Reading is power: How Literacy is Changing Maternal Mortality Rates

By: Ashley Holst

American University - SIS

International Development Master's Candidate

Maternal health was one of the Millennium Development Goals planned to be completed in 2015. The two indicators are maternal mortality ratio and the proportion of births attended by a skilled health personnel. The maternal mortality ratio target was to “reduce by three quarters… the maternal mortality ratio.” In thinking about maternal mortality rate, it occurred to me that a woman is often not given the space or the skills necessary to make a medical decision about her labor plan. Many steps go into the process of receiving new information about best practices for pregnancy and labor, let alone reading and understanding pamphlets or medicine bottles. These thoughts led me to the question: does the ability of a woman to read (and understand suggestions for pregnancy care and labor practices) lead to a lower maternal mortality rate?



http://www.un.org/millenniumgoals/maternal.shtml

The presence of a skilled doctor or midwife is certainly helpful in case of difficult deliveries and can lower the number of very preventable deaths of infection and blood loss, but literacy would come many years prior to a girl even imagines herself a mother, long before delivery that have an effect on maternal well-being. There are several additional variables that have an effect on a woman’s access to health care including availability of services, funds to pay for health care, and basic understanding of the importance of pre-natal care. 

In addition to ability to understand the information for her own well-being, literacy may be a way to understand the amount of value that is placed on a woman. If she is “worth” educating, she may also be “worth” spending money on in case of a labor emergency. In which case, literacy may also be a predictor as to whether a husband or family may decide to spend money on prenatal or labor care. There could be some concern that high MMR may lead to people being less likely to invest in the education of women. It is not logical that a family would make this type of decision based on the probability that she may die 20 years later in child birth. I think it is much more likely that MMR will be lower in a country where women are valued prior to childbirth and where she is educated enough to understand prenatal care and make an educated decision about her labor.

For this analysis I chose to focus on the most recent year with data available during the MDG period (1990-2015). In thinking about the future of the MDGs, we must look at what is happening now, and what improvements have been seen, to identify a possible direction for future policy and development programming. In this case, 2010 was the year with most complete data available for maternal mortality rate (MMR) and female literacy rate. I compiled my data from the World Bank Indicators. MMR is measured as deaths per 100,000 births. MMR ranges from 2 – 1200 deaths (per 100,000 births) and the mean MMR was 182.5 in 2010. On first inspection of the variables, it appeared that MMR may be skewed, but when I compared the relationship with literacy. The correlation of literacy with both the logged and not logged variable were statistically significant. This showed me that logging MMR would not make my analysis more precise and using the unlogged variable would give me a more realistic result.

Female Literacy Rate, reported as a percentage of women over 15 who are literate, proved to be more difficult to obtain. To utilize more cases, I chose to compile years 2009-2011. Few countries had more than one measurement in this period and literacy rates tend to be slow changing, so I felt this was an appropriate method to increase the strength of my analysis. If a country had more than one measurement I found the mean of the years. The compiled data ranged from 12.2% - 99.9% of the female population. The mean female literacy rate was 81%. It should be noted that the United States was not included in the literacy data set, though it is recorded at 99% by the CIA World Factbook. Figure B shows the uncontrolled relationship between Female Literacy Rate and MMR. Uncontrolled, on average a 1 percentage point increase in Female Literacy Rate is associated with a 5% decrease in MMR. This relationship is statistically significant (p<.001).

