The intervention consisted of a comprehensive package that included 68 structured SMS messages, six video messages each lasting approximately 2–4 min, and patient-friendly workbooks (Pandian et al., 2023). These materials were delivered over a 1-year period to participants in the intervention arm and were made available in 12 widely spoken Indian languages to ensure accessibility and cultural relevance. The SMS and videos were carefully designed to highlight potential problems related to stroke risk factors while simultaneously conveying messages of encouragement and hope, with the overarching aim of promoting sustainable behavioral changes and preventing stroke recurrence.

At the baseline visit, each participant received a workbook that featured a variety of interactive stroke-prevention activities, including board games, matching exercises, true-or-false questions with answers, and simple physical activity prompts (Pandian et al., 2023). To facilitate monitoring and engagement, every SMS and video message included a dedicated phone number that patients or their caregivers were asked to call to confirm receipt; each call served as a documented acknowledgment of the delivered message (Pandian et al., 2023).

As part of the structured follow-up, patients in the intervention group underwent vascular assessments at both 6 months and 1 year. Participants in the control group also completed these scheduled follow-up visits in addition to receiving standard routine care. Further detailed information regarding the design, development, and creation of the intervention components is available in the research article describing the intervention development process (Kate et al., 2020a). Confirmation of receipt of the SMS messages and videos among patients in the intervention group was low, with only 17% acknowledging receipt.

The baseline assessment consisted of a comprehensive set of variables that included demographic details, clinical stroke subtype, an extensive profile of vascular and lifestyle-related risk factors, BP measurements, FBS, lipid parameters, total physical activity expressed in metabolic equivalents, as well as information on medication non-compliance. Definitions for behavioral factors were standardized: a “current smoker” was identified as any individual who had used smoked tobacco within the 12 months preceding enrolment in the trial, while “current alcohol intake” referred to alcohol consumption during the same period (Pandian et al., 2023).

Medication non-compliance was evaluated using a structured self-report questionnaire that explored the underlying reasons for missed medication doses, thereby enabling a nuanced understanding of adherence barriers. Educational attainment, captured through the trial's Case Report Form (CRF), was categorized in detail as follows: less than primary school education (< 5th standard); completion of primary school (up to 5th standard); completion of middle school (up to 8th standard); completion of secondary school (up to 10th standard); completion of high school (up to 12th standard); attainment of a diploma, college, or university degree; possession of a postgraduate qualification; and an “unknown” category for cases where information could not be ascertained (Pandian et al., 2023).

The primary outcome comprised a composite measure that included recurrent stroke, high-risk TIA, ACS, and all-cause mortality. A range of secondary outcomes was also evaluated, capturing broader aspects of recovery and health behavior. These included changes in BMI, total physical activity expressed as metabolic equivalent minutes per week, current smoking status, current alcohol consumption, functional status measured by the mRS, and levels of medication non-compliance at the 1-year mark. Outcome assessments for both the intervention and control groups were conducted 1 year after randomization to ensure uniform follow-up. Functional outcomes (mRS) were assessed at 1 year using standardized definitions as per the trial protocol. Outcome assessors were the treating physicians (center principal investigators), who were masked to group allocation across all participating centers.

An interim futility analysis, undertaken by the Data and Safety Monitoring Board (DSMB), ultimately led to the early termination of the trial (Pandian et al., 2023). The trial was halted primarily due to funding limitations and lack of demonstrable effectiveness (Pandian et al., 2023). The DSMB's recommendation was based on a lower-than-expected conditional power and a considerably smaller number of primary vascular outcome events than originally anticipated, which together reduced the likelihood that extending the trial would yield meaningful or conclusive results.

A chi-square test was used to find out the difference between patients in the private and government hospital based on group allocation, baseline characteristics, mRS, and baseline behavioral history such as alcohol intake, smoking, chewed tobacco, drug addiction, and medication compliance. An independent sample t-test was used to find out the difference between private and government hospitals based on SBP, DBP, FBS, BMI, LDL, high density lipids HDL, TCs, TGs, and physical activity score.

