When More Prenatal Care Looks Like Less: Selection, Information and Maternal and Infant Health

We use Taiwan's 2003 SARS epidemic as a natural experiment, instrumenting prenatal care with gestational exposure to SARS-induced fear that reduced visits by over 20 percent. Adequate prenatal care reduces preventable maternal complications by 3.4 percentage points (26 percent), with no effect on infant outcomes. OLS estimates are perversely signed for maternal complications but overstate benefits for infant outcomes, consistent with competing patient- and provider-driven selection. Mechanism evidence points to prenatal care as a longitudinal information system: repeated contact facilitates detection of evolving risks and generates clinical information that informs delivery decisions, including delivery-mode routing and subsequent management at delivery.