Do publicly supported generic pharmacies improve financial risk protection? Findings from a nationally representative cross-sectional survey in India Authors: Neha Purohit, Aarti Goyal, Gaurav Jyani, Biju Soman, Sitanshu Sekhar Kar, Ramesh Verma, Kuldeep Singh, Sandra Albert, Prakash Patel, Saroj Kumar Rana, Sanjay Kumar, Sashi Bushan Biswal, Shankar Prinja Published: 2026 ABSTRACT Abstract AbstractObjectivesThe Government of India launched the Pradhan Mantri Bhartiya Janaushadhi Pariyojana (PMBJP) to expand access to affordable generics through private retail outlets named as Jan Aushadhi Kendras (JAKs). This study examines the association of PMBJP with out-of-pocket expenditure (OOPE), catastrophic health expenditure (CHE) and impoverishment rate (IR) attributable to medicines.DesignA cross-sectional observational study was conducted across nine Indian states in 2022–2023.SettingOutpatient (OPD) and inpatient (IPD) departments of secondary and tertiary government hospitals, private pharmacies and JAKs in 18 districts of IndiaParticipantsA total of 10 336 patients were recruited from OPD (n=2881) and IPD (n=1009) departments of government hospitals as well as pharmacy settings (n=6446). Data on sociodemographics, disease severity, number of generic prescriptions, source of acquiring medicines and medicine-related OOPE were collected through semistructured interviews and periodic follow-ups.Primary and secondary outcomesPrimary outcomes included mean OOPE on medicines, incidence of CHE (≥40% of non-food consumption expenditure on medicines), IR among JAK and non-JAK users were the primary outcomes of the study. Secondary outcomes comprised awareness of JAKs, generic prescribing rates in hospitals and the factors associated with OOPE, CHE and IR.ResultsPatients procuring medicines exclusively from JAKs reported the lower mean OOPE (OPD: ₹172; IPD: ₹275; pharmacy: ₹307), compared with significantly higher spending at private pharmacies (OPD: ₹1085; IPD: ₹3165; pharmacy: ₹1031). After adjusting for covariates, OOPE among exclusive JAK users was significantly lower relative to private pharmacy users by 60.6%–89.3%. Furthermore, matched analysis confirmed 42% lower expenses, compared with private pharmacies. The likelihood of CHE was also significantly greater among private pharmacy users. However, utilisation of JAKs remained limited, mainly due to low awareness, perceived stock shortages and low rates of generic prescribing.ConclusionPMBJP is associated with significant reduction in OOPE and financial hardship, positioning it as an effective cost-containment intervention within India’s universal health coverage framework. Strengthening supply chains, promoting generic prescribing and integrating JAKs with public facilities would further maximise its impact.Keywords: Health Care Costs, HEALTH ECONOMICS, Health policy, PUBLIC HEALTH STRENGTHS AND LIMITATIONS OF THIS STUDY The study used a large, multistate, nationally representative cross-sectional survey, with longitudinal follow-up of participants to estimate 30-day medicine-related expenditures on disorders. The study used a large, multistate, nationally representative cross-sectional survey, with longitudinal follow-up of participants to estimate 30-day medicine-related expenditures on disorders. Propensity score matching was applied to reduce residual confounding and selection bias between users and non-users of Jan Aushadhi Kendras (JAKs), thereby enhancing the internal validity of the results. Propensity score matching was applied to reduce residual confounding and selection bias between users and non-users of Jan Aushadhi Kendras (JAKs), thereby enhancing the internal validity of the results. Mean out-of-pocket (OOP) expenditure on medicines was adjusted for confounding sociodemographic and clinical factors including disease severity and prescription patterns. Notably, EuroQol five-dimension-five-level utility scores were used as an objective and novel proxy measure to adjust for disease severity and its effect on OOP expenditure. Mean out-of-pocket (OOP) expenditure on medicines was adjusted for confounding sociodemographic and clinical factors including disease severity and prescription patterns. Notably, EuroQol five-dimension-five-level utility scores were used as an objective and novel proxy measure to adjust for disease severity and its effect on OOP expenditure. The cross-sectional observational nature of the study limits causal inference, and residual selection bias may remain despite robust matching techniques. The cross-sectional observational nature of the study limits causal inference, and residual selection bias may remain despite robust matching techniques. The analysis was limited to a predefined set of medical conditions for which medicines are available through JAKs, which may restrict the generalisability of findings to patient groups with other diseases. The analysis was limited to a predefined set of medical conditions for which medicines are available through JAKs, which may restrict the generalisability of findings to patient groups with other diseases. Background Access to affordable, quality-assured medicines is central to the achievement of universal health coverage (UHC) and to the broader goals of reducing financial hardship and improving health equity. Despite global commitments under Sustainable Development Goal target 3.8 to ensure equitable access to essential medicines,1 evidence from multiple regions, including South-East Asia, Europe and parts of Africa, indicates that the cost of medicines remains a major contributor to out-of-pocket expenditure (OOPE) and catastrophic health expenditure (CHE), particularly among economically vulnerable populations.2 India presents a particularly salient case in the global discourse on medicine affordability. Despite being home to one of the world’s largest generic pharmaceutical industries, medicines account