The pilot project was implemented in two rural health districts: Dandé in Hauts-Bassins region and Tougan in Boucle du Mouhoun region, which include rural areas with high unmet need for FP ( 23.8% vs. 17.4% in urban areas) and significant potential to increase access to FP. Events that sensitized the community to FP were held in these districts. The pilot implemented task sharing of short-acting methods to community health workers (CHWs) and long-acting methods to four health cadres lower than doctors and clinical officers: registered nurses, registered birth attendants, auxiliary birth attendants, and mobile health workers. Registered nurses are authorized to provide promotional, preventive, curative and rehabilitative services, while registered birth attendants are authorized to counsel, inform, educate and provide comprehensive care as it relates to reproductive health. Auxiliary birth attendants and mobile health workers assist registered birth attendants and registered nurses, respectively, in their authorized tasks. The four are collectively referred to as primary care cadres in this paper and are distinguished as a group from the lay cadre, CHWs.

In total, 79 primary care providers from 26 public health centers in Dandé and Tougan were trained to provide quality FP counseling at facility level, implant and intrauterine device counseling, insertion and removal of these devices, while 128 CHWs affiliated with those health centers were also trained to provide comprehensive FP counseling, prescribe OCPs safely, administer the subcutaneous formulation of injectable contraceptives (DMPA-SC) and deliver key messages at community events like market days ( Table 1). These activities constituted the capacity-strengthening and demand creation components of the pilot intervention ¹⁸ . CHWs began providing services earlier than both groups of primary care providers--by January 2017 after being trained in November and December 2016. Primary care providers in Dandé began providing services in February 2017 (training January to February 2017), while their counterparts in Tougan initiated service provision in April 2017 (training December 2016 to March 2017). By the end of April 2017, all trained providers in both districts had also received follow-up training and supervision.

The objective of the evaluation was to assess whether task sharing long-acting FP services with primary care cadres and short-acting FP services with CHWs is feasible and can increase uptake of high quality, safe and acceptable FP services in Dandé and Tougan districts. The following indicators guided data collection, analysis and interpretation:

A mixed methods descriptive evaluation utilizing service statistics and client exit interviews (quantitative) as well as in-depth interviews (qualitative) was conducted between December 2017 and May 2018—commencing several months after primary care providers began providing LARC methods, and almost a year after trained CHWs began providing pills and injectables. This gave intervention sites sufficient time to provide follow-up supervision and to standardize pilot project procedures before initiation of the external evaluation.

A total of 27 in-depth interviews (IDIs) were conducted with a subset of primary care providers and CHWs, community representatives, as well as district and national level MoH personnel. We conducted 425 client exit interviews with women who received methods from providers trained for the pilot project. Service statistics of FP uptake obtained for Tougan and Dandé districts were analyzed to determine changes in uptake over time.

The quantitative component: family planning clients and service statistics. All users new to their chosen long-acting or short-acting method, and who received FP services from a provider trained for and affiliated with any of the 26 project sites, were eligible and approached to participate in the evaluation. Surveys focused on satisfaction with their chosen method and the services received from their providers. To determine the desired sample size, we estimated the proportion of women who would report that their provider talked to them about the possibility of side effects associated with FP method use (a key indicator of service quality). We used a base estimate of 75% for this indicator, resulting in a minimum sample size of 284 clients to estimate this indicator within 5% with a 95% confidence interval. Based on experience, we assumed that at least 75% of providers would discuss the possibility of side effects, because they were trained and received follow-up supervision for the pilot intervention. We did not anticipate needing to use the more conservative standard, a 50% base estimate.

We obtained service statistics from the MoH on FP uptake before initiation and during implementation of the pilot project intervention (calendar years 2016 and 2017) to allow comparisons of FP uptake before and during the intervention. Dandé’s service statistics covered 11 months of the intervention; Tougan’s statistics covered nine months due to the two-month difference in initiation of the pilot.

The qualitative component: primary care cadres, community health workers, community representatives and government officials. With the assistance of Consortium members, we purposively identified and interviewed pertinent individuals for key informant interviews. Two community representatives from Dandé and two from Tougan along with a total of seven district-level and national-level officials from the MoH ^a provided feedback on sociocultural, normative (community representatives) and high-level, administrative perspectives (MoH officials) on FP services.

