METHODS: We studied the patterns and effect of practice variations (ie, treatments used and access to services) among participants in the INTERSTROKE study, an international observational study that enrolled 13 447 stroke patients from 142 clinical sites in 32 countries between Jan 11, 2007, and Aug 8, 2015. We supplemented patient data with a questionnaire about health-care and stroke service facilities at all participating hospitals. Using univariate and multivariate regression analyses to account for patient casemix and service clustering, we estimated the association between services available, treatments given, and patient outcomes (death or dependency) at 1 month.

FINDINGS: We obtained full information for 12 342 (92%) of 13 447 INTERSTROKE patients, from 108 hospitals in 28 countries; 2576 from 38 hospitals in ten high-income countries and 9766 from 70 hospitals in 18 low and middle-income countries. Patients in low-income and middle-income countries more often had severe strokes, intracerebral haemorrhage, poorer access to services, and used fewer investigations and treatments (p<0·0001) than those in high-income countries, although only differences in patient characteristics explained the poorer clinical outcomes in low and middle-income countries. However across all countries, irrespective of economic level, access to a stroke unit was associated with improved use of investigations and treatments, access to other rehabilitation services, and improved survival without severe dependency (odds ratio [OR] 1·29; 95% CI 1·14-1·44; all p<0·0001), which was independent of patient casemix characteristics and other measures of care. Use of acute antiplatelet treatment was associated with improved survival (1·39; 1·12-1·72) irrespective of other patient and service characteristics.

INTERPRETATION: Evidence-based treatments, diagnostics, and stroke units were less commonly available or used in low and middle-income countries. Access to stroke units and appropriate use of antiplatelet treatment were associated with improved recovery. Improved care and facilities in low-income and middle-income countries are essential to improve outcomes.

METHODS: Data collection included two strategies. First, previous systematic reviews were searched for studies that met the inclusion criteria of the current review. Second, a new search was done, covering the time since the previous reviews, i.e. January 2013 to May 2017. Five search concepts were combined in order to capture relevant literature: stigma, mental health, intervention, professional students in medicine and nursing, and LMICs. A qualitative analysis of all included full texts was done with the software MAXQDA. Full texts were analysed with regard to the content of interventions, didactic methods, mental disorders, cultural adaptation, type of outcome measure and primary outcomes. Furthermore, a methodological quality assessment was undertaken.

RESULTS: A total of nine studies from six countries (Brazil, China, Malaysia, Nigeria, Somaliland and Turkey) were included. All studies reported significant results in at least one outcome measure. However, from the available literature, it is difficult to draw conclusions on the most effective interventions. No meta-analysis could be calculated due to the large heterogeneity of intervention content, evaluation design and outcome measures. Studies with contact interventions (either face-to-face or video) demonstrated attitudinal change. There was a clear lack of studies focusing on discriminatory behaviours. Accordingly, training of specific communication and clinical skills was lacking in most studies, with the exception of one study that showed a positive effect of training interview skills on attitudes. Methods for cultural adaptation of interventions were rarely documented. The methodological quality of most studies was relatively low, with the exception of two studies.