Establishment of a Maternal Newborn Health Registry in the Belgaum District of Karnataka, India

Registration of pregnancy

The MNHR receives inputs from the register of Married Women of Reproductive Age (MWRA,
formerly known as the ECSR) at the PHCs and sub-centers. The MWRA register contains
information regarding couples eligible to receive contraception services, those who
are infertile and those yet to reach menopause. It also identifies those couples wishing
to, and who are likely to conceive, within the following year. The MWRA is updated
every year by a house-to-house survey (occurring over a 3-months period) conducted
by auxiliary nurse midwives (ANMs), Anganwadi workers (AWWs) and accredited social
health activists (ASHAs). In addition to informing public health needs, the MWRA provides
early identification for pregnancies and informs enrollment in the MNHR. Building
on the MWRA, the data collection for the MNHR began in May 2008 in 20 geographic areas,
each incorporating a PHC with ?300 annual deliveries and covering a total population
of about 660,000, which is also encompassed in the MWRA catchment area. Although one
of the primary sources used to identify pregnant women is the MWRA, registry staff
and other health workers attempt to identify all pregnant women, married or not, in
the catchment area and these women are included in the registry.

Inclusion criteria for the Global Network MNHR

In central and southern India, where Karnataka is located, it is common, particularly
among nulliparous women, to stay with relatives and deliver where those relatives
reside. Similarly, pregnant women may enter the catchment area to deliver at the home
of relatives. Recording movement of pregnant women for delivery was one of the major
problems we had to overcome to approach 100% ascertainment. Thus, pregnant women were
eligible for inclusion in the MNHR if they were residents of the study clusters during
pregnancy, regardless of delivery location. Women were also eligible if they lived
outside the geographic area but delivered within the MNHR catchment area. All identified
women who met criteria were included in the registry after they had provided written
informed consent.

Enrollment and follow-up

All women identified as likely to conceive were followed up by monthly visits through
the MWRA. After a missed menstrual period and a pregnancy confirmed by a pregnancy
test, an eligible woman is then invited to be enrolled in the MNHR by ASHA workers.
Once the woman has been enrolled, the MNHR registry administrators (RAs), who are
medical officers of the PHCs, obtain consent and complete data entry forms for the
MNHR. This methodology of screening for pregnancy has also enabled us to document
spontaneous abortions and medical terminations of pregnancy (MTPs) at the community
level. These data had not been previously collected.

Data forms

Three forms are used to collect study data, are available elsewhere 18]. (1) The first, filled out as soon as possible after conception and enrollment, collects
the maternal demographic data and the date of last menstrual period; (2) A perinatal
form, completed at the time of pregnancy loss or delivery, and information relative
to the status of antenatal care, maternal health issues, the mode, time and place
of delivery, characteristics of the fetus/neonate, and any adverse events at or soon
after the delivery; (3) At the 6 week follow-up visit, a third form contains information
about the status of the newborn and the mother through 42 days post-delivery, and
includes information on any major adverse events occurring between birth and the follow
up visit. Protocol deviations, withdrawals from the study, loss to follow-up and time
lag between event, data collection, entry and transmission to the Data Management
Center are also assessed and recorded. The overall administrative structure of the
MNHR is shown in Figure 1.

Figure 1. Administrative organization of the Maternal Newborn Health Registry

Data collection procedures and staffing

After the forms obtained at the three specific times discussed above are compiled
by the RA, the completed forms are then verified and collected by the data entry operators
on fixed days of the week. Data are entered into an electronic database at regional
data centers located at the sub-district level and transmitted to a central data management
center. The quality of data received is regularly monitored by a research coordinator,
who oversees the study, and field research officers, who are generally medical officers
with oversight of field activities, together with RAs and data entry operators at
the regional data centers. Cleaning and editing data and transmission to the Global
Network Data Coordinating Center (DCC), located at a central location in the U.S.
are ongoing activities. Combined data are processed and analyzed at the DCC for data
consistencies and completeness. The DCC sends monthly monitoring reports which are
discussed by the study team (i.e., the data manager, RA’s and field supervisors) and
serve as process and outcome indicators. Inconsistencies and unexpected findings or
trends, when they occur, are verified and corrective steps initiated during monthly
meetings conducted at the study site headquarters in Belgaum.

Quality Assurance and Completion of Data Collection Instruments

Data quality was monitored in a number of ways. The monitoring metrics included: 1)
the number of days taken to complete the forms and record of all relevant events,
2) the proportion of women consented for inclusion in the study from the total of
pregnancy women screened who were eligible, 3) the rates of MTP, miscarriages, stillbirths
and live births, as well as 4) the proportion of follow-up through 42 days postpartum.
The monthly monitoring reports contain these elements which allow the team to further
check and validate the results.

An important feature of the quality measurement is having an estimate of the number
of births that should occur in each cluster. We use an estimate of the crude birth
rate (CBR) within the study clusters as an indicator of completeness of registration
of pregnancy outcomes and as a check to ensure complete capture of all births. The
CBR was computed from the live births and the estimated population of the cluster.
Significant differences of the number of births registered from those expected are
investigated.

The time taken in completion after the event of interest and its entry into the data
management system has also been considered as a quality indicator, along with documentation
of the extent of birth weights recorded and the presence of a healthcare provider
at the time of delivery.