Abstract
Background
Early hospital arrival is essential for timely thrombolysis in acute ischaemic stroke. However, many patients in low-resource settings continue to reach tertiary care centres beyond the recommended treatment window.
Purpose
To evaluate pre-hospital and in-hospital delays among patients with acute ischaemic stroke in Jharkhand and identify factors associated with timely hospital arrival.
Methods
A prospective observational study included 237 consecutive patients with acute ischaemic stroke admitted to a tertiary care hospital in Ranchi (July 2022–August 2024). Patients were classified as early (≤4.5 h) or delayed (>4.5 h), and multivariable logistic regression identified predictors of early hospital arrival.
Results
Only 15 patients (6.3%) reached the hospital within 4.5 h. Most were men, older adults and from rural, socio-economically disadvantaged backgrounds. Most (85.7%) first visited a nearby non-specialist healthcare facility, while awareness of stroke symptoms (14.0%) and thrombolysis (5.5%) was poor. Timely arrival was less likely after visiting a nearby healthcare facility (aOR 0.005, p < .001) but more likely with higher socio-economic status (aOR 17.02, p = .016) and ambulance availability (aOR 6.11, p = .021).
Conclusion
Improving ambulance services, referral pathways and public awareness may reduce delays and improve timely stroke treatment.
Keywords
Introduction
Stroke remains one of the leading causes of death and long-term disability worldwide, and its burden continues to rise in low- and middle-income countries.1, 2 Most strokes are ischaemic in origin, and timely restoration of cerebral blood flow is essential to prevent irreversible neuronal injury.3–5 Intravenous thrombolysis with recombinant tissue plasminogen activator is the most widely accessible reperfusion therapy for acute ischaemic stroke and offers the greatest benefit when administered as early as possible, ideally within 4.5 h of symptom onset.6, 7 The therapeutic effect diminishes rapidly with each passing minute, underscoring the principle that ‘time is brain’. 8 International guidelines emphasise rapid triage, early neuroimaging and streamlined in-hospital workflows to minimise door-to-needle times. 9 However, in many parts of the world, including India, the proportion of patients who reach a stroke-ready hospital within the recommended time window remains persistently low.10, 11 Previous Indian and regional studies have repeatedly highlighted several obstacles: limited public awareness of stroke warning signs, delays in recognising symptom severity, inadequate emergency transport systems and fragmented referral pathways. 4 These challenges are amplified in resource-constrained settings, where rural residence, lower socio-economic status and distance from tertiary hospitals are common barriers.12–16 Jharkhand is a largely rural and tribal state where health facilities are spread unevenly, and specialised stroke services are still developing. These findings will help to identify practical solutions to strengthen emergency response and improve access to timely thrombolysis. In this study, we evaluated not only prehospital delay (time to hospital arrival) but also in-hospital delays following patient presentation. Additionally, we examined multiple contributing factors, including patient demographics, stroke severity and care-seeking behaviour.
Methods
Study Design and Setting
This prospective observational study was carried out in the Department of Neurology at the Rajendra Institute of Medical Sciences (RIMS), Ranchi, a tertiary care hospital. The hospital serves a large rural and tribal population of Jharkhand. The study was carried out for two years, from August 2022 to August 2024, and during this time, all the patients with symptoms of acute ischaemic stroke were screened.
