Globally, stroke is the leading cause of disability and the second leading cause of death [1]. In the last two decades, the lifetime risk of stroke has increased by 50%, with 1 in 4 people estimated to have a stroke in their lifetime [2]. Annually, 6.5 million people die from stroke, constituting approximately 6% of deaths among those aged 15 to 45 years-. Over 143 million healthy life years are lost annually, with 15% lost among those aged 15 to 45 years. India, like other low- and middle-income countries, is experiencing a stroke epidemic. It is the fourth leading cause of death and fifth leading cause of disability in India [3]. Karnataka, a southern Indian state, ranks ninth among all states in India in incidence of stroke. With 109 new patients who had a stroke for every 1,00,000 of its population in 2019, Karnataka had higher incidence of stroke than the national average [4].

Appropriate, timely, and efficient stroke management is crucial for improved outcomes, as standardized care systems have shown to reduce mortality and enhance functional outcomes in acute stroke cases. Effective management requires prompt recognition, rapid transportation to health care facilities, and specialized care to minimize deaths and disabilities [5]. In India, stroke care is mostly urban centric and private sector driven, with shortages of trained specialists and limited access to imaging and diagnostics [6,7]. Even medical college hospitals face a shortage of neurologists, highlighting the need for alternative care models [8]. There is a critical need to develop comprehensive, standardized stroke care systems across different levels of health care delivery for timely intervention, management, and referrals [5,9-11]. While some care models exist, especially in hospital settings, community-based models tailored for India are rare [12].

This quasi-experimental study aims to develop and assess the effectiveness of a Comprehensive Acute Stroke Care Model (CASCADE) in Karnataka, India, on improvement of preparedness, pathways to care, and management practices related to stroke in different health care settings and improvement in survival, disability, and quality of life (QoL) of patients who had an acute stroke.

We plan to conduct a quasi-experimental preinterventional and postinterventional study.

The health care facilities comprise 1 neurospecialty hospital (a caseload of approximately 300 cases per month) and 1 medical college hospital (a caseload of approximately 50 cases per month) at the tertiary level and 1 district hospital (a caseload of approximately 30 cases per month), 1 medical college hospital (a caseload of approximately 50 cases per month), and 2 government hospitals (a caseload of approximately 50 and 30 cases per month) at the secondary level in Karnataka.

The study unit consists of health care facilities including health care personnel and patients who had a stroke. Health care facilities and their personnel will be the unit of study for stroke care preparedness, while all patients who had a stroke for the first time (both ischemic and hemorrhagic) attending the emergency rooms of these hospitals and/or admitted within the hospital during a 1-year period (or until the requisite sample size is obtained, whichever is earlier) will be the unit of study to assess CASCADE effectiveness before and after the intervention.

All aspects of stroke care from the time of onset to discharge and after discharge (for 1 year), namely preparedness, management, pathways to care, and the CASCADE, shall be assessed using semistructured, partially open-ended schedules (Table 1).

Increased survival, reduced disability, and improvement in stroke-specific QoL after introducing the CASCADE compared to the preintervention phase is considered an outcome. An increase in the scores and indicators of hospital preparedness and management practices is considered as the outcome for hospital preparedness.

The study will be conducted for 4 years, which comprises of 6 months for development and implementation of the CASCADE and 1 year each for before and after the intervention.

The primary patient-level outcome for sample size estimation is 28-day poststroke survival. Disability, QoL, rehabilitation processes, management practices, and hospital preparedness indicators are considered secondary outcomes and are assessed at the respective predefined follow-up periods. The sample size is estimated based on the Freedman formula [13] given below, where n is the total sample size required, E is the expected number of events in the study population, and and are proportion survived in group 1 (before intervention) and group 2 (after intervention), respectively, after 28 days of onset of stroke:

