Post-abortion care with misoprostol – equally effective, safe and accepted when administered by midwives compared to physicians: a randomised controlled equivalence trial in a low-resource setting in Kenya

Objective To assess the effectiveness of midwives administering misoprostol to women with incomplete abortion seeking post-abortion care (PAC), compared with physicians. Design A multicentre randomised controlled equivalence trial. The study was not masked. Settings Gynaecological departments in two hospitals in a low-resource setting, Kenya. Population Women (n=1094) with incomplete abortion in the first trimester, seeking PAC between 1 June 2013 to 31 May 2016. Participants were randomly assigned to receive treatment from midwives or physicians. 409 and 401 women in the midwife and physician groups, respectively, were included in the per-protocol analysis. Interventions 600 µg misoprostol orally, and contraceptive counselling by a physician or midwife. Main outcome measures Complete abortion not needing surgical intervention within 7–10 days. The main outcome was analysed on the per-protocol population with a generalised estimating equation model. The predefined equivalence range was –4% to 4%. Secondary outcomes were analysed descriptively. Results The proportion of complete abortion was 94.8% (768/810): 390 (95.4%) in the midwife group and 378 (94.3%) in the physician group. The proportion of incomplete abortion was 5.2% (42/810), similarly distributed between midwives and physicians. The model-based risk difference for midwives versus physicians was 1.0% (–4.1 to 2.2). Most women felt safe (97%; 779/799), and 93% (748/801) perceived the treatment as expected/easier than expected. After contraceptive counselling the uptake of a contraceptive method after 7–10 days occurred in 76% (613/810). No serious adverse events were recorded. Conclusions Treatment of incomplete abortion with misoprostol provided by midwives is equally effective, safe and accepted by women as when administered by physicians in a low-resource setting. Systematically provided contraceptive counselling in PAC is effective to mitigate unmet need for contraception. Trial registration number NCT01865136; Results.

Are the diagnosis and subsequent enrollment of women was done by the physician or midwife in their respective group [So that there could be a more specific enrollment which could avoid the possibility of inclusion of lots of complete abortions]? rrespective of subsequent allocation. Thus, in this respect, these two groups are same.
Are the basic intervention is technically difficult or need expertise or experience so that it might create a difference in between the groups?
No. Only a single dose of oral Misoprostol was applied in both groups and this needs neither physician's knowledge nor midwife's experience.
Were the follow up done by the groups differently?
Not mentioned in details. This is a very important area of this study where you need to ensure whether the woman has achieved the complete abortion. How it was ensured? The authors have mentioned few signs of untoward events and absence of them is considered as the successful outcome. But this needs to be more detailed. Moreover, who has done those follow-up? Respective groups or any other midwifes irrespective of groups as designated for enrollment? If the second option is here, then there exists no difference between the groups if the follow-up was done by a third party.

Result:
The success rate of this procedure in this method is 94.3 % and 95.3% in physician and midwife's group  where as 14% women experienced heavier than normal bleeding in each group [ table-4]. These data evokes a query whether these 14% women in follow-up needed any extra care? Can they be designated as ‗'failure''? If not, why? I am not going into the detailed results as well as discussion part. Those also needs to be more precise and oriented. the case in many Kenyan and other African hospitals. 9). Regarding the secondary outcomes: The authors say they have assessed acceptability, but this is not the case. The only measure given (p. 10, lines 30-32) are whether the women perceived the treatment to be as expected or easier than expected, and if they would recommend it to a friend, both of which are very crude and tell us very little about acceptability. One cannot conclude that the treatment is ‗highly acceptable' (line 52-54 and in the discussion and abstract), nor that contraceptive uptake is high when we only know the outcome immediately posttreatment. On p.10, several of the secondary outcomes (e.g. bleeding, pain, symptoms) are discussed under Main Outcomes, whereas in the Methods section, they are listed as secondary outcomes and they should be placed in this results sub-section. 10). Discussion The stated limitations are correct, but several more need to be added, as mentioned above, e.g. the need for long-term follow-up of contraceptive method use-certainly more than what the index woman is offered at the exit visit. p. 13, lines 22-23: there are also other studies done on other continents that should be noted with some comparison made, and South Africa's experience should also be noted.