Participants and Eligibility Criteria
Adults aged 18 years or older were eligible if they presented with clinical features consistent with acute ischaemic stroke and had neuroimaging confirmation on CT or MRI. Written informed consent was obtained either from the patient or, when the patient lacked capacity, from a family member or legally authorised representative. Socio-economic status was assessed using a simplified approach based on components of the Modified Kuppuswamy scale, including educational status and available socio-economic indicators, and categorised into low and higher socio-economic groups for analysis. 17 In the absence of reliable income data, participants were classified as low socio-economic status if they possessed any of the following: Below Poverty Line (BPL) card, Ration card or Ayushman Bharat card; all others were categorised as higher socio-economic status. Affordability was assessed based on patient-reported financial constraints related to accessing healthcare. Stroke severity was assessed using the National Institutes of Health Stroke Scale (NIHSS). 18 The age cutoff (<50 vs ≥50 years) was selected in accordance with established epidemiological definitions of early-onset stroke, where this threshold is widely used to differentiate younger-onset from typical late-onset stroke populations.19, 20 A total of 1,562 individuals were assessed during the study period. Each patient underwent clinical evaluation, baseline laboratory testing, NIHSS scoring when feasible and neuroimaging. Patients were excluded if they had stroke mimics, intracerebral haemorrhage or other non-ischaemic stroke variants, incomplete diagnostic work-up, uncontrolled blood pressure despite treatment (SBP >185 mmHg or DBP >110 mmHg), as these levels are contraindications to intravenous thrombolysis and could confound the assessment of timely eligibility for reperfusion therapy, missing essential time parameters (e.g., onset-to-door time) or lack of consent. After all exclusions, 237 patients with confirmed acute ischaemic stroke and complete clinical imaging data sets were included in the final analysis.
Data were collected using a structured questionnaire administered to the patient or, when necessary, to an accompanying caregiver after initial stabilisation. Time-related details included the exact moment of symptom onset or last-known-well, the onset-to-door interval, whether arrival occurred within the therapeutic 4.5-h window and the time of day when symptoms first appeared. Stroke onset was categorised into daytime, evening and nighttime to capture potential differences in circadian patterns, patient response behaviour and accessibility to healthcare services across different time periods. The variation of time differs in different studies. 21 Demographic variables included age, sex, educational attainment, socio-economic status, residence (urban or rural) and approximate distance from the hospital. Clinical information included stroke severity assessed by the NIHSS, the range of presenting neurological symptoms and past medical history such as hypertension, diabetes, dyslipidaemia, atrial fibrillation, cardiovascular disease and relevant family history. Behavioural factors such as tobacco use, smoking and alcohol consumption were also recorded. Health-system and access-related variables focused on the patient’s first point of contact after symptom onset (whether they initially visited a nearby facility), availability of ambulance services, particularly in rural areas, awareness of stroke warning signs and thrombolysis, financial or affordability concerns and whether the patient was alone at the time of onset. These combined variables allowed for a comprehensive assessment of individual, behavioural and system-level contributors to early or delayed hospital arrival. The operational definitions of the study variables used in this study are presented in Table 1.
Variable Definition.
Statistical Analysis
The study evaluated patients reaching the hospital after stroke onset by categorising patients into early (≤4.5 h) and delayed (>4.5 h) presenters. To understand which factor might influence this timing, each variable was analysed separately. Continuous measures were compared using independent t-tests, while categorical variables were assessed with Chi-square tests. Any factor showing a p value below .2 in these initial checks was then entered into a multivariable logistic regression model to see which ones independently predicted early presentation. A p value of less than .05 was taken as statistically significant. All analyses were carried out using IBM SPSS Statistics Version 21.0 (IBM Corp., Armonk, NY, USA).
Results
A total of 1,562 patients were screened during the study period. After applying the predefined eligibility criteria, 237 patients with confirmed acute ischaemic stroke and complete data were included in the final analysis. The exclusion criteria included haemorrhagic stroke (n = 815) and delayed presentation beyond 10 days of symptom onset (n = 423). Additional exclusions included unspecified stroke (n = 55), uncontrolled blood pressure despite treatment (n = 20) and other non-ischaemic stroke types such as subarachnoid haemorrhage and cerebral venous thrombosis (n = 12). After accounting for all exclusions, the remaining 237 patients formed the final study cohort (Figure 1).
Screening and Exclusion Flowchart for Patients Evaluated for Acute Ischaemic Stroke.
The baseline demographic, clinical, and healthcare access characteristics of the study population are summarized in Table 2.
Baseline Characteristics of the Study Population (N = 237).
