." In Foundations of Insurance Economics: Readings in Economics and Finance, pp. 190-205. Dordrecht: Springer Netherlands, 1982. Motsepe, Molatelo. "An Industry Level Analysis of Demand for Insurance in South Africa." (2018). University of Cape Town, Faculty of Commerce. Cape Town: Graduate School of Business (GSB). http://hdl.handle.net/11427/28411 Mushonga, Farai B., and Syden Mishi. "Natural hazard insurance demand: A systematic review." Jàmbá-Journal of Disaster Risk Studies 14, no. 1 (2022): 1223. Muthen, Bengt. "Latent variable structural equation modeling with categorical data." Journal of Econometrics 22, no. 1-2 (1983): 43-65. Nasir, H. U. (2024). Actuarial fairness: what does it mean and when is it desirable? School of Graduate Studies Electronic Theses and Dissertations, 154. Prokopjeva, Evgenija, Galina Chernova, Nataliya P. Kuznetsova, Svetlana Kalayda, and Leonid Ivanov. "Insurance instruments in risk management of the manufacturing sector of a region: the case of the Republic of Khakassia (Russia)." Investment Management & Financial Innovations 17, no. 4 (2020): 299. Rao, M. Subba, and R. Srinivasulu. "Contribution of insurance sector to growth and development of the Indian economy." IOSR journal of Business and Management 7, no. 4 (2013): 45-52. Raposo, Mário Lino, Helena Maria Alves, and Paulo Alexandre Duarte. "Dimensions of service quality and satisfaction in healthcare: a patient’s satisfaction index." Service Business 3 (2009): 85-100. Regan, Laureen, and Yeon Hur. "On the corporate demand for insurance: The case of Korean nonfinancial firms." Journal of Risk and Insurance 74, no. 4 (2007): 829-850. Rothschild, Casey. "The efficiency of categorical discrimination in insurance markets." Journal of Risk and Insurance 78, no. 2 (2011): 267-285. Santos, Gustavo. (2023, March 15). Multilevel Regression with R. Retrieved from Towards Data Science: https://towardsdatascience.com/multilevel-regression-with-r-eb3eb7d8de88 Southwell, Brian G., Jessica Otero Machuca, Sabrina T. Cherry, Melissa Burnside, and Nadine J. Barrett. "Health misinformation exposure and health disparities: observations and opportunities." Annual review of public health 44, no. 1 (2023): 113-130. Wood, Glenn L., and Richard L. Morris. "CONSUMER COMPLAINTS AGAINST INSURANCE COMPANIES." Journal of Personal Finance 9 (2010). Wright, Sewall. "On the nature of size factors." Genetics 3, no. 4 (1918): 367. Xiong, Xiaolei, Zhiguo Zhang, Jing Ren, Jie Zhang, Xiaoyun Pan, Liang Zhang, Shiwei Gong, and Si Jin. "Impact of universal medical insurance system on the accessibility of medical service supply and affordability of patients in China." PloS one 13, no. 3 (2018): e0193273.doi:https://doi.org/10.1371/journal.pone.0193273 Yoshida, Kaori, and Robert E. Deyle. "Determinants of small business hazard mitigation." Natural Hazards Review 6, no. 1 (2005): 1-12. Zhou, Fujin, and Wouter Botzen. "Firm level evidence of disaster impacts on growth in Vietnam." Environmental and Resource Economics 79, no. 2 (2021): 277-322. Zhu, Minglai, Chao Kui, and Yuzheng Fang. "Demand for corporate insurance in China: empirical analysis with province-level panel data." Chinese Economy 44, no. 6 (2011): 15-29. Zou, Hong, and Mike B. Adams. "The corporate purchase of property insurance: Chinese evidence." Journal of Financial Intermediation 15, no. 2 (2006): 165-196. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6613076","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":470308820,"identity":"36a9219a-da69-4c97-9025-cc613b87a5b4","order_by":0,"name":"Nikhil Rathee","email":"","orcid":"","institution":"Madras School of Economics","correspondingAuthor":false,"prefix":"","firstName":"Nikhil","middleName":"","lastName":"Rathee","suffix":""},{"id":470308822,"identity":"dad17936-3b11-44fe-b47f-a9ed36b101c4","order_by":1,"name":"Rupel Nargunam","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAArklEQVRIiWNgGAWjYLCCBxUwlgGxWhLOkKwlsY14FzEwmM/IPfggcd42e34G5ocfGAruENYicyMv2SBx2+3EmQ1sxhIMBs8Ia5GQyDGTAGpJMDjAYAb0y2GitJj/SJxz297+APs3orWYMSQ23GbcwMBDrC0875IlEo7dTpxxmKdYIoEoLey5Bz98qLltz9/evvHDhz9EaGFg4IHSzECcQIwGhJZRMApGwSgYBbgAANJrNQ3hiphCAAAAAElFTkSuQmCC","orcid":"","institution":"Madras School of Economics","correspondingAuthor":true,"prefix":"","firstName":"Rupel","middleName":"","lastName":"Nargunam","suffix":""}],"badges":[],"createdAt":"2025-05-07 14:53:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6613076/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6613076/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":84551205,"identity":"3891b707-9481-440c-a16d-40b8e233e468","added_by":"auto","created_at":"2025-06-13 10:23:53","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":39059,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAverage medical cost in Rs\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6613076/v1/cd10c2a2728091e3ac5d0f9c.png"},{"id":84550689,"identity":"406d2733-88f9-4f5b-9210-6abc0bb15a6e","added_by":"auto","created_at":"2025-06-13 10:15:53","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":35835,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAverage waiting time for medical treatment in minutes\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6613076/v1/b51703dca54aeedd2cf097ed.png"},{"id":84550691,"identity":"dddb31a4-df62-44bc-887a-72585be4728e","added_by":"auto","created_at":"2025-06-13 10:15:53","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":36040,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDistribution of hospital in insurance network\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6613076/v1/d016694fa6acaf557759941d.png"},{"id":84550692,"identity":"13620a0f-257b-4b9a-a87d-e5b6393738ac","added_by":"auto","created_at":"2025-06-13 10:15:53","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":35944,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNumber of complaints statewise\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6613076/v1/1b9c72abb96ce05522ea34bc.png"},{"id":99315524,"identity":"d628b78c-a89c-4624-8710-e2bc7b960ff6","added_by":"auto","created_at":"2025-12-31 16:27:02","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1237640,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6613076/v1/8720cda1-4e3f-4ae3-8f3c-0056c115fb5a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eAn Actuarial Perspective on the Discrepancies in Health Insurance Pricing: Evidence from India\u003c/p\u003e","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003eHealth insurance is a form of collectivism, by construct it is designed to enable individuals to pool their risk of incurring medical expenses. Since the health insurance acts as a binding agreement between the insurer and insured, its realised value is influenced by the multiple factors such as unequal distribution of medical resources, lack of human resources, poor infrastructure, and low-quality medical services. These factors contribute to the disparities in delivery of health care under health insurance and overthrow the underlying principle of designing health insurance contracts.\u003c/p\u003e \u003cp\u003eData from the Indian Human Development Survey-II (IHDS-II) reveals that approximately 40.8 percent of households with health insurance have incurred additional medical expenses that exceed the risk cover. In 2011, India had a doctor-to-population ratio of 1:1457 which is lower than the WHO\u0026rsquo;s recommended ratio of 1:1000 (Deo, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). This is emphasizing the critical shortage of human resources in the health sector. Furthermore, India faces numerous obstacles in its health insurance sector, such as resource distribution, infrastructure problems and administrative inefficiencies (Mavalankar et al., 2000) which disrupt the delivery of quality care under health insurance schemes. These unaddressed challenges and associated costs distort the principle of indemnification by which insurance contracts are designed, resulting in health insurance contracts where the expected cost of premiums does not match the true expected cost of medical care.\u003c/p\u003e \u003cp\u003eThe objective of this article is to investigate the aforementioned discrepancies present in the Indian health insurance network by examining the nature of these factors individually and highlighting their impact on the cost of medical care for insured households. A comprehensive analysis of these discrepancies is expected to help the pricing actuary to determine the expected costs more efficiently and determine premium rates which are in line with the fundamental principles of insurance contracts. This study focuses on examining the individual impact of each of these factors on expected cost of health care, to account for the variations in costs incurred by individuals from different socio-economic backgrounds. It is crucial as the presence discrepancies cause fluctuations in medical costs, which would in turn indicate that health insurance contracts are not priced accurately and may tend to deviate from the principle of actuarial fairness.