Does a Better Match Between Pregnancy Condition and Level of Hospital Care Can Save Lives?

Maternal deaths are a pervasive problem in many countries. In 2015, over 300,000 women and adolescent girls died from pregnancy and childbirth complications (WHO, 2016). While maternal mortality is higher in resource-limited contexts, the situation is a growing problem worldwide, including developed countries. Some of the countries in the Organization for Economic Cooperation and Development (OECD), like the United States (US) and Mexico, experience relatively high levels of maternal mortality: the US had 14 maternal deaths per 100,000 live births and Mexico had 41 in 2015. This number situated Mexico at the top of the list among OECD countries. In fact, Mexico City, representing the country’s best health care system, had 61.7 maternal deaths per 100,000 live births. In order to respond to this dire situation, governments and international organizations look for interventions to address basic health needs of pregnant women. The World Health Organization (WHO) suggests that most maternal deaths could be prevented through quality care provided for pregnant women during pregnancy, childbirth and the postnatal period.

Despite the benefits of prenatal care, some regions still face low levels of its utilization. For example, in 2016, only 17% of women in Mexico City who gave birth initiated prenatal care in the first trimester and had an average of 2.5 visits (DGIS, 2017). This issue has called for innovative approaches to increase prenatal care utilization and reduce high maternal mortality rates. The Doctor at your Home (DH) program is an effort of the Mexico City government to address low levels of prenatal care utilization and improve maternal health.

Through the DH, social workers visit households looking for pregnant women without prenatal care and refer them to a health center for prenatal care consults. Whenever women do not attend to the health center, a physician visits her to evaluate the overall health of the pregnancy.

This paper evaluates the impact of this intervention on maternal mortality rates and explores a possible mechanism through which the program prevents maternal deaths: a better match between mothers’ pregnancy condition and level of hospital care. To do this, I will utilize the program’s roll out. In particular, the empirical approach uses variation in the time the program began operating in each Mexico City county to estimate a differences-in-differences model, comparing outcomes in counties implementing DH in a particular year to counties that did not, before and after DH started in a given county. I will estimate these impacts first for the entire population and then separately by age, geography and income group to assess whether this program closed disparities among these groups.

The empirical analysis uses a comprehensive dataset that links the information collected by DH with patients’ clinical records and with birth and death certificates. The data will include all births in Mexico City over 2008-2017 (around 1.3 million births). The data is supplemented with Mexico City hospitals’ discharge records for the same period.