for over 60% of household health expenditure.3 Public-sector provision of medicines in India is uneven across levels of care: while medicines are intended under national and state policies to be provided free of charge,4 particularly for inpatient care in government secondary and tertiary hospitals, coverage remains variable in practice.5 Patients therefore frequently rely on private pharmacies, even when consultations occur in public facilities. Stock-outs, limited formularies and fragmented procurement often necessitate out-of-pocket purchases, resulting in heterogeneous medicine-acquisition pathways.5 Health insurance coverage in India has expanded through publicly financed schemes such as Ayushman Bharat–Pradhan Mantri Jan Arogya Yojana (PMJAY) and a range of state-sponsored schemes.6 However, insurance coverage remains heterogeneous, and most publicly financed health insurance schemes in India primarily cover inpatient care and provide limited or no coverage for outpatient medicines or long-term pharmacotherapy for chronic conditions.6 Consequently, even insured households continue to incur substantial OOPE on medicines, particularly for outpatient care and chronic disease management. India’s pharmaceutical market is dominated by off-patent medicines sold as branded or unbranded generics, while patented originator products constitute a smaller, higher-priced segment concentrated in specialised settings.7 Although therapeutically equivalent, branded generics cover 70% of the market and are priced 20–60% above the unbranded alternative due to brand promotion, asymmetric price information and prescribing incentives.8 Regulatory price controls apply to selected medicines under the National List of Essential Medicines, but significant price variation persists across brands.9 While policy frameworks mandate generic-name prescribing in public facilities and encourage price transparency, enforcement in private-sector settings remains limited and generic substitution is uncommon in routine practice.10 Public-sector medicine procurement in India is conducted through competitive tendering mechanisms at the central and state levels, aimed at securing quality-assured medicines at low cost.11 However, persistent gaps in availability at the point of care, limited outpatient medicine coverage under insurance schemes and continued reliance on private retail markets have constrained the effectiveness of these mechanisms in reducing household OOPE. To address these limitations, the Government of India launched the Pradhan Mantri Bhartiya Janaushadhi Pariyojana (PMBJP) to expand access to affordable unbranded generic medicines through centrally negotiated procurement and distribution via a nation-wide network of dedicated private retail outlets known as Jan Aushadhi Kendras (JAKs).12 PMBJP operates through a regulated network of privately run pharmacies, who are provided with one-time capital investment amount at the outset, while the routine operational costs are borne by the private pharmacies. Medicines are procured from a list of the WHO-Good Manufacturing Practices-certified suppliers and each batch of the drug is tested at accredited laboratories to ensure the quality of the medicines.12 13 As of 2022, the scheme had scaled up significantly, with over 9000 JAKs operating across the country, offering a formulary of more than 1800 medicines and surgical products at 50%–90% lower prices than their branded alternatives.14 In parallel, the Affordable Medicines and Reliable Implants for Treatment (AMRIT) scheme was introduced to provide discounted medicines and medical implants within selected public hospitals, primarily targeting tertiary-care settings.15 The coexistence of PMBJP and AMRIT reflects the incremental evolution of India’s pharmaceutical access policies, with each scheme addressing distinct gaps in medicine availability and affordability across levels of care rather than representing duplication. As one of the largest government-supported pharmaceutical access programmes globally, the PMBJP holds promise for countries facing similar challenges of high OOPE and fragmented private markets. However, evidence to date has focused largely on supply-side performance indicators, such as availability, pricing and stock levels, and demand-side proxy metrics like awareness and perceptions of generics,16,20 with limited empirical data assessing whether PMBJP is associated with measurable reductions in OOPE at household level. In this context, the present study aims to evaluate the association of PMBJP with OOPE on medicines and financial risk protection in India, in order to inform the strengthening of affordable and sustainable pharmaceutical systems. Methods Study design A cross-sectional, multisite investigation was carried out to evaluate the association of PMBJP with OOPE and CHE attributed to medicines. The study applied a multistage random sampling technique to ensure both rigour and representativeness. In the first stage, states were grouped into two strata based on household spending on medicines, as reported in the Consumer Expenditure Survey in the 74th round of the National Sample Survey.21 Four states were randomly selected from each of the two strata defined by high and low expenditure on medicines, resulting in the inclusion of Uttar Pradesh, Gujarat, Tamil Nadu, Kerala, Haryana, Rajasthan, Bihar and Odisha. To ensure regional representation, Assam was purposefully included to represent the northeastern part of the country. In the next stage, within each state, one district each was randomly chosen from two strata of districts created based on the Multi-Dimensional Poverty Index, which accounts for disparities in education, health and living conditions in each district within the state.22 Within each district, one secondary-level and one tertiary-level public health facility were randomly identified. In addition, three JAKs were