Previously titled : Cost per insertion and couple year of protection for post-partum intrauterine devices and implants provided during service scale-up in Kigali , Rwanda

In two high-volume government hospitals, their two affiliated Introduction: health facilities, and two additional health facilities, we developed and implemented postpartum intrauterine device (PPIUD) and postpartum (PP) implant promotional counseling and service delivery procedures between May-July 2017 in Kigali, Rwanda. Between August 2017 and July 2018, 9,073 pregnant women received PPIUD/PP implant promotions who later delivered in one of our selected facilities. Of those, 2,633 had PPIUDs inserted, and 955 had PP implants inserted. The goal of the present analysis is to detail implementation expenditures and estimate incremental costs per insertion and couple years of protection (CYP) for PPIUD and PP implant users. We detail the incremental costs during the implementation from Methods: the health system perspective (including both the implementation costs and the cost of contraceptive methods) and use of standard methods to estimate the cost per insertion and CYP for PPIUD and PP implant users. In addition to the incremental costs of labor and supplies, the costs of promotional activities are included. Research costs for formative work were excluded. A total of $74,147 USD was spent on the implementation between Results: August 2017 and July 2018. The largest expense (34% of total expenses) went toward personnel, including doctoral-level, administrative, data management and nurse counseling staff. Training for PPIUD and implant providers and promoters comprised 8% of total expenses. Recruitment and 1,2 3 2


Introduction
Voluntary family planning (FP) is one of the most cost-effective public health interventions, reducing both maternal and child mortality and improving national economies 1 . However, there is high unmet need for family planning in the developing world. In postpartum periods, 61% of women across 21 low-and middle-income countries experienced unmet need 2 , while 95% of women desire to avoid pregnancy for at least 1 year after delivery 3 . In Rwanda, although only 2% of postpartum women report a desire for another child within 2 years of delivery, the unmet need in the postpartum period is 51% 4 .
To meet women's postpartum fertility goals and improve maternal-child health via birth spacing or limiting 5 , the Rwandan government has made postpartum family planning a key objective of the Rwandan Family Planning 2020 Commitment (Objective 2: 'Scale up the postpartum family planning (PPFP) in all health facilities in Rwanda to increase method choice including access to long term methods…') with the goal of preventing 250,000 unintended pregnancies annually 6 .
Long-acting reversible contraceptive (LARC) methods (the hormonal and copper intrauterine device (IUD) and hormonal implant) are not only the most effective reversible methods (lasting 10 and 3-5 years, respectively, with typical use failure rates <1%/year), but are very cost-effective 7-11 . A copper postpartum IUD (PPIUD) can be inserted immediately after delivery of the placenta, during a cesarean delivery, up to 48 hours after childbirth, or beginning at 4 weeks after delivery 12,13 (https://www.mcsprogram.org/resource/pathway-of-opportunities-for-postpartumwomen-to-adopt-family-planning/). A postpartum (PP) implant can be inserted any time after delivery (https://www.mcsprogram. org/resource/pathway-of-opportunities-for-postpartum-womento-adopt-family-planning/). and the WHO Medical Eligibility Criteria were recently revised for postpartum implants 14 . However, IUDs and implants make up a relatively small share of method use in Rwanda (2.5% and 16.9% of the method mix, respectively) (http://www.familyplanning2020.org/entities/81).
This relatively low uptake is thought to be related to lack of method promotions to both women and their male partners as well as limited provider comfort counseling on and delivering these methods [15][16][17][18][19] . Because baseline knowledge about the LARC methods among potential users is lower than for other methods 20-24 , demand creation strategies must include comprehensive information addressing method benefits, side-effects, and misconceptions 20,25,26 . To address these issues, funding from a Bill and Melinda Gates Grand Challenge Award was received to improve PPIUD supply and demand in Kigali, Rwanda, with supplementary funding from Emory University to provide PP implant services. Briefly, in two large health centers (providing antenatal care (ANC), family planning, and infant vaccination services), their two adjoining referral hospitals (providing routine and complex labor and delivery), and two additional large health centers (providing ANC, family planning, routine labor and delivery, and infant vaccination services), Emory-based non-governmental organization Projet San Francisco (PSF) developed and implemented PPIUD and PP implant promotional counseling and service delivery procedures in August 2017. By July 2018, 9,073 pregnant women received PPIUD/PP implant promotions who later delivered in one of our selected facilities. Of those, 2,633 had PPIUDs inserted, and 955 had PP implants inserted. These published findings represented a significant increase in PPIUD and PP implant uptake versus the 6 months prior to our implementation (p<0.001) 27 .
The goal of the present analysis is to detail expenditures during the implementation and estimate the incremental cost per PPIUD insertion, PP implant insertion, and couple years of protection (CYP) for PPIUD and PP implant users to inform decisionmaking by the Ministry of Health and to estimate the cost of scaling up activities. Importantly, in addition to the costs of labor and supplies, the costs of promotional activities are included when calculating the costs and cost-effectiveness estimates of this intervention because postpartum LARCs are still relatively unknown (this is especially true for the IUD for which baseline knowledge is low 20-24 ) and require a significant investment in demand creation.

