Building on safety, feasibility, and acceptability: the impact and cost of community health worker provision of injectable contraception

This project in Zambia contributes to our understanding of the impact of community-based provision of injectables on method choice and uptake and of the costs of adding DMPA to an established community-based family planning program. The project also illustrates the importance of involving stakeholders from the outset, analyzing costs relevant to scale up, and engaging in policy change dialogue not at the end, but rather throughout project implementation.

Task sharing has been employed as a strategy to address this problem by delegating health care tasks that are usually carried out by doctors and nurses to a lower-level provider who is more accessible to the community. For example, the Government of Zimbabwe recently decided to increase access to antiretroviral treatment (ART) by authorizing trained nurses to prescribe drugs and manage patients in care. 2 Many countries in SSA deploy lower-level cadres in different forms: as government or nongovernment-affiliated, as volunteers or salaried workers, with limited or wide-ranging responsibilities to the communities they serve. Just in the area of family planning, tasks assigned to community health workers (CHWs) can vary. For example, while CHWs in Rwanda are able to resupply clients with both pills and injectables only after a clinical evaluation, in Uganda they are able to initiate and resupply clients with hormonal methods. In Ethiopia, CHWs even insert implants. 3 Thus, CHWs can play an important role in providing family planning services. Moreover, success in Uganda, Ethiopia, and other SSA countries suggests that the role of CHWs need no longer be limited to distribution of condoms and oral contraceptive pills or referral to higher-level providers.
At a June 2009 technical consultation convened by the World Health Organization (WHO), the U.S. Agency for International Development (USAID), and Family Health International (now FHI 360), 30 technical and program experts from 18 countries reviewed evidence and experiences from programs using CHWs to expand access to injectable contraceptives. These experts concluded, ''Given appropriate and competency-based training, CHWs can screen clients effectively, provide DMPA (depot medroxyprogesterone acetate) injections safely and counsel on the side effects appropriately, demonstrating competence equivalent to facility-based providers of progestin-only injectables.'' 4 With the conclusions endorsed by normative bodies such as the International Federation of Gynecology and Obstetrics, the United Nations Population Fund, the International Council of Nurses, and USAID, more countries in SSA initiated pilot studies, began implementing scale-up efforts, engaged in policy change dialogue, or realized policy changes that allow CHWs to provide injectable contraceptives.
More recently, WHO released a set of guidelines that define health worker roles for maternal and newborn health. These guidelines focus on task sharing among various cadres of health care providers to address the critical human resource shortages in many developing countries. Using the latest scientific evidence and the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology, WHO endorsed the ''initiation and maintenance of injectable contraceptives'' by lay health workers using a standard syringe, provided that a strong monitoring and evaluation system is in place. 5 As of August 2012, 13 countries in SSA were undergoing various stages of rolling out CHW provision of injectables. 6 Different paths were followed, as some governments changed policy first and then conducted pilot studies and scale up, while others began with pilots before considering scale up and policy change. 7 The impact of CHW-provided injectable contraception has been measured in countries such as Kenya, Madagascar, Malawi, Nigeria, and Uganda, where these programs reported expanded access to family planning services, increased uptake of family planning methods, reduced workload in clinics, and improved method continuation rates among DMPA users. 8,9 In Zambia, the process to gain approval for CHW provision of DMPA began with a request from the government to conduct a pilot study in hard-to-reach areas, where staff turnover also presents a significant problem. As in other SSA countries, use of family planning services among rural women in Zambia is relatively low; the contraceptive prevalence rate (CPR) for modern methods is 37%, compared with 48% in urban areas. Similarly, unmet need in rural areas is 28% (19% spacing, 9% limiting) versus 23% (13% spacing, 10% limiting) in urban areas. 10 In a July 2009 stakeholder meeting to discuss design of the pilot study, government officials and national stakeholders requested measures of program impact to be included, in addition to local confirmation of the safety, feasibility, and acceptability of CHW provision of injectable contraception.
With approval from the Government of Zambia (GoZ), FHI 360 collaborated with ChildFund Zambia, the local affiliate of ChildFund International (formerly Christian Children's Fund) to design and implement an intervention to introduce injectable contraception into ChildFund's existing CHW family planning program. The ChildFund CHW program has been in operation since 1987. In A growing body of evidence supports the provision of injectable contraceptives by CHWs in hard-toreach areas.
Impact and cost of CHW provision of injectable contraception www.ghspjournal.org addition to providing family planning services, ChildFund CHWs also deliver health education on personal hygiene and safe motherhood as well as sensitize the community-focusing on men-about family planning. CHWs are both men and women, with varying levels of secondary school education, who have been chosen by community members to provide basic services at the community level. They are volunteers but receive in-kind remuneration in the form of materials and equipment (bicycles, raincoats and boots, t-shirts, bags) and costshared animal restocking. Their initial family planning training, conducted over 2 weeks, uses the GoZ curriculum that also includes topics on clients' rights, anatomy and physiology, HIV/ AIDS mode of transmission, family life education, male involvement, distribution and storage of commodities, and more. Although they work for ChildFund, these CHWs are affiliated with and supervised by GoZ health center staff (as well as ChildFund staff), from whom they obtain family planning commodities and to whom they submit records for inclusion in the district data management system. With ChildFund's assistance, we also collected information on additional, or ''incremental,'' costs of adding injectable contraception to their ongoing CHW provision of condoms and oral contraceptive pills. As programs grapple with limited resources, such information is needed to estimate costs of scaling up and to establish that an intervention provides ''value for money.'' Thus, this paper not only presents results on the safety, feasibility, and acceptability of CHW provision of DMPA in the Zambian context, but it also focuses on the impact and costs of adding DMPA to an established community-based family planning program.
The study objectives were to:

Overview
The safety of CHW provision of injectables was measured by DMPA client reports and by a 21-item structured observation checklist (SOC) divided into 2 scales that measured infection prevention (11 items) and injection procedures (10 items Forty practicing CHWs (20 from each district) affiliated with the 16 health facilities in Mumbwa and Luangwa were trained by master trainers from the MOH to safely provide DMPA injections in addition to the family planning services they already provided. CHWs received 5 days of didactic instruction on determining method eligibility (screening), counseling and informed choice, client referral, and provision of oral contraceptive pills, condoms, and DMPA. * Impact and cost of CHW provision of injectable contraception www.ghspjournal.org CHWs also completed a 2-to 4-week clinic-based practicum during which 6 or more DMPA injections had to be successfully administered before the CHW was allowed to provide injections unsupervised. We divided the 21-item structured observation checklist used by evaluators into its 2 main components and calculated a total score for safe injections with regard to infection prevention and to injection procedure. Evaluations of the first and last injections given by each CHW constituted the measures of safety. The training and practicum were conducted on a staggered schedule in Luangwa and Mumbwa between December 2009 and January 2010. By February 2010, all CHWs were certified and their capabilities introduced to their respective communities at an official graduation ceremony involving local leaders and ChildFund representatives. CHWs were then asked to record all provision of condoms, pills, and injectables to their family planning clients for 13 monthsdistinguishing between new acceptors, switchers, and continuing users of all methods-using ChildFund Zambia's family planning register, modified to include DMPA.
The study team-comprising FHI 360, ChildFund Zambia, and Ministry of Health staff, the Family Planning Technical Working Group (FPTWG), { and other stakeholders-selected the following measures of impact: Approximately 9 months into data collection, we interviewed all CHWs and a subset of their DMPA clients (n5253) who received their first injection from a CHW between February and April 2010. By then, CHWs had performed several months of DMPA administration, and enough time had elapsed for these initial DMPA users to receive up to 3 injections from a CHW.
ChildFund Zambia also selected 6 male employees-3 assigned to each district-to personally retrieve family planning data on a monthly basis from the 40 CHWs. These men (given bicycles for transportation) were engaged specifically to cover the distances to and from CHWs' homes, discuss and verify data with the CHWs, and transfer the required information from the family planning register to the data retrieval form developed for the study. This form collected information on: N Method received/used on the first visit in study period N Whether the client was a continuing user or new family planning acceptor 1 N Previous method used N Number of pill cycles or condoms distributed and the scheduled re-injection date for DMPA clients over the course of the 13-month data collection period The 6 data retrievers received instruction from the Lusaka-based project coordinator on use of the data retrieval form and the family planning client register. Data retrievers also met with the project coordinator monthly for the first half of data collection, then every 2 to 3 months thereafter to verify the accuracy of the information and to submit CHW family planning uptake data from February 2010 to February 2011.
We also obtained family planning statistics for pills, condoms, and DMPA distribution from the District Health Offices in Luangwa and Mumbwa for the same 13-month time period. These statistics reportedly included both health facility and CHW provision of family planning methods. All analyses were performed with SAS 9.2. 11 * The training materials used in 2009 can be accessed at: http:// www.k4health.org/toolkits/cba2i/sample-community-healthworker-training-curriculum. However, the content has since been updated to emphasize dual protection to prevent STIs/HIV and a 13-week DMPA re-injection window. Originally the Zambia MOH was following a 12-week policy. 1 During the data collection period, no women stopped and restarted a method after 6 months; we did not determine upon recruitment if continuing users entering the study had restarted a method after 6 months.
Study methods included checklists to measure safe provision, interviews with CHWs and clients, and cost analyses.
Impact and cost of CHW provision of injectable contraception www.ghspjournal.org Global Health: Science and Practice 2013 | Volume 1 | Number 3