In-depth interviews were conducted with primary care providers and CHWs to understand their respective experiences and views of task sharing FP services. Via convenience sampling, we selected one each of registered nurses, registered birth attendants, auxiliary birth attendants and mobile health workers from both districts, as well as four CHWs each from Dandé and Tougan districts.

Family planning client interviews. Data collectors, with assistance from participating primary care providers at public health centers and CHWs in the evaluation catchment areas, identified FP clients who were of reproductive age, adult or emancipated if under 18 years of age (i.e., married), and met the criterion for new contraceptive user (recently initiated an FP method they accepted from an intervention-trained provider between December 2017 and February 2018). At health centers, LARC clients who expressed interest in participating in the evaluation were directed to data collectors posted on-site. In health center catchment areas, CHWs informed eligible acceptors of injectables and OCPs about the evaluation. The name and contact information of CHW clients who were interested in participating were promptly given to data collectors who then contacted clients for an interview within one week to reduce recall bias. Before initiating the interview, data collectors confirmed that each client met eligibility criteria and obtained informed consent.

Surveys were administered in French via electronic tablets, with explanations provided as needed to participants in the local languages (Dioula or Mooré). All data collectors were fluent in the three languages. Family planning clients were interviewed between December 12 ^th, 2017 and February 4 ^th, 2018 in designated areas that provided audio, and if possible, visual privacy within evaluation facility catchment areas. Interviewers obtained information on clients’ sociodemographic characteristics, experiences using contraceptive methods, factors in client decision to use their chosen method, interaction with and information provided by the primary care provider or CHW, and client satisfaction with her choice of method and with the provider.

In-depth interviews. Both types of providers, community representatives and government officials were asked for their opinions on task sharing in general, and specifically, how the process functioned in the Consortium’s pilot project. Interviews captured perceptions of demand for FP services and the role task sharing plays in creating demand, perceived challenges and successes of task sharing, the availability of FP stocks/commodities, community acceptance/non-acceptance of task sharing FP services, primary care provider and CHW workload and motivation, appraisal of provider training and supportive supervision related to task sharing, and recommendations for scale-up of task sharing in Burkina Faso.

Interviews were conducted in French or the local language using a hard copy interview guide (see extended data ¹⁰ ). Interviews were also audio recorded to produce transcripts. Primary care providers were interviewed between December 11 ^th and 21 ^st, 2017 in a private location within the facility. Interviews with the eight CHWs were conducted between December 14 ^th and 24 ^th, 2017 and were also located in a private area within the evaluation facility. The four community representatives were interviewed between December 17 ^th and 19 ^th, 2017 in their homes, while the government officials were interviewed between January 31 ^st and March 19 ^th, 2018 in their offices.

During the evaluation, access to hard copy and electronic data was granted only to staff at ISSP and FHI 360. Informed consent forms signed by evaluation participants were stored in a separate locked drawer or cabinet. Electronic data were stored in password-protected files. Upon completion of the evaluation, all stored materials were destroyed at ISSP. All electronic data were transferred to FHI 360.

Quantitative data. Client survey data were cleaned and analyzed in Stata 15 ¹⁹ . Frequencies, means and crosstabulations were computed. Health Management Information System (HMIS) data on FP uptake were organized in an Excel spreadsheet (Office 365 v.1808 ²⁰ ) for descriptive analysis and were represented graphically to illustrate changes over time.

Qualitative data. Qualitative data gathered through IDIs with providers, community key informants and government representatives were analyzed using an applied thematic analysis approach ²¹ . A team of two qualitative analysts created a structured codebook for each type of interview and tested them on the first few IDIs available for analysis, and coded all transcripts in NVivo 11 ²² . Intercoder reliability was established at 92% ²³ . Analysis memos were developed to summarize findings related to the interview domains.

FHI 360’s Protection of Human Subjects Committee (PHSC) granted this evaluation (Project #: 1106971) research exempt status according to the requirements under 45 CFR 46.101. Burkina Faso’s Comité d’Ethique pour la Recherche Santé (CER-Ethics Committee for Health Research) in the Ministry of Health does not exempt any health research involving human subjects. CER approved the evaluation (Deliberation No. 2017-11-173) without reservations or recommendations. All participants voluntarily agreed to take part in the evaluation following the written informed consent process executed by trained data collectors.

