Medical abortion offered in pharmacy versus clinic-based settings

Background: Medical abortion is usually offered in a clinic or hospital, but could potentially be offered in other settings such as pharmacies. In many countries, pharmacies are a common first point of access for women seeking reproductive health information and services. Offering medical abortion through pharmacies is a potential strategy to improve access to abortion. Objectives: To compare the effectiveness and safety of medical abortion offered in pharmacy settings with clinic-based medical abortion. Search methods: We searched CENTRAL, MEDLINE, Embase, four other databases, two trials registries and grey literature websites in November 2020. We also handsearched key references and contacted authors to locate unpublished studies or studies not identified in the database searches. Selection criteria: We identified studies that compared women receiving the same regimen of medical abortion or post-abortion care in either a clinic or pharmacy setting. Studies published in any language employing the following designs were included: randomized trials and non-randomized studies including a comparative group. Data collection and analysis: Two review authors independently reviewed both retrieved abstracts and full-text publications. A third author was consulted in case of disagreement. We intended to use the Cochrane risk of bias tool, RoB 2, for randomized studies and used the ROBINS-I tool (Risk Of Bias In Non-randomized Studies of Interventions) to assess risk of bias in non-randomized studies. GRADE methodology was used to assess the certainty of the evidence. The primary outcomes were completion of abortion without additional intervention, need for blood transfusion, and presence of uterine or systemic infection within 30 days of medical abortion. Main results: Our search yielded 2030 records. We assessed a total of 89 full-text articles for eligibility. One prospective cohort study met our inclusion criteria. The included study collected data on outcomes from 605 women who obtained a medical abortion in Nepal from either a clinic or pharmacy setting. Both sites of care were staffed by the same auxiliary nurse midwives. Over all domains, the risk of bias was judged to be low for our primary outcome. During the pre-intervention period, the study’s investigators identified a priori appropriate confounders, which were clearly measured and adjusted for in the final analysis. For women who received medical abortion in a pharmacy setting, compared to a clinic setting, there may be little or no difference in complete abortion rates (adjusted risk difference (RD)) 1.5, 95% confidence interval (CI) -0.8 to 3.8; 1 study, 600 participants; low certainty evidence). The study reported no cases of blood transfusion, and a composite outcome, comprised mainly of infection complications, showed there may be little or no difference between settings (adjusted RD 0.8, 95% CI -1.0 to 2.8; 1 study, 600 participants; very low certainty evidence). The study reported no events for hospital admission for an abortion-related event or need for surgical intervention, and there may be no difference in women reporting being highly satisfied with the facility where they were seen (38% pharmacy versus 34% clinic, P = 0.87; 1 study, 600 participants; low certainty evidence). Authors' conclusions: Conclusions about the effectiveness and safety of pharmacy provision of medical abortion are limited by the lack of comparative studies. One study, judged to provide low certainty evidence, suggests that the effectiveness of medical abortion may not be different between the pharmacy and clinic settings. However, evidence for safety is insufficient to draw any conclusions, and more research on factors contributing to potential differences in quality of care is needed. It is important to note that this study included a care model where a clinician provided services in a pharmacy, not direct provision of care by pharmacists or pharmacy staff. Three ongoing studies are potentially eligible for inclusion in review updates. More research is needed because pharmacy provision could expand timely access to medical abortion, especially in settings where clinic services may be more difficult to obtain. Evidence is particularly limited on the patient experience and how the care process and quality of services may differ across different types of settings.