Skip to content. | Skip to navigation

Personal tools

Maternal mortality ratio (MMR)

Filed under:

The number of maternal deaths per 100,000 live birth.

A maternal death (as cited in International Classification of Disease or ICD-10, [WHO, 1992]) is the death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the duration and the site of the pregnancy, and can stem from any cause related to or aggravated by the pregnancy or its management, but not from accidental or incidental causes.  

Late maternal deaths refer to deaths caused by direct or indirect obstetric causes more than 42 days but less than one year after the termination of pregnancy.

Pregnancy-related deaths are deaths while pregnant or within 42 days of the termination of pregnancy, irrespective of the cause.

Maternal deaths fall into two groups, direct and indirect, as follows:

Direct obstetric deaths

Direct obstetric deaths result from obstetric complications of the pregnant state (pregnancy, labor, and puerperium), from interventions, omissions, incorrect treatment, or from a chain of events resulting from any of the above.

Indirect obstetric deaths

Indirect obstetric deaths result from previous existing disease or disease that developed during pregnancy and which was not due to direct obstetric causes, but which was aggravated by physiologic effects of pregnancy.

The maternal mortality ratio is calculated as:

All maternal deaths occurring within a reference period (usually 1 year) x100,000
______________________________________________________
Total # of live births occurring within the reference period

Information on all maternal deaths occurring in a pe­riod (usually 1 year) and information on the total num­ber of live births occurring in the same year.

Including all pregnancies in the denominator gives a true indication of the total population of pregnant and delivering women at risk of maternal death, but researchers and evaluators more commonly use live births since these data are more readily available and are easier to collect.

Where data on the numbers of live births are absent, evaluators can calculate total estimated live births using census data for the total population and crude birth rates in a specified area. Total expected births = population x crude birth rate.


Many countries have three main sources of data with which to calculate the maternal mortality ratio:

  • Vital registration;

  • Service statistics; and

  • Population based surveys or surveillance.

 The serious limitations of these sources, in both the developing and developed world, have been well documented elsewhere. (AbouZahr, 1999; Berg, Danel, and Mora, 1996; Campbell and Graham, 1990).


The maternal mortality ratio is the most widely used measure of maternal deaths. It measures obstetric risk (i.e., the risk of dying once a woman is pregnant). It therefore omits the risk of being pregnant (i.e., fertility, in a population, which is measured by the maternal mortality rate or the lifetime risk) (Graham and Airey, 1987).

Maternal mortality is widely acknowledged as a general indicator of the overall health of a population, of the status of women in society, and of the functioning of the health system. High maternal mortality ratios are thus markers of wider problems of health status, gender inequalities, and health services in a country. The maternal mortality ratio is therefore useful for advocacy purposes, but lacks information on the causes of high maternal mortality or the interventions required to reduce maternal deaths.


Maternal deaths are difficult to investigate because of their comparative rarity on a population basis, as well other context-specific factors, such as reluctance to report abortion-related deaths, problems of memory recall, or lack of medical attribution. Thus, no single source or data collection method is adequate for investigating all aspects of maternal mortality in all settings (WHO, 2006).

Few developing countries have vital registration systems sufficiently complete to provide reliable population estimates (AbouZahr, 1998).

The main drawback of health services data relates to the selectivity of the service-using population. Without detailed knowledge of the catchment population, it is hard to gauge whether the maternal mortality ratio under or over estimates the level for the general population (which also includes non-users of the service). Other problems related to using health services information include inaccuracies in routine registers and omission of deaths occurring outside maternity wards.

Population based surveys are the primary source of information for calculating the maternal mortality ratio in many developing countries. These types of surveys include:

  • RAMOS (Reproductive Age Mortality Surveys) studies seek to identify all female deaths in the reproductive period, using a combination of approaches, such as cross-sectional household surveys, continuous population surveillance, hospital and health center records, and key informants (WHO, 1987).
  • Direct Estimation relies on asking questions about maternal deaths in a household during a recent interval of time, say one to two years. These questions can be asked in the context of a household survey or a census of all households, although as yet experience with the latter is fairly limited (Campbell, 1999).

Both these types of methods provide up-to-date estimates but are time-consuming and costly because they require large sample sizes to obtain single-point estimates with sufficiently narrow confidence intervals to enable monitoring of time trends.

  • The sisterhood method goes some way to overcoming large sample size requirements by interviewing adult respondents about the survival of all their sisters. The indirect method (Graham, Brass, and Snow, 1989) involves fewer questions to respondents but provides a pooled estimate that relates statistically to a point around 10-12 years prior to the survey. The direct method (Stanton, Abderrahim, and Hill, 2000) provides a more current estimate at about 3-4 years prior to the survey, but requires more questions and is more costly and time consuming.

Because of the imprecision in these estimates, modeling methods have also been developed, (WHO, UNICEF, and UNFPA, 2001; AbouZahr and Wardlaw, 2001; UNFPA, 1998b).

Maternal mortality ratios are only a broad indication of the level of maternal mortality, rather than a precise measure, because of the limitations inherent in most measurement methods. The use of confidence intervals around the estimates helps raise awareness that a point estimate is usually too imprecise to be used to monitor trends (AbouZahr and Wardlaw, 2001).  Furthermore, the data sources and collection methods described above have very different strengths and weaknesses and yield estimates of varying reliability. For this reason, surveys to estimate maternal mortality should occur no more frequently than every 5-10 years. Evaluators must take into account the large confidence intervals in interpreting the maternal mortality ratio.

Distinguishing between real and artificial changes in the maternal mortality ratio is complicated because observed differences do not necessarily indicate improved maternal health status (Graham, Filippi, and Ronsmans, 1996). Other important issues to consider include: 

  • Non-sampling errors such as changes in the accuracy of reporting or of classification over time or between districts or populations (Stanton, Abderrahim, and Hill 2000);
  • Changes in the definition of a maternal death between ICD-9 and -10 (WHO, 1977; WHO, 1992). Presentation of the maternal mortality ratio should thus clearly state which version it used. In the case of ICD-10, one must specify which of the three categories (direct and indirect maternal deaths up to 42 days postpartum, late maternal deaths, pregnancy-related deaths) the numerator includes;
  • Aggregate levels may hide wide differentials between population subgroups; and
  • Apparent differences in the maternal mortality ratio between rural and urban areas may simply reflect differences in the pattern (not level) of fertility, with more rural women who are grand multiparous and for whom the risk of death will likely be higher. Other possible confounders include general health status, such as levels of anemia or malaria, and socio-economic factors.

safe motherhood (SM)

Reproductive Health Indicators: Guidelines for their generation, interpretation and analysis for global monitoring.  WHO, 2006.

Filed under:
Navigation