PPIUD/PP Implant program development and operations
The PPIUD/PP implant intervention (described in detail previously 27 ) was developed with input from stakeholders, providers, community health workers (CHW), and couples/clients.

Amendments from Version 1
In this revision, we address comments from two expert reviewers which has improved the manuscript. The parent postpartum IUD/implant intervention is now more thoroughly described. We now more explicitly explain the need for promotional activities for postpartum LARCs and discuss why promotional activity costs are important for implementers to consider (in short, because postpartum LARCs are still relatively unknown they require a significant investment in demand creation). We clarify that we used standardized methods to calculate the incremental costs to provide our intervention beyond existing pre-intervention services (which are now also detailed). We more explicitly describe inclusion of labor and supply costs and clarify that reimbursements for providers include the cost of providers' time/labor and were made in addition to their regular salaries. Research costs were not included in our cost calculation. The payer perspective is now explicitly defined as the health system perspective. We discuss that while we have only included costs necessary for a larger scale-implementation, it is not certain whether the estimated cost outcomes would apply to larger scaleup activities. Given the preference for the implant observed in some studies, if commodity costs for implants were reduced this method could become even more affordable for health systems to implement. This is timely given the updated WHO MEC Criteria for postpartum implants. Key barriers to LARC services related to demand creation are now described, and we highlight the lack of other similar studies examining postpartum LARC method CYP. We present CYP data for other methods from the literature in table format, noting that these studies do not include the costs of demand generation activities and that direct comparison across studies is made difficult by the different methods for costing and service implementation. A discussion of the relatively popularity of the IUD versus the implant is also included.

REVISED
Stakeholders included the Rwanda Ministry of Health, the District Mayors, the Rwandan Family Planning Technical Working Group, and clinic directors. Through formative work between May and July 2017, we evaluated knowledge, attitudes, and practices regarding PPIUD/PP implant services among community health workers and providers and clients/couples. This formative work led to the development of intervention operational procedures and a promotional counseling flipchart to be delivered to women or couples. Promotional counseling was conducted primarily by counselors during ANC, labor and delivery, and infant vaccination services or within the community by CHW. In addition, dedicated promoters were hired to administer promotions. In August of 2017, nurses and midwives working in labor and delivery and family planning departments began training in PPIUD insertions (implant insertion training had been previously provided). Clinic staff and CHWs were trained to promote the PPIUD/PP implant services. Follow-up appointments were scheduled for PPIUD clients within 6 weeks after PPIUD insertion (typically coinciding with the 6-week infant vaccination visit).
Pre-intervention postpartum LARC services were conducted by two national PPIUD trainers located at two of our selected district hospitals. One of these national PPIUD trainers was collecting PPIUD insertion and follow-up data in a logbook specifically for PPIUD services. In the 6-months prior to our intervention (from February-July 2017), n=46 PPIUDs were inserted (average of 7.7 insertions/month) and n=182 PP implants were inserted (average of 30.0 insertions/month) in the selected health facilities. The percent increase comparing monthly PPIUD insertions between February-July 2017 to our intervention period of August 2017-July 2018 was 2,687% for PPIUD and 169% for PP implant.

