To compare wealth status and under-5 child mortality within a country, Bendavid used data from the demographic and health surveys for 1.2 million women living in 929,224 households in 54 developing countries. The women provided information about their children’s survival status. 

 “The people who conduct these surveys, they’re intrepid surveyors,” said Bendavid, who is also a core faculty member of Stanford Health Policy, which is part of the Freeman Spogli Institute for International Studies. “They reach remote villages up the Congo basin and in the Sahel in Niger, and track the heads of households and women for these in-depth interviews.”

The surveys include information about each woman’s birth histories, including detailed birth registries documenting millions of children. With this information, Bendavid could estimate the probability of a child dying before reaching age 5 per 1,000 live births.

Tallying household possessions

Determining each household’s wealth status was not as straightforward as reviewing annual income and tax returns, which don’t exist in the countries involved in the study. “These surveys tally the possessions in the household. What is the floor made of? What is the roof made of?” Bendavid said. “You can get a wide distribution of household possessions that reflects to a large degree the household wealth.”

Next, Bendavid developed a three-tier wealth index using the household assets. The three wealth categories were relative — poorest, middle and wealthiest.

To analyze trends in wealth status and under-5 mortality, Bendavid looked at all developing countries that had completed the surveys in two specific time frames: 2002-07 and 2008-12. The study found that the under-5 mortality rates among the poorest groups had decreased the most rapidly. The average decline was 4.36 deaths each year per 1,000 live births among the poorest, 3.36 among the middle and 2.06 among the wealthiest. Because the poorest group’s mortality rate is decreasing more quickly that the other groups, the gap in child-mortality rates is closing.

This is good news, Bendavid said. However, not all countries followed this same trend. In a quarter of the surveys examined by the study, inequality in under-5 mortality increased over time.

Bendavid found that four factors were present in countries with a narrowing child-mortality gap: government effectiveness, rule of law, control of corruption and regulatory quality. He found that the difference in mortality rates was significantly associated with the governance score: Better governance scores were related to greater convergence in mortality rates among the three wealth groups.

Benefits from controlling communicable diseases

Bendavid said the evidence in this study is consistent with gains in controlling communicable diseases, such as malaria, measles, diarrhea and respiratory illnesses, that preferentially affect the poorest. Over the past decade, international health aid organizations have financed interventions for these diseases at a high rate.

It makes a persuasive case that these improvements have often begun to benefit the poor even more than the better-off.

“Dr. Bendavid’s study is an important contribution to knowledge about child health improvements in the developing world,” said Davidson Gwatkin, a senior fellow at the Results for Development Institute and a senior associate at Johns Hopkins Bloomberg School of Public Health. “It makes a persuasive case that these improvements have often begun to benefit the poor even more than the better-off.” Gwatkin was not involved in the study.

The study also raises questions about the role of foreign aid institutions in low- and middle-income countries. While the aid efforts are making a difference in child-mortality rates in countries with effective governments, the study seems to show that this is not the case in nations with poor governance, Bendavid said.

“We have the technologies, we have the means, we have the know-how to reduce child mortality dramatically,” said Bendavid. “Even for such low-hanging fruit, however, implementation is not always easy. You have to have government that enables basic safety, and the ability to reach poor and rural communities that benefit from these kinds of programs.”

This work was supported by the National Institute of Allergy and Infectious Diseases (grant KOIAI084582), the Doris Duke Charitable Foundation and the Dr. George Rosenkranz Prize for Health Care Research in Developing Countries.

 Information about Stanford’s Department of Medicine, which also supported this research, is available at http://medicine.stanford.edu.

Doctors want to prevent prematurity because of its serious consequences. Premature babies often need long hospitalizations after birth. They are more likely than full-term babies to die in infancy. Many of those who survive face lasting developmental delays or long-term impairments to their eyesight, hearing, breathing or digestive function.

Phibbs’ team analyzed all deliveries covered by the Veterans Health Administration from 2000 to 2012, a total of 16,344 births. They found that 3,049 infants were born to women with PTSD diagnoses. Of these, 1,921 births were to women with “active” PTSD, meaning the condition was diagnosed in the year prior to giving birth, a time frame that the researchers thought could plausibly affect pregnancy.

The researchers examined the effects of several possible confounding factors. Being older, being African-American or carrying twins all increased the risk of giving birth prematurely, as extensive prior research has shown.

The researchers also looked at the effects of maternal health problems (high blood pressure, diabetes and asthma); possible sources of trauma (deployment and military sexual trauma); mental health disorders other than PTSD; drug or alcohol abuse; and tobacco dependence. However, these factors had little influence on risk for premature birth.

The effect of stress

In other words, although pregnant women with PTSD may have other health problems or behave in risky ways, it’s the PTSD that counts for triggering labor early.

“The mechanism is biologic,” Phibbs said. “Stress is setting off biologic pathways that are inducing preterm labor. It’s not the other psychiatric conditions or risky behaviors that are driving it.”

Stress is setting off biologic pathways that are inducing preterm labor.

However, if a woman had been diagnosed with PTSD in the past but had not experienced the disorder in the year before giving birth, her risk of delivering early was no higher than it was for women without PTSD. “This makes us hopeful that if you treat a mom who has active PTSD early in her pregnancy, her stress level could be reduced, and the risk of giving birth prematurely might go down,” said Phibbs, adding that the idea needs to be tested.

Although PTSD is more common in military veterans than the general population, a fairly substantial number of civilian women also experience PTSD, Phibbs noted. “It’s not unique to the VA or to combat,” he said, noting that half of the women in the study who had PTSD had never been deployed to a combat zone. “This is relevant to all of obstetrics.”

The VA has already incorporated the study’s findings into care for pregnant women by instructing each VA medical center to treat pregnancies among women with recent PTSD as high-risk. And Phibbs’ team is now investigating whether PTSD may also contribute to the risk of the mother or baby being diagnosed with a condition that causes doctors to recommend early delivery for health reasons.

The lead author of the study is Jonathan Shaw, MD, instructor in medicine at Stanford. The other co-authors are Steven Asch, MD, professor of medicine at Stanford and chief of health services research for the VA Palo Alto Health Care System; Rachel Kimerling, PhD, psychologist at VAPAHCS; Susan Frayne, MD, professor of medicine at Stanford and staff physician at VAPAHCS; and Kate Shaw, MD, clinical assistant professor of obstetrics and gynecology at Stanford.

The research was supported by the VA Office of Academic Affairs and Health Services Research & Development and by VA Women’s Health Services.

Stanford’s Department of Pediatrics also supported this research.