Table 1. Summary Statistics

	Mean	Std. Dev.	Min.	Max.	Description
Maternal Mortality Rate 2010	176.96	213.39	2.0	960.0	Maternal deaths per 100,000 births
Female Literacy 09-11	80.17	23.00	12.2	99.9	% of the total female population over 15 who are literate
GDP per Capita (US$)	7955.93	11404.23	359.6	71510.2	GDP per Capita in US$
Female Life Expectancy at Birth	71.44	10.25	45.0	84.7	Average life expectancy of a female calculated at birth
Observations	72

I included two additional variables as controls. First, GDP per capita because it is widely agreed that GDP per capita has an effect on both literacy rates and MMR so it was important to include it in the model as a confounding variable [1]. I predict that a higher GDP per capita would lead to a higher literacy rate and lower MMR. I had originally planned to use a third variable, Health Expenditure per Capita. I identified a multicollinearity problem specifically between this variable and GDP per capita. After an analysis of variance inflation model, I chose to eliminate Health Expenditure from the model. While it does not tell me how much on average people spend on health care, GDP gives me a better understanding of wealth. Health Expenditure per capita was also a faulty measurement because it does not account for differences in policy, for instance Health Expenditure in the US is much higher than many European countries. Second, I included Female Life Expectancy. I utilized this variable for one major reason, to identify overall health of women. Countries with a short life expectancy may be have a higher MMR simply due to lower overall health and not due to other cultural or political factors.  I expect that life expectancy and MMR are negatively correlated. This variable is troublesome as a high MMR would lower life expectancy of women, but in this case it is the best proxy variable for overall health and allows me to control for unmeasured health confounds. The results of the multiple regression analysis of MMR and Female Literacy Rate controlling for GDP per Capita and Life Expectancy is shown below in Model 3 of Table 2.

Table 2. Effects of Female Literacy Rate on Maternal Mortality Rate 

DV: Maternal Mortality Rate	(1)	(2)	(3)
Female Literacy	-7.585***		-3.403***
	(-11.228)		(-4.828)
GDP per Capita (log)		-64.177***	-40.138**
		(-6.879)	(-3.241)
Female Life Expectancy at Birth		-11.452***	-9.429***
		(-8.473)	(-7.294)
Constant	789.583***	1543.677***	1455.506***
	(13.963)	(20.225)	(17.871)
Adj. R-Squared	0.63	0.66	0.83
Observations	76	176	75
NOTE: OLS estimates with t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001 Source: World Bank Development Indicators 2010

When the relationships of these variables were analyzed with MMR the results were very interesting. I find that Female Literacy was in fact significantly correlated with MMR in both the bivariate (uncontrolled) and controlled models. GDP per capita and Life Expectancy are also both significantly correlated with MMR. When controlling for the other variables in the model, a 1% point increase in female literacy results in a decrease of 3.4 maternal deaths per 100,000 births. This is an important finding showing that by increasing literacy in a country by a mere 5% could save the lives of almost 20 mothers. GDP per capita and life expectancy also showed positive implications. On average, controlling for GDP and literacy constant a 1 year increase in life expectancy is associated with a decrease of 9.4 deaths per 100,000 births. Controlling for the other variables, a 10% increase in GDP per capita is associated with a decrease of 4 deaths per 100,000 births. Both relationships are statistically significant (p<.05).

According to my findings working to increase the quality of life of women with improve MMR, the 5^th Millennium Development Goal. My results show that a combination of health, economic, and educational development will improve the chance that a women will have a successful pregnancy and labor. As each variable is measured in a different way, it is impossible to say which element of development is most effective, but the results show that any improvement in these three categories will result in lives saved.

Organizations or programs hoping to reduce MMR by offering prenatal supplies, pregnancy or labor educational workshops, or educating women about how to identify signs of a dangerous labor need to consider literacy in their project design. If a women is unable to read and comprehend the information, it will be useless information when the educators are no longer around. Helping to improve literacy rates may also allow a woman to be more aware of and confident in family planning methods which could reduce young, old, and multiple pregnancies (both of which have higher risk of maternal mortality).

Notes:

[1] GDP per capita was positively skewed so the analysis reflects logged values.

Works Cited:

Millenium Development Goals and Beyond 2015. http://www.un.org/millenniumgoals/maternal.shtml

The World Factbook. https://www.cia.gov/library/publications/the-world-factbook/fields/2103.html

World Development Indictors. http://data.worldbank.org/data-catalog/world-development-indicators