Among the 237 patients with acute ischaemic stroke, men constituted 61.2% of the study. Most patients were aged 50 years or older (83.5%). Educational attainment was generally low, with 48.5% being illiterate, and 88.2% belonged to the low socio-economic group. Rural residence was common and was reported in 70.1% of patients. The mean distance was 120 ± 88.3 km and the median 120 km (IQR 48–170). The mean travel time was 3.06 ± 2.13 h with a median of 3.11 h (IQR 1.14–4.10), ranging from 0.08 to 12 h. A family history of vascular risk factors was present in a smaller proportion of participants, including hypertension in 28.7%, diabetes mellitus in 24.1%, ischaemic heart disease in 3.4%, stroke in 12.2% and myocardial infarction in 1.7%. Among comorbidities, hypertension (56.2%) and diabetes mellitus (28.7%) were the most frequent. Dyslipidaemia (2.5%), previous myocardial infarction (3.8%), valvular heart disease (4.6%), atrial fibrillation (1.3%) and other comorbidities (6.4%) were less common. Regarding behavioural factors, smoking was reported in 11.0% of patients, alcohol consumption in 35.4% and tobacco use in 54.9%. Beedi use and hookah use were reported in 5.5% and 2.5% of patients, respectively. In terms of healthcare access and awareness, 68.0% of patients reported living far from a hospital, and 85.7% first visited a nearby healthcare facility before reaching the study centre. Awareness of stroke warning symptoms was reported in only 14.0% of patients, while awareness of thrombolysis was present in 5.5%. Ambulance availability was reported in 39.2% of patients, and 12.6% reported affordability as a barrier to care. Stroke onset occurred mostly during daytime hours (49.0%). Only 6.3% of patients arrived within 4.5 h of symptom onset, whereas 93.7% presented after this period. The most common presenting symptoms were limb weakness (86.5%), slurred speech (68.4%), facial deviation (48.5%) and loss of consciousness (37.6%). Stroke severity at presentation was high: 63.7% of patients had severe or very severe NIHSS scores, 77.6% had total disability on the Barthel Index, 56.5% had moderate to severe impairment on the Glasgow Coma Scale, and 93.7% were dependent on the modified Rankin Scale (mRS 3–5).
The association between demographic, clinical, and behavioral variables and hospital arrival time is presented in Table 3.
Association Between Demographic, Clinical and Behavioural Variables and Outcome: Category-wise Distribution with N (%) and p Values.
When baseline neurological status was compared between early and delayed presenters, no statistically significant differences were observed. The distribution of Glasgow Coma Scale categories was comparable between the groups (p = .163), and functional status assessed using the mRS also did not differ (p = .956). Similarly, stroke severity based on NIHSS scores showed no meaningful variation between early and delayed arrivals (p = .156), indicating that initial clinical severity did not influence time to hospital presentation. In contrast, several socio-economic and health system-related factors demonstrated significant differences. Patients presenting within the 4.5-h window were less likely to belong to lower socio-economic groups (73.3% vs 89.2%; p = .015). Access to ambulance services in rural areas was significantly higher among early presenters (73.3% vs 36.9%; p = .011). The most striking difference was observed in the first point of healthcare contact. Only 13.3% of early presenters visited a nearby healthcare facility prior to reaching the stroke centre, compared with 90.5% of delayed presenters (p < .001), identifying this as the strongest barrier to timely hospital arrival. Awareness of stroke symptoms and thrombolysis was lower among delayed presenters; however, these differences did not reach statistical significance (p = .089 and p = .075, respectively). The distribution of presenting symptoms, including limb weakness, slurred speech, facial deviation, unsteadiness of gait, vomiting, headache and visual disturbances, was similar across both groups. Less frequent symptoms such as seizures and dysphagia were also comparable, with no statistically significant differences observed. Alcohol use, smoking, tobacco consumption, myocardial infarction, valvular heart disease, atrial fibrillation, hypertension, diabetes mellitus and dyslipidaemia were similarly distributed between early and delayed presenters, with no significant associations identified. Among demographic variables, educational status showed a significant association, with higher levels of education more frequently observed among early presenters (p = .025). In contrast, place of residence (urban vs rural), sex and age group did not differ significantly between the groups (all p > .05). Overall, the findings suggest that delays in hospital presentation were primarily influenced by health system and socio-economic factors rather than clinical severity or symptom profile. Limited access to ambulance services, lower socio-economic status and initial consultation at non-specialist healthcare facilities emerged as the key contributors to delayed presentation, whereas clinical characteristics remained largely comparable between early and delayed groups. The independent factors associated with delayed hospital arrival identified by multivariable logistic regression are presented in Table 4.
