Advertisement
Original research article| Volume 97, ISSUE 2, P152-159, February 2018

Feasibility of assessing the safety and effectiveness of menstrual regulation medications purchased from pharmacies in Bangladesh: a prospective cohort study

Open AccessPublished:August 17, 2017DOI:https://doi.org/10.1016/j.contraception.2017.08.002

      Abstract

      Objective

      To assess the feasibility of following up women who purchase mifepristone+misoprostol or misoprostol-only from pharmacies in order to measure the safety and effectiveness of self-administration of menstrual regulation.

      Study design

      A prospective cohort study followed women purchasing mifepristone+misoprostol or misoprostol-only from pharmacies in Bangladesh. Participants were recruited by pharmacy workers either in person or indirectly via the purchaser of the drugs. End users were contacted by phone 2 weeks after recruitment, screened and interviewed.

      Results

      Study recruitment rates by pharmacy workers were low (30%, 109 of 642 women informed about the study), but 2-week follow-up rates were high (87%). Of the 109 end users interviewed, 87 purchased mifepristone+misoprostol and 20 misoprostol-only, while 2 women did not know what drugs they had purchased. Mean self-reported number of weeks of pregnancy was 5.7 weeks. Information provision by pharmacy workers was inadequate (40.4% received none, 8.7% received written information or pictures). A total of 80.5% of mifepristone+misoprostol users were sold the correct regimen versus 9 out of 20 misoprostol-only users. A total of 68.8% did not report experiencing any complications (70.0% misoprostol-only; 69.0% mifepristone+misoprostol users, p=1.0). A total of 94.3% of mifepristone+misoprostol users and 75% of misoprostol-only users reported that they were not pregnant at day 15 (p=.020). However, 7.3% of all users sought additional treatment.

      Conclusions

      Challenges in assessing outcomes of self-managed menstrual regulation medications purchased from pharmacies must be overcome through further development of this methodology. Interventions are urgently needed to ensure that women have access to correct dosages, accurate information and necessary referrals.

      Implications

      This paper assesses the outcomes of women who self-manage menstrual regulation medications purchased from pharmacies. The methodology requires further development, but our study provides preliminary positive evidence on the safety and effectiveness of self-management despite low information provision from pharmacy workers.

