Abstract
BACKGROUND: The modified Rankin Scale (mRS) is widely used to assess functional status after stroke, yet prior work in Zambia has shown low agreement between clinician assessments and patient-reported scores. OBJECTIVE: To identify cultural, linguistic, and other factors contributing to discrepancies between clinician-determined and self-reported mRS scores among adults with stroke in Zambia. METHODS: Consecutive adults hospitalized with stroke were recruited into a mixed-methods cross-sectional study. A neurologist assigned an in-person mRS score. Within seven days, a verbal mRS questionnaire was administered in the participant's preferred language (English, Nyanja, or Bemba). Semi-structured interviews explored how respondents interpreted and justified their answers. Transcripts were analyzed for themes related to comprehension, health perceptions, and observed behavior. RESULTS: Among 51 participants, in-person and self-reported mRS scores matched in 51% of cases (weighted kappa = 0.548). Self-reported scores were 1 point higher in 19% and 1 point lower in 11% of cases; 8% reported scores 2 points higher. Agreement for dichotomized outcomes (good <2; poor >3) was moderate (k = 0.60). Discrepancies arose from translation challenges, caregiver responses based on incomplete knowledge, patients' fear of attempting tasks such as sitting or walking, underestimation of non-motor deficits, and mismatches between patient and clinician interpretations of mRS criteria. CONCLUSION: Clinician and patient-reported mRS scores aligned for only half of participants. Improving translations and better understanding how patients perceive their abilities may enhance interpretation of both clinician-assigned and self-reported mRS scores in this setting.