\u003c/p\u003e \u003cdiv id=\"Sec2\" class=\"Section2\"\u003e \u003ch2\u003e1.1 Pricing Health Insurance and Actuarial Fairness\u003c/h2\u003e \u003cp\u003eTo comprehend actuarial fairness in health insurance, it is essential to understand the significance of the definition of health insurance Pitacco, E. (2014). Health insurance is described as a tool for transferring the risk of healthcare costs to another party in exchange for single or series of payments, called insurance premium. The determinisation of premium is of critical importance and should be determined with fairness. Donahue and Barocas (2021, p. 186) define the cost of insurance as \"pricing in which each participant in the insurance pool pays their expected costs.\" In other words, for health insurance to be actuarially fair, the premium paid in advance should be equal to the expected cost of healthcare.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e1.2 Discrepancies\u003c/h2\u003e \u003cp\u003eDiscrepancies are unexpected differences that suggest something is wrong between two sets of conditions and needs to be explained. A study at the time when foreign direct investment in the Indian health insurance industry was severely restricted, highlights challenges such as the distribution of resources, improper infrastructure, poor healthcare quality, administrative inefficiencies, lack of awareness, and adverse selection (Mavalankar and Bhat 2000). In this study, we focus on factors that deviate the cost of health insurance from being actuarially fair. More specifically, we identify qualitative variables that are expected to have an impact on the pricing of health care costs. In that regard, this study highlights four discrepancies in the delivery of health care under health insurance schemes in the Indian context, namely: Inconsistencies in claims processing, misinformation, quality of medical care, and access to medical services.\u003c/p\u003e \u003cp\u003eThe following subsections give detailed descriptions of the discrepancies addressed by this study:\u003c/p\u003e \u003cdiv id=\"Sec4\" class=\"Section3\"\u003e \u003ch2\u003e1.2.1 Misinformation\u003c/h2\u003e \u003cp\u003eInconsistencies in claims processing and misinformation are not uncommon in the health insurance environment. They are observed in the issuance of insurance and during claim submissions. Misinformation arises from differences in knowledge between individuals and insurers, leading to misunderstandings about coverage and benefits. During claim submission, insured individuals experience inconvenience and dissatisfaction, which result in the ongoing insurance policy being overpriced or risk cover going underutilized regarding their actual risk experience. Leinsdorf et al. (1980) suggest that misinformation affects the equilibrium in insurance markets due to lack of access to accurate information that explains the full disclosures of insurance contracts, leading to market inefficiencies such as mispriced policies, adverse selection, and moral hazard. Further, the delay in timely updation of health insurance provider directories can cause insured individuals to inadvertently select out-of-network providers (Kleban, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section3\"\u003e \u003ch2\u003e1.2.2 Accessibility and Quality of Health care\u003c/h2\u003e \u003cp\u003eAccessibility and quality of health care discrepancies related to the additional expenditure incurred by individuals. These costs are associated with obtaining medical care services in the same vicinity or for access better quality medical services provided by hospitals within the insurance network. When individuals incur significantly higher expenses to access these services, it indicates that the insurance premium does not accurately represent the true medical costs borne by the insured. Consequently, these unaccounted inconsistencies lead to the deviation from actuarially fair health insurances policies, as the premium does not reflect the actual expected cost of medical care. Assessing patient satisfaction from the quality of health care services provided, Raposo et al. (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) finds factors such as cleanliness, temperature, comfort, effective administration, and good nursing behaviour significantly influence patients' positive perceptions of facilities, thereby enhancing overall satisfaction. Xiong et al. (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) examine the role of health insurance within the context of the Universal Medical Insurance System (UMIS) in China, in enhancing both the accessibility and affordability of medical services across diverse populations and regions, suggest that a unified system helps to create a more inclusive and sustainable health insurance system.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"2. Data and Modelling Methodology","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Data\u003c/h2\u003e \u003cp\u003eThe primary focus of this study is to quantify the discrepancies in health care the delivery of health care under insurance schemes. In this regard, a set of variables that constitute the factors mentioned in section 1.2 are chosen. The data sources include the India Human Development Survey II (IHDS-II), which covers over 42,000 households across India and provides data on health insurance coverage, health utilization, and economic status. The Indian Insurance Statistics (IIS) Handbook 2011-12, published by the IRDAI, offers detailed statistics on the insurance market, including policy numbers, consumer complaints, and insurer performance metrics. The National Health Profile (NHP) 2011, compiled by the Central Bureau of Health Intelligence, provides information on health infrastructure, healthcare professionals, disease prevalence, and public health finances. Additionally, data from Tata AIG Ltd. on hospitals registered under ROHINI (Registry of Hospitals in Network of Insurance) offers insights into the accessibility and quality of services from managed hospitals.\u003c/p\u003e \u003cp\u003eTable 1: Table describing the variables used for research.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e \u003ccolgroup cols=\"4\"\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=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSummary Statistic\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eData Source\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSTATEID\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIdentifier for the state of each surveyed household.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNumeric value from 1 to 34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIHDS-II\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCO34\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMedical cost incurred by a household (in-patient)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026minus;\u0026thinsp;6414.4\u003c/p\u003e \u003cp\u003eStandard deviation \u0026minus;\u0026thinsp;34604.72\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 4000000\u003c/p\u003e \u003cp\u003eMin \u0026minus;\u0026thinsp;0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIHDS-II\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eQC8\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMedical Treatment wait time\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026minus;\u0026thinsp;28.10795 minutes.\u003c/p\u003e \u003cp\u003eStandard deviation \u0026ndash; 33.78\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 760 minutes.\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 0 minutes.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIHDS-II\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOG5\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRespondent knowledgeable about health and education expense\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eVery little knowledge (1)\u003c/p\u003e \u003cp\u003eSomewhat (2)\u003c/p\u003e \u003cp\u003eVery knowledgeable (3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIHDS-II\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eQC3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMedical treatment location\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSame Village/Town (1)\u003c/p\u003e \u003cp\u003eAnother Village (2)\u003c/p\u003e \u003cp\u003eOther Town (3)\u003c/p\u003e \u003cp\u003eDistrict Town (4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIHDS-II\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHHEDUC\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHighest adult education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003enone (0)\u003c/p\u003e \u003cp\u003e1st class (1) \u0026ndash; 12\u003csup\u003et h\u003c/sup\u003e class (12)\u003c/p\u003e \u003cp\u003e1-year post-secondary\u003c/p\u003e \u003cp\u003e2-years post-secondary Undergraduate level of educations (15)\u003c/p\u003e \u003cp\u003eAbove Undergraduate level of educations (16)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIHDS-II\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNOC\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNumber of complaints filed against insurance providers.