Assessment of Incremental Costs of Adding DMPA to the CHW Program
The approach to costing the addition of a health service intervention to an existing program is to concentrate on additional or ''incremental'' costs only, since costs of the existing program would have been incurred even without the intervention. Incremental costs of an intervention can be classified according to 3 phases: 1) planning/ designing the intervention; 2) preparing for service delivery; and 3) delivering the new or improved services. Each phase comprises a set of activities, and each activity uses resources, such as time of trainers and providers and medical supplies and equipment. When costs are attached to resources used in the intervention, the total incremental cost of the intervention may be calculated. Stakeholders may be more interested in costs of activities relevant to scaling up, giving less emphasis to costs in the planning and design phase. In our study, these activities included the following: 1. We used key informant interviews, record reviews, and periodic progress reports to identify all intervention-related activities and resources, and we designed Excel-based spreadsheets to organize information on the costs of these resources. Most costs reflect actual expenditures, except for personnel costs, which were estimated using MOH salary scales for positions considered to be equivalent to those of research and project staff who implemented the intervention.
Incremental cost per CYP of CHW provision of DMPA was calculated by dividing the annualized incremental cost of the intervention (that is, costs adjusted for the period of the study) by the number of CYPs attributable to the CHW intervention. This indicator provides a sense of the value of resources needed to protect 1 couple from pregnancy for a year through DMPA provided by CHWs in ChildFund's program.
FHI 360's Protection of Human Subjects Committee and ERES Converge Ethical Review Board in Zambia reviewed and approved this study.

Safety of DMPA Provision by CHWs
On the 11-item scale on infection prevention procedures, CHWs initially carried out about 9 of the items, or 82%, on average. At their final assessment, the average score increased to a perfect 11, or 100%. Thus, there was improvement from the first to the last DMPA injection evaluated during the practicum, but the starting point was high to begin with.
For the 10-item injection procedure scale, the initial average score was 7.4, and at final assessment the score was 9.6. Again, there was improvement, but it is important to note that CHWs who scored low in the initial assessment improved markedly by the last assessment. This is captured by the narrowing of the range of scores from the initial to the final assessments for both infection prevention and injection procedure ( Figure 1).
Only 6 (2%) of the 253 DMPA clients interviewed after 9 months of data collection reported any problems with CHW-provided injections. Three could not describe the problem, 2 reported pain at the injection site, and 1 client described numbness in the arm. There were no reports from clients of abscesses or infections. During monitoring and supervision activities conducted by ChildFund throughout the study period, CHWs also reported to ChildFund and MOH supervisors that they had not found any abscesses or infections.

Acceptability of Method to Clients
The acceptability of DMPA and CHW-provision of the method was ascertained by interviews with 253 clients: Of the 7% of DMPA users who did not want another injection, the main reasons mentioned were desire to have a child (35%), side effects or problems with DMPA use (35%), and husband's disapproval (18%). For the 4 women (1.6%) who could have but did not receive a second injection, the reasons were dissatisfaction with side effects (n52) and husband's disapproval (n52). Thus, CHW provision of DMPA and the method itself proved highly acceptable.
In interviews with the 40 CHWs, 98% reported that it was easy to find women who were interested in receiving DMPA, and 80% felt that it was easy to gain the confidence of the community in their ability to provide DMPA. Although 89% said that their workload increased due to the addition of DMPA to their family planning services, they reported that the increase was acceptable and not a burden.

