By Cynthia C. Harper, PhD, Professor, Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, School of Medicine
The Centers for Disease Control and Prevention (CDC) Healthy People 2020 recognizes unintended pregnancy as a significant public health challenge (51 percent of pregnancies in the United States), and recommends increasing access to long-acting reversible contraceptives (LARC). IUDs (intrauterine devices) and contraceptive implants have the highest efficacy levels (greater than 99 percent), but are seldom used in the United States, compared to other developed countries. National data show that, contrary to evidence-based CDC recommendations, only 38 percent of physicians providing contraception in the United States offer IUDs to adolescents, who are at high risk of unintended pregnancy.
We designed a clinic intervention to educate providers to integrate IUDs and implants into routine contraceptive care. The intervention was based on our formative research that identified priorities in translating evidence to clinical practice. Identified priorities included knowledge and counseling deficits and insertion skills. We developed a CME/CE-accredited intervention and sought to measure the impact of a clinic-level intervention on patient-level contraceptive and pregnancy outcomes. No clinic-based intervention had effectively reduced pregnancy in randomized trials. Our research aimed to address an outstanding scientific and clinical question: whether a randomized intervention could improve access to high-efficacy contraceptives and reduce pregnancy.
We conducted a cluster randomized trial with 40 community clinics across the United States. A cluster design was necessary to avoid contamination that could occur among providers with individual patient randomization. We conducted the study with Planned Parenthood Federation of America clinics. Clinic sites were randomized and allocation concealed until study initiation. After intervention, clinic staff were trained (n=269); we recruited 1,500 women from all sites to measure contraceptive and pregnancy outcomes.
Eligible women ages 18–25, did not want pregnancy within 12 months, and received contraceptive counseling. We collected 12-month follow-up data from questionnaires, pregnancy tests and medical records. The primary study outcome was women’s choice of an IUD or implant. The secondary outcome was a biological outcome, pregnancy incidence during follow-up. We also assessed women’s reports of method counseling and their knowledge of contraceptive effectiveness.
The intervention was multifaceted and designed to improve providers’ knowledge, skills and counseling, resulting in greater patient awareness and access to LARC methods. A unique feature was the emphasis on the team in clinical care; the in-service training was given to all staff. The 3.5 hour CME-accredited session offered didactics on copper and levonorgestrel intrauterine and etonogestrel subdermal contraceptives, as well as a clinician hands-on practicum. The counseling training presented the World Health Organization tiered contraceptive effectiveness chart, and the CDC recommended use of open-ended questions on pregnancy intentions.
The training emphasized patient-centered counseling with shared decision making. Following adult learning theory, we used multiple modes of teaching, including videos, hands-on training, and counseling role plays. Technical assistance for billing was provided. Usual costs for contraception were maintained to test the intervention under real cost conditions. Control clinics followed standard care.
Study participants (802 interventions, 698 control) were 21 years old on average, with 96 percent unmarried. Fifty percent were white, 27 percent Latina and 15 percent black. Thirty-eight percent were uninsured and 27 percent relied on public insurance such as Medicaid. In intervention, 71 percent of women reported their provider discussed LARCs, compared to 39 percent in control (odds ratio (OR) 3.8; 95 percent CI, 2.8 to 5.2). Women’s knowledge of method effectiveness was higher in intervention (44 percent) than control (27 percent) (OR 2.1; 95 percent CI, 1.6 to 2.8). In intervention, 28 percent chose a LARC method compared to 17 percent in control (OR 1.9; 95 percent CI, 1.3 to 2.8).
There were 211 incident pregnancies. Intervention pregnancy rate was 15 per 100 person-years and control was 18.5 (hazard ratio (HR) 0.89; 95 percent CI, 0.64 to 1.24). In an analysis stratified by clinic setting, pregnancy in family planning settings was significantly lower in intervention, 7.9 per 100 person-years, than in control,15.4 (HR 0.54; 95 percent CI 0.34 to 0.85). In the abortion care setting, pregnancy rates were higher overall, with no difference by arm (HR 1.35; 95 percent CI 0.91 to 2.01). More women in intervention choosing LARC were able to obtain them in family planning (73 percent) than in abortion (44 percent) care.
A large majority of young women want to delay childbearing but report unprotected intercourse, as in our study population. Clinic visits are important opportunities for patient education, although providers do not routinely include LARC methods in counseling, even while recommended by the American College of Obstetricians and Gynecologists and the American Academy of Pediatrics. The intervention resulted in increased counseling and greater choice of high-efficacy methods among young at-risk women.
For the first time, this randomized trial showed a significant reduction in pregnancy in the family planning setting. The intervention is appropriate for implementation science and was targeted at the clinic level, which efficiently reached a large patient population: intervention clinics had an annual contraceptive patient volume of over 100,000. Strengths also included a randomized design, biological measures for pregnancy, and high follow-up. Finally, the intervention succeeded in improving a health outcome long-resistant to change in the United States. Results differed for post-abortion contraception, where cost barriers are high. Fewer women choosing LARC post-abortion were actually able to obtain the method than in family planning sites. Insurance coverage, less commonly used in abortion care, is especially important for contraceptives with high up-front costs. In some states, women must return to a family planning clinic to qualify for contraceptive coverage. Contraceptives are essential preventive care to improve women’s health, and these results showed clinicians can successfully integrate high-efficacy methods into clinical practice and reduce pregnancy in family planning clinics.
Implementation Science Stage
Cost-effective, replicable interventions to reduce unintended pregnancy are much needed in the United States, where the health of women and infants is poor relative to comparable countries. Since trial completion, we have brought the intervention to a range of new practices, including hospitals, teen clinics and primary care clinics. Our training team is rated in the CME evaluations as “excellent” and the intervention continues to be effective in ongoing evaluations. We partner with large agencies; for example, we recently worked with the Department of Health in Louisiana for a statewide training of all of their public clinics and hospitals. We offer no-cost contraceptive educational materials and well as clinician tools (beyondthepill.ucsf.org).
Our study findings are available in the following publications:
Harper CC, Rocca CH, Thompson KM, Morfesis J, Darney PD, Westhoff CL, Speidel JJ. Reduced pregnancy in the US from long-acting reversible contraception: a cluster randomized trial. Lancet. 2015;386(9993): 562-8.
Simmons KB, Rodriguez MI. (Commentary). Reducing unintended pregnancy through provider training. Lancet. 2015;386 (9993):514-6.
Thompson KM, Rocca CH, Kohn JE, Goodman S, Stern L, Blum M, Speidel JJ, Darney PD, Harper CC. Public Funding, provider training, and use of highly effective contraceptives: a cluster randomized trial. American Journal of Public Health 106(3):541-546, 2016.
Our research and CME/CE activities were supported by grants from the William and Flora Hewlett Foundation and the JPB Foundation. The National Campaign to Prevent Teen and Unplanned Pregnancy provided a small grant to produce a patient education video used in the intervention. Units for intervention training in intrauterine device insertion were provided by Teva Pharmaceuticals Industries and Bayer HealthCare.
We would like to recognize the UCSF Office of Continuing Medical Education, especially Dr. Robert Baron, associate dean of GME and CME, Mr. Tymothi Peters, director of CME, and Ms. Stacey Samuels. We also thank Ms. Johanna Morfesis and Planned Parenthood investigators, the Bixby Beyond the Pill research and training team, and all study participants