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026ndash; 627\u003c/p\u003e \u003cp\u003eStandard deviation \u0026ndash; 839.67\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 3516\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndian Insurance Statistics Handbook\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNogovdr\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNumber Allopathic doctors in a state\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026ndash; 3494\u003c/p\u003e \u003cp\u003eStandard deviation \u0026ndash; 3836.59\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 14509\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNHP\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAcceNABH\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003estatus of hospitals, which reflects their compliance with the specific healthcare quality and the safety standards set by accreditation bodies such as NABH.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNABH - Pre-entry Level\u003c/p\u003e \u003cp\u003eNABH - Higher level\u003c/p\u003e \u003cp\u003ePre-accredited\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTata AIG general insurance\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eSource: Author\u0026rsquo;s contribution using IHDS, IIS, NHP, TATA AIG\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTables\u0026nbsp;1, 2, and 3 provide an overview of the variables used in this study. Table\u0026nbsp;1 presents the variables collected for the research, while Tables\u0026nbsp;2 and 3 display the transformed quantitative and qualitative variables, respectively. In Table\u0026nbsp;1, STATEID, CO34, OG5, QC8, QC3, and HHEDUC are variables sourced from IHDS-II. Among these, STATEID, OG5, QC3, and HHEDUC are qualitative variables, while CO34 and QC8 are quantitative. The NOC variable, collected from the IIS Handbook 2011-12, is quantitative. The qualitative variables used in this study are Nogovdr, sourced from NHP 2011, and AcceNABH, obtained from Tata AIG Ltd.\u003c/p\u003e \u003cp\u003eAs these variables are collated from different sources and represent varying levels of aggregation. IHDS-II provides household-level data, IIS Handbook 2011-12 offers data from complaint addressal centres, NHP supplies state-level aggregate data, and Tata AIG\u0026rsquo;s data lists individual hospitals with NABH accreditation in each Indian state. Suitable transformation was applied to normalise the varying levels. The NOC variable from the IIS Handbook was proportionally distributed across states based on population coverage by complaint addressal centres. The \"AcceNABH\" variable from Tata AIG was aggregated at the state level to represent the number of hospitals with specific accreditation making them quantitative variable from qualitative. Following these transformations, the variables were combined with IHDS-II data, resulting in a dataset with both household-level and state-level aggregates.\u003c/p\u003e \u003cp\u003eFrom Table\u0026nbsp;1, it is evident that the standard deviation of some quantitative variables is 100 times greater than that of other variables, implying that the variance of these variables would be 10,000 times higher, which could impact the modelling process. To achieve a better fit, natural logarithmic transformations of the variables were applied to all the quantitative variables. The transformed variables are given in Table\u0026nbsp;2.\u003c/p\u003e \u003cp\u003eTable 2: Table describing transformation of the quantitative variables.\u003c/p\u003e\u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Tabb\" border=\"1\"\u003e \u003ccolgroup cols=\"3\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable Name\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSummary Statistics\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003elog_CO34\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_CO34\u0026thinsp;=\u0026thinsp;log (CO34\u0026thinsp;+\u0026thinsp;1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026ndash; 3.57\u003c/p\u003e \u003cp\u003eStd dev \u0026ndash; 4.41\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 14.22\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003elog_QC8\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLog_QC8\u0026thinsp;=\u0026thinsp;log (QC8\u0026thinsp;+\u0026thinsp;1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026minus;\u0026thinsp;2.85\u003c/p\u003e \u003cp\u003eStd dev \u0026ndash; 1.09\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 6.58\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003elog_NABH_P\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_NABH_P\u0026thinsp;=\u0026thinsp;log (NABH_pre_entry\u0026thinsp;+\u0026thinsp;1), where NABH_pre_entry is the total number of hospitals with NABH accreditation entry level.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026ndash; 2.14\u003c/p\u003e \u003cp\u003eStd dev \u0026ndash; 1.67\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 5.15\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003elog_NABH_H\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_NABH_H\u0026thinsp;=\u0026thinsp;log (NABH_Higher\u0026thinsp;+\u0026thinsp;1), where NABH_Higher is the total number of hospitals with NABH accreditation level above entry and below full accreditation.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026ndash; 2.51\u003c/p\u003e \u003cp\u003eStd dev \u0026ndash; 1.70\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 4.80\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003elog_NABH_A\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_NABH_A\u0026thinsp;=\u0026thinsp;log (NABH_accrediated\u0026thinsp;+\u0026thinsp;1), where NABH_accrediated is the total number of hospitals with NABH full accreditation.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026ndash; 2.31\u003c/p\u003e \u003cp\u003eStd dev \u0026ndash; 1.63\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 4.90\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003elog_NOC\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_NOC\u0026thinsp;=\u0026thinsp;log (noc\u0026thinsp;+\u0026thinsp;1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026minus;\u0026thinsp;6.40\u003c/p\u003e \u003cp\u003eStd dev \u0026ndash; 1.20\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 8.16\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 2.30\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003elog_num_doc\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_num_doc\u0026thinsp;=\u0026thinsp;log (Nogovdr\u0026thinsp;+\u0026thinsp;1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean \u0026minus;\u0026thinsp;8.33\u003c/p\u003e \u003cp\u003eStd dev \u0026ndash; 0.832\u003c/p\u003e \u003cp\u003eMax \u0026ndash; 9.58\u003c/p\u003e \u003cp\u003eMin \u0026ndash; 4.04\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003e\u003cem\u003eSource: Author\u0026rsquo;s Contribution using IHDS, IIS, NHP, TATA AIG\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFrom Table\u0026nbsp;1, the qualitative variable has been also transformed into dummy variables such as: OG5 has been divided into OG5_1, OG5_2 and OG5_ 3; QC3 has been divided into QC3_1, QC3_2, QC3_3 and QC3_4; HHEDUC variable has been grouped into 2 categories, above undergraduate level of education and below undergraduate level of education, captured with Education_Group. The transformed dummy variables are given in Table\u0026nbsp;3.\u003c/p\u003e \u003cp\u003eTable 3: Table describing transformation of the qualitative variables.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Tabc\" border=\"1\"\u003e \u003ccolgroup cols=\"3\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCounts\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOG5_1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOG5_1\u0026thinsp;=\u0026thinsp;1, if household claim to have very little knowledge of medical expenses. OG5_1\u0026thinsp;=\u0026thinsp;0, ow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCount of 1s \u0026ndash; 403\u003c/p\u003e \u003cp\u003eCount of 0s \u0026ndash; 4171\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOG5_2\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOG5_2\u0026thinsp;=\u0026thinsp;1, if household claim to have some knowledge of medical expenses. OG5_2\u0026thinsp;=\u0026thinsp;0, ow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCount of 1s \u0026ndash; 1144\u003c/p\u003e \u003cp\u003eCount of 0s \u0026ndash; 3430\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOG5_3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOG5_3\u0026thinsp;=\u0026thinsp;1, if household claim to have very knowledge of medical expenses. OG5_2\u0026thinsp;=\u0026thinsp;0, ow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCount of 1s \u0026ndash; 3027\u003c/p\u003e \u003cp\u003eCount of 0s \u0026ndash; 1547\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEducation_Group\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEducation_Group\u0026thinsp;=\u0026thinsp;1, if the highest adult education is undergraduate level of education and above or else\u0026thinsp;=\u0026thinsp;0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCount of 1s \u0026ndash; 1094\u003c/p\u003e \u003cp\u003eCount of 0s \u0026ndash; 3480\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eQC3_1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026#119876;\u0026#119862;3_1\u0026thinsp;=\u0026thinsp;1, if the medical treatment location is Same Village/Town or QC3_1\u0026thinsp;=\u0026thinsp;0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCount of 1s \u0026ndash; 2863\u003c/p\u003e \u003cp\u003eCount of 0s \u0026ndash; 1171\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eQC3_2\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026#119876;\u0026#119862;3_2\u0026thinsp;=\u0026thinsp;1, if the medical treatment location is Another Village. QC3_2\u0026thinsp;=\u0026thinsp;0, ow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCount of 1s \u0026ndash; 731\u003c/p\u003e \u003cp\u003eCount of 0s \u0026ndash; 3843\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eQC3_3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026#119876;\u0026#119862;3_3\u0026thinsp;=\u0026thinsp;1, if the medical treatment location is Other Town QC3_3\u0026thinsp;=\u0026thinsp;0, ow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCount of 1s \u0026ndash; 723\u003c/p\u003e \u003cp\u003eCount of 0s \u0026ndash; 3851\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eQC3_4\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026#119876;\u0026#119862;3_4\u0026thinsp;=\u0026thinsp;1, if the medical treatment location is District Town. QC3_4\u0026thinsp;=\u0026thinsp;0, ow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCount of 1s \u0026ndash; 257\u003c/p\u003e \u003cp\u003eCount of 0s \u0026ndash; 4317\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003eSource: Author\u0026rsquo;s Contribution using IHDS, IIS, NHP, TATA AIG.\u003c/em\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Empirical Analysis\u003c/h2\u003e \u003cp\u003eAs the discrepancies are qualitative in nature and cannot be measured directly from the data, they are modelled as latent variables using modelling techniques. Structural Equation Modelling (SEM) is used to define the latent variables and Hierarchical Linear Modelling (HLM) is to control for state-specific constant. As the latent variables are constructed from data sources of which some data represent state-level aggregates, the latent variable developed would also comprise of similar attribute. In such circumstances there arises the need to use models which control for state-specific constant, this motivates the use of HLM in this study.\u003c/p\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e2.3.1 Structural Equation Modelling (SEM)\u003c/h2\u003e \u003cp\u003eSEM, also known as path analysis, is a statistical technique used to understand complex relationships between observed and latent variables (Muth\u0026eacute;n, 1983). This method is useful in testing theoretical models to examine the direct and indirect interaction of variables. Latent variables, representing abstract concepts, are inferred from measurable indicators. SEM uses factor analysis to validate these relationships. The general form of the measurement model which relates the observed variables (X) to latent variables (\u0026#120585;) is given as follows: Latent variables, representing abstract concepts, are inferred from measurable indicators. SEM uses factor analysis to validate these relationships. The measurement model links observed variables (X) to latent variables (\u0026#120585;) with the equation\u003c/p\u003e \u003cp\u003eX\u0026thinsp;=\u0026thinsp;Λ_X ξ + \u0026#120575; (1)\u003c/p\u003e \u003cp\u003ewhere, Λ_X represents factor loadings and \u0026#120575; represents measurement errors.\u003c/p\u003e \u003cp\u003eThe fitted model used for construction of the latent variables is given as follows:\u003c/p\u003e \u003cp\u003eM\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:I\\:={\\:\\alpha\\:}_{1}\\:OG5\\_2\\:+\\:{\\alpha\\:}_{2}\\:OG5\\_3\\:+\\:{\\alpha\\:}_{3}\\:Education\\_Group\\:+\\:{\\epsilon\\:}_{MI}\\)\u003c/span\u003e\u003c/span\u003e (2)\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:NetMedQ\\:=\\:{\\beta\\:}_{1}\\:log\\_NABH\\_H\\:+\\:{\\beta\\:}_{2}\\:log\\_NABH\\_P\\:+\\:{\\beta\\:}_{3}\\:log\\_NABH\\_A\\:+\\:{\\epsilon\\:}_{NetMedQ}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$\\:MSA\\:=\\:{\\gamma\\:}_{1}\\:QC3\\_2\\:+\\:{\\gamma\\:}_{2}\\:QC3\\_3\\:+\\:{\\gamma\\:}_{3}\\:log\\_QC8\\:+\\:{\\gamma\\:}_{4}\\:log\\_num\\_doc\\:+\\:{\\epsilon\\:}_{MSA}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e4\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere, MI, NetMedQ, and MSA are the latent variables constructed from the observed variables in Tables\u0026nbsp;2 and 3; OG5_2, OG5_3, and Education_Group for MI, log_NABH_H, log_NABH_P, log_NABH_A for NetMedQ and QC3_2, QC3_3, log_QC8 and log_num_doc for MSA. The coefficients α, β, and γ represent the values associated with each variable, and ε represents the error terms. The values for all the coefficients α\u003csub\u003ei\u003c/sub\u003e, β\u003csub\u003ei\u003c/sub\u003e, and γ\u003csub\u003ei\u003c/sub\u003e, i\u0026thinsp;=\u0026thinsp;1,2,3, of the latent variables MI, NetMedQ, and MSA used in this study are estimated.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.3.2 Hierarchical Linear Modelling (HLM)\u003c/h2\u003e \u003cp\u003eHLM also referred to as multilevel modelling is ideal for analysing clustered data. Unlike ordinary least squares (OLS) regression, HLM accounts for the correlation of data within clusters, which violates the independence of errors assumption (Santos, 2023). HLM effectively handles data structured at multiple levels, allowing for regression equations at each level and accommodating both fixed and random effects. Given that this research involves both aggregate data and household-level data, it suggests that HLM is the most suited and effective model for the analysis.\u003c/p\u003e \u003cp\u003e \u003cb\u003eLevel 1 Model (Within-Group Relationships)\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe general form of the model which describes the relationship between a dependent variable and one or more independent variables within a group is given below:\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:{Y}_{ij}={\\beta\\:}_{0j}+{\\beta\\:}_{1j}{X}_{ij}+{r}_{ij}\\)\u003c/span\u003e \u003c/span\u003e​ (5)\u003c/p\u003e \u003cp\u003ewhere, \u0026#119884;\u003csub\u003e\u0026#119894;\u0026#119895;\u003c/sub\u003e is the outcome variable for individual i in group j, \u0026#119883;\u003csub\u003eij\u003c/sub\u003e is the predictor variable for individual i in group j, \u0026#120573;\u003csub\u003eij\u003c/sub\u003e and \u0026#120573;\u003csub\u003eij\u003c/sub\u003e are the coefficients for group j, which can vary from group to group, \u0026#119903;​\u003csub\u003eij\u003c/sub\u003e represents the residuals at the individual level, assumed to be normally distributed.\u003c/p\u003e \u003cp\u003e \u003cb\u003eLevel 2 Model (Between-Group Relationships)\u003c/b\u003e:\u003c/p\u003e \u003cp\u003eThe general form of the model which explains how the coefficients \u0026#120573;\u003csub\u003e0\u0026#119895;\u003c/sub\u003e and \u0026#120573;\u003csub\u003e1\u0026#119895;\u003c/sub\u003e​ from the Level 1 model vary across groups is given below:\u003cdiv id=\"Equ3\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$\\:{\\beta\\:}_{0j}={\\gamma\\:}_{00}+{\\gamma\\:}_{01}{W}_{j}+{u}_{0j}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e6\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ4\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ4\" name=\"EquationSource\"\u003e\n$$\\:{\\beta\\:}_{01}={\\gamma\\:}_{10}+{\\gamma\\:}_{11}{W}_{j}+{u}_{1j}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e7\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere, \u0026#120574;\u003csub\u003e00\u003c/sub\u003e and \u0026#120574;\u003csub\u003e10\u003c/sub\u003e are the intercepts for \u0026#120573;\u003csub\u003e0j​\u003c/sub\u003e and \u0026#120573;\u003csub\u003eij\u003c/sub\u003e​, respectively, \u0026#120574;\u003csub\u003e01\u003c/sub\u003e​ and \u0026#120574;\u003csub\u003e11\u003c/sub\u003e​ describe how \u0026#120573;\u003csub\u003e0j\u003c/sub\u003e​ and \u0026#120573;\u003csub\u003e1j\u003c/sub\u003e change with a group-level predictor \u0026#119882;\u003csub\u003ej\u003c/sub\u003e​, \u0026#119906;\u003csub\u003e0j\u003c/sub\u003e and \u0026#119906;\u003csub\u003e1j\u003c/sub\u003e​ are the random effects for group j, indicating variation across groups not explained by \u0026#119882;\u003csub\u003ej\u003c/sub\u003e.\u003c/p\u003e \u003cp\u003eThe fitted model used for regressing the natural logarithm of medical cost of the insured on the above specified latent variables is given below:\u003cdiv id=\"Equ5\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ5\" name=\"EquationSource\"\u003e\n$$\\:{log}_{e}{\\left(CO34\\right)}_{ij}={\\lambda\\:}_{0}+{\\lambda\\:}_{1}\\:{MI\\_scores}_{ij}+{\\lambda\\:}_{2}\\:{NetMedQ\\_scores}_{ij}+{\\lambda\\:}_{3}{\\:MSA\\_scores}_{ij}\\:+{{\\lambda\\:}_{4}\\:log\\_NOC}_{ij}+{u}_{0j}+{ϵ}_{ij}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e8\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere, λ\u003csub\u003e0\u003c/sub\u003e is the intercept term and λ\u003csub\u003ei\u0026rsquo;s\u003c/sub\u003e are the predictor variables\u0026rsquo; coefficients, namely MI_scores, NetMedQ_scores, MSA_scores and log_NOC, log\u003csub\u003ee\u003c/sub\u003e(CO34) is the predicted variable; u is random effect error term; ε is residual error term.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003eThe empirical results of the study are presented in this section.