Characteristics of Acceptors and Family Planning Uptake
Data on family planning method uptake were recorded for a total 4,241 family planning clients in both districts during the 13-month data collection period. The average age of women was 28, with a range of 15 to 53 years. Clients had an average of 3.6 living children, with a range of 0 to 14 children. Based on provision of methods by ChildFund CHWs from February 2010 to February 2011, 51 condom clients, 391 pill clients, and 2,206 DMPA clients would be protected from pregnancy for 1 year. Uptake of condoms, pills, and DMPA was greater in Mumbwa than Luangwa, as the majority of family planning clients (73%) were from Mumbwa district, the more populous area, with a higher contraceptive prevalence rate (40% vs. 27%). 10 In both districts, condoms conferred the fewest CYPs, while DMPA conferred the most (Table 1).
With regard to new acceptors, 41% of CHW clients during the 13-month period were new to family planning. In this group of 1,739 women, 85% chose DMPA, 13% chose oral contraceptive pills, and the remaining 2% chose condoms as their first family planning method. Of the continuing users, 63% reported using pills, 30% DMPA, 6% condoms, and the rest unspecified. Of all the family planning clients in the study, 82%, or 3,479 women, obtained DMPA from a CHW some time between February 2010 and February 2011. About 20% were continuing clients who were formerly clinic clients, and 24% were former pill and condom users who switched to CHWprovided DMPA.

Continuing Method Use for Pills and DMPA
To determine continuing method use, we examined data from women who obtained DMPA or pills from CHWs during the study who had the opportunity-based on when they started-to use DMPA or pills for 6 months, 9 months, or 12 months (Figure 2). Continuation for DMPA users was always higher than for pill users (using mutually exclusive groups of women with 6, 9, and 12 month follow-up data), with a significant difference noted at the 12-month mark-63% vs. 47% (Chi-square P,.001).  Table 2 reports incremental costs of intervention activities relevant to future scale up. The total incremental costs were US$37,300 (in 2010 dollars), while the annualized costs adjust some of the cost items to reflect that their effects would extend beyond the initial year of the intervention. For example, the ToT workshop produced 10 persons capable of training CHWs to provide DMPA, but only 4 of them participated in the pilot; therefore, only 40% of the ToT costs are applicable to the pilot. For both training activities, the effects are assumed to last for 2 years (requiring periodic refresher training), and so we applied one-half of the training cost (including practicum expenses) to the 1-year period of service delivery in the pilot program. Other activity costs are unchanged. The estimate of annualized incremental cost (which also serves as the numerator of the incremental cost per CYP ratio) was US$24,322. The denominator of the cost per CYP ratio reflects the number of CYPs that can be attributed to the intervention. All CYPs (373) from new DMPA acceptors are included, along with 20 additional CYPs that represent increased contraceptive protection contributed by women who switched to DMPA from the less effective pills and condoms. The remaining CYPs (1,813) can be attributed to the intervention if we assume that these continuing users would not have returned to the clinic for DMPA services. Figure 4 shows the cost per CYP that can definitively be attributed to CHW provision (US$61.89) and the change in cost per CYP at different levels of DMPA continuation. If we assume that 50% of existing DMPA users continued with their method solely because of the improved access afforded by CHW provision, cost per CYP would be lower, at US$21.24. If all users continued because of CHW provision, cost per CYP would be US$11.03.

Safe, Feasible, and Acceptable
This pilot study contributes to the body of evidence on CHW provision of injectable contraception. As detailed in the Results section of this article, the findings establish the safety, feasibility, and acceptability of CHWs providing DMPA in the Zambian context.

Impact on Method Use and Choice
The findings also demonstrate the impact of providing DMPA through CHWs on method use and choice, namely that a sizable number of women became new acceptors of all the methods provided by CHWs. Women were also able to switch to a more desirable method and/or service delivery setting.

Other Factors Affecting Uptake
The project cannot take full credit for the increase in DMPA use in the 2 districts following initiation of the pilot study since the method was also available at 44 health clinics. However, it was a noteworthy increase, made all the more striking by the fact that provision by the 40 CHWs accounted for more than half of the CYPs reported by the District Health Offices from  February 2010 to February 2011. This proportion reflects a shortage of health center workers, who were outnumbered by the CHWs, but also points to a pent-up demand for DMPA and women's desire for easier access to this very popular method.