The respondents in the preimplementation phase of the CASCADE are considered group 1 and those in the postimplementation phase as group 2. The primary outcome for CASCADE effectiveness includes improved hospital preparedness and improved patient treatment outcomes (death, disability, and QoL). The timing and sequence of implementation at each site will be documented in case of minor variations in the timing of intervention implementation across participating sites due to administrative or operational factors beyond the control of the investigators. These factors will be incorporated into the statistical analysis plan. Site-level and time-related variations will be adjusted for during analysis to account for potential differences in implementation timing across hospitals. We hypothesize that the treatment outcomes in group 2 will be better than those in group 1. The Kaplan-Meier method will be used for estimating the survival functions of the two groups at follow-up. The log-rank test will be used to compare the survival distributions between the two groups for each time point. Finally, the Cox proportional hazards regression analysis will be performed at each of the 4 time points separately in which the grouping variable is the intervention (CASCADE) in addition to potential confounders. Relevant patient-level covariates, including stroke severity, National Institutes of Health Stroke Scale score, timing of presentation, referral pathway, and other clinical characteristics, will be considered as potential confounders in adjusted analyses. The QoL before and after the implementation of the CASCADE will be compared using an independent sample 1-tailed t test or the Mann-Whitney U test depending on the normality assumption. Linear mixed models will be used in which QoL is the dependent variable and the CASCADE is the main independent variable. In the model, the group and the 4 time points are fixed factors and participants are considered as random. The modified Rankin Scale (mRS) scores before and after implementation of the CASCADE will be compared using the Mann-Whitney U test at each time point after adjusting for multiple comparisons using the Bonferroni method. Survival rate by different categories of stroke (hemorrhagic and ischemic) and by different severities of stroke (Glasgow Coma Scale and mRS based) will be used for evaluation of effectiveness of the model.

To address the possibility of time-varying confounding in the quasi-experimental study design, the study incorporates collection of several patient-level and system-level variables during both preintervention and postintervention phases. Patient-level variables include stroke severity, National Institutes of Health Stroke Scale score, mRS score, timing of presentation (including arrival within the golden hour), referral pathways, investigations performed, and treatments received. These variables will help assess differences in case mix and care processes across study periods.

In addition, hospital preparedness assessments conducted during both phases will capture system-level and organizational changes, including patient attendance patterns, staffing, infrastructure, referral mechanisms, service availability, and operational changes occurring over time.

These variables will be considered during analysis and interpretation to contextualize observed differences between preintervention and postintervention phases and reduce the impact of time-varying confounding. The objective is to achieve improved survival rate at every level of case severity in proportion to the level of the stroke care facility.

For improved hospital preparedness, the specific scoring system developed and indicator data will be compared using an independent samples t test or the Mann-Whitney U test depending on normality.

Ethics approval for this study was obtained from the institutional ethics committee of the National Institute of Mental Health and Neurosciences on June 17, 2022, #NIMHANS/35^th IEC (BS&NS DIV.)/2022. This trial was registered under the Clinical Trials Registry-India (CTRI/2023/04/051370) April 6, 2023. Any protocol modifications will be reported to the ethics committee and the trial registry. Administrative approvals were obtained from all the participating hospitals. Informed consent will be sought from the patient or caregiver ensuring priority for care, privacy, and confidentiality. Informed consent will be administered either in the local language (Kannada) or English as preferred by the respondents. Illiterate individuals will have the consent form read to them in the presence of a literate witness and will be requested to provide a thumb impression. In cases where the patients are aphasic, unconscious, or otherwise unable to provide informed consent due to severe disability or impaired decision-making capacity, written informed consent will be obtained from the primary caregiver or legally acceptable representative accompanying the patient in the presence of a witness. The electronic data collection system includes a field to identify whether the respondent is the participant or the caregiver. Where caregiver consent is obtained, the caregiver’s relationship with the participant will also be recorded. Wherever feasible, if the participant subsequently regains communication ability or decision-making capacity, the study procedures will be explained directly to the participant and their consent or assent for continued participation will be obtained.

We plan to communicate to the participants and general public, scientific audiences, and policymakers about the results of the trial in various platforms including mandatory reporting to the trial registry, publishing in scientific journals, and providing a plain language summary as well as fact sheets.

This first-of-its-kind endeavor, the CASCADE seeks to reduce stroke-related mortality and morbidity while improving the long-term survival of patients with QoL for both the patients and their caretakers. This protocol aims at estimating current management practices and pathways to care and to evaluate the effectiveness of the developed CASCADE, which includes triage algorithms, diagnosis, investigations, treatment, referral, discharge-related interventions, and an assessment of health care facility preparedness for acute stroke care across different levels of health care. The outcomes assessed are improvement in survival, disability, and QoL of patients who had an acute stroke.