\u003c/p\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Descriptive Analysis\u003c/h2\u003e \u003cp\u003ePlots of the latent variables, namely average medical cost insured by households, average waiting period for medical treatment, distribution of ROHINI with NABH accreditations, and Number of complaints respectively in India are given in Figs.\u0026nbsp;1, 2, 3 and 4. respectively\u003c/p\u003e \u003cp\u003eIn Fig.\u0026nbsp;1, the average medical cost incurred by the household participating varies from INR500 to INR24000, with maximum in Delhi INR23,968.30 and lowest by Chandigarh and Manipur. In Fig.\u0026nbsp;2 the average waiting period before receiving medical treatment varies from 13 to 68 minutes, with Jammu and Kashmir (including the territory of Ladakh) showing the maximum average waiting time of 67 minutes and lowest in Punjab 13 minutes. In Fig.\u0026nbsp;3, the state-wise distribution of hospitals with NABH accreditation is displayed. Tripura has no hospitals with NABH accreditation while Uttar Pradesh shows the maximum number of hospitals with NABH accreditation totalling to 93,884. In Fig.\u0026nbsp;4, the number of complaints filed state-wise against non-life insurance ranging from 5 to 3,600 is displayed. The union territory Daman and Diu show has the least number of complaints equal to 9, while Maharashtra has the maximum number of complaints filed totalling to 3,516.\u003c/p\u003e \u003cp\u003eComparing Figs.\u0026nbsp;1 and 2, we see that Jammu and Kashmir, Kerala, Tamil Nadu, West Bengal and Karnataka show darker colouring which indicates that the average waiting time is more than the average medical cost, whereas the colouring is lighter in Madhya Pradesh. This suggests that medical care cost and average waiting time are positively related that may be attributed to high demand of the services which impacts the duration of waiting time to receive medical attention. Similarly, Figs.\u0026nbsp;3 and 4, suggest that Rajasthan, Gujarat, Maharashtra, Karnataka, Bihar showing contrast in the colouring shades which suggest that the number of hospitals with NABH accreditation and number of complaints filed against non-life insurance schemes have a negative relationship. Figures\u0026nbsp;2 and 3 suggest that the average waiting period for medical treatment and the distribution of hospital in the insurance network are negatively correlated for the states Uttar Pradesh, Bihar, Madhya Pradesh, Jammu and Kashmir. Kerala, West Bengal Andhra Pradesh, Assam, Bihar, Karnataka show contrast in the colouring shades which can be inferred as the number of hospitals in an insurance network is high than the average waiting time to receive medical care is less.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Latent Variable Development \u0026ndash; Structural Equation Modelling\u003c/h2\u003e \u003cp\u003eThe SEM is fitted on the following three latent variables: Misinformation (MI), Network Medical Quality (NetMedQ), and Medical Service Availability (MSA). The estimates of the fitted model are given in Table\u0026nbsp;4.\u003c/p\u003e \u003cp\u003eTable 4. Estimators of parameters from SEM\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Tabd\" border=\"1\"\u003e \u003ccolgroup cols=\"6\"\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=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLatent Variable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIndicator\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEstimate\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eStd. Err\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ez-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eP-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOG5_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOG5_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-1.325\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.179\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-7.414\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEducation_Group\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.091\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-5.238\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNetMedQ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_NABH_H\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_NABH_P\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.981\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e247.339\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_NABH_A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.959\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e244.920\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eQC3_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eQC3_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.917\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.486\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-1.884\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.060\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_QC8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8.170\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.570\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.179\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elog_num_doc\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-21.462\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6.691\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-3.208\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003cem\u003eSource: Author\u0026rsquo;s Contribution using IHDS, IIS, NHP, TATA AIG\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFrom the estimated values of the fitted model given in Table\u0026nbsp;4, we see that the coefficient α\u003csub\u003e2\u003c/sub\u003e of OG5_3 and α\u003csub\u003e3\u003c/sub\u003e of Education_Group are negative and significant at the 1% level. OG5_2 at 1.000 serves as a scaling factor for the latent variable MI. The interpretation for α\u003csub\u003e2\u003c/sub\u003e and α\u003csub\u003e3\u003c/sub\u003e is as follows: if the insured individual claims to be very knowledgeable about medical and education expenditure (OG5_3\u0026thinsp;=\u0026thinsp;1), the MI variable decreases by 1.325 units compared to those with no knowledge; however, if the highest adult education in the household is a undergraduate level of education's degree or above (Education_Group\u0026thinsp;=\u0026thinsp;1), the MI variable decreases by 0.091 units compared to households with lower education levels, repectively.\u003c/p\u003e \u003cp\u003eFor NetMedQ, the estimated coefficients β\u003csub\u003e2\u003c/sub\u003e of NABH_P and β\u003csub\u003e3\u003c/sub\u003e of NABH_A are positive and significant at the 1% level. Log_NABH_H at 1.000 serves as a scaling factor for the latent variable NetMedQ. The interpretation of β\u003csub\u003e2\u003c/sub\u003e and β\u003csub\u003e3\u003c/sub\u003e as follows: if the number of hospitals with pre-entry level accreditation increases by 1%, the NetMedQ variable increases by 0.981 units. Similarly, a 1% increase in hospitals with pre-accreditation leads to a 0.959 unit increase in Network Hospital Quality, respectively.\u003c/p\u003e \u003cp\u003eFor MSA, the estimated coefficient γ\u003csub\u003e2\u003c/sub\u003e of QC3_3 is significant at the 10% level, while the coefficients γ\u003csub\u003e1\u003c/sub\u003e, γ\u003csub\u003e3\u003c/sub\u003e, and γ\u003csub\u003e4\u003c/sub\u003e are significant at the 1% level, with positive and negative signs respectively. QC3_2 at 1.000 serves as a scaling factor for the latent variable MSA. The interpretation of γ\u003csub\u003e2\u003c/sub\u003e, γ\u003csub\u003e3\u003c/sub\u003e and γ\u003csub\u003e4\u003c/sub\u003e as follows: if persons get treatment in other town the MSA will decrease by 0.917 units, a percentage change in the waiting time of medical treatment (log_QC8) leads to an 8.170 unit increase in MSA. However, a 1% increase in the number of doctors (log_num_doc) results in a 21.462 unit decrease in MSA.\u003c/p\u003e \u003cp\u003eAfter estimating these parameters, the scores for the latent variables MI, MSA, and NetMedQ were computed as MI_scores, MSA_scores, and NetMedQ_scores. These scores are used in the hierarchical linear model (HLM) for further analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Regression \u0026ndash; Hierarchical Linear Modelling\u003c/h2\u003e \u003cp\u003eSince HLM can accommodate for both fixed and random effects in a model, it is used in the furtherance of this analysis. The level 1 model which accounts for within-group relationship are the part of fixed effects and level 2 model explain how the coefficients (intercept and slopes) from the Level 1 model may vary across groups, which are capture by random effects.\u003c/p\u003e \u003cp\u003e \u003cb\u003eRandom Effects\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe estimates of the fitted level 2 model from HLM are given result in Table\u0026nbsp;5\u003c/p\u003e \u003cp\u003eTable 5: Random Effects estimates from HLM.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Tabe\" border=\"1\"\u003e \u003ccolgroup cols=\"4\"\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=\"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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGroup\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eComponent\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEstimate\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eStd. Dev.