Table 1 Differences and similarities between the 2007 and 2012 national health facility assessments

Sampling

All hospitals and health centers in district capitals; 40 % stratified systematic random sample of lower level health centers and posts; sampling frame restricted to facilities that had attended at least one delivery in the last year n = 450

Census of all facilities that had attended at least one delivery in the last year n = 946

Weighting

Data required probability weighting to account for sampling at lower level health centers and posts and adjusted for non-response

Self-weighting

Data collectors

Range of health professionals

Medical students from the University of Eduardo Mondlane

Questionnaires

Paper draws from 3 of 17 questionnaires: 1) summary of service statistics for each facility, 2) maternal death reviews, 3) inventory of infrastructure, equipment, supplies and drugs for maternal and neonatal health

Paper draws from 2 of 6 questionnaires (similar to those used in 2007–08): 1) summary of service statistics, 2) inventory of infrastructure, equipment, supplies and drugs. These were adapted from AMDD tools [30]. Maternal death reviews were not administered.

Primary sources of data

Information on maternal deaths was collected from maternity, postpartum and gynecology logbooks, operating theatre, discharge and morgue registers, maternal death reviews/audits. Data collectors also asked staff for other facility-specific sources of information on maternal deaths. Information on institutional deliveries was collected from the relevant subset of registers. The use of monthly summaries of events (births or deaths) was discouraged due to concerns for accuracy.

Same as 2007–08

Service statistics reference period and number of events

12 months (November 2006—October 2007). The number of unweighted maternal deaths was 2,199, unweighted institutional deliveries = 312,537.

3 months (August-October 2012). Number of maternal deaths = 459, institutional deliveries = 161,986.

Overall data collection period

November 2007-January 2008

November-December 2012

Mozambique IRB approval and privacy safeguards

National Committee for Bioethics in Health IRB 00002657, Ministry of Health of Mozambique. All patient data were de-identified and analyzed in the aggregate; no names were collected.

Department of Reproductive, Child and Adolescent Health, within the Mozambique Ministry of Health’s National Directorate for the Promotion of Health and Disease Control. Like 2007, no names or identifying characteristics of patients were collected.