      Keywords

      1. Introduction

      Women seeking abortion in low- and middle-income countries (LMICs) often turn to pharmacies as they are accessible, offer anonymity and commonly sell medical abortion (MA) pills (misoprostol or mifepristone+misoprostol) without prescription [
      • Sneeringer R.K.
      • Billings D.L.
      • Ganatra B.
      • Baird T.L.
      Roles of pharmacists in expanding access to safe and effective medical abortion in developing countries: a review of the literature.
      ]. Availability of these pills in communities has revolutionized access to safer abortion and reduced morbidity and mortality from unsafe abortion in recent years [
      • Hyman A.
      • Blanchard K.
      • Coeytaux F.
      • Grossman D.
      • Teixeira A.
      Misoprostol in women's hands: a harm reduction strategy for unsafe abortion.
      ,
      • Miller S.
      • Lehman T.
      • Campbell M.
      • Hemmerling A.
      • Brito Anderson S.
      • Rodriguez H.
      • et al.
      Misoprostol and declining abortion-related morbidity in Santo Domingo, Dominican Republic: a temporal association.
      ,
      • Åhman E.
      • Shah I.H.
      New estimates and trends regarding unsafe abortion mortality.
      . Women can safely self-administer MA if they have clear and accurate information to know how to take the pills safely and effectively, and when and where to seek help in case of complications or ongoing pregnancy [
      • WHO
      Expanding health worker roles to help improve access to safe abortion and post-abortion care: WHO launches new guideline.
      ]. However, pharmacy workers in LMICs lack accurate knowledge about MA, do not provide adequate information to clients [
      • Sneeringer R.K.
      • Billings D.L.
      • Ganatra B.
      • Baird T.L.
      Roles of pharmacists in expanding access to safe and effective medical abortion in developing countries: a review of the literature.
      ] and commonly sell the drugs without packaging or instructions for use, particularly as misoprostol-only is not indicated for MA.
      There is a dearth of evidence assessing the safety and effectiveness of self-administered MA when purchased without prescription from pharmacies. Such data are difficult to obtain; recruiting and following up women who self-administer drugs are challenging because pharmacies and women may be unwilling to report on a form of provision and use that is not permitted, and low client volumes hinder the identification of women who purchase and self-administer these medications. Studies that have attempted to evaluate safety have measured complications among women presenting at clinics for postabortion care after taking pharmacy-purchased MA pills [
      • Giri A.
      • Srivastav V.
      • Suwal A.
      • Sharma B.
      A study of complications following self-administration with medical abortion pills.
      ,
      • Nivedita K.
      • Shanthini F.
      Is it safe to provide abortion pills over the counter? A study on outcome following self-medication with abortion pills.
      ]. However, these studies suffer from substantial selection bias as samples do not include women who have complete, uncomplicated abortions without the need for follow-up care or those who manage adverse events outside of the formal health system. One feasibility study in Mexico attempted to use pharmacy-distributed flyers containing a phone number to recruit women prospectively into a study intended to assess clinical outcomes but suffered an extremely low response rate [
      • Brahmi D.
      • Maxwell L.
      • Collado M.E.
      • Schiavon R.
      • Montaño P.C.
      • Sánchez R.C.
      Pharmacy recruitment of misoprostol users in Mexico: a feasibility study.
      ].
      Addressing these methodological challenges is essential in order to obtain robust evidence on the clinical outcomes of women who self-administer MA purchased from pharmacies. Such evidence can inform: WHO guidelines on health worker roles, policy makers on the safety of reducing restrictions on the availability of MA, and design and evaluation of interventions aiming to improve the availability and safety of MA in communities.
      This prospective cohort study aimed to assess the feasibility of following up women who self-administer misoprostol-only and mifepristone+misoprostol purchased from pharmacies and to establish the clinical outcomes of women who self-administer these medications purchased from pharmacies in Bangladesh.

      2. Materials and methods

      2.1 Study setting

      In Bangladesh, although abortion is legal only to save a woman's life, menstrual regulation (MR) services have been part of Bangladesh's family planning program since the 1970s [
      • Chowdhury S.N.M.
      • Moni D.
      A situation analysis of the menstrual regulation programme in Bangladesh.
      ]. MR is officially recognized in Bangladesh as an interim method for establishing nonpregnancy. MR services use manual vacuum aspiration up to 12 weeks after a missed period [
      • Vlassoff M.
      • Hossain A.
      • Maddow-Zimet I.
      • Singh S.
      • Bhuiyan H.U.
      Menstrual regulation and postabortion care in Bangladesh: factors associated with access to and quality of services.
      ] or the mifepristone+misoprostol combined regimen up to 9 weeks [
      • Government of the People's Republic of Bangladesh Directorate General of Family Planning
      Minutes of 63rd meeting of the National Technical Meeting.
      ]. Pregnancy is not confirmed prior to administering the procedure or medications. Misoprostol is available on prescription for other indications (e.g. postpartum hemorrhage). There is little regulation of pharmacies in Bangladesh, and many medications are sold over the counter without prescription [
      • Babu M.
      Factors contributing to the purchase of over the counter (OTC) drugs in Bangladesh: an empirical study.
      ], including mifepristone and misoprostol. Information provided by the Directorate General of Drug Administration shows that 32 brands of misoprostol-only are registered for sale by 28 pharmaceutical companies and that 12 mifepristone-only brands are registered for sale by 12 companies in Bangladesh [
      Directorate general of drug administration, Bangladesh.
      ]. Mifepristone and misoprostol are available packaged together, indicated for menstrual regulation, by at least eight different companies. In this study, the term “MR medications” is used to refer to misoprostol-only or mifepristone+misoprostol. The term “pharmacy” refers to any outlet selling medications whether or not its staff are registered or trained to do so.