randomly selected, and for each JAK, one nearby private pharmacy was purposively chosen based on geographical proximity to facilitate comparability in catchment population and to ensure comparative analysis of medicine expenditure and CHE. Sample size and participant selection In order to estimate the required sample size to detect the extent of change in OOPE, 5% level of significance, with 85% power and 5% error margin was used. The sample was estimated individually for recruitment at the health facilities and the pharmacies. In health facilities, the participant sample size was determined based on an expected reduction in the prevalence of CHE from 12% to 9.5%, resulting in a final sample size of 3224. For pharmacies, the total sample size was estimated at 6288. This sample was proportionally allocated between JAKs and private pharmacies in a 1:3 ratio. This oversampling was undertaken to improve the quality and feasibility of matching of private pharmacy users with JAK users, accounting for heterogeneity in sociodemographic and clinical factors. Since the type and severity of illness are key determinants of OOPE on medicines, only patients diagnosed with a set of predefined health conditions were included. These conditions were selected based on the following criteria: considerable population prevalence, well-defined case definitions with limited diagnostic ambiguity, presence of disease-specific relevant medicines and consumables in PMBJP drug lists and representation of both acute and chronic diseases. Based on these considerations, the selected conditions included diabetes mellitus, hypertension, congestive heart failure, osteoarthritis, cervical cancer, breast cancer, schizophrenia, chronic obstructive pulmonary disease and pneumonia. The participant selection for outpatient and inpatient settings was based on age (≥18 years), confirmed diagnosis of any of the selected disorders and willingness to participate in the study. There was no disease criterion for selection of participants at the pharmacies. A consecutive sampling approach was followed to recruit the sample at all the facilities. Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research. Data collection Data collection was carried out between April 2022 and March 2023 by trained local field teams using the computer-assisted personal interview method. Data were gathered from participants visiting outpatient departments (OPDs) and inpatient wards at the selected secondary and tertiary care facilities, as well as from six pharmacies—comprising three JAKs and three private pharmacies in each state (figure 1). Approach to data collection. IPD, inpatient department; OPD, outpatient department. Non-hospitalised participants were interviewed on the day of recruitment, whereas hospitalised patients or their caregivers were interviewed daily during their hospital stay to document their choice of pharmacy and expenditure on medicines. Following recruitment (for outpatients) and discharge (for inpatients), participants were followed up telephonically at three intervals: 24 hours, 15 days and 30 days, to collect information on subsequent pharmacy access and medicine-related OOPE. Data collection tools Three semistructured tools, tailored to facility settings, were developed for data collection. Each tool comprised seven main sections capturing sociodemographic details, information on consumption expenditure, clinical history, self-assessed baseline health status (using the EuroQol five-dimension- five-level (EQ-5D-5L) instrument),23 prescription audit, source of medicine procurement and medicine-related expenditure incurred on the day of enrolment (Sections S1, S2, S3). To acquire information on 30-day OOPE from the date of enrolment for non-hospitalised patients and the date of discharge for inpatient participants, three additional sections were included to gather follow-up data on medicine access and related expenses. The reasons for non-utilisation of JAKs were elicited on the 15th day interview. The tools were digitised using Research Electronic Data Capture (REDCap) software (REDCap Mobile App (Mobile application software). Vanderbilt University. https://projectredcap.org/software/mobile-app/) to enable streamlined and standardised data collection. Data analysis Data were analysed using SPSS V.21 (SPSS, Chicago, Illinois, USA), Stata V.13 (StataCorp, College Station, Texas, USA) and Microsoft Excel (Microsoft Corporation, Redmond, Washington, USA). Descriptive statistics were generated to summarise the sociodemographic and clinical characteristics of outpatient, inpatient and pharmacy-based respondents. Health-related EQ-5D-5L based quality of life scores were derived using the Indian value set.24 Based on these scores, disease severity was categorised as mild (EQ-5D-5L value ≥0.8), moderate (EQ-5D-5L value of 0.56–0.79) or severe (EQ-5D-5L value ≤0.55). The mean OOPE on medicines was calculated by disease condition, disease severity, facility type and number of prescribed medicines. All OOPE values are reported in Indian National Rupees (₹). To reduce selection bias and ensure comparability between participants recruited at JAK and non-JAK facilities, we employed propensity score matching (PSM). Propensity scores were estimated using a logistic regression model that included type and severity of disease, number of prescribed drugs, area of residence (rural–urban), prescription of generic medicines, household socio-economic status (wealth quintile) and educational status of the respondent as covariates. The propensity score represented the probability of receiving the treatment/exposure, conditional on the observed covariates. Matching was performed using a calliper width of 0.2 of the SD of the logit of the propensity score. Balance diagnostics, including standardised mean differences (SMDs) before and after matching, were assessed to evaluate the success of the matching