Author’s response to reviews Title: Comparative satisfaction of receiving medical abortion service from nurses and auxiliary nurse-midwives or doctors in Nepal: Results of a randomized trial Authors:

1. Results: It will be helpful to include a brief statement on why some women were excluded from the present analysis, e.g. due to missing data? Response: Thank for spotting this omission. We have revised the text as follows: Of 1,295 women screened for eligibility, 535 were randomly assigned to a doctor and 542 to a nurse or ANM. Nineteen women were lost-to-follow up in the former group and 27 were lost-tofollow up or did not complete the acceptability interview in the latter group. This study is, therefore, based on 516 women in the doctor‟s group and 515 women in the nurse or ANM group. Methods 2. Under Study Procedures the authors mention the "acceptability form" two times before it is described in the last paragraph in the section. It would be helpful on p.6 to briefly describe this form as an interviewer-administered questionnaire to measure women's experiences and satisfaction with MA. Response: Thanks for the helpful suggestion. We have revised the text as follows: In addition to the clinical assessment, an “Acceptability Form” was administered by the female research assistants to ascertain women‟s experiences of symptoms and their satisfaction with MA. Women were asked about their experiences/perceptions regarding pain, amount of bleeding, duration of abortion process, side effects and privacy. For each of these five conditions, women were asked to respond either „less than expected‟, „same as expected‟, „more than expected‟, or „no comment‟. Questions measuring the coverage and quality of counselling and communication included: (1) “was the method explained clearly to you before taking tablets”; (2) “were the signs of complications explained clearly to you”; (3) “did you receive counselling on contraceptive use”; and (4) “did the provider give you an opportunity to ask questions”. Preferences of sex of the provider and the length of practice were also ascertained. In addition, all women were asked one direct question on „satisfaction‟: “how satisfied are you with the method”. 3. Conceptual framework a. The authors mention that the framework was developed drawing on evidence from previous studies, and it would be helpful to add citations of the evidence linking the included factors to the satisfaction outcome. Later in the paper it comes across that expectations are thought to drive satisfaction, but this is not well described in the conceptual framework. Adding the literature here will be helpful. Response: We have revised the text on Conceptual Framework, shown above in response to Point 1 of Reviewer #1. b. The authors describe a proximate determinants framework, and it would be helpful to create a simple figure to display the proximate determinants more clearly. Response: Figure 1 has been added to show the proximate determinant framework more clearly. c. In the last sentence, the authors mention that preference for a male or female provider was included as a predictor. Since you also have data on the sex of the provider who actually treated the woman, it might be more informative to look at concordance between the preference and the provider she received. It is likely the concordance between preference and the sex of the provider she saw that drives satisfaction rather than just her preference. Response: Thank you for the insightful suggestion. The sex of actual MA provider was no recorded in the Form for women obtaining the service. However, we know that all nurses and auxiliary nurse-midwives were females and 79% of doctors were male. Among 1027 women, only 4 women expressed preference for a male doctors, while 35.6% had no preference and 64% preferred a female provider. Given that there was no difference in satisfaction by provider type, the study confirm, albeit indirectly, that sex of the provider has no bearing on the satisfaction with MA. We have added the text: Providers‟ age was significant, but not the length of experience in providing abortion. We also probed if preference for the sex of provider had any bearings on the reported satisfaction with MA. Among 1,027 women included in the study, 366 (35.6%) had no preference, 657 (64%) preferred a female provider and the remaining four women preferred a male provider. Given little (0.4%) preference for male provider, we could only compare women with „no preference‟ to those with a preference for a female provider and find that the former were more likely to report higher satisfaction than the latter. We were unable to extend the analysis further to examine if the concordance between the preference for sex of provider and of the actual provider resulted in greater reported satisfaction because the study did not document the sex of actual MA provider for each woman obtaining service. However, we know that doctors were mostly men (79%) and nurses and ANMs were all women. The study finds no difference in the level of satisfaction by provider type and thus indirectly confirms that the sex of the actual provider did not influence the reported satisfaction of MA. Results 4. As mentioned in the previous comment, it would be nice to see some further analysis of concordance in preference for a provider of a specific sex and the sex of the provider she was assigned. Right now, the authors draw the conclusion that preference for female providers drives discomfort with the vaginal examination in the doctors group, but this is not measured directly. Based on the description of variables in the Methods and Discussion, it sounds like you could create a variable that has four categories: 1) wanted female provider, received female provider, 2) wanted male provider, received male provider, 3) wanted female provider, received male provider, and 4) wanted male provider, received female provider. Categories 3 and 4 are the discordant groups, and you might expect satisfaction to be lower for these women. Including this analysis would provide more support for your conclusions about the role of the sex of providers in client satisfaction. Response: Thank you. We considered the suggestion, but it was not possible to pursue the analysis for the reasons given above in point c. Only 4 women expressed preference for a male provider.