      2.2 Recruitment and data collection

      Between November 2015 and March 2016, we conducted a prospective cohort study among women using MR medications purchased from pharmacies. We used random systematic sampling to select 110 pharmacies from a list of outlets that are supplied by a large Bangladeshi pharmaceutical company known to distribute misoprostol and the mifepristone+misoprostol combination pack in two of the country's eight administrative divisions (Dhaka and Mymensing). Field workers visited the pharmacies to explain the study and obtain informed consent. In each pharmacy, one pharmacy worker participated in the study. If more than one pharmacy worker was present, we invited the most senior to participate first. We reimbursed pharmacy workers 1000 taka (US$12.70) upfront for participating.
      We asked pharmacy workers who agreed to take part to invite any person purchasing MR medications to participate in the study, using a script. We gave each pharmacy worker 40 information slips to distribute to women or men who purchased mifepristone+misoprostol or misoprostol-only for any indication. The information slip contained a unique ID, a brief explanation of the study, instructions to send an SMS to the study phone line and an explanation that callers would receive 100 taka (US$1.28) phone credit.
      Pharmacy workers instructed interested clients to SMS the study phone line immediately, and the pharmacy worker also offered to send the SMS on their behalf. Upon receiving an SMS, a trained research assistant phoned the client immediately and asked to speak to the end user of the drug in order to check study eligibility and invite them to participate in a telephone interview 2 weeks later. Eligibility criteria were: women using MR medications purchased from selected pharmacies for MR by themselves or by proxy, aged 18–49 years, providing verbal informed consent, contacted before the end user took the pills and taking MR medications for the first time in their current menstrual cycle. We excluded women who were not taking MR medications for the first time in their current menstrual cycle because we aimed to recruit women at the point of purchase and then follow up outcomes; we would not have been able to collect adequate data on prior attempts and the woman's MR completion status at the point of making their purchase.
      We conducted follow-up telephone interviews with the woman who took the pills 15 days after they took the first pill. Women who gave verbal consent during this call were interviewed using a structured questionnaire, which covered: information received from the pharmacy, MR regimen used, MR outcomes, side effects and complications. We developed questionnaires in English, which were then translated into Bengali, reviewed, revised and back-translated. Pregnancy was assessed through self-report because women were interviewed by phone and provision of pregnancy tests may have affected behaviors relating to self-administration. To identify the regimen used, participants were asked the brand of drug provided, the number of pills provided and the number of pills taken. Women were not asked whether they had confirmed the pregnancy through a pregnancy test, but gestational age was estimated through women's self-report of weeks of pregnancy at the time of pill-taking. Women who reported needing medical help during the interview were referred to the Marie Stopes Bangladesh call center for advice or onward clinical referral.

      2.3 Sample size

      We calculated a sample size of 432 misoprostol-only clients and 400 mifepristone+misoprostol clients (a) to measure a 15% difference in regimen effectiveness between misoprostol-only (75%) and mifepristone+misoprostol (90%) and (b) to compare the effectiveness of combination MR medications provided with and without prescription (which was estimated at 50%). The sample size calculation for misoprostol-only clients was increased to 550 to account for clients who used misoprostol for other indications, estimated at 23% (unpublished data). The planned sample size was further increased to 825 misoprostol-only and 600 mifepristone+misoprostol clients to account for 33% expected loss to follow-up. The intended sample size was designed to have 80% power with α of .05. An average of 12 MR clients per month was expected per pharmacy (unpublished data). Allowing for 50% refusal among MR clients and 30% refusal among pharmacies, we estimated that we would need to approach 110 pharmacies to complete recruitment within 3 months. Pharmacy workers were instructed to keep a tally sheet of all clients purchasing MR medications in order to estimate response rates.

      2.4 Analysis

      We double-entered data into Epi Data (version 2.0.3.15), exported data into SPSS (version 18) and conducted descriptive analysis. We compared outcomes between misoprostol-only users and mifepristone+misoprostol users. We intended to compare outcomes between women who did and did not have a prescription for the combined regimen, but the number of participants with a prescription was too small for this analysis to be conducted. Comparisons between groups were made using χ2 tests, or Fisher's Exact Test when the sample size was less than five per cell.

