RESUMEN
BACKGROUND: African women face high rates of depression, particularly during pregnancy or postpartum or after a recent HIV diagnosis. The Patient Health Questionnaire-9 (PHQ-9) depression screening tool has been quantitatively validated and extensively used to identify depression and link individuals to care. However, qualitative work is necessary to identify important opportunities to improve PHQ-9 question comprehension and performance among Kenyan women. METHODS: We administered the Kiswahili or English PHQ-9 (based on preference) to 29 pregnant and postpartum women in Thika, Kenya. Following administration, we conducted cognitive interviews with a purposive sample of 20 women. We used analytic memos and data matrices to identify themes around scale acceptability, comprehension, and decision and response processes. RESULTS: Most participants preferred to answer the PHQ-9 in Kiswahili (N = 15; 52%). Among the 20 interview participants, 12 (60%) had scores ≥5, indicating depressive symptoms. Overall, participants found the scale acceptable as an interviewer-administered tool. Participants reported few problems related to comprehension but had difficulty answering items not relevant to their lives (e.g., "watching television") and double-barreled items (e.g., "poor appetite or overeating"). They were hesitant to endorse items related to "duties as a wife and mother" and suicidal ideation. Most participants had difficulty distinguishing between response options of "several days" and "more than half the days". CONCLUSIONS: We detected several problems related to PHQ-9 comprehension, decision processes, and response processes. We provide recommended changes to instructions and item wording to improve PHQ-9 validity among Kenyan women.