This is a facility-based, cross-sectional study conducted in six public hospitals in Baghdad between March 1, 2010 and the June 30, 2010. It consists of a near-miss criterion-based clinical audit implemented according the WHO near-miss approach for maternal health [6
]. Selection criteria for the facilities included were 1) to be a public hospital and 2) to have more than one thousand deliveries per year according to the information provided by the health and biostatistics department in the Iraqi Ministry of Health. We included all of the hospitals fitting these criteria in Baghdad: a total of six. These public hospitals were distributed all over Baghdad city to serve a population of approximately seven millions. Three of these hospitals were general hospitals with obstetric units and the rest were major maternity hospitals. Four out of six hospitals had intensive care units (ICUs), although the remaining two hospitals had a "close observation unit" to monitor and treat women with post-operative and post-delivery complications, run by specialized obstetricians. These hospitals receive referrals from midwives, health centers, private hospitals as well as unbooked patients (patients self-referring themselves to the hospitals). No fees are paid by patients for the services provided.
A study coordinator among the hospital coordinators was designated in each of these six facilities and the overall study coordination was performed by the Centre of Training and Human Development in the Iraqi Ministry of Health. These six selected study coordinators were trained on a two-day course in January 2010 on identifying severe life-threatening conditions, maternal near-miss events and deaths, and how to implement data collection.
The WHO consensus on maternal near-miss definition was used to define the cases [8
]. Definitions and abbreviations used in this paper are found in Table . Data were collected on a daily basis by the coordinators using hospital records or staff interviews and the forms were filled while the women were still in the hospital. Cases were defined according to potentially life-threatening conditions including severe postpartum hemorrhage, severe preeclampsia, eclampsia, sepsis and ruptured uterus, whereas organ or system failure depending on certain clinical criteria, laboratory markers and management proxies were used to identify the near-miss cases among potentially life-threatening conditions (Additional file 1
: Table S1). The maternal outcome, gestational age and neonatal outcome were also collected during the hospital stay or by the 7th day postpartum, whichever came first.
Table 1 Maternal near-miss terminology and indicators[8,9]
Using the data collected, various indicators have been calculated. In line with maternal mortality ratio, maternal near-miss ratio was the number of near-miss cases per 1000 live births. Moreover, maternal near-miss mortality ratio, which is the ratio between maternal near-miss cases and maternal deaths, was calculated. For this indicator, higher ratios indicate better care, meaning more women survived as a near miss rather than becoming maternal deaths. Also, mortality index was calculated, where the number of maternal deaths was divided by the number of women with life-threatening conditions (maternal near miss and maternal deaths) and was expressed as a percentage. Higher indices indicate that more women with life-threatening conditions die (low quality of care), whereas lower indices signify better quality of care.
Access to hospital and intra-hospital care were assessed by the proportion of near-miss cases and maternal deaths presenting within 12 hours of hospital stay versus after 12 hours, the latter indicating the quality of care provided within the hospital. Intensive care unit (ICU) use among our study population was collected as well.
We also collected data on the coverage of selected evidence-based interventions used for prevention and treatment of the main causes of maternal deaths. This was part of a criterion-based clinical audit approach used to assess the quality of care and included interventions related to the prevention and treatment of postpartum hemorrhage, severe preeclampsia and eclampsia, use of antibiotics for infection prophylaxis during caesarean section (C-section) and treatment of sepsis. In addition, we measured the use of corticosteroids for fetal lung maturity.
Data were sent monthly to the country coordinator and subsequently entered into an online data entry system. This online data management system was based on the Google platform and data entered in the online form were stored in an online spreadsheet, which incorporated a comprehensive set of consistency rules to provide concurrent data quality check. The inconsistencies that were identified generated queries to the study coordinators. The standard data tables based on the WHO near-miss approach was automatically generated using Microsoft Excel as the data were entered. The online data entry system was password protected.
The study was approved by the ethics committee of the local supervising committee of the Arab Board for Health Specializations (ABHS). All data were obtained from medical records and did not identify participants, therefore each site was granted a waiver of individual informed consent.