Long-acting reversible contraceptives (LARCs), including implants and the copper intrauterine device (IUD), present positive characteristics, including high effectiveness, long duration of action, few contraindications, and reversibility with a prompt return to fertility ¹ . They do not depend on user adherence and do not require daily attention or pericoital use. Additionally, compared to short-acting methods like pills and injectables, LARCs require fewer visits for re-supply. At the same time, while developments with injectable contraceptives allow for self-injection in a growing number of countries ² , LARCs must be inserted by a trained provider. Implants and IUDs also require removal by a trained provider at the end of their labeled duration of use or at any time of the user’s choosing. As such, ensuring access to removals plays a critical role in informed choice in family planning ^{3, 4} .

With the growing popularity of implants in sub-Saharan Africa, the need for removals is accelerating ⁵ . Yet some data indicate that service capacity for implant removal may lag behind that for insertion ^{5, 6} . While ease of removal has increased with 1- and 2-rod implant technologies compared to the six-rod Norplant systems, implant removals remain more technically demanding than insertions, especially in cases involving difficult removals such as when the implant was deeply inserted, has migrated, has broken, or has become encapsulated. Removal may also become challenging in cases where clients have gained or lost weight, making the implant more difficult to locate ⁷ . Clinical challenges are closely related to programmatic challenges associated with scaling-up the availability of both implant insertion and removal services: providing sufficient training to health care providers, including in difficult removal techniques such that providers are confident in their removal skills; ensuring that appropriate equipment and supplies are available for both routine and difficult removals; and establishing necessary referral mechanisms as needed, including for difficult removals or where providers may insert but not remove implants ^{5, 7} . A few recent studies which document the client perspective in relation to implant removals in sub-Saharan Africa reiterate these challenges ^{8– 13} .

Compared to implants, IUD utilization tends to be lower in sub-Saharan Africa, even though copper IUDs have longer efficacy and have been available for decades. Recent targeted efforts to re-invigorate IUD use have often been focused on specific populations, including postpartum/postabortion users and integrated with HIV services, rather than the general population ¹⁴ . This may begin to change as hormonal IUDs become more widely available, particularly in the public sector ^{15, 16} . However, while service delivery patterns of IUD uptake have been documented in the literature, there are few studies describing women’s experiences seeking IUD removal ^{8, 17} or provider experiences providing care.

To our knowledge, few studies capture provider experiences providing implant and IUD insertions and removals in sub-Saharan African settings. Available evidence includes a study of medical and nursing students’ self-assessments of their ability to provide implants at the community level in the Democratic Republic of Congo ⁹ and qualitative studies documenting insufficient training, lack of equipment and time, and unclear referral pathways for difficult removals as barriers to providing implant removals in Botswana ¹³ , lack of experience and lack of refresher trainings as barriers to inserting IUDs cited by midwives in Ghana ¹⁸ , and insufficient space to provide IUDs mentioned by providers and stakeholders in Ethiopia ¹⁷ . None of these studies offer a side-to-side comparison of provider experiences with implant and IUD removals.

This study conducted in Senegal assesses potential challenges with LARC removals from the perspective of providers to inform potential adjustments. Senegal was the first African country to introduce Norplant, starting with a pre-introductory trial in 1996 ¹⁹ . Second-generation implants like Jadelle (2-rod levonorgestrel implant) and Implanon (now Nexplanon, a 1-rod etonogestrel implant), introduced in the country in 2009 and 2014, respectively, are currently available. As of 2017, implants represented 31% of the method mix and IUDs 8%, with over 90% of LARCs obtained through the public sector ²⁰ . The specific objectives of this study were to describe provider training and clinical experiences with implant and IUD removals.