      2.5 Ethics

      We obtained ethical approval from the Bangladesh Medical Research Council and the Marie Stopes International Ethical Review Committee.

      3. Results

      Fig. 1 details recruitment processes and loss to follow-up. Of 110 pharmacies approached, 76 agreed to take part (69%). Almost a third of clients during the study period were not informed about the study by the pharmacy worker (31%). Of the 642 pharmacy clients who were informed about the study by a pharmacy worker, the response rate was 30%. Clients were recruited from 22 of the 76 pharmacies. Among those pharmacies that recruited clients, the total number of respondents ranged from 1 to 18 per pharmacy, with a mean of 5 (median: 4) respondents recruited per pharmacy over the 3-month study period. The follow-up rate was 87%, among whom 109 respondents were eligible for a complete interview.
      Respondent characteristics are displayed in Table 1. Respondent characteristics were not significantly different between those who purchased the drugs themselves and those who purchased drugs through a proxy purchaser (data not shown). The majority of women had used mifepristone+misoprostol (79.8%) versus 18.3% who used misoprostol-only. Two respondents did not know what drug they had purchased. Only two respondents had a prescription, and three did not state whether they had a prescription (all mifepristone+misoprostol purchasers).
      Table 1Characteristics of respondents (n=109)
      n (%)
      Age
       Under 2013(11.9)
       21–2537(33.9)
       26–3035(32.1)
       Over 3024(22.0)
      Education
       None or incomplete primary18(16.5)
       Primary complete16(14.7)
       Secondary incomplete48(44.0)
       Secondary complete or higher27(24.8)
      Employment status
       Housewife81(74.3)
       Employed19(17.5)
       Student8(7.3)
       Unemployed1(0.9)
      Marital status
       Married107(98.2)
      Number of living children
       013(11.9)
       136(33.0)
       236(33.0)
       3+24(22.1)
      Who purchased the pills
       End user55(50.5)
       Husband44(40.4)
       Other relative, friend or neighbor10(9.2)
      Prescription status
       No prescription104(95.4)
       Had a prescription2(1.8)
       No response3(2.8)
      Type of medication purchased
       Misoprostol-only20(18.3)
       Mifepristone+misoprostol87(79.8)
       Not known2(1.8)
       Mean weeks of pregnancy5.7 weeks (range: 2–12 weeks)
      The mean self-reported number of weeks of pregnancy of the clients at the time of taking the pills was 5.7 weeks (SD 1.8) (Table 1) and ranged from 2 to 12 weeks. Most women (89.9%) did not have their eligibility for taking the medication assessed by the pharmacy worker (criteria included: pain in lower abdomen, intrauterine device inserted, severe asthma, long-term treatment for a medical condition, heart disease, bleeding problems) (Table 2). In an open-ended question about the information received from the pharmacy worker, women reported receiving very little information about the MR medications; 40.4% reported receiving no information. Few (8.4%) received written information or pictures on how to take the drugs. However, 97.2% stated that they administered the drugs as advised by the pharmacy staff. Most women (80.4%) reported that they were not informed of any warning signs, and 50.5% were not told what to do in the event of complications.
      Table 2Information provision and dosages sold by pharmacy workers (n=109)
      n (%)
      Information provided by pharmacy worker
       None44(40.4)
      Information on warning signs
      Open-ended multiresponse question.
       None86(80.4)
      What to do in case of danger signs/complications
      Open-ended multiresponse question.
       Not told what to do52(50.5)
      Eligibility criteria checked
       No eligibility criteria checked98(89.9)
       Written information or pictures provided showing how to take pills9(8.4)
      Was the regimen taken advised by the pharmacy worker?
       Yes106(97.2)
       No1(0.9)
       Don't know2(1.8)
      Mifepristone+misoprostol dosage sold by pharmacy worker (n=82)
       Combination mifepristone+misoprostol pack including 5 pills70(80.5)
      Misoprostol-only dosage sold by pharmacy worker (n=20)
       Under 800 mcg4(20.0)
       800 mcg5(25.0)
       1000 mcg2(10.0)
       1200 mcg5(25.0)
       2400 mcg4(20.0)
      a Open-ended multiresponse question.