Incremental PPIUD/PP implant program costs
We used a standard, comprehensive micro-costing approach as recommended to calculate the incremental cost of the PPIUD/PP implant intervention from the health system perspective 28 . Using standardized data collection tools, resource use data was collected from expenditure records, study case report forms, and interviews with program implementers. Costs of labor, promotions, and supplies are included as detailed below, and no research costs are included. Thus, the costs included are the incremental costs required to implement the promotional counseling and service delivery intervention above the minimal existing preintervention postpartum LARC services described above.
Part-time salaries and fringe were provided for three Emory staff and the PSF Director. PSF-based personnel included a dedicated physician with part-time support from two project physicians, two study coordinators, a senior nurse counselor, a data manager, and two promotions managers. Per diems were provided for trainees during training activities. Training costs included the costs of training providers to insert PPIUDs during a 2-day didactic training and mentored practical certification process, and the costs of training PPIUD/PP implant promotional agents. Field travel included travel for Emory-based staff and transportation for local staff. Field travel was required to transport staff to trainings (which would be recurring during future implementation stages) and the implementation clinics. Other field expenses included wire transfer fees, transcription and translation services, and meals during trainings. Transcription and translation services were required to produce implementation tools in two of the main languages spoken in Rwanda (Kinyarwanda and French).
Recruitment/reimbursement expenses began in February/March 2018 and included: PPIUD client transport reimbursement for follow-up visits ($2.29 United States Dollars [USD]/client), reimbursements for CHW promoters ($0.57 USD/client presenting their referral when requesting a PPIUD or PP implant), reimbursements for providers ($1.20 USD/PPIUD and $0.57 USD/PP implant insertion), and reimbursements to the selected facilities for administrative costs associated with implementing the PPIUD/PP implant program ($57 USD/ facility/month). CHW and clinic provider reimbursements used the Rwandan performance-based-financing (PBF) system as a guide 29 . Reimbursements for providers included the cost of providers' time/labor to provide insertions. This was provided to them in addition to their regular salary (the average monthly salary for family planning or labor and delivery nurses is $124-364 USD, depending on their education). Communications expenses included internet and phone airtime for staff. Field consumables/office supplies included specula, forceps, batteries, logbooks, chargers for tablets, PPIUD kits and various office supplies. Tablets were used to collect data from logbooks for quality assurance/control.
We also included the cost of methods (estimated from the prices incurred by the United Nations Population Fund (UNFPA) in 2015 of $0.37 USD per copper T380 IUD and $8.93 USD per Jadelle levonorgestrel rod implant (http://mshpriceguide.org/ en/home/), and converted to 2018 USD ($0.39 and $9.49 USD, respectively). Expenditures are reported by activity in 2018 USD.
Only implementation costs related to service provision were included (i.e., we did not include research costs for formative work conducted between May and July 2017). Thus, the expenses presented represent the frontline incremental costs required to implement the program between August 2017 and July 2018 from the health system perspective. No discounting of costs was performed given the short time horizon. We follow the Consolidated Health Economic Evaluation Reporting Standards 28 .

PPIUD/PP implant program outcomes
Outcomes of interest include the number of PPIUDs and PP implants inserted and the cumulative couple years of protection (CYP) for PPIUD and PP implant users. CYP is a commonly used estimate of the length of contraceptive protection against pregnancy provided per unit of that method and is estimated at 4.6 for the Copper T380 IUD and 3.8 CYP for Jadelle (5 year) implant 30 (https://www.usaid.gov/what-we-do/global-health/ family-planning/couple-years-protection-cyp). Using the incremental cost measures and outcomes of interest, we calculated  the cost per PPIUD inserted, cost per PP implant inserted, cost per CYP for PPIUD users, and cost per CYP for PP implant users. No discounting of outcomes was performed given the short time horizon of the 12-month implementation.

Ethical considerations and consent
The Emory University Institutional Review Board (IRB) and the Rwanda National Ethics Committee (RNEC) approved the research component of the project (IRB 00001497). Written informed consent was obtained from all participants prior to enrollment. The Emory University IRB determined the programmatic service delivery component of the project (PPIUD promotions and insertions performed in government clinics) was exempt from review.

Results
Raw data for this study are available in Dataset 1 31 .

Incremental PPIUD/PP implant program costs
Program costs are summarized in Table 1. A total of $74,147 USD was spent on the implementation between August 2017 and July 2018. The largest expense (34% of total expenses) went toward personnel, including doctoral-level (MD and PhD) staff, and administrative, data management and nurse counseling staff. Trainings for PPIUD and implant promotional counselors and PPIUD providers comprised 8% of total expenses. Recruitment and reimbursements comprised 6% of expenses. Costs of implants to the government comprised 12% of the expenses, much higher than the cost of IUDs (1%).