REVIEWER
Jennifer Blum Gynuity Health Projects, USA REVIEW RETURNED 21-Feb-2017 GENERAL COMMENTS abstract: in population missing end date for data collection "2016" main outcome measure: its unclear to me why the medication will work differently if administered by a nurse versus a physician. What is the medical rationale for this possible difference? In the methods, the authors need to clearly state which tasks were done by nurses versus MDs and show that there was no overlap -if yes, then I am not sure what the outcome is again, is it nurse led, MD supported versus MD led, nurse supported? In either case, I don't understand the reasoning to think that the efficacy of the pills, swallowed by a woman, will be different if the pills are handed to her by one type of provider in a hospital versus another. Strengths: The first strength has already been demonstrated in a number of studies. Here are some publications to consider and add to the reference list: lines 34 -40 can be shorten for precision, probably you can eliminate info on "emancipating space for study forms" Page 6, line 7 -please provide additional details on the potential unexpected symptoms/burden. inclusion and exclusion -it seems like eligibility/incomplete abortion status was determined before allocation. If yes, please state which type of provider confirmed eligibility -confirming eligibility is a key task in PAC -so its helpful to know who did this.
Can the authors explain why women were asked to stay at the clinic for 4 -12 hrs after swallowing miso? Most studies no longer do thisand it isn't common in clinical practice either.
results: I am not sure that the discussion about socio-economic status possibly or not affecting the results is a useful discussion since again, its just swallowing a few pills, why would there possibly be an effect? I suggest this section of the discuss be shortened and more concise.
Discussion: I also think the discussion could be shorter and more precise and again, would benefit from a review of relevant references. I think for the sake of recognizing the PAC consortium/community, its helpful to make clear that post abortion contraception counseling is an element of PAC, not an addition. For clarity the three paragraphs in the discussion starting line 18 -line 4 on the next page can probably be edited into one concise paragraph. Page 13 lines 12 -20 can also be edited and merged into one concise paragraph on the situation in Kenya, the guideline and their potential impact on populations of women and girls throughout the country. This would leave more space for the discussion to reflect back on the results of the paper -and maybe discuss more how these services can be provided outside of tertiary level hospitals throughout Kenya (and already are) Page 13, lines 27 -32. I think many people will disagree that there is scarce evidence that trained midwives can provide PAC. USAID recommends it and its recommended in the WHO guidelines on health worker roles in abortion care (2015) -see 4. Title: The title of the manuscript needs to be a bit informative. Misoprostol can be applied by different routes while managing incomplete abortion. Among them, the authors here had chosen the oral route and this need to be reflected in the study-title.
ANSWER: We chose to use misprostol, only, to keep it as short as possible. We consider other information (design and the context) to be important as well in the title. The administration (orally), is clearly written in the abstract under the sub-heading Intervention. We are awaiting the decision from the BMJ-Editor, if to add this information in the title.