procedure. Post-matching outcome comparisons were conducted using appropriate regression models that accounted for the matched nature of the data. To explore the correlates of OOPE on medicines, multiple linear regression analyses were conducted separately for outpatient, inpatient and pharmacy respondents using the ordinary least squares estimation method. To correct for right-skewed distributions in expenditure data, continuous variables were log-transformed. Categorical variables were appropriately recoded into dummy variables. The regression model followed the standard linear form: where Y is the outcome variable, Xi is the value of the ith predictor and e is the error. We used medicine expenditure as a dependent variable, while remaining variables of patient’s age, gender, education level, occupation, religion, place of residence, marital status, caste, status of health insurance, type of family, type of facility, number of drugs prescribed and type and severity of diseases were used as predictors to understand their influence on the response variable. The assumptions regarding normality of residuals and regressors; homoscedasticity and multicollinearity among were assessed using Kolmogorov-Smirnov test, Breusch-Pagan test and variance inflation factors, respectively. Financial risk protection was evaluated for those accessing JAKs, non-JAK or both types of pharmacy outlets using two key indicators: CHE and impoverishment rate (IR). CHE was defined as spending on medicines exceeding 40% of a household’s non-food consumption expenditure, consistent with established thresholds in the literature.25 26 IR was also calculated in terms of increase in poverty headcount, defined as the percentage of households falling below the poverty line after accounting for medicine-related OOPE. The poverty line is equivalent to subsistence expenditure, adjusted for household size.26 To identify the factors associated with CHE and IR, multiple logistic regression analyses were performed, incorporating the sociodemographic and clinical explanatory variables. Sensitivity analysis was conducted to estimate the prevalence of CHE at different thresholds of total and non-food expenditure. Results Sociodemographic and clinical profile of participants A total of 2881, 1009 and 6446 participants were recruited from the OPD, inpatient department (IPD) and pharmacies, respectively. After excluding the participants with incomplete data (n=207), the analysis was undertaken on the remaining sample. The sociodemographic and clinical profile of patients is presented in online supplemental table S1. The mean age of the respondents recruited in OPD, IPD and pharmacies was 52.28 years, 53.47 years and 44.38 years, respectively, with about 55% of the sample being male respondents. Around 29.8%, 38.6% and 27.3% of respondents recruited in OPD, IPD and pharmacy settings, respectively, reported having some form of health insurance. Approximately one-third of the respondents at all facilities were classified as having mild, moderate and severe disease (online supplemental table S1). We found that the mean number of medicines prescribed per patient was 5.1 in outpatient settings and 9.2 in inpatient settings. The corresponding rates of generic prescription were 18% and 6% in outpatient and inpatient settings, respectively. With respect to medicine availability, 41.9% of prescribed outpatient and 29.5% of inpatient medicines were dispensed directly through public health facilities. An additional 11.4% and 10.0% of medicines were obtained from JAK and other subsidised pharmacies, respectively. The remaining 48.5% of outpatient and 37.7% of inpatient medicines were purchased from private retail pharmacies. Out-of-pocket expenditure on medicines and associated factors OOPE on medicines varied significantly by point of care and source of medicine procurement. The lowest mean OOPE was consistently observed among patients who procured medicines exclusively from JAKs (OPD: ₹172 (SD=236.7), IPD: ₹275 (SD=254.8) and pharmacy: ₹307 (SD=344.8)) (table 1, figure 2). In contrast, OOPE was substantially higher among those who relied solely on private pharmacies (OPD: ₹1085 (SD=2033.4), IPD: ₹3165 (SD=4440.4) and pharmacy: ₹1031 (SD=2298.2)). Table 1Mean out-of-pocket expenditure on medicines for respondents recruited at outpatient clinics, inpatient wards and pharmaciesMean OOPE (SD)OPDIPDPharmacyMean OOPE on medicines534 (1502.6)2231 (3829.6)907 (1982.3)Type of pharmacy visitedJAK172 (236.7)275 (254.8)307 (344.8)Non-JAK1085 (2033.4)3165 (4440.4)1031 (2298.2)Both1005 (956)3038 (3557)987 (1034.7)Type of facilityTertiary hospital736 (1753.2)2599 (4272.2)Secondary hospital62 (186.1)1169 (1684.7)Type of settingOPD842 (1803.6)IPD2369 (4123.1)Type of diseaseHypertension/diabetes547 (1002.8)1654 (5020.8)873 (1517)Cancer1013 (4302)2489 (3844.2)1889 (4661.1)Others439 (1837.6)2044 (3278.5)941 (2492.3)ENT684 (1424)Multiple334 (833.2)2208 (3565.1)1072 (1985.2)Age (years)≤19435 (601.7)1893 (2020.9)784 (1900.7)20–29495 (1228.1)1632 (1552.6)917 (1974)30–39515 (913.9)2637 (5098.7)881 (2003.7)40–49655 (2171.1)1930 (2787)953 (2323.2)50–59613 (1777.7)2150 (4135)872 (1552.8)60–69418 (839)2442 (4182.6)1020 (2232.9)≥70428 (878.1)2336 (3379.5)830 (1524)GenderMale611 (1527.3)2304 (3465.8)924 (2183.6)Female441 (1467.1)2133 (4270)887 (1715.3)Education statusIlliterate516 (907.2)2557 (4425.6)977 (2213.4)literate541 (1669)2048 (3441.3)888 (1913.7)Occupation statusEmployed624 (1772.5)2317 (3620.3)899 (2180.7)Unemployed470 (1272.9)2175 (3961.9)912 (1851.8)ReligionHindu579 (1605.5)2180 (3687.7)909 (1977.9)Muslim409 (970.6)2384 (4476)912 (2148.3)Sikh260 (490.9)7800 (10612.7)744 (606.4)Christian105 (377)2306 (3424.3)845 (1374.5)CasteSC/ST487 (984)2226 (4195.4)842 (2074)Other backward classes735 (2165.7)2493 (3929)863 (1931.6)General410 (956.7)2037 (3565.4)953 (1990.9)Marital statusNever married445 (715.1)2306 (3424.3)843 (1992.7)Married516 (1348.1)2226 (4195.4)920 (1990.1)Separated/widow1033 (3681)2493 (3929)939 (1638.7)Insurance statusNone611 (1700.3)2701 (4219)906 (1977.1)Government funded252 (702.3)1421 (3247.1)947 (1538.8)Private597 (1107.1)1596 (2321.6)772 (3166.2)Type of familyNuclear437 (1521.6)1688 (3239.8)904 (1907.2)Joint607 (1484.3)2524 (4085.8)909 (2036.8)Wealth quintilePoorest353 (2073.6)2563 (4507.1)781 (1434.8)Poor419 (935.7)2653 (3282.5)825 (1800.1)Middle693 (2191)2488 (4428.8)865 (1912.3)Rich548 (898.6)1624 (2420)907 (2170.9)Richest579 (1089.8)1825 (4001.1)1157 (2431.1)Area of residenceRural521 (1742.2)2392 (3842)907 (1911.5)Urban560 (1078.5)1795 (3803.7)910 (2066.1)Slum182 (240.3)755 (752.2)607 (327.5)Number of drugs≤4376 (739.7)701 (790.4)727 (1652.5)5–8705 (1715.4)1570 (1887.5)1058 (2045.8)>8715 (2948.8)3418 (5309.7)2233 (3891.7)Diseases severityMild277 (702.5)1355 (2489.8)875 (1964.1)Moderate465 (1759.1)2552 (3943.5)853 (1867.8)Severe891 (1821.2)2795 (4620.3)996 (2110)Findings at OPD, IPD and pharmacy settings are not directly comparable to each other due to differences in sociodemographic and clinical factors.ENT, ear, nose, throat; IPD, inpatient department; JAK, Jan Aushadhi Kendra; OOPE, out-of-pocket expenditure; OPD, outpatient department; SC/ST, scheduled caste/scheduled tribe. Mean out-of-pocket expenditure on medicines for respondents recruited at outpatient clinics, inpatient wards and pharmacies Findings at OPD, IPD and pharmacy settings are not directly comparable to each other due to differences in sociodemographic and clinical factors. ENT, ear, nose, throat; IPD, inpatient department; JAK, Jan Aushadhi Kendra; OOPE, out-of-pocket expenditure; OPD, outpatient department; SC/ST, scheduled caste/scheduled tribe. Out-of-pocket expenditure on medicines by pharmacy/facility and socio-economic status. IPD, inpatient department; JAK, Jan Aushadhi Kendra; OPD, outpatient department. The type of pharmacy accessed emerged as a significant factor in multi-variable regression analysis, with OOPE being 11.2%–89.3% lower among exclusive JAK users, as compared with the patients using mixed or private pharmacy sources (online supplemental table S2-S4). Furthermore, after appropriate matching (as indicated by SMDs <10% in matched and unmatched sample; as well as Rosenbaum bounds sensitivity analysis with estimated treatment effect remaining statistically significant up to Γ=1.8), patients recruited at JAK pharmacies incurred 42% lower expenses on medicines than those recruited at private pharmacies (online supplemental table S5-S6). Facility type was also associated with differential spending: patients treated at tertiary-level facilities incurred significantly higher OOPE than those at secondary hospitals: 190.1% higher in OPD and 21.4% higher in IPD (online supplemental table S2, S3). The number of drugs purchased was positively associated with OOPE in all settings. In addition, patients with moderate or severe illness incurred significantly higher OOPE compared with those with mild conditions, ranging from 6.5% to 41.1% depending on setting and severity level. Additionally, inpatients incurred 57% higher OOPE than outpatients within the pharmacy cohort. Socio-economic status was also associated with OOPE on medicines. Patients belonging to the poor, rich and richest wealth quintiles incurred significantly higher OOPE compared with those in the poorest quintile, particularly in pharmacy settings, with excess OOPE of 10.2%, 9.1% and 33.4%, respectively. However, in this sample and context, insurance status was not associated with lower OOPE on medicines. Catastrophic health expenditure and impoverishment rate The overall prevalence of CHE was 9.2% among outpatients, 35.4% among inpatients and 8.3% among patients recruited at pharmacies. Online supplemental table S7 presents the prevalence of CHE at different threshold cut-offs (10%, 25% and 40%) of total and non-food expenditure. IRs were 5.7%, 21.9% and 6.3%, respectively. CHE was lowest among those purchasing medicines exclusively from JAKs: 0% (OPD), 5.9% (IPD), 2% (pharmacy). In contrast, CHE prevalence was significantly higher among patients using private pharmacies: 17.1% (OPD), 50% (IPD) and 16.4% (pharmacy). Compared with JAK users, patients purchasing medicines from private pharmacies had significantly higher odds of experiencing CHE across all settings (OPD: OR=3.11, IPD: OR=48.40, pharmacy: OR=9.46). The odds remained elevated among those accessing both JAK and private pharmacies (IPD: OR=21.0, pharmacy: OR=7.45). Similar trends were observed for impoverishment, particularly among patients recruited at pharmacies. Patients attending tertiary care hospitals were at markedly greater risk of CHE (OPD: OR=21.21, IPD: OR=3.83) and impoverishment (OPD: OR=15.74, IPD: OR=3.02) compared with those at secondary-level facilities (table 2, online supplemental S8-S9). Table 2Factors associated with CHE and impoverishment in outpatient and inpatient departments of secondary and tertiary hospitals, and pharmaciesVariablesCategoriesORCHEIROPDIPDPharmacyOPDIPDPharmacyType of pharmacy visitedJAKNon-JAK 3.113† 48.398† 9.458† 1.0951.474 4.615† Both1 21.072† 7.451† 11 2.653† Type of facilitySecondary hospitalTertiary hospital 21.208† 3.734† 15.738† 3.074† Type of settingIPDOPD 0.591† 0.619Age (years)≤1920–29 0.983* 0.997 0.989† 0.967* 1.007 0.99* 30–3940–4950–5960–69>70GenderMaleFemale1.0270.6790.9770.6211.2940.588Area of residenceRuralUrban0.9000.65 0.546† 0.392† 0.508 0.451* Slum11.6600.4240.247.9461ReligionHinduMuslim 0.290† 1.346 0.57† 0.177† 0.8310.755Sikh0.4481.021111.0001Christian0.5641.0480.7740.5770.8820.479Education statusIlliterateliterate 0.636* 0.9870.8000.692 0.408† 0.674*Occupation statusEmployedUnemployed0.6741.1291.0781.3570.8921.621*CasteSC/STOBC0.9540.8931.3361.7140.6140.766General0.6131.237 1.836† 1.0700.9140.976Marital statusNever marriedMarried0.9830.7811.1633.6101.3031.709*Separated/widow 4.581† 1.1131.527 11.054* 0.5531.607Type of familyNuclearJoint 1.497* 1.1050.869 16.068† 5.747† 3.611† Insurance statusNoneGovernment funded0.6850.8291.250.7780.6850.882Private0.7421.0811.4021.0241.0500.900Wealth quintilePoorestPoor0.392 0.402† 0.584† 0.499 0.064† 0.172† Middle 0.308† 0.130† 0.35† 0.028† 0.007† 0.034† Rich 0.080† 0.045† 0.059† 0.003† 0.001† 0.004† Richest 0.009† 0.026† 0.033† 0.000† 0.001† 1Number of drugs prescribed≤45–8 1.133† 1.104† 1.246† 1.255† 1.156† 1.238† >8Diseases severityMildModerate 1.977* 1.2771.0801.3562.0410.775Severe 2.603† 1.7471.167 3.049† 3.019* 1.120Findings at OPD, IPD and pharmacy settings are not directly comparable to each other due to differences in sociodemographic and clinical factors.Bold values denote significant at p value <0.05.