      Women who used the combination mifepristone+misoprostol regimen reported purchasing brands that are sold as a combination pack of 200 mcg mifepristone and 800 mcg misoprostol. Among women who purchased the mifepristone+misoprostol regimen without prescription (n=87), 80.5% were sold five pills, the correct number for the WHO-approved mifepristone+misoprostol regimen (Table 2). Sixty-nine percent took the correct regimen (200 mcg mifepristone followed by 800 mcg misoprostol after a 24-h interval) [
      • World Health Organization
      Safe abortion: technical and policy guidance for health systems.
      ] (data not shown).
      Among the 20 women who purchased misoprostol-only, 4 were sold less than 800 mcg, 5 were sold 800 mcg, 7 were sold between 800 and 2400 mcg, and 4 were sold 2400 mcg. The recommended misoprostol-only regimen for MR up to 12 weeks is an initial dose of 800 mcg of misoprostol, with subsequent doses of 800 mcg administered as needed [
      • World Health Organization
      Safe abortion: technical and policy guidance for health systems.
      ]. Nine women were therefore sold a correct dosage of the drug (800 or 2400 mcg), but additional purchases may have been required for those who purchased less than 2400 mcg (Table 2).
      Among all participants, expected self-reported common side effects were cramping (79.4%); feeling very unwell with nausea, vomiting or abdominal pain (39.3%); and fever (19.6%). Self-reported potential complications were rarer, with 68.8% not experiencing any complications described in Table 3. The most common self-reported potential complications were heavy, prolonged bleeding (12.7%) or minimal or no bleeding (10.8%). The only side effect or complication that differed between misoprostol-only and mifepristone+misoprostol users was minimal or no bleeding (22.7% vs. 5.8%, p=.021), possibly reflecting underdosing and ongoing pregnancy. Eight (7.5%) women had sought additional treatment; of these, five reported that they received an injection or tablet, one reported receiving a blood transfusion, and two received other unspecified treatment (data not shown). The source of treatment was a pharmacy for five of the women and a general practitioner for two of the women (data not shown).
      Table 3Experience and outcomes of women using pharmacy-purchased MR, by drug regimen
      All users (n=109)Misoprostol-only users (n=20)Mifepristone+misoprostol users (n=87)
      n (%)n (%)n (%)p value
      Experience of side effects in the 15 days since taking the first pill
       Fever21(19.6)2(10.0)18(20.7).351
       Feeling very unwell: nausea, vomiting, abdominal pain42(39.3)8(40.0)33(37.9)1
       Cramping85(79.4)17(85.0)66(75.9).557
      Experience of potential complications in the 15 days since taking the first pill
       None75(68.8)14(70.0)60(69.0)1
       Minimal or no bleeding
      Comparison between misoprostol-only and mifepristone+misoprostol users is significant (p<.05) using Fisher's Exact Test.
      11(10.0)5(22.7)5(5.8).021
       Heavy, prolonged bleeding that made you feel faint or lasted beyond 4–6 h of last tablet14(12.7)2(9.1)12(14.0)1
      Low blood pressure2(1.8)0(0.0)2(2.3)1
      Foul-smelling vaginal discharge1(0.9)1(4.5)0(0.0).19
      Severe low abdominal pain that persisted or spread and was not reduced by pain medicines6(5.5)0(0.0)6(7.0).59
      Shoulder pain along with severe pain in your abdomen1(0.9)0(0.0)1(1.2)1
      Effectiveness of MR (self-reported)
       Thinks she is no longer pregnant
      Comparison between misoprostol-only and mifepristone+misoprostol users is significant (p<.05) using Fisher's Exact Test.
      98(89.9)15(75.0)82(94.3).020
      Continuing side effects at day 15
       Experiencing heavy bleeding11(10.1)1(5.0)8(9.2)1
       Fever15(13.8)3(15.0)11(12.6).73
       Bad cramping or pain5(4.6)0(0.0)4(4.6)1
      Additional treatment sought
       Woman sought additional treatment after taking the tablets8(7.3)0(0.0)7(8.0).341
      a Comparison between misoprostol-only and mifepristone+misoprostol users is significant (p<.05) using Fisher's Exact Test.
      At day 15, 89.9% of participants reported that they were not pregnant. This was higher among mifepristone+misoprostol users (n=82; 94.3%) than misoprostol-only users (n=15; 75.0%) (p=.020). However, some participants were experiencing symptoms on day 15 that suggested the need for medical attention, including fever (13.8%), heavy bleeding (10.1%) and bad cramping or pain (4.6%) (no difference between drug types).