Discussion
Our implementation provided services at a cost per insertion of $25 and $77 for the PPIUD and PP implant, respectively, and CYP of $5 and $20 for the PPIUD and PP implant, respectively. Understanding the cost per PPIUD/PP implant inserted can help to inform decision-making by the Ministry of Health and to estimate the cost of scaling up PPIUD/PP implant service implementation activities. Since cost per CYP is a standard and commonly used measure, our estimates of cost of CYP also help the government to determine contraception funding priorities.
For comparison, in a previous study conducted in Rwanda, 478 PPIUDs were inserted over 15 months in 12 sites at an incremental cost of $95,004 USD. After amortization of training costs over three years, investigators estimated outcomes of $110/PPIUD inserted and $24/CYP for the PPIUD 15 .
Though few additional postpartum contraception studies exist for comparability, other studies (summarized in Importantly, these studies did not include the cost of demand creation activities. The cost of promotional counseling activities is important for implementers to consider when evaluating postpartum and LARC-focused interventions because postpartum LARCs are still relatively unknown and require a significant investment in demand creation. Studies support that such promotional activities should also educate men [35][36][37][38] , as done in this study, which incurs additional costs above focusing promotional activities on women alone. Once social diffusion is achieved and the target population is knowledgeable about postpartum LARC methods, demand creation activities can decrease.
We focused our promotions on the IUD because it is the least well-known method in sub-Saharan Africa, including in Rwanda 21,39-42 , which explains the relatively high uptake of the IUD relative to the implant. Other LARC implementation studies have observed that the implant is more popular than the IUD 43,44 but that this trend shifts after focused IUD educational and counseling efforts, community-based and media efforts, and provider refresher IUD trainings 24,45 . Thus, though the IUD is less well-known versus the implant in much of sub-Saharan Africa 21,39,40,42,46 and providers may have lower baseline comfort promoting and inserting IUDs 17-19 , concerted promotional counseling and training efforts can be successfully employed as was achieved in these examples and our study to increase IUD demand.
There is very limited literature on the cost of the PP implant, making our findings timely especially in light of the WHO Medical Eligibility Criteria (MEC) updates related to the postpartum implant 14 . Women who are <6 weeks postpartum and breastfeeding can use the implant with a MEC of 2 (meaning the method is generally recommended) while all other women can use the implant regardless of breastfeeding with a MEC of 1 (meaning no restrictions on use). Given the preference for the implant observed in some studies 43,44 , if the commodity costs for implants were reduced this method could become even more affordable for health systems to scale-up.

Limitations
Similar to the other studies cited here, we included costs from the health system perspective only; however, we recognize that more detailed costing analyses including the societal perspective including women's time and the value of their time would be informative and may strengthen evidence to increase LARC services (since women are saved time traveling to clinic for OCP refills or 3-monthly injectables). It would have also been informative to estimate the cost per promotional method employed (e.g., promotions occurring during ANC, labor and delivery, infant vaccination, or delivered in the community by CHW), but as many women received multiple promotions from several places and our promotional strategies evolved over time, this was not possible in the present study. Given our short time horizon, we did not amortize our training costs as in another PPIUD/PP implant in Rwanda 15 , though the education provided during trainings may translate into service provision over several years in the future; amortization would have decreased our estimated costs per insertion and CYP. We did not collect the data needed to divide the costs of consumables such as specula by their number of uses to arrive at per insertion costs. Additional supplies such as alcohol pads and gauze were among government supplies used and were not measured or included in our calculations. It is not certain whether the cost outcomes estimated here would apply directly larger scale-up activities. We hypothesize that economies of scale may be gained, for example when training a larger number of nurses simultaneously, but it remains to be seen whether quality services can be provided for the same (or reduced) CYP at scale. Finally, our results are most generalizable to sub-Saharan African countries.

Conclusions
There is consensus in the international community that greater investment in postpartum family planning, and the IUD in particular, is needed. We have developed a successful, multi-level intervention that increases PPIUD and PP implant uptake that has relatively low costs per insertion and CYP. Future analyses will explore whether the intervention is cost-effective (or potentially cost-saving).

