The effects of interventions to integrate long‐acting reversible contraception with treatment for incomplete abortion: Results of a 6‐year interrupted time series analysis in hospitals in mainland Tanzania and Zanzibar

Abstract Objective To evaluate an intervention that aimed at strengthening voluntary access to long‐acting reversible contraception (LARC) within postabortion care (PAC) in hospitals in mainland Tanzania and Zanzibar. Methods From 2016 to 2018, we conducted PAC quality improvement interventions, emphasizing family planning (FP) counseling and voluntary access to LARC. Researchers conducted an interrupted time‐series analysis of service statistics compiled from 2014 to 2020 using segmented linear mixed effects regression models to assess the interventions' effect on postabortion contraceptive uptake. Results The intervention in mainland Tanzania was associated with an immediate 38% increase in postabortion LARC uptake, a trend that declined from late 2016 to mid‐2020 to 34%. In Zanzibar, the intervention was associated with a gradual increase in LARC uptake that peaked in late 2018 at 23% and stabilized at approximately 15% by mid‐2020. Whereas the interventions in mainland facilities did not generate significant changes in postabortion FP uptake overall, the launch of interventions in Zanzibar in mid‐2016 was associated with a precipitous rise in that outcome over time, which plateaued at approximately 54% by 2019. Conclusion Increased voluntary uptake of postabortion contraception was associated with the introduction of training in PAC, including FP, and quality improvement interventions and gains were sustained over time.


| INTRODUC TI ON
Unintended pregnancies and subsequent abortions are problems worldwide and a driver of maternal mortality and morbidity, particularly in countries such as Tanzania, where there are social, cultural, and legal constraints to accessing comprehensive sexual and reproductive health services. Tanzania has one of the world's highest maternal mortality ratios (410 per 100 000 live births) and a ratio of 21 abortions per 100 live births. 1,2 Accordingly, its government has expanded its program on postabortion care (PAC), which helps save the lives of women suffering from abortion complications by providing emergency treatment and offering family planning (FP) counseling, including voluntary access to contraception, in the same visit.
Most women requiring PAC in Tanzania are treated at the tertiary level, despite recent efforts aimed at decentralizing the service from district and regional hospitals to primary care settings. 3  which includes Mwanza and Geita. In these regions, the health system has a typical three-level pyramid structure with tertiary and regional and district facilities, intermediate health centers, and, closest to communities, primary healthcare centers. Before the project, hospitals provided the widest range of PAC care including vacuum aspiration for routine incomplete abortion and advanced surgical care for severe complications (e.g. laparotomy), and a wide range of FP methods were available in separate hospital settings. Health centers were able to provide treatment for routine complications of abortion and short-term contraceptive methods. In health centers, also, FP services were provided separately. At lower levels of care, PAC services during this period experienced lapses in sustainment owing to intermittent patient flows and staff turnover and logistical challenges. According to policy, only surgical treatment, not misoprostol, could be used for treating abortion complications. In Zanzibar, which has received appreciably less external support to organize a regional PAC program, the vast majority of PAC services were provided at the Regional Referral Hospital in Unguja, with far fewer cases recorded per month throughout the network of district hospitals on both islands. Even though both types of facilities could treat abortion complications and provide postabortion contraception, the regional referral facility, Mnazi Moja Hospital, received most cases with severe complications. FP services in both settings were offered separately from PAC and limited to short-acting methods.
In Zanzibar, cases of routine incomplete abortion during this time were occasionally recorded at Primary Healthcare Units. At variance with the mainland PAC program, misoprostol has been recognized as an appropriate treatment for abortion complications in Zanzibar for some time, even though there were no official guidelines on this until PAC-FP helped to develop them in 2018. The intervention used an iterative five-step approach to embed the processes in district-level healthcare management structures, hospitals, and points of care where PAC is available. This is illustrated in Figure 1.

| The intervention
During step 1, PAC-FP worked with district-level counterparts to conduct organizational capacity assessments (OCAs). OCAs illuminated district-level barriers and facilitators to achieving project goals and established stakeholder consensus on next steps. In mainland Tanzania, where the PAC program was more mature and uptake of short-acting methods of contraception was already high, stakeholders prioritized integrating LARC into the postabortion method mix. 9 Whereas in Zanzibar, whose PAC program had received little external assistance, stakeholders chose to strengthen FP integration more gradually, focusing first on better counseling and availability of FP services, and the development of clinical guidelines on PAC that promote access to all types of postabortion contraception. 10 These guidelines also give instruction on the strengthening of PAC, including postabortion FP, in primary care facilities, which had not been emphasized in prior work to help develop the Zanzibar PAC program.
In both settings, OCA groups adopted a phased implementation approach that started in hospitals in 2016.
In step 2, OCA groups disseminated the results at selected hospitals and, in turn, facilitated facility-level capacity assessments with the support of PAC-FP. With this, hospitals' existing QI teams developed site-specific action plans. In mainland Tanzania, this included updates to product placement and requisition, and integration of existing PAC and LARC trainings into modules and tools. In Zanzibar, QI teams also focused on FP product placement and requisition reforms, and established dedicated private settings in each hospital for PAC provision. Stakeholders in Zanzibar opted to adapt and use the centralized PAC and LARC training curriculum from the mainland as an initial step in developing region-specific clinical PAC guidelines.
During steps 3 and 4, the QI teams implemented their action plans, receiving catalytic financial, material, and technical support from PAC-FP, including PAC and LARC training. PAC-FP monitoring and evaluation staff routinely compiled data on action plan achievement, the contextual factors affecting performance, and changes made to action plans as a result. PAC-FP convened OCA and QI teams for periodic meetings in each region where teams shared, interpreted, and used results to inform efforts to accelerate progress and solve problems.
In step 5, OCA teams reflected on lessons and considered matters related to extending the intervention to primary care settings, engaging communities, and other strategic issues.
The project completed all steps by October 2018.

| The evaluation
This study reports on the effect of the above cycle of implementation in eight hospitals in mainland Tanzania and six hospitals in Zanzibar.
The analysis presented herewith is confined to this subset of the 65 F I G U R E 1 Five-Step Framework for QI and Service Integration TA B L E 1 Sociodemographic and PAC service delivery characteristics for pre-intervention (January 2014 to July 2016), intervention (August 2016 to October 2018) and sustainment (November 2018 to June 2020) phases (mainland Tanzania) a

| DISCUSS ION
Our study sought to assess changes in the integration of treatment for abortion complications and FP services that were associated with a multi-year, phased intervention that aimed at strengthening systems and improving the quality of PAC in two distinct regions of Tanzania. In Geita and Mwanza, where PAC service strengthening has long-standing historical roots, PAC-FP interventions aimed at LARC integration were successful, catalyzing a sharp increase in the provision of IUDs and hormonal implants in the immediate term, and, in the long run, resulting in a more evenly balanced contraceptive method mix accessed by PAC patients that was sustained over time.
It should be noted, however, that the incidence of postabortion

ACK N OWLED G M ENTS
We are grateful to the MOHCDGEC in Tanzania for their leadership and support during data collection. Special appreciation also goes to AMCA Concern for implementation of the endline study. We are grateful to the respondents who participated in the study. This manuscript was edited by Amy Agarwal. The work was funded by the United States Agency for International Development (USAID) under associate cooperative agreement AID-OAA-A-00050. The opinions expressed are those of the authors and do not necessarily reflect the views of USAID, or the United States government.

CO N FLI C T S O F I NTE R E S T
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.