      4. Discussion

      This study assessed the feasibility of following up women self-administering misoprostol or mifepristone+misoprostol purchased from pharmacies in Bangladesh to examine their clinical outcomes.

      4.1 Feasibility of the study methodology for following up women who self-administer MR medications purchased from pharmacies

      This study has highlighted methodological challenges for conducting research on pharmacy provision of MR medications. Using pharmacy workers to recruit clients into the study did not yield the planned sample size. Pharmacies did not inform all MR clients about the study, which may have been because the recruiting pharmacy worker was not present or was too busy to inform clients. We recruited most respondents from a small number of pharmacies, and those pharmacies that were better able to recruit clients may systematically differ from pharmacies with low to no recruitment with regard to the information that they provided to clients. Therefore, the experiences of our respondents may not be representative of all pharmacy clients. In future research, staggered payment to pharmacy workers for study participation and regular monitoring may improve study engagement, and including all pharmacy workers at each outlet may help to reduce the number of clients not informed about the study. Collecting detailed data on the characteristics of pharmacies that are included in similar studies in the future would also enable better understanding of how pharmacy characteristics might influence information provision and women's experiences.
      According to the pharmacy workers, purchasers were reluctant to participate in the research. This contrasts with our previous experience interviewing MR clients in clinics, which may be because pharmacies are intended to be a faster, anonymous option or because clinics are viewed as more trustworthy than pharmacies. Response rates in this study were low (30%), and those who agreed to participate likely differed from those who did not, which may have affected our estimates of information provision, regimen taken and reporting of outcomes. Future research needs to ensure that recruitment processes are nonthreatening and convenient. Recruitment rates could be increased by placement of research assistants in pharmacies, investing in recruitment from a much larger sample of pharmacies or selection of pharmacies known to have higher client volumes.
      Participant follow-up was more successful than recruitment (87%). However, the study's reliance on self-report means that estimates of side effects, complications and completion may not be aligned with clinical definitions. Measures to assist self-report of outcomes would strengthen further studies attempting to distinguish between complications and side effects and to assess completion. In future studies, longer follow-up times may also help to verify self-report of MR completion. Similarly, since pregnancy tests were not used to verify pregnancy status, we cannot be certain that women were pregnant before they took the medications. In future studies, women should be asked how they knew that they were or were not pregnant to assist with verification of self-reported pregnancy status and self-reported completion rates.