The proposed CASCADE is more comprehensive and encompasses the full continuum of care for stroke across various levels of health care. Most studies assess either one or few components of care during a hospital stay such as a critical path for nursing care, an emergency room management algorithm, and a hospital unit physician’s order sheet [14]. Very few studies have looked at follow-up care [15,16]. As far as we understand, there are no studies conducted that assess the quality of stroke care, which includes infrastructure, imaging facilities, or readily available stroke care units. This assumes importance along with the lack of availability of specialists at different levels of health care. There are only 2300 neurologists for a population of 1.4 billion, and they are mostly urban centric. Adding to this, we have an immensely skewed health care system along with radiological services. Furthermore, the lack of organized stroke care networks in urban areas makes effective acute stroke intervention difficult. The CASCADE shows promise by including the full continuum of care from the time of occurrence of stroke until 1 year after the stroke across different levels of health care. This includes pathways of care prior to admission, in-hospital triage, care and management practices, patient and sample journeys within health care settings, and posthospitalization care.

Other options to achieve the objectives were considered before deciding on the quasi-experimental study design. Stroke registries are a known resource for developing stroke care models [17] but lack the comprehensiveness required for our study due to limitations such as nonuniform data collection; method used; and unreliable information on management practices, care pathways, and rehabilitation. These limitations make stroke registries unsuitable for our objectives. Given the objectives, a cluster randomized trial or stepped-wedge design could have been a more appropriate design. It is known that the cluster or stepped-wedge designs are suitable for implementation on a large scale [18]. In our study, patients who had a stroke are recruited from a limited number of 6 strategically selected health facilities in Bengaluru and Kolar. These facilities were chosen based on patient admission trends and existing referral pathways. Considering this and the practical feasibility in evaluating the model in real-world settings, a quasi-experimental study design was chosen. Quasi-experimental predesign and postdesign may be influenced by time trends and other external factors. However, the effect of changes in stroke systems of care itself is one of the key aspects being evaluated in this study. Changes in survival, disability, and QoL may be influenced by hospital-level factors, patient case mix, referral patterns, staffing, diagnostics, and organizational processes. Several of these changes may occur because of gaps identified during baseline assessments and may therefore become part of the intervention exposure itself rather than only representing a source of bias.

At the same time, factors that are not directly related to the intervention but are likely to change during the study period will be documented and considered during analysis. A separate hospital preparedness assessment tool will be administered during both preintervention and postintervention phases to capture facility-level and system-level factors, including patient attendance, infrastructure, staffing, service availability, referral mechanisms, diagnostics, and other organizational changes. These data will help contextualize observed improvements and assess whether changes may be related to broader system-level modifications occurring alongside the intervention.

This study has several strengths. First, the model includes both individual-level and macrolevel interventions. Unlike existing models, which are primarily designed for tertiary care settings, the CASCADE is being developed for implementation across various health care levels, incorporating algorithm-driven management and referral of stroke cases. Second, the model addresses the entire continuum of care, from the onset of stroke to posthospital care, focusing on survival, disability, and QoL for 1 year after the stroke. Third, it includes comprehensive assessments of both prehospital and posthospital care pathways, which is a unique strength. Additionally, the study evaluates within-hospital patient journeys, acknowledging their impact on treatment outcomes.

Allocation of sample size is based on the caseload in each of the identified hospitals. Quality of data collection and processes are monitored by the independent monitoring team members and the expert advisory group. This will validate applicability and results of the CASCADE.

This study has certain limitations that need mention. First, the effectiveness of the CASCADE will be assessed in only a few settings in Kolar and Bengaluru, which might not be representative of all the settings within Kolar and Bengaluru. However, the model developed would be replicable (with appropriate modifications to suit local needs), given the uniform presentation of stroke cases and the hospitals selected largely represent hospitals similar to those in Bengaluru and Kolar.

Second, being a quasi-experimental study, there might be some residual confounding for the effect measures assessed due to unknown confounders. Given the inability to incorporate randomization into the study design, this limitation is unlikely to be completely negated. Furthermore, the current methodology adopted is likely to provide valuable insights into how the CASCADE might work in real-world situations where confounders cannot be completely controlled. Finally, this being a multisite study, it is likely that coordination and cooperation of study sites become a challenge. The study team and the institution have a strong working relationship over the years with these hospitals. The study institution also has a significant presence in the field, with Kolar being their designated public health observatory.

If effective, the CASCADE is likely to be adopted widely. Developed through a robust methodology, the preimplementation phase identifies current care practices and gaps, guiding the model’s refinement. Finalized with expert and stakeholder input, the intervention phase will assess the model’s impact on survival, disability, and mortality. Given its evidence-based development and validation, the model is replicable and could become standard practice in India and other low- and middle-income countries.