Data availability
Underlying data are available from Harvard Dataverse. Dataset 1: Replication Data for an interim evaluation of a multi-level Introduction -most hormonal IUD products are registered for 5 years duration. Please clarify in text.
Introduction -you state that IUDs and implants make up "a relatively small share of method use in Rwanda." However, I'm not sure it's fair to lump these two methods together as 17% for implants seems like a relatively high proportion. Rwanda was not included in the analysis in this recent paper by Jacobstein (2018 ), but still it would be good to more fully acknowledge the growing popularity and use of implants in SSA.
Note some minor typos -e.g. name of Project San Francisco -"project" is misspelled.
Discussion section -I would include mention of this study among PP women in Rwanda: Dulli et al. (2016 ). Related to this suggestion, it sounds like promotion of PPIUD and PP implants in your intervention occurred in part in clinics which offer child immunization services. You may want to highlight this because FP-Immunization integration is considered a "promising" High Impact Practice by USAID, UFNPA and others. See here: . https://www.fphighimpactpractices.org/briefs/family-planning-and-immunization-integration/ Discussion -you noted that in your intervention you "focused your promotions on the IUD" -please 1 2 1.
Discussion -you noted that in your intervention you "focused your promotions on the IUD" -please clarify the meaning here, as it is important to clarify if the IUD was promoted in the context of full method choice.
A summary of my concerns about the methods are as follows: The authors say that they use "a standard, comprehensive micro-costing approach as recommended to calculate the net cost of the PPIUD/PP implant intervention from the payer perspective." They reference the CHEERS economic standards described in the paper by Husereau et al. However, in  In the third paragraph of the Introduction-the category of LARCs also includes the hormonal IUD (the LNG-IUS). Suggest revision to acknowledge this; it could just be a footnote.
Introduction -citation 7 -This is a white paper form 2005. there are many more recent papers that document the effectiveness and cost-effectiveness of FP methods. Suggest replacing with a more recent citation.
Note that the analysis in the paper in citation #18 by Tumlinson et al was updated in a more recent paper. I encourage you replacing the data/citation here this this more recent work. See: Rademacher (2016) . et al.
In the Ethics section, it was confusing how the focus group discussions and surveys related to the costing component, if at all. Consider revising to clarify.
I realize that you describe the intervention in-depth in another paper, but I suggest that you clarify in this article what the baseline service delivery model was for PPFP at the sites before the intervention was introduced. That will allow you describe and calculate the costs of the incremental new PPIUD and PP-implant intervention being evaluated.
The Neukom et al paper that is cited (Citation #18) used a dedicated provider model. Related to point above, it would be good to clarify if dedicated providers were employed as part of the intervention, or it PPIUD and PP-implant were added to the scope of work for existing providers at the facilities. If it was the latter, this would not be a fair comparison to the costs described in the Neukom et al paper as the models are different. And again, the key is to calculate the incremental costs of the intervention you described -so only the incremental costs of having providers add PPIUD/PP-implants to their service package.
Thank you for the chance to review. Again, this is an important topic, and I hope the authors will consider revising and re-submitting. 1.

2.
These recommendations are described in more detail below, along with more minor suggestions. Once these concerns are addressed, I strongly recommend indexing. Thank you for the opportunity to review this paper. :

Abstract
The presents the results of the parent study/intervention which was designed to introduction increase demand for LARCs within a small number of facilities in Kigali. I encourage the authors to re-write this paragraph to better lay the foundation for the specific goal of this current paper. The authors may want to consider a short statement of the high unmet need during the postpartum period and the low prevalence of LARCs and the value of better understanding CYP for LARCs after factoring in the cost of promotional/demand-creation activities.
Similarly, the section of the abstract falls a little bit short; it would be helpful to indicate methods that the authors utilized a standardized method for calculating net cost of the intervention. :

Introduction
Overall the introduction is very well written and pleasantly concise. However, I would recommend insertion of a short paragraph that helps the readers to understand a key challenge of LARCs: promotional or demand creation activities are often necessary to increase uptake. Few prior studies (to my knowledge) have been able to calculate a CYP for a LARC that includes the cost of these demand creation activities. This is an important strength of this paper and should be highlighted.
An additional recommendation for the introduction involves the description of the parent study/intervention. In the second sentence of paragraph four, the authors briefly describe the results of the parent study in terms of the enormous increase in uptake. This is important information; however, it feels as if the authors have cut and paste from the abstract of the prior paper and -on first glance -it was confusing as I didn't realize the authors were describing the intervention study and I mistakenly thought the results of the current paper were being summarized in the introduction. I recommend revisions so that this paragraph does a better job of explaining that there was a parent study/intervention that was found to be enormously successful in increasing the uptake of LARCs in select Kigali facilities and now the authors are writing this paper with the goal of understanding the CYP of LARCs, factoring in the cost of these very effective promotional activities. :

Methods
I would consider moving the first paragraph (ethics) to the end of the methods section, if possible. I was confused to read about focus group discussions in the first sentence, since I didn't yet understand that formative research was done prior to designing effective promotional activities, neither of which are the real focus of this paper.
Are there available data on the cost of the supplies needed for insertion and the cost of provider's time for insertion? For example, in a paper I wrote (which the authors cite, reference #17) that included CYP of various methods, we included the cost of supplies ($1.24) and labor (2.91) when calculating CYP for Sino-Implant ($12.10 total direct cost). I see the authors include "reimbursements for providers" but it's not clear if this is the cost of labor (and, if so, it seems low). I also see consumables and supplies which appear to include specula and forceps, but I imagine there may be other supplies needed, for example alcohol pads, gauze/bandage, etc. It should not be difficult to obtain this data if not currently in hand.
: Excellent and concise presentation of exciting results! Results