Introduction:
The introduction part is very descriptive, elaborate as well as over-informative, which are often a bit unrelated with the purpose of the study. Yes, incomplete abortion is a very commonly encountered problem and possibility of a medical management of this complication by ground level healthcare provider is of immense help. With this background, a short, brief and informative introduction is needed.
ANSWER: Thanks, we have shorten the introduction section, to make it more stringent (13 lines shorter). Page 3 and 4.
6. Methods: a. What are the diagnostic criteria of ‗' incomplete abortion‖'? The authors have included ‗'women with signs of incomplete abortion (i.e., bleeding, contractions and lower abdominal pain during pregnancy, open cervical os, and partial expulsion).''. But these signs are never alone can diagnose even a pregnant state if you do not do a simple urine for pregnancy testing initially. Moreover, this is often difficult to differentiate between complete abortions with incomplete abortion even after clinical examination. But authors here do not take the help of ultrasonography before enrolling. These points need to clarify here in details. b. What were the basic differences in enrollment, intervention and follow-up in between two groups, while managing the incomplete abortions? Are the diagnosis and subsequent enrollment of women was done by the physician or midwife in their respective group [So that there could be a more specific enrollment which could avoid the possibility of inclusion of lots of complete abortions]? c. Enrollment was done by two midwifes initially irrespective of subsequent allocation. Thus, in this respect, these two groups are same. Are the basic intervention is technically difficult or need expertise or experience so that it might create a difference in between the groups? d. Only a single dose of oral Misoprostol was applied in both groups and this needs neither physician's knowledge nor midwife's experience. Were the follow up done by the groups differently? Not mentioned in details. This is a very important area of this study where you need to ensure whether the woman has achieved the complete abortion. How it was ensured? The authors have mentioned few signs of untoward events and absence of them is considered as the successful outcome. But this needs to be more detailed. Moreover, who has done those follow-up? Respective groups or any other midwifes irrespective of groups as designated for enrollment? If the second option is here, then there exists no difference between the groups if the follow-up was done by a third party.
ANSWER: a. Of course this is important to state, and that was such an obvious criteria that we did note state it in detail. We have added that they did a pregnancy test at the facility. It is also true that it could be difficult to sort out differential diagnoses to incomplete abortion, such as complete abortion. However, that's the true situation in many places in Kenya. Many facilities don't have an ultrasound as a standard equipment or the routine to use it. If they need to use an ultrasound they need to transfer the women to another ward (if it is available), and the woman has to pay for this additional exam. We have done some clarifications regarding inclusion and exclusion criteria. a. Line 229-235, p 6. b. None. Four midwives at each facility were trained, two to be the head research assistants, and two to stand-in (responsible for the first screening, enrolment, the allocation and follow up at 7-10 days). Already described on page 7, however we have revised and included the two stand-in. Line 262-263, p 7, Line 271, p 7. c. Please see response 6b. To provide the misoprostol was not considered to be a technical procedure. But, both groups went through the same training program with misoprostol, because before the current training, the superior regime to treat a first trimester incomplete abortion was MVA. d. Please, see response no 6a-c. Misoprostol was applied in both groups and it may be true, that this needs neither physician's knowledge nor midwife's experience. However, in these settings (like in many other settings in worldwide) regulations restrict misoprostol prescription and supervision to physicians. In addition, these kind of settings lacks modern equipment's and skills how to use them, and therefore is ultrasound rarely used in this objective. Each and every country needs their own studies that applies to the unique context before new medical routines are implemented. Kenya does not differ in that context -this study has been very well implemented at the two facilities, the midwives continues (up to date) to treat first trimester incomplete (as the superior regime). ANSWER: Heavier than normal bleeding could not be considered as a failure for complete abortion. This is self-reported statements and was considered a normal variation. Among the women who reported heavier bleeding than normal bleeding (14%), only 15% were diagnosed with incomplete abortion (needed repeated dosage of misoprostol or MVA). 85% were complete abortion.
Reviewer: 2 Andrzej Kulczycki University of Alabama at Birmingham (UAB), USA 8. Please state any competing interests or state ‗None declared': None declared ANSWER: Please, see response no. 3 9. The paper concerns an important topic, and adds evidence that midwives can treat incomplete abortion with misoprostol. It is encouraging to see more papers on this. In the checklist, the items noted 'no' should have the appropriate modifications made in the text. These and some additional edits should be recommended and should be manageable for the authors: ANSWER: Please, see response no. 1 10. Title: The work is done in 1 hospital (see also point 5) in 1 city in Kenya, so this must be inserted, because it is currently misleading to suggest this work was done throughout the country or that it is somehow representative of Kenya.
ANSWER:The study was conducted at two large public hospitals, line 191-192, p 5. This area is considered to be -a‖ low-resource setting (mentioned in the title), similar to other areas in Kenya, also with regard to public hospital/facilities. The third largest city does not necessary mean that it is not a low-resource setting. We can add Kisumu, if the BMJ-Editor agree to this suggestion.
11. Abstract and Use of English there and throughout manuscript: some editing is needed throughout to improve on the writing and make it read better, e.g. the very first entry in the Abstract, ‗Objective' has a second sentence that starts off with: ‗Secondly to assess...; and ‗Main Outcome measures' has a second line that starts: ‗Secondary, women's....' ANSWER: Thanks. We have made some adjustments in the abstract and also adjusted the English throughout the MS. Before we submitted the MS, we used Scribendi for proof editing of the language.
12. Introduction: First 3 sentences all need correcting. These are all estimates. The key thoughts are here, but they are rather disjointed and the Introduction should be strung together more coherently. ANSWER: Please, see response no.5.
13. Top of p.5: Aims: a. The first aim has seen a few studies in SSA and these should be cited. It is good that there should be more studies on the topic, but earlier work should be credited. b. Second aim is listed as acceptability, but the paper does not get at this. You only have very vague questions with self-reported answers that allude to acceptability. Also, these are not analyzed (in part because they cannot be, they lack depth), they can only be described. The second aim should be revised. It can be broadened to include other secondary outcomes, but these should be presented more modestly-you only have very crude measures .
