Introduction of subcutaneous depot medroxyprogesterone acetate through use of community-based distributors in Zambia

I am pleased to see more information about the CBD pilot that was done in Zambia. It is highly important that these findings be shared more publicly and widely, contributing to the growing knowledge base surrounding DMPA-SC introduction. That being said, this article would benefit from more clearly articulating and detailing the Zambian experience in several ways. Furthermore, the main conclusions stated in the discussion session are not directly supported by the results. Specific suggestions are listed below:
 
Background:

• More information is needed about how CBDs conduct their community-based activities and identify clients in need. Where/how did they find the 12,808 clients served?

Response: By design, the CBD model under SARAI in Zambia is implemented in a way that CBDs operate in the peripherals of the catchment area of a health facility to distribute FP commodities and messaging. Each CBD is accountable to a nearby supervising facility and they submit monthly distribution and stock reports to the supervisors. In the three months of the pilot each of the 161 CBDs served an average of 27 clients with various FP commodities. CBDs conduct health education sessions which mostly focus on family planning as well as HIV prevention. Health facilities also encourage revisit clients to seek services from a CBD nearest to them when returning for a refill. This text has been included in the manuscript edit
 Methods:

• Is the participant consent (that was waived) referring to the participating CBDs? Please clarify.

Response: The waiver applied to DMPA-SC users

• Was there a particular reason why the 3 districts were chosen for the pilot?

Response: Yes, the three districts were chosen because they had all facilities meeting the ‘model facility’ criteria as defined by the project in consultation with the ministry of health

• What are the qualifications/training to become a CBD? Is this cadre of health worker considered volunteers or formal employees of the public health system? What are their regular duties?

Response: CBDs are volunteers selected from within the community by community members. Basic qualifications are ability to read and write, reside within the catchment of the community. Once CBDs are recruited, they become part of the Neighborhood health committee which is a recognized structure by the ministry of health for community-based health service delivery and they are responsible to the public health facility that encompasses their catchment areas.

• Who were the master trainers? What were their qualifications?

Response: These were SRH experts from key departments of MoH

• What does the standard MOH injection supervision checklist look for? Did this need to be adapted for DMPA-SC?

Response: The checklist is a step by step procedure for administration of DMPA-SC

• How regular were the supervisory field visits?

Response: Once a week for facility-based and twice a month for district health office-based supervisors.

• Did CBDs also provide DMPA-IM in communities prior to the pilot? It seems to suggest so, but it would be helpful to state explicitly.

Response: Yes, CBDs were providing DMPA-IM prior to the pilot, text will be added to clarify that.

• Each CBD was given 10 doses of DMPA-SC per month? Or just initially after training?

Responses: The 10 doses of DMPA-SC were issued to CBDs as post training initial stock because previous data indicated that each CBD distributed an average of 10 doses of DMPA per month. More doses were issued to CBDs who needed more, depending on demand in their communities.

• How was the data on the characteristics of the CBDs collected? This is not described anywhere despite the data being shown in the results.

Response: Data about CBDs was part of the training database for CBDs which was kept by the project. This data was collected during the initial selection process for DMPA-IM training and was updated during the DMPA-SC training.
 Results:

• Do the clients served include both men and women? Please clarify. If so, method distribution by gender would be useful to understand method choice of DMPA-SC among the other options.

Response: Yes, clients served during the 3 months of the pilot includes men, male clients were only provided with male condoms hence, under the method switching analysis, only female clients were included as they’re the only ones capable of switching to any other method.

• How were “revisits” identified? Is this determined by the provider or the client? Does this only pertain to methods available at facilities (as opposed to traditional methods, for example)? Similarly, what is the definition of a “new acceptor” and how are they identified? Were these definitions consistently applied?

Response: The context taken for each type of visit was based on modern contraceptive methods. Therefore, a revisit client is a client who have used one or more of the modern contraceptive methods before their current visit to a CBD or a health facility to seek other Family Planning services.  A New acceptor is a client receiving a modern contraceptive method for the first time ever. These definitions were consistently used by all CBDs and data verifiers, it formed part of the both training for CBDs.

• Related to the above, why are there 3% of newly accepting clients? Weren’t CBDs only allowed to service returning FP clients? What explains this discrepancy?

Response: The 3% of new clients were those who were provided with counselling and assessed for eligibility at the health facility but still opted to be provided with DMPA-SC, which was only administered by CBDs at the time hence referrals were made by health facilities staff to CBDs.