      4.2 Safety and effectiveness of MR medications purchased from pharmacies

      There is a dearth of prospective research assessing the process and outcomes of pharmacy-purchased, self-managed MR medications. The study implementation itself illustrates the factors hindering communication of adequate information to users when MR medications are purchased from pharmacies: they lacked space for privacy, pharmacy workers were too busy or unavailable, and drugs were sometimes purchased by friends or family members. Women's ability to distinguish side effects from complications or to self-assess completion without adequate information may also hinder care-seeking after self-administration.
      Our study found that pharmacy workers did not adequately screen clients for eligibility before selling the medications or provide adequate information about warning signs and complications. These findings are in line with previous studies of pharmacy practice for MR in Bangladesh [
      • Huda F.A.
      • Ngo T.D.
      • Ahmed A.
      • Alam A.
      • Reichenbach L.
      Availability and provision of misoprostol and other medicines for menstrual regulation among pharmacies in Bangladesh via mystery client survey.
      ]. Despite this, 97% of participants stated that they had taken the drugs as advised by pharmacy workers, suggesting that although very little advice was provided, women felt that they had been given some direction on how to use the pills. Most women who were sold mifepristone+misoprostol took the correct regimen, and 94.3% of users reported that the medications were effective at 2 weeks. A lower proportion of misoprostol-only users were given an adequate quantity of the drugs, and effectiveness at 2 weeks was lower (75%), although small numbers in this group mean estimates should be treated with caution. The estimated completion rates in our study are within the range of expected effectiveness of clinic provision of mifepristone+misoprostol (95%–98%) [
      • World Health Organization
      Safe abortion: technical and policy guidance for health systems.
      ] and just below the range for misoprostol-only (78%–90%) [
      • Faúndes A.
      • Fiala C.
      • Tang O.S.
      • Velasco A.
      Misoprostol for the termination of pregnancy up to 12 completed weeks of pregnancy.
      ,
      • Von Hertzen H.
      • Piaggio G.
      • Wojdyla D.
      • Marions L.
      • My Huong N.T.
      • Tang O.S.
      • et al.
      Two mifepristone doses and two intervals of misoprostol administration for termination of early pregnancy: a randomised factorial controlled equivalence trial.
      ,
      • Blum J.
      • Raghavan S.
      • Dabash R.
      • Ngoc N.T.N.
      • Chelli H.
      • Hajri S.
      • et al.
      Comparison of misoprostol-only and combined mifepristone–misoprostol regimens for home-based early medical abortion in Tunisia and Vietnam.
      ]. Lower-than-required doses of misoprostol may have been used as a result of misoprostol not being indicated for MR because instructions are not included in packaging. It could also reflect purchasers' inability to pay for a high-enough quantity of the drugs upfront. The continued use of misoprostol-only following the registration of mifepristone+misoprostol in 2013 may be due to cost, poor knowledge of the new regimen among purchasers or pharmacy workers, or lack of availability of mifepristone+misoprostol and should be studied further. Increasing availability of high-quality combination regimen products is important for reducing complications from incomplete abortion [
      • World Health Organization
      Safe abortion: technical and policy guidance for health systems.
      ].
      Further work is also needed to develop and test interventions to improve the quality of pharmacy provision of MR medications. While training can be effective in increasing knowledge, turnover of staff and low availability of trained persons in pharmacies can limit the sustainability and scalability of such interventions. Support mechanisms for both pharmacy workers and for women directly are needed to ensure that users have adequate information for safe self-management; this may include pharmacy worker training or individual detailing visits, written and pictorial instructions for self-management in product packaging, tools to support self-assessment of completion, and hotline support for women and pharmacy workers with phone numbers printed on product packaging.

      5. Conclusions

      Understanding the experiences and outcomes of women who purchase MR medications from pharmacies is critical given the increasing scale of this practice. This feasibility study offers important lessons on appropriate methodologies to recruit and follow up pharmacy users and to collect data on the outcomes of women who purchase MR medications from pharmacies. Further development of this methodology is needed, particularly in countries where misoprostol is commonly used alone, as inadequate quantities of the drugs appeared to be commonly sold. The study has also shown that increasing access to accurate information and referrals in case of complications, as well as access to correct quantities of these drugs (in particular mifepristone+misoprostol), is required to reduce potential harms for women caused by incorrect dosing and administration. Measures to support safe self-management might include registration of the combination regimen (where not yet available), pharmacy worker training, hotline/call center support, informational materials in product packaging, and user tools to assess eligibility and completion.

      Acknowledgment

      This research was funded by UKaid from the Department for International Development. The funders had no role in the study design, data collection, analysis, interpretation or write up.