Additional Benefits of CHW Provision
The benefits of serving clients at the community level go beyond the increase in DMPA use. For example, it can lighten the burden in health centers, leaving nurses more time to provide services that require a higher level of training. For their part, CHWs can learn new skills and build capacity by bringing expanded family planning services to the community. For clients, the reduction in travel and wait times and better access to family planning, including DMPA, can result in fewer women lost to follow-up. Indeed, the use of CHW-provided DMPA at 12 months in this study is significantly higher than that for oral contraceptive pills and is slightly higher than the worldwide norm: typical 1-year continuation rates for DMPA (and pills) are usually between 50% and 60%. 12

Cost
Since, as stated above, we do not know the true impact of CHW provision on continuing use of DMPA among family planning users who initiated at clinics, we cannot say with certainty whether the incremental cost per CYP is closer to US$61 or to US$11. Our results showed that cost per CYP declines rapidly with small increases in the proportion of users for whom the convenience of CHW resupply improves DMPA continuation. Ministry-level support for scaling up community-based provision of DMPA was very strong in the immediate post-study phase. This was due in large part to deliberate efforts to engage key stakeholders and influence decision makers from the earliest stages of the research process in 2009 and to maintain that engagement throughout the entire study. Without such concerted efforts to involve stakeholders from the beginning of the study-and the FPTWG's pivotal role in recommending policy changes and scale up in Zambia-it is unlikely that the translation of this study's findings into practice would have occurred so rapidly.

Study Tour to Sustain Momentum
However, the general elections immediately following the October 2011 stakeholder meeting led to changes in leadership within the MOH and the creation of a new Ministry of Community Development, Mother & Child Health. These changes decelerated the momentum of the early scale-up process. FHI 360, ChildFund Zambia, and USAID/Zambia worked to orient new leadership to the project and cultivate a renewed sense of ownership. As part of these efforts, a delegation of Zambian stakeholders traveled to Rwanda for a south-to-south tour to observe Rwanda's robust community-based family planning program and engage with stakeholders around important policy-level and operational issues. The study tour effectively improved country-level ownership of the replication process and allowed professional bodies, donors, implementers, and key personnel from both ministries to learn from Rwanda's experience and develop plans for moving ahead with CHW provision of DMPA in Zambia. Many stakeholders are strongly advocating a policy change in 2013.

Replicability to Other Settings
It should be noted that this pilot study was implemented within an established CHW program, operated by an NGO that already had a family planning program in place, and with clients who actively sought these services from their CHWs. As such, the intervention was carried out in what could be considered an ideal setting. For that reason, our positive results may not be replicable to the same degree, especially if similar conditions and political will are absent. Nevertheless, this pilot demonstrated: 1) the value of investing in a program where the need for DMPA is ably addressed by a trained cadre of lower-level family planning providers, and 2) the successful expansion of CHW provision of DMPA through effective and continuous collaboration of research, practice, and advocacy.

Sustaining Commitment
In Zambia, the potential for this practice to be widely replicated and sustainable is increased in part by a resurgent global interest in family planning. Country-level commitments and support arose out of the highly visible 2012 London Summit on Family Planning. Among them, the GoZ pledged to increase contraceptive prevalence through various strategies, including reducing barriers to task sharing and doubling budget allocations for family planning. Most recently at the 2013 Women Deliver conference, the First Lady of Zambia, Her Excellency Dr. Christine Kaseba-Sata, emphasized her commitment to creating a supportive environment for task sharing and ensuring the scale up of CHW provision of family planning, including DMPA. Scale up has already begun in 3 districts with USAID/Zambia funding and is expected to continue, with appropriate adaptations that will facilitate large-scale expansion, especially as government ownership and funding increase.

Indications of Future Demand
Prospects for increasing CHW-initiated and resupplied DMPA in Zambia are very promising. (For a description of the scale-up work already underway in Zambia, see the box.) Once the word spread that certain CHWs were providing DMPA, women flocked to them for the method, including residents of Nangoma, Mumbwa, where the health center is affiliated with the Catholic Church and family planning services are not provided. In Luangwa, some women (not included in our data capture) came to our trained CHWs from neighboring Zimbabwe and Mozambique, since the GoZ provides these services free-ofcharge to anyone. Therefore, it was easy for CHWs to find women who wanted to use DMPA.

LIMITATIONS OF THE STUDY
As is often the case, family planning statistics obtained at the district level may be incomplete, but to the best of our knowledge all health centers in the districts were included. Nevertheless, it is possible that not all clinics and all CHWs reliably and consistently submitted their monthly forms to the DHO for the time period under study. As such, the data may underrepresent the true contributions of clinics and/or CHWs who were not involved in the pilot.