We conducted a cross-sectional, mixed-method study including a tablet-based survey and in-depth interviews (IDIs) with providers from all public sector facilities offering LARCs in two districts of Senegal (Dakar Centre and Kolda). The study was initially planned for three districts representing different geographic and cultural contexts; work in the third district was cancelled due to the onset of the coronavirus disease 2019 (COVID-19) pandemic. Providers who had been trained to insert implants and had been at their current facility for at least one year were eligible. We established a listing of all public facilities offering LARCs within the two districts and identified eligible providers within them. Sample size was driven by the number of facilities; we aimed to survey between one and two providers from all facilities offering LARCs depending on the number of providers that were eligible and available. Following evidence that 80% saturation can be achieved within 8 IDIs, we purposively selected four facilities per district that had at least two eligible providers and represented different levels of the health system (health centers and health posts) and invited one provider from each of these facilities to participate in an IDI, for a total of 8 providers ²¹ .

The study team worked with facility managers to gain access to health facilities, identify all eligible providers, and introduce the study team to the providers. Senegalese female research assistants hired as consultants approached selected providers in-person and interviewed consenting providers individually and in private at health facilities in French between January 22 and March 18, 2020. Trained research assistants captured survey data on tablets; two separate research assistants with advanced degrees and prior qualitative experience conducted, audio-recorded, and transcribed IDIs in French. Research assistants recorded observations as field notes that were reported into the transcripts. Participants had limited knowledge about the research assistants and transcripts were not returned to participants for review.

Survey and IDI topics included training, client counseling at insertion and removal, and clinical experiences with insertions and removals. The survey also explored if providers had ever encountered specific obstacles to removals, while IDIs included additional themes related to equipment and supplies, referral dynamics, and fee structure. Both instruments included questions regarding difficult removals, defined as removal of non-palpable implants and of IUDs with non-visible strings. We developed the data collection instruments based on the existing literature and adjusted them through peer review and extensive discussions with Senegalese study staff and partners. To identify potential problems in the design of questions and likely response options, research assistants conducted cognitive interviews with 10 providers who met eligibility criteria but were recruited outside of the study area. The cognitive interviews were conducted using a rapid iterative approach, observed by one of the investigators (AB), and documented through notes. To inform final revisions prior to data collection, research assistants conducted a pre-test of both instruments with another set of providers with characteristics similar to those of study participants. The final study materials used can be found as Extended data ²² .

Providers gave written consent prior to interviews. The Comité National d’Ethique pour la Recherche en Santé in Senegal (SEN19/65, approved November 7, 2019) and FHI 360’s Protection of Human Subjects Committee in the United States (1383816-2, approved June 19, 2019) approved the study.

We performed descriptive analysis of survey responses using Stata version 15 (RRID:SCR_012763). Typed transcripts were uploaded into NVivo 12 (RRID:SCR_014802) for coding and applied thematic analysis. Two analysts (MML and HD) coded transcripts using a codebook combining a priori codes identified based on informational needs and data-driven codes that emerged from the initial reading of transcripts, with periodic verification of intercoder agreement. Three analysts (MML, HD and VL) then developed analytic memos exploring patterns in the data.

We found that all providers who had been trained in implant insertion had also been trained on removal, and that all but one provider offering IUD insertion had received training on IUD removal. Moreover, close to half of providers had received some training on removals in the two years preceding the survey. Provider confidence with both implant insertion and removal was high, though qualitative interviews noted increased opportunities to practice removals during pre-service education and in service as potential areas for improvement. Confidence levels with implant insertion tended to be higher than for IUD insertion, and providers were overall more confident about IUD removals than IUD insertions. This finding may reflect inherent technical differences between procedures. In that sense, it is generally consistent with another study reporting on average time spent performing these different procedures in Nigeria, which found similar amounts of time required for inserting implants and IUDs (7–11 minutes), but more time to remove implants (14–21 minutes) compared to IUDs (4 minutes) ²³ . Additionally, this finding may be partly explained by the fact that demand for implants is higher than IUDs in the population overall, so providers have more regular opportunities to practice their skills. Innovations in the implant trocars, such as with Nexplanon, have made problematic deep insertions less likely ⁷ . IUD insertions (and removals) may also require the need for more dedicated, private facility space, which was noted as a barrier to IUD provision in Ethiopia ¹⁷ .