• The subheading and text describing Figure 5 is confusing. What is the denominator? It seems to be anyone switching from another method to DMPA-SC and not necessarily FP clients who are revisits, but this isn’t what is conveyed in the text.

Response: The denominator is 2,043, which is the total number of clients who switched from other methods to DMPA-SC. For a client to qualify to be a switcher of FP methods, they must be revisits because a new acceptor would not be on any method prior to their vist.
Discussion:

• Based on the data shown, I don’t think you can conclude that “CBDs were able to safely administer, appropriately store, and dispose of DMPA-SC.” You did not show any information on measures of safety; lack of AE reports does not necessarily reflect the safety of administration, but rather the safety of the drug itself. How did you know that the products were appropriately stored? What data do you have to show that storage was proper? That disposal was also proper? What monitoring was done to check that CBDs were following instructions? Is this from the supervisory visits? If so, then the data from those visits should be reported.

Response: The author has included a write up a reference to the assessment conducted by the National Family Planning Technical Working Group

• Be careful about use of overly general language, such as “performed better.” It is not clear that administering higher numbers of doses per month reflects “better” performance. Does this truly reflect better service delivery? Reach? Coverage? Or more efficient service provision? Or simply poorer quality services to more people?

Response: The comparative reference here only refers to quantities, text to be edited to omit usage of better.

• How did you conclude that DMPA-SC has “the potential to reduce unmet need?” Since CBDs can only attend to returning clients, then this doesn’t seem to be a case where the contraceptive user base is being expanded. You also did not disaggregate client type by age, so it’s not clear how you can conclude that unmet need among adolescents is “especially” addressed.

Response: The context which the author is considering for this conclusion is that when information DMPA-SC was attractive to new users, though not part of the design of the pilot. The conclusion has been edited to only make reference to the addition of more options for clients resulting from the inclusion of DMPA-SC to the method mix that existed.
Other:

• The dataset for the CBD characteristics is not provided.

Response: The database will be added to the link for datasets used.

CONDITIONALLY Approved
General appreciation:
While this paper contributes to the growing body of evidence regarding the feasibility and acceptability of introducing the provision of DMPA-SC at the community-level in Sub-Saharan Africa, by adding a new country (Zambia) to the list of environment where this strategy has been piloted to improve access to and use of family planning services, major revisions are required before it is suitable for indexing.
 
Overall, the manuscript does not satisfactorily address its stated goal of evaluating the “feasibility” of introducing DMPA-SC at the community level. This stems partly from the fact that the research design is different from previous pilots exploring the same research questions, which included primary data collection (including systematic interviews of both DMPA-SC clients and CBDs, plus focus group discussions or / and in-depth interviews with key stakeholders). The research presented here only engages with routine service statistics and this creates limitations in the scope and validity of the findings which should be directly addressed in the manuscript. This is not to say that only pilots using primary data collection should be presented, however the authors need to be realistic in presenting what the available datasets can and cannot contribute to the existing body of evidence on the introduction of DMPA-SC for provision by community-based distributors.
 
Introduction:

• Figures and percentages cited would benefit from scientifically recognized sources such as DHS and/or other population-based surveys.

Response: the author cited credible and recognized sources with up to date data pertaining to Zambia e.g. Zambia Demographic Health Survey 2013-14 and Integrated Family Planning Scale-up Plan 2013 - 2020.
 

• Reference (4) in the first paragraph seems incorrect (not addressing barriers to access in Zambia).

Response: True! Correct reference in number 5 (Integrated Family Planning Scale-up Plan 2013 – 2020), to be edited.
 

• Additionally, regarding ref (4), there is at this point enough literature on the introduction of DMPA-SC at the community level in Sub-Saharan Africa for the authors to engage directly with this corpus, rather than experiences conducted in Europe.

Response: Noted! However, the author used ref 4 as a source of information for the product (DMPA-SC) and not in relation to provision at any setting.
 

• In general, the Introduction needs to engage with methods and findings from the aforementioned corpus and position the presented research accordingly. Feasibility of introducing DMPA-SC at the community level is now fairly well established in multiple low-income environments, so it is important for the authors to clarify what their research add to the existing body of research/for future developments of the model.

Response: Well noted, text has been added
 

• How exactly is “feasibility” going to be measured? According to which criteria?

Response: Client and environmental safety assessments carried out by the Family Planning Technical Working Group of the ministry of health and stakeholders.

Methods:

• First paragraph under “Pilot description” is very unclear (“with established and having established”?)