      References

        • Sneeringer R.K.
        • Billings D.L.
        • Ganatra B.
        • Baird T.L.
        Roles of pharmacists in expanding access to safe and effective medical abortion in developing countries: a review of the literature.
        J Public Health Policy. 2012; 33: 218-229
        • Hyman A.
        • Blanchard K.
        • Coeytaux F.
        • Grossman D.
        • Teixeira A.
        Misoprostol in women's hands: a harm reduction strategy for unsafe abortion.
        Contraception. 2013; 87: 128-130
        • Miller S.
        • Lehman T.
        • Campbell M.
        • Hemmerling A.
        • Brito Anderson S.
        • Rodriguez H.
        • et al.
        Misoprostol and declining abortion-related morbidity in Santo Domingo, Dominican Republic: a temporal association.
        BJOG. 2005; 112: 1291-1296
        • Åhman E.
        • Shah I.H.
        New estimates and trends regarding unsafe abortion mortality.
        Int J Gynecol Obstet. 2011; 115: 121-126
        • WHO
        Expanding health worker roles to help improve access to safe abortion and post-abortion care: WHO launches new guideline.
        2015
        • Giri A.
        • Srivastav V.
        • Suwal A.
        • Sharma B.
        A study of complications following self-administration with medical abortion pills.
        Nepal J Obstet Gynaecol. 2015; 10: 20-24
        • Nivedita K.
        • Shanthini F.
        Is it safe to provide abortion pills over the counter? A study on outcome following self-medication with abortion pills.
        J Clin Diagn Res. 2015; 9: QC01-QC04
        • Brahmi D.
        • Maxwell L.
        • Collado M.E.
        • Schiavon R.
        • Montaño P.C.
        • Sánchez R.C.
        Pharmacy recruitment of misoprostol users in Mexico: a feasibility study.
        Contraception. 2014; 90: 299
        • Chowdhury S.N.M.
        • Moni D.
        A situation analysis of the menstrual regulation programme in Bangladesh.
        Reprod Health Matters. 2004; 12: 95-104
        • Vlassoff M.
        • Hossain A.
        • Maddow-Zimet I.
        • Singh S.
        • Bhuiyan H.U.
        Menstrual regulation and postabortion care in Bangladesh: factors associated with access to and quality of services.
        Guttmacher Institute, New York2012
        • Government of the People's Republic of Bangladesh Directorate General of Family Planning
        Minutes of 63rd meeting of the National Technical Meeting.
        ([Dhaka, Bangladesh])2014
        • Babu M.
        Factors contributing to the purchase of over the counter (OTC) drugs in Bangladesh: an empirical study.
        Int J Third World Med. 2007; 6: 1-8
      1. Directorate general of drug administration, Bangladesh.
        (n.d.)
        • World Health Organization
        Safe abortion: technical and policy guidance for health systems.
        ([Geneva])2012
        • Huda F.A.
        • Ngo T.D.
        • Ahmed A.
        • Alam A.
        • Reichenbach L.
        Availability and provision of misoprostol and other medicines for menstrual regulation among pharmacies in Bangladesh via mystery client survey.
        Int J Gynecol Obstet. 2014; 124: 164-168
        • Faúndes A.
        • Fiala C.
        • Tang O.S.
        • Velasco A.
        Misoprostol for the termination of pregnancy up to 12 completed weeks of pregnancy.
        Int J Gynecol Obstet. 2007; 99: 172-177
        • Von Hertzen H.
        • Piaggio G.
        • Wojdyla D.
        • Marions L.
        • My Huong N.T.
        • Tang O.S.
        • et al.
        Two mifepristone doses and two intervals of misoprostol administration for termination of early pregnancy: a randomised factorial controlled equivalence trial.
        Lancet. 2007; 369: 1938-1946
        • Blum J.
        • Raghavan S.
        • Dabash R.
        • Ngoc N.T.N.
        • Chelli H.
        • Hajri S.
        • et al.
        Comparison of misoprostol-only and combined mifepristone–misoprostol regimens for home-based early medical abortion in Tunisia and Vietnam.
        Int J Gynecol Obstet. 2012; 118: 166-171