Implementing stroke unit care in selected hospitals in Rwanda

Background: The burden of stroke in low-and middle-income countries (LMICs) has risen sharply in recent years and the rate of increase is set to accelerate due to socio-demographic and lifestyle changes related to the industrialization and a rise in many modifiable vascular disease risk factors. Aims: The first aim was to establish, for countries like Rwanda, how much stroke is a major problem. The second aim was to explore whether the existing stroke services were well prepared. The third aim was to develop and implement a relevant service improvement in Rwanda. Methods: First, I conducted systematic reviews of the literature on the epidemiology and impact of stroke, and the available stroke services in Africa. Second, I conducted a systematic review of the literature and analyzed the INTERSTROKE study data to identify the stroke key performance indicators (KPIs) that have been described in stroke care and assessed their association with patient outcomes. Finally, I selected the stroke unit care KPIs relevant to Rwandan and other LMIC settings, and used several strategies including site champions, provision of educational materials, feedback on usual care, training hospital staff on stroke KPIs, consensus discussions on service improvement by local staff, and discussions with the hospital directors to promote the implementation of the selected KPIs in two hospitals in Rwanda. Results: Stroke was found to be common and important in Africa. However, the provision of stroke care was below the recommended standards. After adjustment for case mix and stroke onset-hospital arrival interval, I found a consistent trend of associations between my implementation intervention and improved delivery of stroke KPIs and patient outcomes. Conclusion: Several common KPIs of stroke unit care can be implemented in hospitals in Rwanda. However, there are some major challenges that need to be addressed for optimal implementation of stroke unit care. Chapter abstracts Chapter 1: Epidemiology and impact of stroke in Africa: a systematic review of the literature Background: Stroke is the second most common cause of death, and the third most common cause of disability-adjusted life-years (DALYs) worldwide, but there is limited information on the stroke burden in Africa. Aim: To describe the epidemiology (incidence, prevalence, mortality, one month-case-fatality) and impact (disability, quality of life, and cost) of stroke in Africa. Methods: I performed a systematic review which included full-text manuscripts published between January 1980 and June 2017 that described the epidemiology or impact of stroke. I searched Medline, Embase, PubMed, and African Journals Online (AJOL) databases, and screened references from bibliographies. There was no language restriction. To determine the estimates of stroke epidemiology and impact variables in Africa, the overall means with standard deviation (SD) were calculated. Results: I identified 44 eligible studies among which 21 were hospital based and 23 were community based. The majority (30/44) of the studies were conducted in urban settings. Overall, the crude mean per 100,000 population was 122.4 (SD: 68.1) for the incidence, 539.1 (SD: 381.5) for the prevalence and 84.7 (SD: 30.15) for mortality. The age-adjusted mean per 100,000 population was 162.7 (SD: 117.5) for incidence, 788.3 (SD: 536.7) for the prevalence and 192.7(SD: 155.2) for mortality. The overall mean rate for stroke one-month case fatality was 30.4% (SD: 11.7%). It was also reported that 30.6% of stroke survivors had moderate or severe disability at one-year post stroke, and at 29 months post stroke, the stroke survivors in general had neither poor nor good quality of life (the mean for the health-related quality of life was 71%). The overall mean for the in-hospital care cost was found to be 1971 (SD: 1108) United States Dollars (USDs). Conclusions: This review provides an overview of the epidemiology and impact of stroke in Africa, despite the paucity of available data. I found that stroke was common and important in Africa. Robust high-quality studies are needed to help policy makers and health care professionals to control the stroke burden in Africa. Appropriate preventive and therapeutic measures should be promoted to decrease the incidence of stroke, improve the outcomes, and maintain the survivors’ quality of life in Africa. Chapter 2: Stroke care in Africa: a systematic review of the literature Background: Appropriate systems of stroke care are important to manage the increasing death and disability associated with stroke in Africa. Information on existing stroke services in African countries is limited. Aim: To describe the status of stroke care in Africa. Methods: I undertook a systematic search of the published literature to identify recent (January 1st, 2006-June 20th, 2017) publications that described stroke care in any African country. Results: My initial search yielded 838 potential papers, of which 38 publications were eligible representing 14/54 African countries. Across the publications included for my review, the proportion of stroke patients reported to arrive at hospital within three hours from stroke onset varied between 10─43%. The median time interval between stroke onset and hospital admission was 31 hours. The reported proportions of stroke patients who received brain imaging within three hours of stroke onset varied between 0% and 13%, and the overall proportion of patients who received brain imaging varied between 13% and 36%. Only twenty-three stroke units in Africa were reported, and two studies indicated that stroke unit admission was associated with a decrease in in-patient case fatality rate of 17-30%. Access to in- and out-patient rehabilitation services was reported to be very low. Poor awareness of stroke signs and symptoms, shortages of medical transportation, health care personnel, and stroke units, and the high cost of brain imaging, thrombolysis, and outpatient physiotherapy rehabilitation services were reported as major barriers to providing best-practice stroke care in Africa. Conclusions: This review provided an overview of stroke care in Africa and highlighted the paucity of available data. Stroke care in Africa usually fell below the recommended standards with variations across countries and settings. Combined efforts from policy makers and health care professionals in Africa are needed to improve, and ensure access to organised stroke care in as many settings as possible. Mechanisms