Response: To be edited to read ‘Voluntary FP services are routinely offered with established community-based distribution programs in which CBDs provided Intramuscular depot medroxyprogesterone acetate (DMPA-IM), oral pills and condoms.’
 

• Authors need to detail the CBD profiles before the beginning of the pilot: are they volunteers? Do they have any clinical background or training prior to the start of the pilot?

Response: comment well noted, the same CBDs profiled during and post DMPA-SC pilot were trained to provide DMPA-IM prior to the pilot period, additional training was provided as alluded to during the introduction of DMPA-SC. Text will be added to provide clarity on CBD profiles prior to the pilot.
 

• Description of “model facilities” is confusing, explain how this was a “community-based distribution” model: where did the CBD operate? Did they work alone or in groups? Door-to-door or during community events? Multiple models of CBD exist in SSA (including some where CBD operate in the courtyard of health facilities) so this section needs to give more details as to what CBD operating in Zambia are specifically capable of doing/allowed to do.

Response: Observation well noted! The text will be edited as follows: ‘The pilot was implemented only in public health facilities that met the criteria of a “model facility” under SARAI. The project defined a model facility as one that encompasses the key components and systems needed at the community level to respond effectively to the family planning needs of the population. These systems must include, skilled service providers; expanded method mix; client-centered care offering age appropriate information, outreach to vulnerable populations; supervision; mentoring for health care providers; youth and adolescent reproductive health services; community involvement through organized and trained community members who mostly are volunteers and data collection and analysis to track progress. Model facilities were selected because the project had already established necessary structures at facility and district levels to monitor and supervise community-based distribution of FP commodities. By design, the CBD model under SARAI in Zambia was implemented in a way that CBDs operate in the peripherals of the catchment area of a health facility while submitting monthly reports to the supervisors based at the facility. All CBDs (163) in all model public health facilities (29) supported by SARAI were selected to participate in the pilot.

• In addition, the fact that these CBD came from “model facilities” creates a bias in the generalizability of the findings that should be addressed in the Discussion section.

Response: True! The reason for selecting CBDs from model facilities was because the project had trained the CBDs in provision of other FP methods including DMPA-IM. These facilities also had trained facility staff in CBD supervision and data compilation. Part of the recommendation for the roll out of the model was that a facility must meet the criteria of a model facility. A section has been added to declare this as a limitation of the pilot.
 

• How reliable are routine statistics (stock management and service data) in Zambia? Was special attention paid to data collection and reporting under the pilot? In that case, any comparison with pre-pilot routine statistics may be flawed due to over/under-reporting.

Response: Data on stock management is part of the Health Management Information System which is compiled by all public health facilities on a monthly basis. CBDs are supervised by qualified health facility staff, trained in stock management and facility reports includes data generated by CBDs as part of commodity accountability. The project in collaboration this MoH staff conducted data quality audits prior and during the pilot in all health facilities.
 

• Adverse event: the authors need to explain what counts as “adverse events” (side-effects from using the method? Allergic reaction? Pregnancy?) While severe adverse events have rarely been recorded during DMPA-SC introduction pilot in DRC, side-effects are a key reason for methods discontinuation and should be discussed in this manuscript as CBD are often poorly trained in side-effect explanation and management.

Response: Well noted, notes will be added to qualify the definition in context. An adverse event in this context refers to severe side effects from using the method. CBDs were trained to report all adverse events (including other mild side effects associated with DMPA) to the facility as soon as they occur. Facility staff were also trained to record and report to project staff immediately.
 
Results:

• The first statement is incorrect (and 17% does not match the 16% indicated on Figure 1). The methodology used cannot measure an “increase in the share of DMPA injectable users” since available data does not include method substitution.

Response: The first statement refers to clients who accessed various methods (including condoms and oral contraceptives) distributed by CBDs during the pilot which is 12, 818 and DMPA-SC distribution accounted for 16% of that number. The 17% increase is in reference to table 2 and only includes DMPA-IM and SC distribution. The increase indicated is pertaining to quantities distributed for the two periods, before and during the pilot and not a proportional share.

• This is a recurring issue with the findings of this article. Table 2, for example, suggests that they may have been at least some substitution between DMPA-IM and DMPA-SC but the authors never discuss this point.

Response: Well noted, the author discusses this point under figure 5 which demonstrates the proportion of clients who switched from DMPA-IM to DMPA-SC. However, text has been added as part table 2 analysis to allude to this finding.
 

