In Pakistan, the vital registration system is weak, and population-based data on the maternal mortality ratio are limited. This study was carried out to collect information on maternal deaths from different existing sources during the current year—2007 (prospective) and the past two years—2005 and 2006—(retrospective), identify gaps in information, and critically analyze maternal deaths at the community and health-facility levels in two districts in Pakistan. The verbal autopsy questionnaire was administered to households where a maternal death had occurred. No single source had complete data on maternal deaths. Risk factors identified among 128 deceased women were low socioeconomic status, illiteracy, low-earning jobs, parity, and bad obstetric history. These were similar to the findings of earlier studies. Half of the women did seek antenatal care, 34% having made more than four visits. Of the 104 women who died during or after delivery, 38% had delivered in a private facility and 18% in a government facility. The quality of services in both private and public sectors was inadequate. Sixty-nine percent of deaths occurred in the postpartum period, and 51% took place within 24 hours of delivery. The study identified gaps in reporting of maternal deaths and also provided profile of the dead women and the causes of death.

Objective: To develop and validate a screening instrument for neurological disorders in children aged 6 months to 2 years in the community. Methods: A comprehensive parent-administered instrument was developed to screen for hearing, vision, seizures, motor deficits and development in Indian children aged 6–24 months. This was tested for reliability and validated in the hospital setting by comparing with pre-decided gold standards. It was then used in a community survey in a two-phase design in which all screen positives and a random sample of screen negatives were validated. Result: The screening instrument had overall sensitivity, specificity and positive predictive values of 95.8, 68.1 and 76.1%, respectively, in the hospital setting. In the field setting, these figures changed to 95.4, 51.8 and 20.6%, respectively. The reasons for this are discussed. Conclusion: Community surveys must use a two-phase design to get the true prevalence. A falsely high prevalence will be computed if only a single-phase design or hospital validation is used.

No abstract