Can trained field community workers identify stroke using a stroke symptom questionnaire as well as neurologists? Adaptation and validation of a community worker administered stroke symptom questionnaire in a peri-urban Pakistani community.

BACKGROUND Stroke is a leading cause of morbidity and mortality worldwide. There is a paucity of data from South Asia where stroke is highly prevalent. Validated tools administrable by community health workers (CHWs) are required to identify stroke in the community in a resource-strapped region such as this. METHODS The study was conducted in a transitional slum in Karachi, Pakistan. Questionnaire to Verify Stroke-Free Status (QVSFS) was adapted and translated into Urdu. Two CHWs, trained by a neurologist, selected 322 community-dwelling subjects using purposive sampling. Each CHW collected data independently, which was validated by a vascular neurologist who directly examined each participant. To assess the effect of audit and feedback, data from the final 10% of the subjects were collected after a second training session for the CHWs. Sensitivity, specificity, and Cohen kappa were determined for the CHW-administered questionnaire against neurovascular assessment. RESULTS Mean age of participants was 56.5 years with 71% of participants being women. The sensitivity and specificity of the questionnaire of detecting stroke was 77.1% (confidence interval [CI], 64.1-86.9) and 85.8% (CI, 83.5-87.5), respectively. The chance-corrected agreement using the Cohen kappa statistic was .51 (CI, .38-.60). Kappa ranged from .37 to .58 for each of the 7 stroke symptoms. Hemianesthesia (72.9%) and hemiplegia (64.6%) were the most sensitive symptoms. The performance and agreement improved from moderate to substantial after audit and feedback. CONCLUSIONS We found a reasonable sensitivity and specificity and moderate agreement between CHW-administered QVSFS and assessment by a vascular neurologist.