access is also likely to afford more women the opportunity to obtain care without the additional indirect costs associated with traveling to a geographically distant abortion provider. The strengths of this study are its statistical power, the large number of providers, and its setting in multiple facilities. A limitation of the study is its nonrandomized design, although the use of propensity score matching allowed for statistical adjustments to address this limitation. Additionally, this study had a low follow-up rate (70%), but this was not unexpected because of the sensitive nature of abortion, which may have deterred women from continuing participation in the study after the procedure. This follow-up rate is also similar to those in other US abortion-related studies with comparable follow-up periods (14---28 days).22,37,46 Although postprocedure complications may have been missed among patients for whom we did not have follow-up data, given the nondifferential follow-up rates between provider groups, we would expect unidentified complications to be equally distributed between groups, leaving the risk difference unaffected. A further limitation of the study is that the health care provider who initially identified a complication was not blinded to the type of provider who performed the abortion. However, we hypothesize that complaints from patients cared for by newly trained NPs, CNMs, and PAs would be more aggressively evaluated if the provider type was known to the health care provider evaluating the patient. Therefore, any bias caused by lack of blinding would have resulted in an overestimate of the risk difference. Our results confirm existing evidence from smaller studies that the provision of abortion by NPs, CNMs, and PAs is safe21,22 and from larger international13 and national47 reviews that have found these clinicians to be safe and qualified health care providers. The value of this study extends beyond the question of who can safely perform aspiration abortion services in California because it provides an example of how research can be used to answer relevant health workforce policy issues. As the demand for health care providers increases under US health care reform,48 one part of the solution for all health care, including abortion care, is to allow all qualified professionals to perform clinical care 0.87 0.70 0.83 0.45 Adjusted 0.29 Adjusted propensity Score-matched sample 1.29 Unadjusted 0.33 -2.0 -1.0 1.0 0.0 2.0 Risk Differences for Complications Between Provider Types, % Risk difference 95% CI 1.10 1.33 Note. CI = confidence interval. Both adjusted models included patient age, race/ethnicity, insurance type, gestational age, gravidity, history of cesarean section, positive test for a sexually transmitted infection, an indicator for extreme obesity, an indicator for chronic illness, and an indicator for psychiatric conditions. 2.0 is also the delta. FIGURE 2—Unadjusted, adjusted, and adjusted propensity score–matched risk differences in overall complication rates of first-trimester aspiration abortion by nurse practitioner, certified nurse midwife, and physician assistant providers compared with physician providers in California. RESEARCH AND PRACTICE March 2013, Vol 103, No. 3 | American Journal of Public Health Weitz et al. | Peer Reviewed | Research and Practice | 459 to the fullest extent of their education and competency.49,50 j About the Authors Tracy A. Weitz, Sheila Desai, Ushma D. Upadhyay, and Molly F. Battistelli are with Advancing New Standards in Reproductive Health (ANSIRH), Bixby Center for Global Reproductive Health, and the Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco. Diana Taylor is with ANSIRH, Bixby Center for Global Reproductive Health, and the School of Nursing, University of California, San Francisco. Jeff Waldman is with Planned Parenthood Shasta Pacific, Concord, CA. Eleanor A. Drey is with the Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco. Correspondence should be sent to Tracy A. Weitz, PhD, MPA, Advancing New Standards in Reproductive Health (ANSIRH), Bixby Center, University of California, San Francisco, 1330 Broadway, Suite 1100, Oakland, CA 94612 (e-mail: weitzt@obgyn.ucsf.edu). Reprints can be ordered at http://www.ajph.org by clicking the “Reprints” link. This article was accepted November 26, 2012. Contributors T. A. Weitz and D. Taylor developed the study concept and design. T. A. Weitz, D. Taylor, and E. A. Drey supervised the overall study and analyzed the interpretation of results. S. Desai oversaw the acquisition of data. U. D. Upadhyay and S. Desai analyzed the data and provided statistical expertise. T. A. Weitz, U. D. Upadhyay, S. Desai, and E. A. Drey drafted the article, and J. Waldman advised on critical revision of the article for intellectual content. M. F. Battistelli provided administrative, technical, and material support. Acknowledgments This research was supported by grants from private foundations including the John Merck Foundation, the Educational Foundation of America, the David & Lucile Packard Foundation, and the Susan Thompson Buffet Foundation. In addition, the research was conducted under a legal waiver from the California Health Workforce Pilot Project Program, a division of the Office of Statewide Health Planning and