The comparison between government and private hospitals was conducted as a post-hoc exploratory analysis. The original SPRINT INDIA trial was powered to assess differences between intervention and control groups, not to detect differences between hospital sectors; therefore, these analyses were not based on a pre-specified sample size calculation for sector comparison (Pandian et al., 2021).

The primary outcomes at 1 year in the SPRINT-INDIA trial, which includes a composite end point of recurrent stroke, high-risk TIA, ACS, and all-cause mortality, were assessed using chi-square tests. The secondary outcomes assessed at 1 year after randomization were changes in BMI, physical activity total metabolic equivalent (min per week), current smoking, current alcohol intake, mRS, and medication non-compliance. The laboratory-measured secondary outcomes were SBP and DBP (mm Hg), FBS (mg/dL), LDL (mg/dL), and triglycerides (mg/dL) at 1 year after randomization.

At 1 year follow-up, the comparison between the two types of hospitals for the primary and secondary outcomes in the SPRINT-INDIA trial was analyzed using the chi-square test for the categorical variables and the t-test for the continuous variables, both overall at 1 year and within the intervention and control groups. The comparison between the two types of hospitals in terms of the patients' work status prior and after stroke analyzed using the chi-square test. Generalized Estimation Equation (GEE) analysis included an interaction term (type of institutions^*group) to assess outcomes between groups, adjusting for age, education, stroke characteristics, medical history, and NIH Stroke Scale, while accounting for within-center correlation using an unstructured working correlation structure. For the categorical variables adjusted odds ratio and 95% confidence interval were presented. For continuous variables, β coefficient with 95% confidence interval was presented. All the analyses were conducted using a complete-case approach, including only participants with available data at the respective follow-up time points. No imputation method were applied for missing data. All p values < 0.05 was considered significant. The statistical analysis was conducted using SPSS version 29.0.

Table 1 shows the data for baseline characteristics at the time of start of study. More than half of the patients in both government and private hospitals had less than high school education [717 (58.7%) and 926 (50%) respectively; p < 0.001]. Ischemic strokes were more prevalent in private hospitals than government hospitals [1,579 (86.9%) vs. 974 (79.7%); p < 0.001].

According to TOAST classification large artery atherosclerosis prevalence is more in government hospitals 340 (35.9%) than private institutions 462 (30.4%) but according to OCSP classification partial anterior circulatory syndrome is reported to be more in private than government institutions [863 (54.7%) vs. 413 (42.4%); p < 0.001 respectively]. Majority of the patients admitted in private hospitals had HTN [1,279 (70.4%)], diabetes mellitus [789 (43.4%)], dyslipidemia [359 (19.8%)], coronary artery disease [199 (11.0%)], and obesity [168 (9.3%)] as compared to government hospitals. On the other hand, most of the patients in government hospitals as compared to private hospitals had TIA (3.5 vs. 2%; p = 0.032) and non-valvular atrial fibrillation (2.9 vs. 1.6%; p = 0.005).

The proportion of patients in private hospitals as compared to government hospitals were taking more medications like diabetic medications [704 (38.8%) vs. 392 (32.7%); p = 0.001] and anticoagulants [221 (12.2%) vs. 108 (8.9%); p = 0.005]. On the other hand, majority of the patients in government hospitals as compared to private hospitals had taken antihypertensive medications [845 (69.2%) vs. 1,184 (65.2%); p = 0.040].

Even though there exists no significant difference in terms of gender (p = 0.279) and age (p = 0.181) major portions of patients in private and government hospitals were males [1,305 (71.9%) vs. 900 (73.6%)] with more than 50 years old [1,202 (66.2%) vs. 780 (63.8%)].