ANSWER: a. That's true. However we have chosen (according to reference limitations) references addressing the same topic (incomplete abortion) and similar contexts, as well as the intervention design -midwife/Physcians‖. When you add these limitations, there are not many studies within this objective. However, we have revised the reference list, accordingly to your comments. b. We agree to your comments, the questions are vague. We have added this as a limitation and we have revised the objective and abstract, to put the highlight mainly on the primary objective/outcome. However, to know the acceptability among the subjects is also a very difficult objective to measure. We considered that self-reported statements was one way to find out how they perceived the treatment. Over 90% reported that they; felt safe, perceived the treatment as expected/easier as expected, and will recommend it to a friend or a relative. Since all the measures were in favors to the treatment (without any doubts), we still consider these aspects to be a warranty for the acceptance of the treatment. Page 3 (limitations) and revised Abstract, page 2.
14. Methods: a. p. 5, parag. e. Yes, it could be indeed. However, women seeking care in these public facilities are considered to be of low SES, and with hence pronounced financial reimbursements barriers. Our analyses showed similar distribution of provided contraceptive methods, between midwifes and physicians (Table 5). f. Yes, and this is already mentioned as a limitation (page 3, and in the discussion). The contraceptive uptake was recorded at follow-up (7-10d). Line 278, p 8 g. We have already mention a reference, line 325, page 9. However we have added one more. One additional ref added (26) h. Please, see response no 10. i. Please see response no 18.
15. Results, p. 9: a. Losses to follow-up are adequately explained and Fig. 1, tables 1-3 are clear. What does the statement: physicians were more experienced than the midwives in PAC‖ mean specifically? What sort of experience did the midwives have? e.g. were they just assisting physicians, as is the case in many Kenyan and other African hospitals. b. Regarding the secondary outcomes: The authors say they have assessed acceptability, but this is not the case. The only measure given (p.10, lines 30-32) are whether the women perceived the treatment to be as expected or easier than expected, and if they would recommend it to a friend, both of which are very crude and tell us very little about acceptability. One cannot conclude that the treatment is ‗highly acceptable' (line 52-54 and in the discussion and abstract), nor that contraceptive uptake is high when we only know the outcome immediately post-treatment.
c. On p.10, several of the secondary outcomes (e.g. bleeding, pain, symptoms) are discussed under Main Outcomes, whereas in the Methods section, they are listed as secondary outcomes and they should be placed in this results sub-section.
ANSWER: a. Please, see responses no. 6a-c. The provider's previous experiences in PAC is described in Table 1. They were asked to state there working experience as shown in the Table. No further analyses were made. The superior regime to treat incomplete abortion (first trimester was MVA in the both facilities). b. Please see responses no 13b and 14f. Contraceptive up-take was recorded at follow-up (7-10d). Please see response no 18. We also did a follow-up at 3-month, and we are still analyzing this data and it will be presented elsewhere (PhD-stud). c. Thanks, we have revised the structure for the outcomes. We still believe that some outcomes are better placed under main outcomes. Therefore, we have divided table 4 into two tables, (Table 4 and 5). c. Page 3, limitations and 20. Introduction: I think this could be tighter and include an updated literature review. Use of terms like 'high quality PAC" should be explained. They can probably also update the literature on midlevel provision of MA, or just eliminate the Nepal reference altogether from this paper (line 37.) The statement on lines 56 -57 "Data are limited by the scarcity of RCTs and biases of the cohort studies" should be further clarified -what are the authors referring to, exactly, here. lines 34 -40 can be shorten for precision, probably you can eliminate info on "emancipating space for study forms" ANSWER: Please see previous response no 5. Introduction is revised, p 3-4.
21. Method: a. Page 6, line 7 -please provide additional details on the potential unexpected symptoms/burden. b. inclusion and exclusion-it seems like eligibility/incomplete abortion status was determined before allocation. If yes, please state which type of provider confirmed eligibility -confirming eligibility is a key task in PAC -so its helpful to know who did this. c. Can the authors explain why women were asked to stay at the clinic for 4 -12 hrs after swallowing miso? Most studies no longer do this -and it isn't common in clinical practice either.
ANSWER: a. Please see response no 6a-c. b. Please see responses 6a-c. c. Please see response 10. This was a new regime, approved at the Ministry of health specifically for this study and the providers required this procedure for safety (stay 4-12 h). Today, after the implementation, the providers does no longer keep the women for this safety-check :).
22. Results: I am not sure that the discussion about socio-economic status possibly or not affecting the results is a useful discussion since again, its just swallowing a few pills, why would there possibly be an effect? I suggest this section of the discuss be shortened and more concise.
ANSWER: Please see response no 10. There is a clear association between women with low SES and severe complications after procuring an unsafe abortion. Since this is a low-resource setting, lacking modern equipment's, routines and skills how to use them, we consider SES as a relevant topic relevant.
23. Discussion: I also think the discussion could be shorter and more precise and again, would benefit from a review of relevant references. I think for the sake of recognizing the PAC consortium/community, its helpful to make clear that post abortion contraception counseling is an element of PAC, not an addition. For clarity the three paragraphs in the discussion starting line 18line 4 on the next page can probably be edited into one concise paragraph. Page 13 lines 12 -20 can also be edited and merged into one concise paragraph on the situation in Kenya, the guideline and their potential impact on populations of women and girls throughout the country. This would leave more space for the discussion to reflect back on the results of the paper -and maybe discuss more how these services can be provided outside of tertiary level hospitals throughout Kenya (and already are).Page 13, lines 27 -32. I think many people will disagree that there is scarce evidence that trained midwives can provide PAC. USAID recommends it and its recommended in the WHO guidelines on health worker roles in abortion care (2015) -see table page 6 and elsewhere within this document. These guidelines also should be included in the reference list -ANSWER: We have deleted text and revised some parts, as you suggestedthanks, very helpful! Additional information, PAC-consortium is the base for this trial (used in the training program, and we have used the reference (14), and additional references are 30 & 31 (used for the PAC-training).
Contraceptive counselling is part of PACwe clearly describe the elements in PAC (Introduction). However, post-abortion contraceptive service (especially in this region), is still a crucial part of PAC. High unmet need for contraception is still a challenging work (not only because of knowledge-gap, but also because of stigmatizing attitudes surrounding contraceptive-use and its misconceptions, which is present among health care providers, stakeholders, and among publics in general). Therefore, we believe this is an element in PAC which deserves to be highlighted (not separated