\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSTATEID\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(Intercept)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.187\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.479\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResidual\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e17.098\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.135\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eSource: Author\u0026rsquo;s Contribution using IHDS, IIS, NHP, TATA AIG\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe STATEID group accounts for the variation in the baseline level of log_CO34 across different Indian states. The estimated variance of 2.187 signifies a significant difference in the starting point of log_CO34. The standard deviation of 1.479 indicates that, on average, states deviate from the mean intercept by about 1.479 units of log_CO34. This substantial heterogeneity suggests that there might be state-specific factors influencing the baseline levels of log_CO34. These factors could potentially include variations in policies, economic disparities, or differences in healthcare infrastructure across states. The Residual group captures the remaining unexplained variation in log_CO34 that the model could not account for through the fixed effects or the random intercepts by statewise categorization.\u003c/p\u003e \u003cp\u003e \u003cb\u003eFixed Effects\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe estimates of the fitted level 1 model from HLM are given in Table\u0026nbsp;6\u003c/p\u003e \u003cp\u003eTable 6: Fixed Effects estimates from HLM.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Tabf\" border=\"1\"\u003e \u003ccolgroup cols=\"4\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEstimate\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStd. Error\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003et value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMI_scores\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.5969\u003csup\u003e**\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.2344\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.546\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003elog_NOC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.8305\u003csup\u003e***\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.2315\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.588\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMSA_scores\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-30.8388\u003csup\u003e**\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e12.1783\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-2.532\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNetMedQ_scores\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.4908\u003csup\u003e**\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.2440\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-2.012\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eSource: Author\u0026rsquo;s Contribution using IHDS, IIS, NHP, TATA AIG\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe estimated coefficient λ\u003csub\u003e1\u003c/sub\u003e of MI_scores is significant at the 5% level, indicating a positive relationship between MI and log_CO34. This implies that a one unit increase in misinformation leads to a 0.59% increase in the cost of medical care by the insured. The estimated coefficient λ\u003csub\u003e4\u003c/sub\u003e of log_NOC is significant at the 1% level, showing that a 1% increase in the number of complaints results in a 0.83% increase in medical care costs. The estimated coefficient λ\u003csub\u003e3\u003c/sub\u003e of MSA_scores is significant at the 5% level. This shows a negative relationship between medical service availability and medical care costs, where a one unit increase in MSA_scores leads to a 30.83% decrease in medical care costs. Lastly, the estimated λ\u003csub\u003e2\u003c/sub\u003e of NetMedQ_scores is significant at the 5% level. This suggests that a one unit increase in the number of hospitals with good health care quality standards leads to a 0.49% decrease in medical care costs.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe results of model fitting given in Tables\u0026nbsp;4, 5 and 6 indicate the presence of discrepancies health care provided by insurance schemes. This significantly influences the medical costs incurred by households participating in health insurance policies. The following discussion delves into the specific impact of each discrepancy on actuarial fairness within health insurance.\u003c/p\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Impact of discrepancies in medical care services under health insurance schemes in India\u003c/h2\u003e \u003cdiv id=\"Sec17\" class=\"Section3\"\u003e \u003ch2\u003e4.1.1 Misinformation (MI)\u003c/h2\u003e \u003cp\u003eIn this study, the differences in knowledge between insured and insurer is defined as misinformation. This latent variable comprises of the variables which correspond to level of knowledge of medical information and education background. This is quantiatively represented in Table\u0026nbsp;4 as: OG5_1, OG5_2, OG5_3, and Education_Group.\u003c/p\u003e \u003cp\u003eThe estimated results indicate that the presence of misconceptions and lack of knowledge on health conditions and the related medical treatments. Households with very little knowledge represented by OG5_1, spend less on medical care, often choosing cheaper, lower quality services. Those with some knowledge represented by OG5_2, have the capacity to spend more on higher quality, more expensive health care services. Highly knowledgeable households represented by OG5_3 make cost-effective choices for quality care. The categorization across levels of education shows that households with an undergraduate level of education or higher incur lower costs. The results of our analysis confirm the presence of information gap which leads to inaccurate risk assessments, causing health insurance mispricing and making it actuarially unfair (Kleindorfer \u0026amp; Kunreuther, 1980). Hence, awareness programs on diseases, health care and access to basic education can lead to a better management of medical expenses, leading to fairer health insurance costs which aligning premiums with actual healthcare expenses and risks.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section3\"\u003e \u003ch2\u003e4.1.2 Network Medical Care Quality (NETMEDQ)\u003c/h2\u003e \u003cp\u003eIn this study, the quality of health care provided by the empanelled hospitals within the insurance framework are examined. The latent variable assesses the quality of medical care within an insurance network is constructed taking into account its crucial role in measuring healthcare costs and insurance fairness. It includes three indicators, namely, log_NABH_H, log_NABH_P, and log_NABH_A, which are tabulated in Table\u0026nbsp;4.\u003c/p\u003e \u003cp\u003eThe estimated results show a positive relationship with NetMedQ which suggest that increase in the number of NABH accredited hospitals improve medical care provided by the network of empanelled hospitals. We see that a higher quality network tends to reduce healthcare costs by providing effective and efficient care, reducing the need for repeated treatments. Failure to maintain high-quality care in the network leads to overpriced insurance policies, as households might seek more expensive out-of-network services (Raposo, Alves, \u0026amp; Duarte, 2009), resulting in actuarial unfairness. To ensure fairness, insurers must maintain rigorous quality standards and increase the number of accredited hospitals, aligning premiums with the true value of services provided.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section3\"\u003e \u003ch2\u003e4.1.3 Medical Service Availability (MSA)\u003c/h2\u003e \u003cp\u003eThe accessibility to quality health care is a crucial determinant in determining the costs associated with medical services. The latent variable used in this study measures the accessibility to medical services in an Indian state. This latent variable comprises of the variables which correspond to the geographical location of the health care provider and number of allopathy doctors available in the respective Indian state. This is quantiatively respresented in Table\u0026nbsp;4 as: QC3_2, QC3_3, log_QC8 and log_num_doc.