Extreme Heat and Maternal and Child Hospitalization

With rising concerns of global warming and climate change, growing literature examines the effect of temperature on various economic outcomes. We examine the relationship between exposure to extreme heat during pregnancy and maternal and child hospitalization. We use the universe of inpatient discharge records from four states (AZ, MD, NY, WA) in the United States and focus on hospitalizations surrounding childbirths.

We make three contributions to the literature on extreme temperature. First, we use relative temperature to define extreme heat. Specifically, we focus on temperature three standard deviations above the historical mean for each county-month. Our measure of extreme heat improves upon previous studies that predominantly use absolute temperature, by accounting for regional differences in the degree of adaptation to heat. Second, we examine long-term effects of extreme temperature. To the best of our knowledge, our analysis provides the first estimates of the impact of extreme temperature on future hospitalizations. Our ability to examine the impact of in-utero weather on future child hospitalizations (beyond birth outcomes) allows us to examine potential mechanisms through which in-utero temperature may affect children’s future economic outcomes, as found in previous papers. Third, we identify mothers and track their hospital utilization during their pregnancy, which helps us further understand identifying mechanisms through which in-utero temperature may affect child’s health at birth as well as future outcomes. Most papers in the literature of in-utero shocks are unable to look at mothers during pregnancy due to data limitations.

We calculate the number of days each pregnancy is exposed to different levels of temperature deviations, and we estimate the relationship between these temperature deviations and hospitalizations (in addition to birth outcomes) within a county-month-sex-race cell, while flexibly controlling for time trends. We find that exposure to extreme heat in the second trimester increases maternal hospitalizations during pregnancy due to pregnancy complications. We find that the child is then more likely to have dehydration at birth, and consequently they are more likely to be re-hospitalized for a variety of causes, such as jaundice and respiratory conditions. This is consistent with the medical literature that shows that dehydration is associated with a variety of illness.

We show that our measure of extreme heat is not correlated with observable characteristics such as sex and race, suggesting that our results are not driven by differential exposure to extreme heat along these dimensions. Our results are robust to including temperature deviations from 2 years ago, suggesting that our results are also not driven by differential time trends across counties or differential fertility based on previous temperature shocks. In addition, our results are robust to controlling for contemporaneous temperature, and they are not sensitive to various regression specifications, such as different functional forms of time trends.