• Delete repetition of “Figure 2 below…”

Response: Well noted
 

• The analysis of age distribution would have been more interesting in terms of changes recorded in method distribution among different age groups after the introduction of DMPA-SC.
   

Response: comment noted, the age disaggregates have been added as figure 2.

• We suggest deleting Figure 2 since the graph does not add relevant information that is not already included in the text.

 Response: The figure has been deleted

• The analysis of “New” vs “Re-visit” acceptors needs to be presented in the Methods section.

Response: This aspect was not part of the design, it occurred during the pilot that some new clients were provided with the service by CBDs as referrals from health facilities. The author saw it fit to add this part of analysis in order to present a true picture of the findings.
 

• How reliable is the “New Acceptors” indicators? Routine statistics analysis conducted in other countries (See for example Track20 initiative) suggest that it is poorly recorded (“new” meaning “new to DMPA-SC”, “new to injectable”, “new to this facility”, “new to modern contraceptives” more or less interchangeably).

Response: The determination of New Acceptors was done by professional health workers according to the definition in the national HMIS procedures manual i.e ‘a new acceptor is a client receiving a modern contraceptive method for the first time ever’. New acceptors were assessed by professional health care workers at facilities but were referred to CBDs for service provision as DMPA-SC was only provided at community level during the pilot period.
 

• Again, define AE, what level of severity would have justified recording as “AE”?

Response: Well noted, notes will be added to qualify the definition in context. An adverse event in this context refers to severe side effects from using the method. CBDs were trained to report all adverse events (including other mild side effects associated with DMPA) to the facility as soon as they occur. Facility staff were also trained to record and report to project staff immediately.
 
Discussion:

• Since CBD in Zambia apparently already provided DMPA-IM, the authors need to explain what could be the added-value of introducing DMPA-SC in the range of methods they provide.

Response: The value added by the introduction of DMPA-SC to the mainstream of FP products in Zambia is to allow clients have a wide range of choices. The other factor is that the packaging of DMPA-SC is comprehensive thereby making it easy to store especially at community level.
 

• What was the distribution of methods by age group prior to the introduction of DMPA-SC? Are there other factors (in the pilot or in the study design) that may explain the high proportion of youth clients?

Response: While there were no deliberate messaging strategies aimed at the youth only, it was noted from the pilot results that the response was higher from the youth than any other age group. This trend is however did not only present itself in DMPA-SC distribution but also before the pilot was conducted.
 

• Since “new acceptors” only represent 3% of all recorded clients (of that number is reliable) and method distribution strongly suggest some method switching/substitution, the authors need to better explain how the introduction of DMPA-SC at the community level might reduce unmet need? The data presented here is insufficient to support that claim.

Response: The assumption taken by the author is that DMPA which has been the most preferred method by clients, if provided in another package such as DMPA-SC which according to the survey, clients indicated that it was less painful, could lead to more clients opting for FP.
 

• The comparison with results from studies conducted in other countries needs to be nuanced. None of the study mentioned measured “feasibility” exclusively in terms of volume of DMPA-SC provided at the community level (and a “better” performance of 5 vs 3 doses per month in the highly controlled environment of a pilot project seems hardly significant).

Response: Comment well noted, the use of the word better has been dropped.
 

• The methodology used in Senegal, Niger and Uganda (and several other countries not mentioned in the paper) and the findings were more complex than suggested here and deserve to be fully engaged when comparing to results presented in this manuscript.

Response: comment well noted and attended to
 
Conclusion:

• How was “safely” measured (absence of AE?). This is never discussed in the manuscript.

Response: Yes, the author is relying on the absence of AE to determine that all clients seen during the period were safely provided with the method. The national Family Planning Technical Working Group, which was independent of the pilot carried out onsite assessments by observing CBDs providing the actual service and how they disposed resultant waste.
 

• The second sentence (on potential to reach underserved population) is not adequately supported by findings from the manuscript.

Response: Well noted, the conclusion the author is driving at is that of DMPA-SC administered at community level having potential to reach the people who are located in areas far from health facilities.
 
Writing and figures:

• We suggest using an English editor to review the manuscript for typos, as well as reword some sentences/paragraphs.
   
• Review the manuscript to avoid use of passive voice as much as possible.
   
• We believe the wording should be “injectable contraceptives” and not “contraceptive injectables”.
   
• Reword: “de-identified data for analysis that is part of the routine data collection for which IRB approval was obtained.”

Response: The script has been scrutinized for grammatical and phrasing related errors.