\u003c/p\u003e \u003cp\u003eThe estimated results show that QC3_2 and log_QC8 show a positive relationship with MSA, suggesting that when medical services are more accessible and when more healthcare providers such as hospitals or clinics are available, there is a positive impact on the perception towards medical care. However, we see that the estimates of QC3_3 and log_num_doc show a negative relationship with MSA, this indicates the presence of challenges in accessibility to quality health care. This is most common in rural parts of India where medical services are predominantly available in distant locations and it reduces overall accessibility. Hence, enhancing service availability through expanded infrastructure and clearer communication about coverage limitations could help mitigate these issues (Xiong, et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section3\"\u003e \u003ch2\u003e4.1.4 Inconsistencies in claims processing (log_NOC)\u003c/h2\u003e \u003cp\u003eIn this study, the inconveniences caused by the insurer and the insured at the time of submission of health insurance claims is called as inconsistencies in claims processing. This variable is pivotal in determining the fairness of health insurance pricing as indemnification is the underlying principle of insurance. It is quantitively measured by the logarithmic transformation of the number of complaints from an Indian state and is given in Table\u0026nbsp;6.\u003c/p\u003e \u003cp\u003eAs shown in Table\u0026nbsp;6, a proportional increase in medical costs is associated with a corresponding rise in the number of complaints. An increase in complaints also indicates which can result in the mispricing of health insurance, either making it overpriced or underpriced. Consequently, as inconsistencies in the claims process increase, so do medical costs, leading to actuarial unfairness in health insurance. Therefore, it is essential to streamline the claims processing by including measures that allow for transparency, active grievance redressal which will minimize inconsistencies and lead to fairness in health insurance.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThis study begins by highlighting critical findings from the Indian Human Development Survey-II (IHDS-II), which reveal that approximately 40.8 percent of households with health insurance have incurred additional medical expenses exceeding their risk cover. With a doctor-to-population ratio of 1:1457 in 2011, far below the WHO\u0026rsquo;s recommended 1:1000, the shortage of healthcare resources in India is evident (Deo, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Furthermore, systemic issues such as uneven resource distribution, poor infrastructure, and administrative inefficiencies (Mavalankar et al., 2000) continue to disrupt the delivery of quality care under health insurance schemes. These unaddressed challenges lead to a mismatch between the expected cost of premiums and the actual cost of medical care, raising concerns about the fairness of actuarial principles underlying these insurance policies.\u003c/p\u003e \u003cp\u003eBuilding on these foundational observations, the study employed SEM and HLM to examine how discrepancies like misinformation, accessibility, hospital quality, and claims process inconsistencies impact medical costs and the fairness of health insurance pricing. The results of the study show statistically significant estimates of these discrepancies which lead to the mismatch of experience adjustments used in actuarial valuation in comparison to the realised costs incurred by insured households.\u003c/p\u003e \u003cp\u003eMisinformation, driven by knowledge gaps, leads to differences in medical spending. Households with less skills obtained through education tend to incur lower initial costs, possibly sacrificing quality, while more informed households make cost-effective decisions. Reducing misinformation through education can align consumer perceptions with reality, supporting actuarial fairness. The quality of care within insurance networks, as indicated by NABH accredited hospitals, enhances health outcomes and reduces costs. Improved facilities in hospitals and accountability in the role of medical practitioners could help insurers maintain fair premium structures, ensuring customers receive the true value for their premiums. Medical service availability significantly impacts costs associated with good quality healthcare. Limited access to services often leads to higher expenditures due to travel or higher local prices, resulting in unfair high premiums. Expanding infrastructure and clear communication about coverage could mitigate these issues. Inconsistencies in the claim process, reflected by rising complaints, suggest inefficiencies that increase medical costs. Including customer friendly policies such as transparency, grievance redressal within the scope of processing health insurance claims can reduce these inconsistencies.\u003c/p\u003e \u003cp\u003eIn summary, this study highlights the importance of experience studies in actuarial science and the need for updating experience assumptions to maintain actuarial fairness in pricing health insurance. The discrepancies present in health insurance costs, including misinformation, inadequate healthcare resources, and systemic inefficiencies, significantly impact the alignment between insurance premiums and actual medical costs. Addressing these issues by improving public knowledge on healthcare schemes and practices, enhancing the quality of healthcare provided, and expanding accessibility to medical services, can lead to more equitable insurance pricing that truly reflects the risks and costs faced by households. By adopting these strategies, policy makers and insurers can foster greater trust and satisfaction among policyholders which could lead to client retention and at large a healthier society. The scope of this analysis is restricted to the IHDS-II data. Future research should continue to identify and mitigate additional discrepancies within the health insurance sector, while also exploring the ethical and legal implications of these disparities, especially in relation to vulnerable populations, to promote equity and efficiency in the health insurance sector.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eCOMPETING INTERESTS\u003c/h2\u003e \u003cp\u003eThe authors do not have any financial competing interest in the publication of this paper. This research was done solely for academic purposes alone.\u003c/p\u003e \u003c/p\u003e\u003cp\u003e \u003ch2\u003eETHICS DECLARATION\u003c/h2\u003e \u003cp\u003eNot applicable.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFUNDING\u003c/h2\u003e \u003cp\u003eThere was no funding involved in the design of study, data collection, analysis and interpretation as this research was a part of the dissertation report carried out as part of the curriculum of the M.A. Actuarial Economics program at the Madras School of Economics, Chennai, India.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eNikhil Rathee: Writing \u0026ndash; original draft, Conceptualization, Data curation, Formal analysis, Methodology, Investigation. Rupel Nargunam: Writing \u0026ndash; review \u0026amp; editing, Supervision, Methodology, Formal analysis, Data curation, Conceptualization, Visualization, Validation.All authors reviewed the manuscript\u003c/p\u003e\u003ch2\u003eACKNOWLEDGEMENTS\u003c/h2\u003e \u003cp\u003eOnly the authors worked upon this paper and there were no other individuals who have contributed towards this article. The usual disclaimers apply.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data sources include the India Human Development Survey II (IHDS-II), which covers over 42,000 households across India and provides data on health insurance coverage, health utilization, and economic status. The Indian Insurance Statistics (IIS) Handbook 2011-12, published by the IRDAI, offers detailed statistics on the insurance market, including policy numbers, consumer complaints, and insurer performance metrics. The National Health Profile (NHP) 2011, compiled by the Central Bureau of Health Intelligence, provides information on health infrastructure, healthcare professionals, disease prevalence, and public health finances. Additionally, data from Tata AIG Ltd. on hospitals registered under ROHINI (Registry of Hospitals in Network of Insurance) offers insights into the accessibility and quality of services from managed hospitals.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eABEL-SMITH, BRIAN. \u0026ldquo;Health Insurance in Developing Countries: Lessons from Experience.\u0026rdquo; Health Policy and Planning 7, no. 3 (1992): 215\u0026ndash;26. http://www.jstor.org/stable/45090068.\u003c/li\u003e\n\u003cli\u003eAl-Rawashdeh, Firas. 2016. Determinates of demand on purchasing insurance policies: Case of Jordan. International Review of Management and Marketing, 6, 1094-1102.\u003c/li\u003e\n\u003cli\u003eArnold, Jens Matthias, Beata Javorcik, Molly Lipscomb, and Aaditya Mattoo. \u0026quot;Services reform and manufacturing performance: Evidence from India.\u0026quot; The Economic Journal 126, no. 590 (2016): 1-39.\u003c/li\u003e\n\u003cli\u003eDeo, Madhav G. \u0026quot;Doctor population ratio for India-The reality.\u0026quot; Indian Journal of Medical Research 137, no. 4 (2013): 632-635.\u003c/li\u003e\n\u003cli\u003eDesai, Sonalde, Amaresh Dubey, B. L. Joshi, Mitali Sen, Abusaleh Shariff, and Reeve Vanneman. \u0026quot;India human development survey.\u0026quot; College Park, Maryland: University of Maryland (2008).\u003c/li\u003e\n\u003cli\u003eDonahue, Kate, and Solon Barocas. \u0026quot;Better together? how externalities of size complicate notions of solidarity and actuarial fairness.\u0026quot; In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, pp. 185-195. 2021.