Multivariate Logistic Regression Analysis for Factors Associated with Delayed Hospital Arrival.
The multivariable logistic regression analysis identified several independent factors associated with timely hospital arrival (≤4.5 h). Visiting a nearby non-specialist healthcare facility prior to reaching the stroke centre emerged as the strongest determinant. This variable showed a marked negative association with timely arrival, with an adjusted odds ratio (aOR) of 0.005 (95% CI: 0.001–0.055; p < .001), indicating that patients who initially sought care elsewhere were significantly less likely to reach the hospital within the therapeutic window. Access to ambulance services in rural areas was positively associated with timely presentation. Patients with ambulance availability had significantly higher odds of early arrival (aOR: 6.11; 95% CI: 1.32–28.29; p = .021). Low socio-economic status was independently associated with delayed presentation, with patients from lower socio-economic groups showing significantly lower likelihood of timely arrival (aOR: 17.02; 95% CI: 1.68–171.98; p = .016). Overall, these findings indicate that system-level and socio-economic factors, rather than clinical characteristics, play a dominant role in determining prehospital delay in this setting. Consistent with the multivariable analysis, univariate analysis showed that early presenters were less likely to belong to lower socio-economic groups. A pronounced difference was observed in the first point of medical contact: only 13.3% of patients who arrived early had visited a nearby healthcare facility prior to reaching the stroke centre, compared with 90.5% of those with delayed presentation (p < .001).
Discussion
In our study, the pattern of vascular risk factors mirrored established Indian data, with male predominance and high rates of hypertension, diabetes and tobacco use. These observations are consistent with large national stroke-risk assessments, including the work of Ram CVS et al., which similarly identify these factors as major contributors to India’s ischaemic stroke burden. 22 Limb weakness was the most frequent presenting symptom in our study, followed by slurred speech and facial deviation. This pattern is consistent with observations from other Indian clinical-profile studies, where focal motor deficits dominate the initial presentation of ischaemic stroke 23 and multicentric analyses highlight motor deficits as the predominant initial presentation in Indian ischaemic stroke cases. 24 Educational attainment in our study was generally low, which is consistent with national data. Lower education has been linked to poorer symptom recognition, reduced awareness of treatment options and suboptimal stroke-related decision-making, patterns also highlighted in the GBD-based Indian stroke-burden analysis by Behera et al. 25 Finally, the NIHSS pattern in our study, with most patients falling into the moderate to very severe range, matches what many Indian studies have reported. Higher NIHSS scores at admission are well known to predict worse in-hospital outcomes and long-term disability, a relationship also shown by Ruby et al. 26 A striking finding was that 85.7% of patients first sought care at a nearby non-stroke-centre facility before coming to the stroke centre. This indirect pathway was a powerful negative predictor of timely arrival and good outcome (p < .001), emphasising initial contact leading to significant delays. Nearly one-third of the patients who first went to non-stroke centres reported stopping at spiritual healers or local practitioners after symptom onset. Edakkattil et al. 27 found that only 15.5% of patients arrived within 4.5 h, with low awareness and poor transport contributing to the delay. A similar pattern was reported by Gupta et al. from a rural tertiary centre in North India, where 87.4% of patients first visited lower-level facilities and the median onset-to-presentation time exceeded 7 h. 28 International data also illustrate the scale of this problem. In Denmark, patients who contacted a general practitioner or another health professional instead of emergency medical services (EMS) experienced major system delays, adding a median of about 490 min in one subgroup, whereas nationwide analyses show that bypassing primary care leads to much faster access to reperfusion therapy. 29 Beyond the first point of care, the literature on inter-hospital transfers also shows clear time penalties. Some observational studies find comparable outcomes between the two, suggesting that factors such as distance, transport time and regional network structure can influence which approach works best. 30 These findings fit well with our own observation that any detour before reaching a stroke-ready centre can seriously worsen outcomes. Socio-economic status was another clear factor; patients from higher SES backgrounds were far more likely to arrive on time (aOR 21.95; p = .012). A study reinforces the idea that socio-economic context, access to information, resources and timely transport play a major role in how fast patients seek and receive stroke care. 31 A systematic review also reports the same pattern in patients with lower education or income, and those living in rural areas, who are far more likely to arrive late and less prepared for reperfusion therapy. 