Contact FutureLearn for Support
Skip main navigation
We use cookies to give you a better experience, if that’s ok you can close this message and carry on browsing. For more info read our cookies policy.
We use cookies to give you a better experience. Carry on browsing if you're happy with this, or read our cookies policy for more information.
Clinic Health Records, Ethiopia
Clinic Health Records, Ethiopia

Developing better data collection and research

The Lancet Maternal Health Series papers by Wendy Graham et al, Oona Campbell et al, Dorothy Shaw et al, and Marge Koblinsky et al, highlight the limited data on maternal health outcomes and on the inputs to, and processes of, the health system. They also touch on what future opportunities there are for progress in collecting such data.

What are the main limitations to the data collected on maternal health?

There are many gaps in the data collected on maternal mortality and morbidity. There are also gaps in data collected on the capabilities of facilities and the care women receive. Both types of information are needed to accelerate progress in maternal health. The limitations of current data on maternal health outcomes and maternal health services fall into four main categories1,2,3:

  1. Data are non-existent
  2. Data are inaccurate or badly measured
  3. Data are old and often incomplete
  4. Data are not disaggregated by key characteristics such as sub-national locality or socio-economic characteristics

Some countries have no data on maternal mortality at all, and rely on modelled estimates instead, and data on morbidity are even scarcer. These data are best obtained via civil and vital registration systems for mortality data, and medical records for morbidity. In addition, many countries do not record what services their facilities can provide to women (e.g. can a health centre provide basic emergency obstetric care (BEmOC)), or what content of care women receive (e.g. what percentage of women delivering in a specific health centre get active management of the third stage of labour (AMSTL)). These data are best collected via routine health management information systems (HMIS) (or periodic surveys to assess health facility service provision) and medical records respectively. All types of data need to be collected at the national, sub-national and health facility levels to ensure we have a better understanding of the profile of maternal health outcomes and of maternity services.

Data that are collected on maternal health can also be inaccurate for a variety of reasons, including the use of incorrect definitions and indicators, or biased collection methods. In the absence of good health facility or medical records, we often resort to surveying women, using population-based surveys such as the Demographic and Health Survey. These surveys are immensely useful, not least because they also tell us about women who do not use health facility services, but they rely on women’s recall and knowledge. For example, when we ask women about where they delivered or who attended their birth, we cannot be sure they are remembering correctly or whether their report is an accurate reflection of the capability of the facility or of the birth attendant’s cadre or skills. Similarly, while it seems reasonable to ask women about the content of antenatal care they received (e.g. whether they were given iron tablets), it seems less valid to ask them about medical procedures during childbirth, as they are less likely to know what was done.

Data which does exist are often old and do not reflect the current context. This may be because surveys are only done periodically, say every five years, or because routine data are not processed and compiled in a timely way.

In addition, data usually have critical gaps, which fail to give the detail needed to fully assess the facilities and providers available to women. Local level data and data disaggregated by characteristics of women are useful for reducing inequity and targeting of resources. We can rarely disaggregate survey data to the district level, where most health services are managed, as surveys are often designed to represent the whole country or the regions, but are too small to represent districts. Routine data can provide local level data, but these can be incomplete data on a micro scale, such as poor data from a particular hospital, or it could be gaps in data on a district or regional level. Routine data often lack information that can be used to assess inequities such as socio-economic status.

What is needed to improve the collection of maternal health data?

Research and information is a key component of the post-2015 maternal health agenda. Two types of data capture are needed to improve maternal health overall6.

  1. Accurate measurement of maternal health outcomes, and reliable disaggregated population data
  2. Improved record keeping of delivered maternity care, and what care is available Such data would enable monitoring of inequalities and sociodemographic factors, which may be useful to inform policy and practice.

In Step 1.6, we discussed the need for standardized definitions and methods of determining and recording causes of death, and categories of illness and illness severity. We need better civil and vital registration systems that accurately and comprehensively document pregnancy outcomes – births, stillbirths, neonatal deaths, and maternal deaths in many low- and middle-income countries. The Maternal Death Surveillance and Response, a global strategy that aims to identify and respond to maternal deaths, is one approach. Advancements can also come from real-time tracking of pregnancies and their outcomes, including for those outside the formal healthcare system.

In their Lancet paper, Koblinsky and colleagues outline some examples of key indicators for measuring burden and ability of health systems to provide quality maternal health care (see Table 1). For example, signal functions assessments for routine maternal and newborn care need to be more widely adopted and collected via public-sector and private-sector facilities (given that these providers conduct many deliveries).4 The list presented is not exhaustive, and does not include important issues such as delays in treatment, timely referrals, use for financial incentives, women’s satisfaction and specific provider skills, but it does use data that can be collected at scale across a range of settings4. In the original article, Koblinsky and colleagues also give examples of how these indicators have been collected and used. A consensus needs to be reached on the measurement and coding of maternal mortality and morbidity, and on which coverage, quality, and timeliness indicators can effectively be tracked at scale for all countries for global monitoring.

Table 1 Examples of Indicators for Measuring Burden & Ability of Health Systems to Provide Quality Maternal Health Care

Alt text for screen readers

What opportunities for progress are there for collecting data on maternal health?

To obtain good quality evidence, we must agree what data should ultimately be collected. The World Health Organization, Maternal Health Taskforce, and the US Agency for International Development, along with flagship Maternal and Child Survival Program reached a consensus on maternal health indicators for global monitoring and individual country reporting. They agreed on 12 indicators, and are further discussing four priority areas for indicator development and testing. A further commitment was also made to a similar process to take place to agree upon indicators to monitor social, political, and economic determinants of maternal health and survival, highlighted in the End Preventable Maternal Mortality strategies.5

Secondly, the implementation of software such as the District Health Information System 2 (DHIS2), an electronic data management platform, to record indicators of facility capability, provision of services and selected health outcomes, should contribute to improving the availability of data, and therefore the accountability of service providers. This will be essential in guiding intervention research and set implementation priorities for coverage of good quality maternity services, as well as maternal mortality and morbidity outcomes, including direct and indirect causes and risk factors. An alternative is to develop and improve health facility service provision assessments, such as the World Health Organization’s Service Availability Readiness Assessments.

A major challenge remains with collecting data on the care individual women receive. Innovations such as m-Health, and the telecommunications revolution have promise to enable good quality, evidence based, maternity care to reach women in some of the most remote parts of the world, while also having good record keeping potential. Advancements such as this should contribute to the acceleration of achieving global equity in maternal health.

Share this article:

This article is from the free online course:

The Lancet Maternal Health Series: Global Research and Evidence

London School of Hygiene & Tropical Medicine