\u003c/li\u003e\n\u003cli\u003eEckstein, David, Vera K\u0026uuml;nzel, and Laura Sch\u0026auml;fer. The global climate risk index 2021. Bonn: Germanwatch, 2021. https://bvearmb.do/handle/123456789/1306\u003c/li\u003e\n\u003cli\u003eGollier, Christian. \u0026quot;The determinants of the insurance demand by firms.\u0026quot; Toulouse School of Economics (LERNA and IDEI). Toulouse: Toulouse School of Economics, 2007. http://tse-fr.eu/pub/7278\u003c/li\u003e\n\u003cli\u003eHeras, Antonio J., Pierre-Charles Pradier, and David Teira. \u0026quot;What was fair in actuarial fairness?.\u0026quot; History of the Human Sciences 33, no. 2 (2020): 91-114.\u003c/li\u003e\n\u003cli\u003eHoyt, Robert E., and Ho Khang. \u0026quot;On the demand for corporate property insurance.\u0026quot; Journal of Risk and Insurance (2000): 91-107.\u003c/li\u003e\n\u003cli\u003eHuta, Veronika. \u0026quot;When to use hierarchical linear modeling.\u0026quot; The quantitative methods for psychology 10, no. 1 (2014): 13-28.\u003c/li\u003e\n\u003cli\u003eKajwang, Ben. \u0026quot;Contribution of Industrial All Risks Insurance to the Growth of Manufacturing Firms in Kenya.\u0026quot; Journal of Developing Country Studies 5, no. 1 (2020): 36-48.\u003c/li\u003e\n\u003cli\u003eKleban, Jaclyn. \u0026quot;Countering Misinformation in the Health Care System: The Case for Stricter Regulations Within Health Insurance Provider Directories.\u0026quot; CARDOzO L. REv. 41 (2019): 1185.\u003c/li\u003e\n\u003cli\u003eKleindorfer, Paul R., and Howard Kunreuther. \u0026quot;Misinformation and equilibrium in insurance markets.\u0026quot; In Economic Analysis of Regulated Markets, pp. 67-90. London: Palgrave Macmillan UK, 1980.\u003c/li\u003e\n\u003cli\u003eKrummaker, Simone. \u0026quot;Corporate Demand for Insurance: Empirical Evidence from Germany.\u0026quot; (2016). https://openaccess.city.ac.uk/id/eprint/24066/\u003c/li\u003e\n\u003cli\u003eLevy, Helen, and David Meltzer. \u0026quot;The impact of health insurance on health.\u0026quot; Annu. Rev. Public Health 29, no. 1 (2008): 399-409.\u003c/li\u003e\n\u003cli\u003eLieberthal, Robert D., and Robert D. Lieberthal. \u0026quot;Defining Health Insurance.\u0026quot; What Is Health Insurance (Good) For? An Examination of Who Gets It, Who Pays for It, and How to Improve It (2016): 3-32.\u003c/li\u003e\n\u003cli\u003eLinnerooth-Bayer, Joanne, and Reinhard Mechler. \u0026quot;Insurance against losses from natural disasters in developing countries.\u0026quot; IIASA, Background paper for UN WESS, July (2008).\u003c/li\u003e\n\u003cli\u003eMavalankar, Dileep, and Ramesh Bhat. \u0026quot;Health insurance in India: opportunities, challenges and concerns.\u0026quot; Ahmedabad: Indian Institute of Management (2000): 1-16.\u003c/li\u003e\n\u003cli\u003eMayers, David, and Clifford W. Smith Jr. \u0026quot;On the corporate demand for insurance.\u0026quot; In Foundations of Insurance Economics: Readings in Economics and Finance, pp. 190-205. Dordrecht: Springer Netherlands, 1982.\u003c/li\u003e\n\u003cli\u003eMotsepe, Molatelo. \u0026quot;An Industry Level Analysis of Demand for Insurance in South Africa.\u0026quot; (2018). University of Cape Town, Faculty of Commerce. Cape Town: Graduate School of Business (GSB). http://hdl.handle.net/11427/28411\u003c/li\u003e\n\u003cli\u003eMushonga, Farai B., and Syden Mishi. \u0026quot;Natural hazard insurance demand: A systematic review.\u0026quot; J\u0026agrave;mb\u0026aacute;-Journal of Disaster Risk Studies 14, no. 1 (2022): 1223.\u003c/li\u003e\n\u003cli\u003eMuthen, Bengt. \u0026quot;Latent variable structural equation modeling with categorical data.\u0026quot; Journal of Econometrics 22, no. 1-2 (1983): 43-65.\u003c/li\u003e\n\u003cli\u003eNasir, H. U. (2024). Actuarial fairness: what does it mean and when is it desirable? School of Graduate Studies Electronic Theses and Dissertations, 154.\u003c/li\u003e\n\u003cli\u003eProkopjeva, Evgenija, Galina Chernova, Nataliya P. Kuznetsova, Svetlana Kalayda, and Leonid Ivanov. \u0026quot;Insurance instruments in risk management of the manufacturing sector of a region: the case of the Republic of Khakassia (Russia).\u0026quot; Investment Management \u0026amp; Financial Innovations 17, no. 4 (2020): 299.\u003c/li\u003e\n\u003cli\u003eRao, M. Subba, and R. Srinivasulu. \u0026quot;Contribution of insurance sector to growth and development of the Indian economy.\u0026quot; IOSR journal of Business and Management 7, no. 4 (2013): 45-52.\u003c/li\u003e\n\u003cli\u003eRaposo, M\u0026aacute;rio Lino, Helena Maria Alves, and Paulo Alexandre Duarte. \u0026quot;Dimensions of service quality and satisfaction in healthcare: a patient\u0026rsquo;s satisfaction index.\u0026quot; Service Business 3 (2009): 85-100.\u003c/li\u003e\n\u003cli\u003eRegan, Laureen, and Yeon Hur. \u0026quot;On the corporate demand for insurance: The case of Korean nonfinancial firms.\u0026quot; Journal of Risk and Insurance 74, no. 4 (2007): 829-850.\u003c/li\u003e\n\u003cli\u003eRothschild, Casey. \u0026quot;The efficiency of categorical discrimination in insurance markets.\u0026quot; Journal of Risk and Insurance 78, no. 2 (2011): 267-285.\u003c/li\u003e\n\u003cli\u003eSantos, Gustavo. (2023, March 15). Multilevel Regression with R. Retrieved from Towards Data Science: https://towardsdatascience.com/multilevel-regression-with-r-eb3eb7d8de88\u003c/li\u003e\n\u003cli\u003eSouthwell, Brian G., Jessica Otero Machuca, Sabrina T. Cherry, Melissa Burnside, and Nadine J. Barrett. \u0026quot;Health misinformation exposure and health disparities: observations and opportunities.\u0026quot; Annual review of public health 44, no. 1 (2023): 113-130.\u003c/li\u003e\n\u003cli\u003eWood, Glenn L., and Richard L. Morris. \u0026quot;CONSUMER COMPLAINTS AGAINST INSURANCE COMPANIES.\u0026quot; Journal of Personal Finance 9 (2010).\u003c/li\u003e\n\u003cli\u003eWright, Sewall. \u0026quot;On the nature of size factors.\u0026quot; Genetics 3, no. 4 (1918): 367.\u003c/li\u003e\n\u003cli\u003eXiong, Xiaolei, Zhiguo Zhang, Jing Ren, Jie Zhang, Xiaoyun Pan, Liang Zhang, Shiwei Gong, and Si Jin. \u0026quot;Impact of universal medical insurance system on the accessibility of medical service supply and affordability of patients in China.\u0026quot; PloS one 13, no. 3 (2018): e0193273.doi:https://doi.org/10.1371/journal.pone.0193273\u003c/li\u003e\n\u003cli\u003eYoshida, Kaori, and Robert E. Deyle. \u0026quot;Determinants of small business hazard mitigation.\u0026quot; Natural Hazards Review 6, no. 1 (2005): 1-12.\u003c/li\u003e\n\u003cli\u003eZhou, Fujin, and Wouter Botzen. \u0026quot;Firm level evidence of disaster impacts on growth in Vietnam.\u0026quot; Environmental and Resource Economics 79, no. 2 (2021): 277-322.\u003c/li\u003e\n\u003cli\u003eZhu, Minglai, Chao Kui, and Yuzheng Fang. \u0026quot;Demand for corporate insurance in China: empirical analysis with province-level panel data.\u0026quot; Chinese Economy 44, no. 6 (2011): 15-29.\u003c/li\u003e\n\u003cli\u003eZou, Hong, and Mike B. Adams. \u0026quot;The corporate purchase of property insurance: Chinese evidence.\u0026quot; Journal of Financial Intermediation 15, no. 2 (2006): 165-196.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Pricing Health Insurance, Actuarial Fairness, Discrepancies, Misinformation, Claim process inconsistencies, Access to Medical Services","lastPublishedDoi":"10.21203/rs.3.rs-6613076/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6613076/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study investigates the actuarial fairness of pricing health insurance policies by examining discrepancies within the Indian health insurance and their impact on medical costs. Health insurance helps an individual in financing their medical expenses, for a fee called premium which is typically the reflection of expected cost of medical services. However, discrepancies such as misinformation, accessibility to health care services, hospital quality and Inconsistencies in claims processing, increase costs associated with health care of individuals participating in the health insurance, thereby challenging the fairness of these policies. Structural Equation Modelling (SEM) has been employed to develop latent variables representing these discrepancies and Hierarchical Linear Modelling (HLM) has been used to assess their effect on the cost of medical care. The findings of this study suggest the presence of discrepancies in the Indian health insurance sector and the results support the significant impact on the increase in medical expenses. The study concludes with policy recommendations aimed at enhancing the efficiency, effectiveness and fairness of the health insurance policies in India\u003c/p\u003e \u003cp\u003e \u003c/p\u003e","manuscriptTitle":"An Actuarial Perspective on the Discrepancies in Health Insurance Pricing: Evidence from India","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-13 10:15:48","doi":"10.21203/rs.3.rs-6613076/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1e418c1b-e292-4a03-b3c2-09e512a670cf","owner":[],"postedDate":"June 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-28T17:23:48+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-13 10:15:48","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6613076","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6613076","identity":"rs-6613076","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}