Current service guidelines in Senegal include guidance on side effect counseling at insertion but no recommendations that providers counsel on where and when to seek removals. While providers reported informing clients of where to go for LARC removal, our findings indicate that discussion on when to seek a removal, including the option to remove the method at any time, may be more limited. CIMCs and non-bleeding side effects are a key reason for discontinuation, including for LARCs ^{24, 25} . While almost all providers offering LARCs reported counseling clients on CIMCs and other side effects, we found variability in the specific topics covered during counseling, leading to gaps in providing clients with comprehensive information. Counseling on irregular bleeding and amenorrhea, two likely CIMCs associated with implant use, was common but not universal; only a little over half of providers reported counseling IUD users on the likelihood of experiencing heavy bleeding, a common side effect with this method. High quality family planning counseling, including information on possible side effects and how to manage them, is associated with method continuation ^{26, 27} . In addition to method benefits and mechanisms of action, counseling materials could be strengthened by the addition of comprehensive information on possible side effects and on the option to remove LARCs at any time. Including job aids such as the NORMAL tool should be considered when updating materials ^{28, 29} .

Quantitative and qualitative findings indicate that many providers counseled clients seeking removals to keep their method. Qualitative findings further highlight differential attitudes toward various reasons for removals, with providers for example being inclined to remove implants when a client presents seeking pregnancy or with weight gain or hypertension but preferring to first offer reassurance counseling or treatment for CIMCs. Potential provider misconceptions about method eligibility warrant attention, such that providers discussed high blood pressure as a contraindication to implant use in qualitative interviews while this is not been substantiated by evidence ³⁰ . Reassurance by providers may contribute to acceptability of CIMCs ³¹ ; however, even if potential concerns around health effects can be alleviated, CIMCs can disrupt women’s domestic, religious, or work life, though some may see benefits to reduced bleeding, such as amenorrhea ^{15, 31, 32} . Counseling job aids and other tools for reassurance counseling and side effects management when clients are interested in keeping the method should be made available. Messaging around voluntary discontinuation should also be reinforced through formative supervision and onsite coaching updates.

Our findings also provide some evidence of systemic challenges, including lack of consumables (particularly for implant removal) and providers being too busy to attend to clients. Similar constraints were documented through the experiences of implant users in Ghana ¹⁰ . In addition, we found that providers may be reluctant to perform a removal for clients who only had their method for a short period of time. This perspective was also documented in other studies, potentially due to concerns about wastage of products ¹⁰ . Systemic barriers and potential biases warrant attention as part of a comprehensive approach to support quality removal services and ensure full, free, and informed choice.

Currently limited evidence exists on difficult LARC removals in low- and middle-income country settings, though issues with nonpalpable/deeply inserted implants and broken or bent rods are not uncommon ^{10, 33, 34} . In this study, about three quarters of providers had encountered removals involving deeply inserted implants and one third of providers who had ever removed an IUD had encountered removal cases with non-visible strings. Although many providers reported having successfully performed removals in these situations, there was also evidence of referrals. Quantitative findings indicate gaps in confidence compared to regular removals and qualitative interviews reveal some difficult experiences. More research is needed to understand the prevalence of difficult removals and quality and continuity of care for women seeking removals.

While this study brings valuable insights to improve removal services, some limitations must be acknowledged. A possibility of recall bias and courtesy bias exists whereby providers may report inaccurately based on their memory of past events or of what they may perceive as undesirable attitudes or behaviors. The selection of providers for this study is mainly purposive and not intended to support statistical generalizability; however, the results from this study provide evidence that can help program managers in Senegal identify challenges and opportunities for strengthening LARC service provision.

This study conducted in two districts of Senegal illuminates many similarities in provider experiences with implant and IUD removal services and demonstrates strong capacity to manage regular LARC removals. The Direction de la Santé de la Mère et de l’Enfant at the Ministry of Health led the presentation of results to stakeholders in Senegal, including regional and district health teams and implementing partners. Several recommendations emerged from this dissemination, including strengthening counseling materials with anticipatory guidance on CIMCs and non-bleeding side effects, developing guidance on side effect management, emphasizing quality of care for removals through formative supervision and onsite coaching updates, and addressing gaps in the supply chain, notably around the availability of consumables. In order to ensure high quality care, support for management of difficult removals also warrants ongoing attention in Senegal and other countries.