*Significant at p value <0.05.†Significant at p value <0.001. online supplemental Table S10 presents absolute risk differences (%) for catastrophic health expenditure and impoverishment rate across categories of variablesCHE, catastrophic health expenditure; IPD, inpatient department; IR, impoverishment rate; JAK, Jan Aushadhi Kendra; OBC, other backward classes; OPD, outpatient department; SC/ST, scheduled caste/scheduled tribe. Factors associated with CHE and impoverishment in outpatient and inpatient departments of secondary and tertiary hospitals, and pharmacies Findings at OPD, IPD and pharmacy settings are not directly comparable to each other due to differences in sociodemographic and clinical factors. Bold values denote significant at p value <0.05. Significant at p value <0.05. Significant at p value <0.001. online supplemental Table S10 presents absolute risk differences (%) for catastrophic health expenditure and impoverishment rate across categories of variables CHE, catastrophic health expenditure; IPD, inpatient department; IR, impoverishment rate; JAK, Jan Aushadhi Kendra; OBC, other backward classes; OPD, outpatient department; SC/ST, scheduled caste/scheduled tribe. Wealth status demonstrated a strong protective effect. The odds of CHE and IR decreased progressively with increasing income across all settings, with the richest quintile having up to 99.9% lower risk of financial hardship than the poorest. The number of drugs prescribed was positively associated with both CHE and IR across all groups. Patients prescribed five to eight medicines had significantly higher odds of CHE (OPD: OR=1.13, IPD: OR=1.10, pharmacy: OR=1.25) and IR (OPD: OR=1.26, IPD: OR=1.16, pharmacy: OR=1). Disease severity was also predictive: outpatients with severe illness were over 2.6 times more likely to incur CHE and three times more likely to be impoverished (OR=3.05, p<0.01) compared with those with mild conditions. Among inpatients, severe illness significantly increased the odds of both CHE (OR=1.79) and IR (OR=2.91), compared with mild illness (table 2, online supplemental S8). Additional factors such as urban residence, higher education and joint family structure also influenced the extent of financial risk protection. Patient-reported barriers for access to JAKs While over half of the prescribed medicines were obtained from government-funded or subsidised sources, a considerable share continued to be procured from private pharmacies. Among participants who neither received medicines from public health facilities nor accessed JAKs, the most commonly reported reasons were lack of awareness and perceived unavailability of required medicines at these outlets (figure 3). Additionally, around 13% of these outpatient and inpatient participants also reported perceived quality of medicines and distance to JAKs as barriers to access. Reasons for non-utilisation of JAKs. JAK, Jan Aushadhi Kendra. Discussion The study provides empirical evidence of association between use of JAKs and the financial risk protection, particularly in mitigating OOPE and CHE on medicines. Across outpatient, inpatient and pharmacy settings, patients who procured medicines exclusively through JAKs consistently incurred significantly lower OOPE. Notably, the individuals relying exclusively on JAKs incurred up to 89.3% lower expenditure compared with those using non-JAK pharmacies for access to medicines. Furthermore, after appropriate matching, patients recruited at JAKs incurred 42% lower expenses on medicines than those recruited at private pharmacies. These findings are consistent with the existing global evidence demonstrating that access to low-cost generics can substantially alleviate household financial burden, especially in low- and middle-income countries.27 28 The study further highlighted that reliance on private pharmacies was associated with increased financial risk. Patients purchasing medicines from private sources had up to 48 times higher odds of experiencing CHE compared with those using JAKs, with a similar pattern observed for impoverishment. Notably, none of the outpatients who obtained medicines exclusively from JAKs experienced CHE, reinforcing the scheme’s potential as a cost-containment intervention within India’s public health framework. Despite this positive association with reduction in OOPE and CHE, the findings indicate underutilisation of JAKs. Approximately half of outpatients and inpatients who purchased medicines from private pharmacies cited lack of awareness about PMBJP or difficulty locating JAKs as the primary barriers. These challenges highlight opportunities for strengthening PMBJP by enhancing its visibility, expanding outreach efforts and improving its integration into routine public healthcare services, as suggested by earlier studies.17 18 Strengthening the linkages between public health facilities and JAKs may improve awareness and utilisation of the scheme. The study observed that a considerable proportion of JAK users also accessed non-JAK pharmacies. This suggests potential issues with medicine availability, which may