A total of 425 new FP clients were interviewed for the evaluation (see underlying data ¹⁰ ). The average age of participants was 27.8 years with women ranging from 17 to 49 years old ( Table 2). Most women were married, had not attended school and had already given birth. The average age of the youngest child was about one year old. More than 75% wanted to have a baby sometime in the future, with 57% wanting that to occur more than two years from the time they were interviewed. Another 15% wanted to give birth within the next two years.

About 47% of participants were LARC acceptors recruited from clinics and the remaining 53% were clients of CHWs who chose injectable contraception or pills. Just over 46% were using a contraceptive method for the first time. Injectable contraception and implants were used most commonly, followed by IUDs and pills ( Table 2).

With regard to the key outcome variable, 85.7% (95% CI: 82.3% to 89%) of clients reported that the provider discussed possible and normal side effects associated with use of their chosen method. This figure was significantly greater than our base estimate of 75%, which allowed us to obtain a more than adequate sample size.

Feasibility of the pilot intervention was assessed by asking primary care providers and CHWs about changes in their workload, the integration of clinic activities with task sharing of IUD and implant services, and if task sharing created or exacerbated stockouts of commodities. Government officials were asked if they thought the task sharing intervention worked well, while community representatives were asked if they felt the intervention met the needs of the people.

Over half of providers interviewed stated that their workload increased as a result of participating in the pilot intervention. Among primary care providers, some described spending more time counseling clients or having more clients; some mentioned that having few providers in the health facilities increased their workload. Nevertheless, most said that they were not over-burdened by the increases and some mentioned they were gaining useful experiences through the pilot.

CHWs tended to describe larger or more burdensome increases in their workload and noted that they had responsibilities that were different from and in addition to their existing work. Others noted that they experienced challenges with transportation, accessing clients and not being paid for their work in the pilot.

Regarding service integration in primary care facilities, it appeared that increasing access to LARC provision may have overwhelmed some facilities but did not affect others. Some providers stated that providing LARCs did not affect the clinics’ other activities and that they were able to provide FP methods during consultations for other issues or refer to the maternity unit. Others mentioned that women who came to facilities to obtain IUDs or implants during the pilot had to wait or come back when a provider was free. Some respondents said that the pilot increased wait times at the clinic.

When asked about stockouts, a few primary care providers described current or prior issues with having stockouts of implants and basic materials. A few primary care providers mentioned that women went to other facilities for FP methods due to stockouts. Only two CHWs mentioned issues with stockouts; the majority mentioned that they did not have any problems or avoided stockouts by being proactive about ordering additional supplies before they ran out.

MoH officials noted that the pilot intervention increased the availability of and demand for LARCs, and they considered that a success. One government official noted that in the regions where the program was active, there was someone capable of providing LARCs at every health center, while in the communities, there was a health worker who could offer pills and injectables. Ministry officials also noted the greatest challenge as the lack of financial incentives for providers, especially CHWs. One government official also noted that provider attrition was a problem, with trained providers leaving and being replaced with untrained counterparts.

Community representatives felt that the intervention met the needs of the people, particularly because women did not need to travel far for FP services. Community sensitization to FP that accompanied the pilot intervention was also noted as helpful, as it motivated men to get on board, which empowered women to plan their families.

Acceptability as measured by client satisfaction. Client reports reflected satisfaction with services and with providers. About 84% of FP acceptors reported being very satisfied with their method (with slightly fewer IUD users very satisfied), while 12% were somewhat satisfied and less than 1% were not satisfied with their chosen method ( Table 3). With regard to satisfaction with overall services, pill users were most apt to be very satisfied, followed by injectable, implant and IUD acceptors. A similar picture emerged when clients were asked how satisfied they were with their particular provider’s services.

Examination of satisfaction indicators by provider type ( Table 4), indicated that the clients of CHWs and primary care providers were equally satisfied with their chosen method and overall services. However, when asked how this experience of initiating a new FP method compared with their usual experience of receiving health care services, more clients of primary care providers reported that it was a better experience than clients of CHWs (75.0% vs. 62.1%, p=0.004).