Table 2 shows the baseline and behavioral history in patients among the private and government hospitals. The majority of the private hospital patients admitted had a good mRS [1,106 (60.9%)]. Most of the government hospital patients were using alcohol [265 (21.7%) vs. 333 (18.3%); p = 0.047] and tobacco smoking [269 (22%) vs. 249 (13.7%); p < 0.001] compared to private school patients. BMI, LDL, total cholesterol, TGs, and fasting blood sugar were high in private hospital patients (25.574 ± 0.14, 105.513 ± 9.22, 173.844 ± 4.66, 154.529 ± 2.81, and 138.1406 ± 0.53, respectively). On the other hand, SBP, DBP, and HDL were high in government hospital patients (134.961 ± 8.58, 83.411 ± 0.41, and 50.352 ± 4.51, respectively).

At 1 year the study focused on the 3,038 patients that were followed up. No significant differences observed in the primary outcomes between the private and government hospitals as seen by the composite primary outcome [private: 50 (2.8%) vs. government: 45 (3.7%); p value 0.215; Table 3].

Table 4 presents the comparison of secondary outcomes based on type of hospital. Majority of the private hospital patients had a good mRS score 1,720 (94.7%). The higher proportion of patients in private hospitals consumed alcohol [private: 103 (5.7%) vs. government: 35 (2.9%); p < 0.001], smoked [private: 89 (4.9%) vs. government: 35 (2.9%); p < 0.001] than in government hospitals which is in contrast to the baseline characters where higher number of patients in government hospitals consumed alcohol and smoked tobacco. Greater number of patients in private hospitals chewed tobacco than government hospitals [private: 68 (3.7%) vs. government hospital: 16 (1.3%); p < 0.001]. Fasting blood glucose, LDL levels, TGs levels, and BMI were found to be higher in private hospital patients (119.324 ± 0.35, 87.113 ± 3.12, 141.67 ± 1.5, 25.604 ± 0.50 respectively) whereas HDL levels were higher in government hospital patients (51.81 + 22.68) which is similar to the baseline characters in the comparison groups (Table 4).

At 1 year, there were no significant difference in primary outcome between intervention and control group in patients who were admitted to private as well as government hospitals. There was no patient who had high risk of TIA in intervention group as compared to control group in private hospitals [intervention: 0 (0%); control: 4 (0.4%); p = 0.051; Table 5].

Table 6 presents comparison of intervention and control group when stratified by private and government hospitals. The greater number of patients in control group has poor functional outcome in control group as compared to intervention group in government hospitals [intervention: 57 (9.2%); control: 79 (13.1%); p = 0.032]. A smaller proportion of patients in private hospitals reported alcohol use in the intervention group compared with the control group (3.7 vs. 7.5%; p = 0.002). Similarly, current smoking was less frequent among intervention recipients (3.6 vs. 6.1%; p = 0.027), and medication non-compliance was significantly lower in the intervention group (6.1 vs. 11.5%; p < 0.001). There were no significant differences between the intervention and control groups in blood pressure, fasting blood glucose levels, lipid profiles, BMI and physical activity scores among patients admitted to both private and government hospitals.

There is a significant difference between private and government hospital patients based on their work status prior to stroke (p < 0.0001) and post stroke (p < 0.0001) (Table 7). Prior to stroke, the majority of patients in both private and government hospitals were full-time paid workers [private hospitals: 871 (45.9%) vs. government hospitals: 613 (50.2%)], while the unemployed were fewer in both hospitals [private hospitals: 72 (3.9%) vs. government hospitals: 54 (4.4%)]. After the stroke, the number of unemployed people increased in both hospitals [private hospitals: 406 (22.4%) vs. government hospitals: 526 (43.0%)], while the number of full-time paid workers decreased [private hospitals: 388 (21.4%) vs. government hospitals: 84 (6.9%).

At 1-year, private institutions showed non-significantly lower odds of primary composite outcome [adjusted OR 0.80 (95% CI: 0.42–1.50); p value: 0.485] and higher odds of good functional outcomes [adjusted OR 1.91 (95% CI: 0.90–4.05); p value: 0.091] vs. government institutions. Private institution patients were more likely to cease alcohol [adjusted OR 3.12; (95% CI: 1.52–6.39); p value: 0.0002] and smoking [adjusted OR 1.99 (95% CI: 1.29–3.07); p value: 0.002]. There were 26% less current smoker in private institutions than government institutions [adjusted OR 0.74 (95% CI: 0.71–0.78); p value: < 0.0001].

