Administration of a pictorial questionnaire to screen for stroke among patients with hypertension or diabetes in rural Ghana

Stroke prevention is an important public health goal in low-and-middle income countries (LMIC) due to its high mortality and morbidity in these settings. Epidemiological data from high-income countries have demonstrated a potent predictive association between self-reported stroke symptoms and risks of future strokes, incident cognitive impairment and all-cause mortality. To utilize a pictographic version of the 8-item Questionnaire for Verifying Stroke Free Status (QVSFS) to screen for stroke symptoms and determine its predictors among hypertensive and diabetic patients in a rural hospital within a LMIC. Between July and October 2016, 500 consecutive patients with hypertension and/or diabetes mellitus encountered in clinic at the Agogo Presbyterian Hospital, a district level health institution in Ghana were enrolled. A validated pictographic version of the QVSFS was administered by trained research assistants to all study participants and demographic and clinical information on hypertension and diabetes control were collected. The neurologist/specialist physician examined all patients neurologically using a structured questionnaire and reviewed medical charts for objective documentary evidence of stroke. Predictors of stroke symptoms were assessed using a multivariate logistic regression model. Among the cohort, median (IQR) age was 58 (51–66) years, 399 (79.8%) were women, 388 (77.6%) had hypertension, 92 (18.4%) had diabetes and hypertension, and 20 (4.0%) had just diabetes. 30 (6.0%) patients had confirmed stroke/TIA after physician evaluation, 151 (30.6%) had at least one symptom of stroke (SS) without prior diagnosis of stroke/TIA and 319 (63.8%) reported no stroke symptoms. Independent predictors of reporting stroke symptoms were increasing age - aOR: 1.38 (1.16–1.64) for each 10years older and HBA1C–OR: 1.74 (1.16–2.61) for each 2% increase. Subjects with previous stroke/TIA diagnosis had significantly higher mean±SD systolic blood pressure (148.0±21.2mmHg vs. 136.0±17.3mmHg vs. 138.3±19.7mmHg, p=0.009) and median (IQR) HBA1C (8.8% (6.7–10.2) vs. 8.2 (6.4–10.6) vs. 7.2 (6.3–8.0), p=0.05) compared with SS and asymptomatic groups. One in three patients with hypertension and/diabetes in rural Ghana report stroke-like symptoms. Routine use of the pictographic version of the QVSFS could assist in identifying patients with or at high risk for stroke who may benefit from appropriate prevention treatments. •The 8-item QVSFS could be used to screen for strok