32 The median travel time of 3.11 h in our study aligns closely with previous studies reporting median prehospital delays of approximately 3–4 h, suggesting a consistent pattern of delayed hospital arrival in stroke patients across diverse settings.33, 34 The mean distance from the hospital was 120 ± 88.3 km, with a median of 120 km (IQR 48–170 km), which is consistent with prior studies reporting that patients residing ≥100 km or even >60 km from tertiary care centres experience significant delays in accessing stroke care.35, 36 Findings from community-based and hospital-based analyses in South India tie lower education and rural residence to late arrival. 6 Registry data from Scandinavian countries further show clear socio-economic differences in prehospital stroke care, including variations in triage and monitoring. These disparities suggest that access and priority can differ right from the patient’s first interaction with the emergency system. 13 Rural ambulance access also showed a clear advantage in our study (aOR 7.73; p = .015). This is consistent with evidence from a multicentre study in Thailand, where patients who used an ambulance were far more likely to arrive within the treatment window (OR ≈ 0.31), while those referred from another hospital had much higher odds of presenting after 4.5 h. 37 In India, the use of EMS is still very low. One tertiary-care study found that only 29.8% of emergency patients arrived by ambulance, and national policy reports continue to show large geographic differences and persistent gaps in access across the country. 38 Awareness was very low in our study, only about 5.5% knew about thrombolysis and around 14% recognised stroke symptoms. In Korea, Kim et al. showed that patients who understood stroke symptoms and knew about thrombolysis tended to reach the hospital much sooner. Their study found that both general stroke awareness and specific knowledge of IV tPA independently reduced prehospital delays, clearly demonstrating that better awareness increases the chances of arriving in time for thrombolysis. 39 Wu et al. 40 also found that patients with better stroke knowledge and more positive attitudes were significantly more likely to reach the hospital early in Hubei Province. Their findings reinforce the idea that poor awareness is a key and modifiable barrier to getting patients into the reperfusion window on time. A 2024 community survey from a South Indian tertiary centre similarly observed that lack of recognition of warning signs and emergency responses ‘often leads to delays’ in reaching care within the golden hour. 41 Taken together, these findings suggest that the very low rate of thrombolysis in our study is not explained by system barriers alone. Poor awareness of stroke symptoms, the critical 4.5-h treatment window and the availability of thrombolytic therapy itself appear to play an equally important role. Hypertension was present in 56.2% of our patients and diabetes in 28.7%, which is very similar to what has been reported in Indian hospital-based AIS studies. Across the country, hypertension consistently emerges as the leading comorbidity, usually ranging between 54% and 75% and diabetes is found in roughly one-third of patients. A study from Navi Mumbai documented hypertension in 54% of ischaemic stroke patients and diabetes in 29%, which is similar to our findings. 42 Our study population represents the typical Indian cardiometabolic stroke profile rather than an unusually high-risk group. Although the proportion of tobacco users in our sample (54.9%) seems high at first glance, it is entirely plausible in the Indian setting, where smokeless tobacco (SLT) use is widespread and remains a major contributor to vascular risk. 43 An official Indian government monograph also lists stroke, along with heart disease and oral cancers, as major health outcomes attributable to SLT use. 44 In the nationwide tenecteplase registry, hypertension was documented in 60.4% of ischaemic stroke patients, very similar to our study, yet only 16.3% had severe strokes (NIHSS ≥15), with most cases falling in the mild-to-moderate range. In contrast, our severity profile aligns more closely with high-acuity inpatient series from South Asia, where roughly one-third of patients present with NIHSS scores above 15 and late arrival is common. 45 Our finding that most patients arrived with severe or very severe strokes fits well with their delayed presentation and the frequent detours through non-specialist facilities.