In government institutions, intervention improved good functional outcomes vs. control [adjusted OR 1.56 (95% CI: 1.34–1.81); p value: < 0.0001]. In private institutions, intervention halved medication non-compliance [adjusted OR 0.50 (95% CI: 0.37–0.66); p value: < 0.0001]. There was significantly deduction in alcohol and current smoker in intervention group than control group in private institutions. The SBP and fasting blood sugar levels was decreased in intervention group in government institution patients (p = 0.003, p < 0.0001 respectively). Whereas the BMI value was decreased in private institutions in intervention group (p = 0.0002; Tables 8, 9).

Patients managed in private hospitals exhibited a higher burden of major vascular risk factors, including HTN, DM, CAD, obesity, elevated BMI, and dyslipidaemia—features consistent with cardio metabolic syndrome. Correspondingly, these patients were more frequently prescribed antidiabetic medications and anticoagulants, reflecting both a greater documented comorbidity burden and more intensive risk stratification. In parallel, private hospital patients demonstrated higher educational attainment, urban residence, and independent home ownership, collectively indicating relatively higher socioeconomic status (SES). These findings are consistent with earlier regional analyses from the SPRINT India trial and with prior work linking education and socioeconomic factors to stroke risk profiles and outcomes (Lindmark et al., 2022; Verma et al., 2025b).

Several observed differences are influenced by the geographic distribution of participating centers. Most private hospitals (9 of 14) were located in South India, where catchment populations are predominantly urban, whereas government hospitals largely served rural communities. This distribution contributed to clear contrasts in SES, education, lifestyle behaviors, and comorbidity patterns across healthcare sectors. These institutional patterns mirror broader national trends in healthcare access and utilization. For instance, per-day hospitalization expenditure is substantially higher in private facilities (INR 6,788) than in government hospitals (INR 2,833), partially explaining preferential private-sector utilization by higher socioeconomic groups (Garg et al., 2022). Contemporary Household Consumption Expenditure Survey (HCES) data further demonstrate persistent socioeconomic gradients, with rural poverty exceeding urban poverty in 2022–2023 (7.2 vs. 4.6%; Research, 2024).

Population-level data reinforce these observations. National Family Health Survey-4 (NFHS-4) data indicate that individuals from higher socioeconomic groups account for a substantial proportion of the national burden of elevated BMI, obesity, DM, and HTN [International Institute for Population Sciences (IIPS), 2017; Corsi and Subramanian, 2019]. In parallel, a mini-review examining social determinants of obesity identifies SES—encompassing income, education, and access to resources—as a key driver, with higher burden observed among higher-SES groups in developing settings (Dinsa et al., 2012). Together, these findings suggest that the greater burden of modifiable vascular risk factors among private hospital patients reflects both socioeconomic patterning of cardio metabolic risk and potentially more systematic screening and diagnostic practices within private healthcare settings.

Stroke subtype distribution also differed between healthcare sectors. Based on OCSP classification, patients treated in private hospitals more commonly presented with partial anterior circulation syndromes and co-existing CAD, whereas those managed in government hospitals more frequently exhibited large artery atherosclerosis, TIA, and non-valvular atrial fibrillation. A higher proportion of ischemic stroke was observed among private hospital patients. Similar patterns have been reported from government medical colleges in southern India, where ischemic strokes occur at more than twice the rate of haemorrhagic strokes and anterior circulation strokes predominate—findings consistent with the South India–weighted private hospital cohort in the present study (Sridharan et al., 2009). These distributions align with the higher cardiometabolic burden observed among private hospital patients and are supported by a Gujarat-based retrospective study identifying HTN, dyslipidaemia, DM, prior CAD, and smoking as major contributors to stroke risk (Vaidya and Majmudar, 2014). National estimates of stroke burden similarly demonstrate strong associations between ischemic stroke and HTN, hyperglycaemia, and dyslipidaemia (Behera et al., 2024). Data from the Global Burden of Disease Study further indicate that individuals from higher SES, despite greater exposure to dietary and behavioral risks, experience comparatively better health outcomes and lower stroke incidence (GBD 2019 Stroke Collaborators, 2021).