stem from structural gaps in supply or physician prescribing behaviours favouring non-generic alternatives. Generic prescribing rates were found to be low (18% in outpatient and 6% in inpatient settings), which highlights a critical barrier to the utilisation of PMBJP. Furthermore, the findings aligned with previous research indicating inconsistent availability of medicines at JAKs,17 potentially influenced by patient demand shaped by prevailing prescribing practices. Nevertheless, the scheme is associated with reduced medicine-related OOPE, with patients recruited from JAKs incurring 42% lower expenditure compared with their matched counterparts recruited from private pharmacies. To realise the scheme’s full potential, the Drugs Technical Advisory Board has recommended the formation of a subcommittee to explore the feasibility of permitting PMBJP pharmacists to substitute prescription medicines listed under Schedules H, H1 and X with generic equivalents.29 This policy measure is aimed at improving access to affordable medicines and enhancing the overall effectiveness of the scheme. While the government has also mandated the prescription of generic medicines by physicians, implementation has faced resistance due to concerns regarding the perceived quality of generics.30 To build trust among both prescribers and the public, there is a need for greater transparency in the regulatory processes for generic medicines and to build further evidence on their therapeutic equivalence with branded alternatives.31 32 Furthermore, the coexistence of parallel initiatives such as AMRIT,33 which provide both generic as well as non-generic medicines at subsidised rates may inadvertently create competition with PMBJP. While both the schemes share the overarching objective of reducing OOPE on medicines, their coexistence without integration may lead to fragmentation of access. Policy coherence across these schemes is critical, and integrating PMBJP with AMRIT through unified logistics, pricing and distribution strategies may improve efficiency and reach. Another key finding was the variation in OOPE by level of care. Patients attending tertiary care facilities incurred nearly twice the OOPE compared with those treated at district hospitals. This could be due to more intensive diagnostics and prescriptions at higher levels of care. However, it may also reflect differential operational models with state-managed secondary facilities often benefiting from centralised free drug supplies, whereas tertiary institutions follow distinct governance systems for drug procurement and supply chain management. To address these disparities, efforts could focus on integrating workflows across departments responsible for supply chain management in both tertiary and secondary facilities at state level. Additionally, establishing formal linkages between tertiary care institutions and the PMBJP, including the establishment of on-site JAKs, could serve as an effective strategy to reduce this gap. Harmonising electronic health records and logistics data with JAK inventory systems could ensure real-time tracking, reduce stockouts and enhance last-mile availability. Digital supply chain innovations such as those used in Tamil Nadu Medical Service Corporation could be adapted nationally to improve responsiveness and transparency.34 Importantly, health insurance coverage, whether public or private did not significantly reduce medicine-related OOPE or financial hardship in OPD and pharmacy settings. This mirrors previous findings that most Indian insurance schemes offer limited or no coverage for outpatient consultations and medicine costs.35 36 The exclusion of outpatient medicines in insurance programmes, as PMJAY limits their financial protection, especially for patients with chronic diseases, who incur recurring medicine expenses. In this context, PMBJP serves as a crucial supply-side intervention that complements insurance-based schemes and can enhance progress toward UHC. Additionally, the impact of insurance on OOPE could not be differentiated in the inpatient setting, which may be attributable to the study sample being confined to public sector facilities. In these settings, the scope and quality of services are largely standardised, resulting in comparable access to medicines for both insured and uninsured patients, all of whom benefit from universally applicable supply-side interventions, such as free medicine and diagnostic schemes. Furthermore, patients with moderate to severe conditions and those prescribed multiple drugs experienced significantly higher OOPE. These findings not only underscore the cumulative financial burden of illnesses but also necessitate targeted investments for early diagnosis and management of diseases, as well as rational prescribing practices. Reinforcing provider training on essential drug lists and generic alternatives may also reduce unnecessary polypharmacy. India’s experience with PMBJP offers important lessons for the global health community. As many low- and middle-income countries confront rising OOPE on medicines, centrally negotiated procurement and generic drug distribution can serve as an effective policy lever for financial risk protection. However, the success of such an initiative depends on their integration into broader health systems, alignment with physician prescribing patterns and consistent public engagement. The study should be interpreted in light of certain limitations. The observational design restricts causal inference, and potential selection bias may remain, as patients choosing JAKs could differ systematically from those using private pharmacies. To mitigate this concern, we adjusted for key factors, including disease type, disease severity and the number of prescribed medicines, all of which are known to influence OOPE. Furthermore, PSM was employed to minimise residual confounding and selection bias, thereby