Study Finds Taiwan’s 2003 SARS Outbreak Led Pregnant Women to Skip Prenatal Visits, Worsening Birth Outcomes

Researchers are drawing new lessons about the value of prenatal care by revisiting an unusual moment in recent public health history: Taiwan’s 2003 outbreak of severe acute respiratory syndrome (SARS). In a new analysis, economists Chien-Tzu Cheng, Shin-Yi Chou, and Hsien-Ming Lien treat the epidemic as a “natural experiment” that abruptly altered health-care seeking behavior—particularly among pregnant patients—because of widespread fear of infection.

During the SARS episode, prenatal visits in Taiwan fell sharply. The authors quantify that drop at more than 20%, attributing it to concerns about exposure in medical settings rather than to changes in underlying medical need. That fear-driven reduction creates a useful research tool: gestational exposure to SARS-related anxiety can be used as an “instrument” to isolate changes in prenatal care that are plausibly unrelated to a mother’s baseline health risk, allowing the team to estimate the causal effects of receiving adequate care.

Using administrative health data from Taiwan’s National Health Insurance system, the study finds that adequate prenatal care substantially improves maternal health, but not in the way many conventional comparisons suggest. Specifically, sufficient prenatal care lowers the risk of preventable maternal complications by 3.4 percentage points—an estimated 26% reduction relative to baseline. By contrast, the researchers do not detect a measurable impact on infant outcomes in their data.

A key contribution of the paper is showing how standard statistical approaches can mislead. When the authors apply ordinary least squares (OLS)—a common method that compares outcomes between people who did and did not receive more prenatal care—the estimated relationship for maternal complications points in the wrong direction. Meanwhile, OLS suggests overly large benefits for infants. The authors interpret this pattern as evidence of selection bias operating in opposite ways: both patients and providers may influence who receives more care, but not necessarily for the same reasons. Higher-risk pregnancies may generate more visits (making care look less effective for mothers), while other dynamics in the clinical system may make infants appear to benefit more than they truly do in simple comparisons.

To understand why prenatal visits matter for mothers, the researchers look beyond visit counts and toward what those encounters produce: information. Their mechanism evidence supports the idea that prenatal care functions as a longitudinal monitoring and information system. Repeated contact with clinicians helps detect risks that emerge or evolve over time, builds a richer clinical record, and provides signals that shape decisions at delivery. Those decisions include how patients are “routed” into different delivery modes and how clinicians manage care during childbirth.

The project also notes the institutional scaffolding behind the work. Cheng reports support from a PSTC International Fellowship at Brown University’s Population Studies and Training Center, funded in part by the U.S. National Institutes of Health (training grant T32 HD007338 and center grant P2C HD041020), as well as a MEXT/JSPS KAKENHI award (Grant 25K16665). Lien reports funding from Taiwan’s National Science and Technology Council (Grant 113-2410-H-004-107-MY2). The authors thank Taiwan’s National Health Insurance Agency for providing administrative data, and they note that the views expressed are their own rather than those of the National Bureau of Economic Research.

Why it matters

Prenatal care is often justified by its presumed benefits for babies, but this study suggests one of its clearest payoffs may be preventing avoidable harm to mothers. The findings also warn that simple comparisons can distort the apparent value of care when higher-risk patients are more likely to seek or be referred for more visits. As health systems plan for future outbreaks or crises that suppress routine care, the results highlight the importance of maintaining continuous prenatal monitoring—and of measuring its effects with methods that can separate causation from selection.