Medication utilization patterns reflected these clinical and systemic differences. Patients treated in private hospitals received a higher number of medications, including antiplatelet agents, antidiabetic drugs, and anticoagulants, consistent with greater comorbidity burden and more intensive secondary prevention strategies. In contrast, patients managed in government hospitals were more frequently prescribed antihypertensive medications, aligning with higher BP levels and lower HDL cholesterol observed in this group. Comparable trends have been reported in a comparative study from Dhaka, which observed higher rates of polypharmacy in private hospitals than in government hospitals, alongside greater adherence to essential drug lists in private hospitals (Samad et al., 2024). Evidence from other LMIC settings similarly highlights constraints in drug availability and service provision in government healthcare systems, while also noting occasional overuse of medications in private care (Basu et al., 2012).

Behavioral risk profiles also differed markedly between sectors. Alcohol use, smoking, and tobacco chewing were more commonly reported among patients treated in government hospitals. National analyses identify smoking and tobacco use as leading behavioral contributors to stroke risk in India (Feigin et al., 2024). These findings are consistent with NFHS-5 and GATS-2 data, which show higher consumption of tobacco and alcohol among lower socioeconomic and rural populations—groups that predominantly access government healthcare hospitals [International Institute for Population Sciences (IIPS), 2021; Tata Institute of Social Sciences (TISS), 2018]. Additional studies confirm that both smoked and smokeless tobacco use are less common among individuals with higher educational attainment and socioeconomic position (Thakur et al., 2013). These disparities likely reflect differences in access to preventive care, health literacy, and socio-cultural practices in rural settings.

International evidence further supports these socioeconomic gradients. A 2024 Swedish stroke registry reported significantly worse outcomes among patients with lower educational attainment and income compared with those with university-level education and higher income, underscoring the consistent influence of socioeconomic position on stroke outcomes across diverse healthcare systems (Lindmark et al., 2024). Hospital sector in this analysis should be interpreted as a structural marker encompassing socioeconomic position, healthcare access, service availability, and continuity of follow-up, rather than as an independent causal determinant of outcomes. The observed differences likely reflect the complex interaction between patient characteristics and healthcare system context.

The impact of the SPRINT India intervention varied across healthcare sectors, reflecting differences in baseline risk profiles, healthcare delivery structures, and patient engagement. Consistent with the overall SPRINT India findings—which did not demonstrate a significant reduction in the composite primary outcome but showed improvements in medication adherence and lifestyle behaviors—patients treated in private hospitals in the intervention arm exhibited comparatively better primary outcome profiles than controls. This may reflect greater responsiveness to structured secondary prevention strategies among patients with higher baseline health literacy and healthcare access. A meta-analysis of multifactorial secondary prevention programmes demonstrated greater reductions in systolic blood pressure among intervention recipients compared with controls, supporting the effectiveness of structured interventions in improving vascular risk markers following stroke or TIA (De Simoni et al., 2013).

Although patients treated in private hospitals demonstrated better functional outcomes overall, the intervention was associated with significant improvement in functional recovery among patients treated in government hospitals compared with standard care. Constraints in drug availability, rehabilitation services, and follow-up within government healthcare systems may contribute to poorer baseline functional outcomes; however, the improvements observed among government hospital patients receiving the intervention appear to be largely driven by behavioral modification. This finding aligns with systematic reviews demonstrating that adherence-focused and multimodal secondary prevention interventions improve vascular risk control and post-stroke health outcomes (De Simoni et al., 2013; Lawrence et al., 2015). These results suggest that structured secondary prevention strategies may yield meaningful functional gains, particularly in resource-constrained government healthcare settings.

