Self-reported abortion-related morbidity: a comparison of measures in Madhya Pradesh, India.

CONTEXT Globally, abortion-related morbidity places a largely unmeasured physical, social and economic burden on women, yet little research has examined population-level morbidity, particularly in developing countries. METHODS Data were drawn from a representative 2002 survey in Madhya Pradesh, India, of married women aged 15-39 who had at least one child; the analysis examined 966 abortion attempts associated with 737 pregnancies. The prevalence and severity of self-reported morbidity were determined using a symptom-based measure and a bed rest-based measure. Comparative and concurrency analyses assessed how the measures captured and categorized morbidity, and how their results differed between rural and urban settings and by provider and method used. RESULTS Both measures found relatively high levels of abortion-related morbidity: The symptoms measure recorded morbidity in 58% and 46% of abortion attempts in rural and urban areas, respectively, and the bed-rest measure did so in 38% and 29% of attempts in these areas. With either measure, the proportion of attempts resulting in severe morbidity was higher for rural than for urban women. A substantial proportion of attempts were categorized as resulting in moderate or severe morbidity under the symptoms measure and in no morbidity under the bed-rest measure (16-20%), and a significant proportion were categorized as leading to severe morbidity under the first measure and to no or moderate morbidity under the second measure (6-17%). Differences also appeared in how the measures assessed morbidity according to the provider and method used. CONCLUSIONS Each measure captured a distinct dimension of abortion-related morbidity, which suggests that multiple measures should be used to more fully assess the burden of morbidity among women in developing countries.