Community perspectives. Both community representatives from Dandé described positive community attitudes towards CHWs providing short acting methods. They reported that community attitudes had changed and become more positive towards FP use and that misperceptions had decreased due to the pilot.

One representative from Tougan noted that some women were hesitant because CHWs were not formal healthcare providers and recommended that they be accompanied by a formal healthcare provider to build trust in their abilities. The other representative stated that community attitudes were mixed. For example, the representative mentioned that some men are supportive while others are not, and some community members have negative attitudes towards FP in general.

Nevertheless, both community representatives from Tougan agreed that there had been an increase in FP use since the beginning of the pilot.

Provider perspectives. Overall, primary care providers and CHWs expressed positive views of task sharing; no participants aired negative viewpoints. Some providers stated that task sharing allows for increased accessibility of contraceptive services to women who are not located close to a clinic, leading to positive reproductive health outcomes and allows women to receive contraceptive services more quickly.

Two CHWs stated that they like that they are able to provide contraceptive services and/or that they enjoy the work that they do.

Providers were similarly satisfied with their ability to provide services in the form of FP methods and counseling without having to refer clients to another provider. Some also mentioned the satisfaction of making a positive impact on their community. Nearly all providers stated that satisfaction with their job increased due to the pilot.

Quality of care assessed from the client’s perspective is fairly high ( Table 5). Over 90% of clients said that the provider spoke to her in a friendly way and over 98% of those who asked their provider questions (about 50% of clients) reported that all their questions were answered satisfactorily. Further, more than 90% of clients of both primary care providers and CHWs reported that they talked to them about all four of the contraceptive methods in question. Other methods such as condoms, emergency contraceptive pills, Standard Days Method, withdrawal, folk remedies/herbs and the lactational amenorrhea method were also mentioned by 10% or more of FP clients, but overwhelmingly, counseling was focused on injectables, pills, implants and the IUD (see underlying data).

Over three quarters of FP initiators recounted that counseling included discussion of advantages and disadvantages, danger/warning signs, possible side effects (including menstrual irregularities) and instructions on what to do if problems or side effects are experienced ( Table 5). In general, clients of CHWs were less likely than primary care clients to report that these issues were discussed. Notably, CHW clients were more likely to have made the decision to use their new method alone, while clients of primary care providers were more likely to have shared that decision with their provider.

Very few FP clients reported any outcomes indicative of unsafe contraceptive method administration or use. All incidents of abscesses or infections were reported by injectable clients, and those events were reported by less than 1% of this group.

Health Management Information System monthly report forms recorded, among other phenomena, indicators of unsafe provision of the four FP methods under study: needle sticks to providers and to clients, complications of implant or IUD insertion, and unspecified undesirable effects associated with pill, injectable, implant or IUD use. Many more unspecified undesirable effects were reported compared to complications with implant insertion and needlestick injuries to a provider or a client ( Table 6).

As illustrated by Figure 1 and Figure 2, there is no general pattern of overall and continuing increase in FP uptake among new acceptors after the initiation of and during the pilot intervention. This is especially evident in Tougan. However, in both Tougan and Dandé, there is a notable increase in implant uptake shortly after the initiation and during the first two to four months of the pilot intervention. These increases coincide with the two occasions in the year (May and November) that the government provides free contraception. Otherwise, the general pattern in both districts displays periodic increases (and decreases) in implant and injectable uptake among new users. The uptake among pill and IUD acceptors shows small increases since the pilot began, but it is relatively flat compared to the more popular methods of implants and injectables.

We cannot provide transcripts or summary notes of the interviews because of the very real possibility of participant identification. Quantitative data for FP clients is available via Harvard Dataverse.

Harvard Dataverse: External evaluation of a pilot intervention to increase access to family planning services in Burkina Faso. https://doi.org/10.7910/DVN/PSIK4Q ¹⁰

This project contains the following underlying data:

Harvard Dataverse: External evaluation of a pilot intervention to increase access to family planning services in Burkina Faso. https://doi.org/10.7910/DVN/PSIK4Q ¹⁰

This project contains the following extended data:

Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).