Public perceptions of abortion complications

Background: Misinformation contributes to the perception that abortion has substantial health risks, despite the known safety of medication and aspiration abortion. We lack detailed information about which health risks the public believes are most likely. Objective: This study aimed to describe public perception of short- and long-term risks of abortion. Study Design: We conducted a cross-sectional survey of US residents aged ≥18 years using Amazon Mechanical Turk (MTurk). We collected information regarding participant demographics, reproductive history, political views, and position on abortion restrictions. We provided participants with a list of 9 short-term and 15 long-term possible complications and asked them to indicate whether they occurred never (0%), very rarely (<1%), rarely (1%–5%), occasionally (5%–20%), or frequently (>20%) following abortion. We used descriptive statistics to understand our population demographics and to capture the perceived incidence of all complications. We created a binary indicator of answering all risk estimates incorrectly vs at least 1 estimate correctly, separately for all long-term possible complications, and the 2 short-term risks of infection and bleeding. We determined the proportion of individuals who responded incorrectly to all questions in each category and used multivariable logistic regression to identify factors associated with incorrect perceptions about the risks of abortion. Results: For all listed complications, participant (N=1057) estimates of risk were higher than the known incidence. For both short-term risks of bleeding and infection, over 40% of participants reported that these outcomes occur occasionally or frequently. Similarly, for both long-term risks of depression and anxiety, over 60% of respondents reported that these outcomes occur occasionally or frequently after abortion. Participants reported that possible complications known to not be associated with abortion, including hair loss, future pregnancy complications, breast cancer, and cosmetic disfigurement, occurred at least rarely. Nearly one-quarter of participants responded that death occurs occasionally or frequently (in over 5% of abortions), and 79% of participants responded that breast cancer can result from abortion. One-quarter (24.9%) of participants incorrectly overestimated both short-term outcomes of infection and bleeding, whereas 19.5% answered all long-term complication questions incorrectly, including outcomes that never occur. On multivariable analyses, we identified that the participants most likely to incorrectly identify risks of abortion identified as Asian or Black race/ethnicity, were from rural communities, or believed that abortion should have more legal restrictions. Conclusion: The public perceives abortion to be much riskier than it actually is. This information can be used to develop targeted clinical and public health efforts to disseminate the true risks of abortion.