GENERAL COMMENTS
This is a manuscript rewritten to show the effectiveness of oral misoprostol to manage incomplete abortions prescribed by the midwives. Authors described in details the maternal morbidity related to unsafe abortion in the long introduction part. This is a study related to incomplete abortions. Then, how unsafe abortion is related to it? Do you consider incomplete abortion and unsafe abortion are same? No. These are two entirely different issues. Did you enroll the women with unsafe abortion here? Not at all and not all unsafe abortions are incomplete and vice versa. The authors randomized the subjects group and physician group and showed the result is comparable in both groups. However, it is not the physician or the midwives, who enrolled or follow up the women. As the study says-‗'the research assistants were responsible for the screening, the follow-up visit including assessment of abortion status at follow-up and contraceptive uptake, and performing MVA when necessary''. So what was the function of the two groups? Only to handover the tablet to the women? So, how a drug will act differently inside the body if given by a doctor or nurse or by anybody? One of my colleagues asked this question beforehand and I am now also raising the same issue. Then, why will the result be different? There is no need to conduct a study for this.

GENERAL COMMENTS
The following issues may be noted / clarified to improve the report.
1. Please justify why analysis of primary outcome was done for the per protocol data set rather than the intention to treat dataset.
2. Some details of the generalized estimating equation model used to estimate the risk difference in primary outcome between the two groups is desirable. Also elaborate how the adjusted risk difference was calculated.
3. The discussion may be reorganized. Shift the second and third paragraphs (that mention the limitations and constraints) towards the end before conclusion.
Introduce a short paragraph to highlight that misoprostol adverse event profile in PAC in this study matches the known adverse profile of the drug and there were no striking differences between the two categories of providers in this regard.
Probably a more immediate future research need is to see if the results of this study are replicable in a primary health care setting, which (as the authors themselves highlight) is somewhat different from a district hospital setting. Table 3 information can be succinctly incorporated in the main text and this table omitted. The adverse event profile listed in Appendix 1 can instead be brought in as a table since safety aspect is always important in any pharmacological intervention.

Page 2, Line 33
Correct the timeline in the abstract.

Page 3, Lines 12-13
The very nature of the outcome measure requires that the data would be captured from self-reported statements during the followup visits. Therefore this is not a limitation. Instead you may include the point that both midwives and physicians were working in the same health facility and therefore there could have been interaction between the two sets of providers in post-abortion care.

Page 5, Line 44
Expand on the acronyms JOOTRH and KCH at first occurrence.
12. Page 6, Line 45-52 Do you mean ‗protocol' or ‗case report form' in this paragraph?
14. Page 8, Line 55-58 The equivalence margin has already been stated earlier. No need to repeat again. Please delete.
15. Page 9, Line 25-38 Flow of study participants has been presented appropriately in Figure 1. No need to repeat here.

Page 10, Lines 23-24
It is preferable to substantiate that there were no differences in the secondary outcomes between the Groups by a adding a column of p value for intergroup comparison in Table 4 and Table 5. 17. Page 11, Line 3 Instead of writing ‗The study design strengthen the reliability and the validity by the sample size …'. You may simply write ‗The strengths of the study lie in its large sample size …'.
18. Page 13, Lines 12-16 The unsafe abortion consequences are not particularly relevant to discussion of the results of this study. May omit this.
Comments concluded