to routinely monitor usual care (i.e. registries or audits) are also needed to inform policy and practice. Chapter 3: Key performance indicators of quality stroke care and their association with patient outcomes: a systematic review of the literature and meta-analysis Background: Translating research evidence into clinical practice often uses key performance indicators (KPIs) to monitor quality of care. However, information on KPIs for stroke care is limited. Aims: To identify the stroke KPIs used in large registries, and to estimate their association with patient outcomes. Methods: I sought publications of recent (January 2000-May 2017) national or regional stroke registers reporting the association of KPIs with patient outcome (adjusting for age and stroke severity). I searched Ovid Medline, Embase and PubMed and screened references from bibliographies. I used an inverse variance random effects meta-analysis to estimate associations (odds ratio; 95% Confidence Interval) with death or poor outcome (death or disability) at the end of follow up. Results: I identified 30 studies (324,409 patients) eligible for the qualitative review. Among these, only 22 were eligible for the meta-analysis. The commonest KPIs were swallowing/nutritional assessment, stroke unit admission, antiplatelet use for ischemic stroke, brain imaging, anticoagulant use for ischemic stroke with atrial fibrillation, lipid management, deep vein thrombosis prophylaxis and early physiotherapy/mobilization. Lower case fatality was associated with swallow/nutritional assessment (OR: 0.78; 0.66-0.92), stroke unit admission (OR: 0.79; 0.72-0.87), antiplatelet use for ischemic stroke (OR: 0.61; 0.50-0.74), anticoagulant use for ischemic stroke with atrial fibrillation (OR: 0.51; 0.43-0.64), lipid management (OR: 0.52; 0.38-0.71), and early physiotherapy or mobilization (OR: 0.78; 0.67-0.91). Reduced poor outcome (death or disability) was associated with adherence to swallowing/nutritional assessment (OR: 0.58; 0.43-0.78) and stroke unit admission (OR: 0.83; 0.77-0.89). Adherence to several KPIs appeared to have an additive benefit. Conclusions: I found that the most frequently reported KPIs for stroke care were swallowing assessment, stroke unit admission, antiplatelets for ischemic stroke, brain imaging, anticoagulants for ischemic stroke with atrial fibrillation, lipid management, deep vein thrombosis (DVT) prophylaxis, and early mobilization. Adherence to common KPIs was consistently associated with a lower risk of death or disability after stroke. Policy makers and health care professionals should implement and monitor those KPIs supported by good evidence. Chapter 4: Stroke services in African and other low and middle-income countries in an international study Background: Stroke key performance indicators (KPIs) have been used to monitor service improvement in high-income countries (HICs), but information regarding their utility in low and middle-income countries (LMICs) is limited. Aims: To explore the association of recording of stroke KPIs with patient outcomes in LMICs, generally and African settings in particular. Methods: I analysed data collected from the INTERSTROKE case control study (conducted between January 2007 and August 2015). I had full data for 12343 participants and analysed 9766 from LMICs. I calculated the odds ratios (OR) with 95% confidence intervals (CI) for the associations of KPIs with 30-day patient outcomes using univariate and multivariate regression analyses to account for patient casemix. I also used the Bonferroni correction method to control the familywise error rate and considered 0.006 as the p-value of significance. Results: In LMICs, availability of a stroke unit (OR: 0.71, 0.60-0.83; p<0.0001) or a stroke specialist (OR: 0.74, 0.64-0.85; p<0.0001), receiving antiplatelet therapy for ischemic stroke (OR: 0.69, 0.56-0.85; p=0.001), and the availability of acute (OR: 0.70, 0.57-0.87; p=0.001) and post-discharge (OR: 0.41, 0.34-0.48; p<0.0001) rehabilitation were independently associated with lower risk for 30-day case fatality. Early brain imaging (OR: 0.71, 0.58-0.88; p=0.001) and stroke unit availability (OR: 0.62, 0.55-0.71; p<0.0001) were independently associated with lower risk of death or severe disability at 30 days. There was a dose dependent relationship of number of KPIs recorded with a better outcome. For the African countries alone, I had no enough evidence to show an association between any of the stroke KPIs investigated and patient outcomes. Conclusions: In LMICs, achieving several common KPIs was associated with a statistically significant reduction in post-stroke death or disability. Policy makers and health care professionals should be encouraged to implement the commonly established stroke KPIs even in settings with limited resources. Chapter 5: Implementing stroke unit care in Rwanda: a two-hospital before and after implementation trial Background: Stroke unit care has become established as the central component of a modern stroke service to improve patient outcomes, but it requires several resources. This raises the question of whether stroke unit care is feasible and applicable to low and middle-income country settings. Aims: To explore the feasibility and effectiveness of implementing stroke unit care in selected hospitals in Rwanda. Methods: I used a before and after implementation trial design. The clinical intervention consisted of 11 key stroke unit care elements that were identified from the results of chapters three and four, the World Stroke Organisation (WSO) recommendations and the Rwandan clinical guidelines. The implementation intervention consisted of identification of site champions, provision of educational materials, face-to-face educational seminar (including feedback on usual care, training on stroke KPIs and local consensus discussions), and discussions with the study hospital directors. Results: Overall, after case mix adjustment for stroke severity (using six simple variables), stroke type and stroke onset-hospital arrival interval, I found a consistent trend of associations between the intervention and an increase in participants who received the KPIs investigated and an increase in better patient outcomes. However, the results were statistically significant for only the use of standardized assessment tools (OR: 2.98, 1.36-6.51; p=0.006), swallowing assessment recorded (OR: 5.73, 2.08-15.74; p=0.001), mobilization recorded (OR: 2.30, 1.16-4.56; p=0.017) and multidisciplinary team meetings recorded (OR: 9.04, 2.74-29.86; p