Behavioral outcomes showed a clear sector-specific response to the intervention. In private hospitals, the intervention was associated with lower rates of alcohol use and current smoking, alongside a greater likelihood of cessation of both behaviors. Medication non-compliance was also reduced by nearly half among intervention recipients compared with controls. These findings suggest that patients with higher baseline health literacy and access to healthcare services may be especially receptive to adherence-enhancing interventions, highlighting the importance of sustained behavioral reinforcement even in settings with relatively robust baseline care.

Physiological responses to the intervention also varied by healthcare sector. Among patients treated in government hospitals, the intervention was associated with significant reductions in SBP and FBG, indicating improved control of major vascular risk factors. Similar benefits have been demonstrated in community-based interventions such as COBRA-BPS, which leveraged community health workers to improve blood pressure control in rural South Asia. In India, government hospitals provide free antihypertensive and antidiabetic medications through national programmes such as the National Health Mission and the National Programme for Prevention and Control of Non-Communicable Diseases, which likely synergised with the intervention to improve vascular risk-factor control (Ministry of Health and Family Welfare (India), 2023; National Health Mission (India), 2023).

In contrast, among private hospital patients, the intervention was associated with a significant reduction in BMI. Given the higher baseline BMI observed in this group, this effect may reflect better access to structured lifestyle modification resources—such as dietetic counseling and physical activity programmes—commonly available in private urban hospitals. The mHealth components of the intervention may have further amplified these advantages (Verma et al., 2025a).

Overall, these exploratory findings suggest context-specific benefits of the SPRINT India intervention across healthcare sectors—improving functional recovery and vascular risk control in government hospitals, while reinforcing behavioral change and medication adherence in private hospitals. These patterns likely reflect differences in baseline risk burden, healthcare access, and patient capacity to engage with long-term secondary prevention, rather than differences in intrinsic intervention efficacy.

Substantial post-stroke employment loss was observed across both healthcare sectors. Patients treated in government hospitals experienced poorer functional recovery, contributing to a higher proportion of unemployment following stroke. While hospital type alone is not a universal predictor of post-stroke employment outcomes, the quality of care processes, access to specialized services, and patients' socioeconomic b ground play critical roles in functional recovery and workforce reintegration. Government medical institutions in India often lack structured long-term rehabilitation services, leading to persistent disability and limited functional gains. A hospital-based survey informed by the World Stroke Organization reported that dedicated stroke units, trained stroke-specific staff, and specialized diagnostic and therapeutic services were largely concentrated in private hospitals, highlighting important gaps in stroke care infrastructure within the government sector (Marshall et al., 2015).

Several limitations warrant consideration. A substantial proportion of participants did not complete follow-up, largely due to early termination of the trial after interim futility analysis. This attrition may have reduced statistical power and introduced potential bias. Engagement with intervention components was limited, with only 17% of participants confirming receipt of the SMS messages or videos. Such limited exposure may have reduced the measurable impact of the intervention and potentially weakened comparisons between healthcare sectors. Detailed engagement metrics stratified by hospital type were not available, limiting our ability to assess differential uptake of the digital intervention across sectors, which may have influenced observed differences (Verma et al., 2023). Although multivariable adjustments were performed, residual confounding from unmeasured sociodemographic and health-system factors cannot be ruled out. Direct measures of income, insurance coverage, discharge counseling practices, and structured follow-up systems were not systematically captured, limiting our ability to disentangle hospital-level effects from broader socioeconomic determinants. SES also intersected with educational attainment, limiting clear separation of their independent effects. Although outcome assessors were masked to treatment allocation, variation in routine clinical documentation or follow-up practices across centers cannot be entirely ruled out. Several outcomes relied on structured self-report, including medication adherence, which was not assessed using a validated adherence instrument, thereby introducing the possibility of recall and reporting bias. In addition, the predominance of urban-based centers may have limited rural representation; urban participants likely derived greater benefit from digital interventions due to higher literacy and healthcare access, while limited follow-up among rural patients—predominantly treated in government hospitals—may have led to under-ascertainment of recurrent stroke and TIA events. Despite these limitations, the prospectively defined 1-year follow-up provides valuable longitudinal data rarely available in standard care, strengthening outcome assessment.