strengthening the internal validity of the findings. The use of nearest-neighbour matching achieved satisfactory covariate balance, with all SMDs below 10%, indicating adequate comparability between treated and control groups in the matched sample. Despite robustness to moderate unobserved confounding as indicated by Rosenbaum bounds analysis, the findings remain susceptible to bias from unmeasured factors inherent to observational study designs. Furthermore, while the while the study provides nationally relevant insights into financial protection among JAK and non-JAK users, the utilisation of the scheme may under-represent pathways dominated by private care. It is important to note that many existing cross-sectional surveys evaluating the impact of government policies on household expenditures and financial risk protection have not adequately considered quality of services (prescriptions) and disease severity as independent correlates of OOPE/ CHE.37 38 From this methodological perspective, our study provides an important contribution by demonstrating the feasibility of using EQ-5D-5L–derived scores as a proxy measure of disease severity. The availability of Indian value set for EQ-5D-5L along with its demonstrated ability to discriminate between groups with varying levels of clinical severity across heterogeneous conditions strengthens its suitability for incorporating disease severity into analyses of OOPE in future surveys and is recommended for adoption in subsequent research.24 Furthermore, the present study was restricted to a predefined list of disease conditions for which medicines are available at JAKs. While this focus was consistent with the study’s objectives the findings are generalisable to conditions for which standardised medicine baskets are available under the PMBJP but may not extend to diseases requiring specialised or patented therapies or alternate care pathways. Lastly, we estimated CHE using medicine-specific out-of-pocket spending and applying 40% household non-food consumption expenditure threshold, as no standard thresholds exist for medicine-only expenditure. However, this focused approach strengthens the analysis by isolating the financial burden of medicines, and sensitivity analyses with alternative thresholds further confirmed the robustness of the findings. Future research should assess the adequacy and comprehensiveness of the basket of medicines mandated under the PMBJP, and explore provider as well as consumer-level factors, including prescribing practices, perceptions of generic medicines and pharmacy-level dispensing behaviours, which may shape utilisation of JAKs and, in turn, influence financial risk protection outcomes. In addition, comparative analyses of PMBJP alongside alternative pharmaceutical system-strengthening approaches, such as incentivising private pharmacies to dispense low-cost generic medicines, improving medicine availability within government health facilities and expanding insurance coverage to include outpatient medicines and long-term pharmacotherapy, could further inform evidence-based policy reforms. Conclusion The findings highlight the important role of the PMBJP as a policy initiative associated with greater access to affordable medicines in India, as reflected by lower OOPE and reduced prevalence of CHE. To strengthen the potential benefits of the scheme, its reach could be expanded through wider geographical coverage and deeper integration within the health system, alongside stronger enforcement of generic prescribing practices supported by prescriber training and public awareness efforts. Additionally, integrating PMBJP inventory systems with hospital logistics platforms and harmonising parallel initiatives under a unified framework could enhance operational efficiency. Aligning facility-level procurement systems and public insurance schemes with PMBJP objectives would further amplify financial protection and ensure sustainability. As countries strive towards UHC, India’s experience with PMBJP offers valuable insights into advancing access to essential medicines through sustained public investment in low-cost generic medicines. Supplementary material Acknowledgements We are grateful to the field investigators as well as participants of the surveys for their contributions to the data collection process. We are also thankful to Ms Priya Abraham for customising the digital data collection instrument, and the data management supervisors in the respective states. Footnotes Funding: The work is supported by a grant from the Department of Health Research, Ministry of Health and Family Welfare, Government of India via grant ID: F.NO.T.11011/02/2017-HR.The funder had no role in the conduct of the study; data collection, management, analysis and interpretation of the data; preparation, review or approval of the manuscript; and decision to submit the manuscript for publication. Prepub: Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2025-113839). Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: The research study was performed in accordance with the Declaration of Helsinki and was approved by the Institutional Ethics Committee, Post Graduate Institute of Medical Education and Research, Chandigarh, India, via reference ID: PGI/IEC/2021/001453. Written informed consent was taken from all the participants of the study prior to the interview. Data availability free text: The datasets generated and/or analysed during the current study are not publicly available due to permissions required from the Government of India but are available from the corresponding author on reasonable request and following approval from the Government of India. Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. Data availability statement Data are available upon reasonable request. References Supplementary Materials Data Availability Statement Data are available upon reasonable request.