Validity and reliability of postpartum morbidity questionnaires in Benin

We describe the validity and reliability of midwife-administered postpartum questionnaires in home and clinic settings. Women identified prospectively before or following hospital delivery (n = 476), September 2004-January 2005, were re-contacted at 6 months postpartum for home interview and medical examination. Reliability was measured by comparing women's responses to the same questions at home and in clinic interviews. Validity was measured by comparing questionnaire responses with results of medical examination. Reliability of responses to questions comparing home and clinic interview was very good (κ > 0.6) for resumption of menstruation and occurrence of hemorrhoids, moderate (0.4 < κ ≤ 0.6) for weight loss and incontinence, and poor (κ ≤ 0.4) for burning sensation or pain on urination and exhaustion or fatigue. The home and clinic interviews had poor validity for detecting common postpartum morbidities: anaemia (sensitivity 33.7%, specificity 65.7%), urinary incontinence (5.1, 98.1%), urinary tract infection (2.1, 94.5%), prolapse (18.2, 91.2%); but good validity for hemorrhoids (71.4, 86.9%). In this setting, questionnaire-based interviews were neither reliable nor valid tools for measuring morbidity at 6 months postpartum. A medical examination is required to identify and measure the levels of morbidity up to 6 months postpartum.