Conclusion
In this study, timely arrival for thrombolysis was exceptionally rare, with only 6.3% of patients reaching the hospital within the therapeutic window. The dominant contributors to delay were system-level barriers, particularly detours to nearby non-specialist facilities, limited ambulance access in rural regions and socio-economic disadvantage rather than clinical severity or comorbidity burden. Public awareness of stroke symptoms and thrombolysis was critically low, further compounding delays. Together, these findings highlight substantial and modifiable gaps in the regional stroke-care pathway. Strengthening EMS, minimising diversion to non-stroke facilities and implementing sustained population-level awareness programmes are essential to improve timely hospital arrival, enhance access to thrombolysis and reduce preventable stroke-related disability in resource-constrained settings.
Limitations
This study has several limitations that should be considered. It is based on data from a single tertiary-care government hospital, and its observational design limits the ability to control for all potential confounders. Although a broad range of demographic, clinical and system-related variables was captured, several important influences, such as family decision-making, cultural attitudes towards medical emergencies and practical transport barriers, were not measured and may have contributed to delays in ways that could not be quantified. Another key limitation is the incomplete documentation of prehospital time intervals. Beyond the overall onset-to-door time, consistent data were not available for time taken to reach the first healthcare facility, delays within that facility or inter-facility transfer intervals. Similarly, in-hospital process metrics, such as door-to-imaging and door-to-needle times, were not consistently recorded. In addition, time-related variables were based on patient or caregiver recall, introducing the possibility of recall bias. Objective income data were not available, and affordability was assessed using self-reported measures. Referral patterns may also have influenced the study population. As a tertiary-care centre, the hospital frequently receives patients with more severe or complicated strokes, while individuals with milder symptoms may remain at peripheral facilities or not seek tertiary care. This may have resulted in a skewed severity profile and could limit the generalisability of the findings to other healthcare settings. Furthermore, several variables relied on self-reported information, such as awareness of stroke and thrombolysis, as well as behavioural factors including smoking and alcohol consumption, introducing the potential for recall bias and social desirability bias. The regional context also represents an important limitation. Jharkhand has a predominantly rural and tribal population with distinct healthcare access patterns compared to urban regions with more developed EMS. Therefore, extrapolation of these findings to settings with established stroke systems of care should be undertaken with caution. Finally, the relatively small proportion of patients arriving within the therapeutic window resulted in wide confidence intervals for several predictors. Larger, multicentre studies are needed to improve the precision and generalisability of these findings.
Footnotes
Author’s Contribution
Tannu Kumari and Surendra Kumar conceived and designed the study. Anupa Prasad, Amit Kumar, Ganesh Chauhan, Lakhan Majhee, and Vivek Verma contributed to the study design, methodology, interpretation of data, and reviewed the manuscript. Yuvraj Sinha contributed to data acquisition, data curation, and data validation. Surendra Kumar supervised the study and provided guidance and resource. Tannu Kumari performed the data analysis, interpreted the findings, and drafted the manuscript. All authors reviewed and approved the final version of the manuscript.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Disclaimer
The views expressed in this article are those of the authors and do not necessarily reflect the views of their affiliated institutions or funding agencies.
Funding
The authors received no financial support for the research, authorship and/or publication of this article.
Patient Consent
Written informed consent has been obtained from all participants or their caregivers.
Statement of Ethics
This study is a part of data derived from the ongoing FeSSH study, which has received approval from the Institutional Ethics Committee of Rajendra Institute of Medical Sciences (RIMS), Ranchi [IEC Reg No.-ECR/769/INST/JH/2015/RR-21].
