An intervention to improve the quality of medication abortion knowledge among pharmacists in India

To test an infographic two‐pager on medication abortions (MA) aimed to improve pharmacists counseling in India.

In other settings, researchers have used other approaches, such as mobile SMS messaging, to improve user knowledge. 6 Of the existing interventions for pharmacists regarding MA, all have been geared at intensive pharmacist-based trainings and most have followed a harm reduction strategy. 7,8 Harm reduction, an evidence-based framework, prioritizes approaches to lessen harm and maintain health in contexts where practices and/or policies may forbid, stigmatize, or force common activities underground. For example, in Nepal, researchers conducted harm reduction orientation and refresher trainings with pharmacists on safe use and provision of misoprostol and mifepristone and compared their reported change in knowledge of MA from baseline to follow-up with those pharmacists in a comparison district. They found trained pharmacists reported providing information on safe use of mifepristone and misoprostol for early first trimester MA more often compared to their counterparts in the comparison district. 8 Evidence from a MA operations study in central Zambia that also used a harm reduction training approach with pharmacists found at end line more pharmacists reported referring to a healthcare facility while fewer reported selling ineffective and unsafe abortion medications as compared to their baseline reports. 9 Other studies, while not focused on MA, have documented some effectiveness of materials and trainings to improve pharmacist quality of care provision of reproductive and other healthcare services. Most of these have included pharmacist training, in addition to written information or other resources such as referral systems. [10][11][12][13][14][15][16] While these studies document that pharmacists are more willing to discuss and provide reproductive health services, such as emergency contraception and condoms, and refer their customers to appropriate care following the intervention, none are able to assess longer-term impact of their interventions on pharmacist behaviours or on client-related reproductive health outcomes. 4 A recent intervention in Nepal found that pharmacists trained in MA provided safe and satisfactory care, which was sustained for about 5 years. 17 Given the increasing evidence that pharmacies are the first and primary point of contact for 11.5 million MA users in India, coupled with the clear dearth in tested pharmacist-based interventions to improve the quality of MA care, our team developed a simple informational and graphic handout (infographic) with information on MA dosing, timing, expected symptoms, side effects, and when to seek care. In India, MA is most commonly purchased and taken through a combination pack ("combi-packs") of one mifepristone (200 mg) and four misoprostol tablets (200 mcg each), thus two types of medication to be taken with specified time spacing must be explained. To our knowledge, this study is the first to develop and test a simple knowledge-based tool for pharmacists to improve quality of care of MA in India. This paper will describe the intervention and evaluate the impact of the intervention on pharmacists' knowledge about MA and its practices.

| MATERIALS AND METHODS
For the study intervention, we designed a two-page (one sheet of paper, back and front) information sheet with graphics (infographic), guided by a harm reduction framework. The goal of this infographic was to provide pharmacists with information about MA to improve the quality of their interactions with clients purchasing MA. The infographic was designed in Hindi (local language of study area) and informed by in-depth qualitative interviews that the research team conducted in 2016-2017 with pharmacists and clients from the same study area. Our team's formative research in India demonstrates lowquality provision of MA despite the growing amount and variety of MA packs available in pharmacies. 18  Pharmacists were approached, the study purpose was explained, and verbal consent was obtained after which the baseline survey data were digitally collected on tablets using Survey CTO software (Dobility, Cambridge, MA). The informed consent also included consent for mystery clients to visit the pharmacy at some point in the next 6 months before the end-line survey.
Of the 283 interviewed, 117 pharmacists were assigned randomly to the intervention group and 166 to the control group.
Marital status for women was indicated through wearing socially recognized symbols of married women, and for men it was conveyed to the pharmacist verbally in the course of the mystery client's conversation. Unmarried women also conveyed this verbally to the pharmacist in their conversation. The user profiles and interaction script were informed by the initial qualitative results of this study, to ensure that the scenario generated was realistic and reflected the concerns of the user populations. Mystery clients did not purchase the medication from the pharmacists, but rather, after all the counseling had been completed, asked the price, and declined, saying that they did not have enough money to purchase it. Immediately after the interview, mystery clients completed a short quantitative survey describing the interaction that took place with the pharmacist.

| Survey measures
Changes in seven outcomes that were measured in the survey both pre-and post-intervention were explored (Table 1).

| Mystery client measures
A number of indicators related to the nature and quality of information provided to mystery clients were explored ( Table 2).
T A B L E 2 Mystery client experiences by intervention and control pharmacists (logistic regression).

Quality indicator OR (95% CI)
Did the provider ask you when you/your wife last had her menstrual period or how many weeks pregnant you were?

| Survey analysis
Pre-and post-intervention differences in primary outcome measures among the intervention group were first tested. Next, a series of difference-in-differences models on the outcomes that showed evidence of change were run. Finally, pharmacists' perspectives on the infographic were described.

| Mystery client analysis
Differences were explored between the responses of mystery clients who visited a pharmacist who received the intervention and those who visited a pharmacist who did not receive the intervention, using logistic regression models and controlling for mystery client profile type.

| RESULTS
Results from the pre-post survey with pharmacists followed by key findings from the mystery client debrief interviews are below. We ran difference-in-differences models on three of the outcomes that were statistically significantly different between pre and post surveys (Table 3). Since so few pharmacists reported the incorrect answer, the numbers were too small to calculate a difference in the correct answer to the questions asking gestation age, confirming the woman was pregnant, and the appropriate mifepristone dose.

| Pre/post survey
The difference-in-differences estimator did not show evidence of a treatment effect of the intervention on pharmacists' knowledge of the appropriate time gap between mifepristone and misoprostol.
The intervention led to an increase in odds that a pharmacist knew there should be no time gap between the doses of misoprostol and the appropriate number of misoprostol tablets, route, and timing after mifepristone.
Most pharmacists showed the infographic to some respondents, with only 24.11% (n=21) reporting they showed it to no clients (Table 4). Similarly, most pharmacists gave the infographic to clients, with only 25.29% (n=22) giving it to no clients. Almost all (n=95, 91.35%) the pharmacists found the infographic handout useful, the majority thought it helped them provide better care (n=88, 84.62%), and 93.27% (n=97) would like something like it to give to clients.

| Mystery clients
Pharmacists in the intervention arm had higher odds (odds ratio [OR] 9.23, P<0.001) of showing MA instructions to the mystery clients (Table 2). They were also marginally significantly more likely to ask the mystery client for the gestational age and to provide the correct information about the appropriate mifepristone and misoprostol dose

| DISCUSSION
We find that a simple infographic, provided without extensive additional training or resources, is effective in increasing pharmacists' knowledge about MA. This finding is encouraging. It suggests that low-cost, easily distributable tools may be effective in improving knowledge about MA. However, this improvement in knowledge did not translate into many changes in practices in this study, although most changes were in the hypothesized direction. The one item that was significant (pharmacists being more likely to show the client instruction) suggests that intervention pharmacists might have seen the value of providing more information yet felt uncomfortable providing that on their own. Prior interventions with pharmacist providing MA have also struggled to make substantial changes in behaviour. 9 Previous interventions aimed to improve knowledge of pharmacists providing MA in South Asia have included more intensive training. The present study expands this to show that even a simple infographic without additional training can also improve knowledge and maybe behaviours. 8 The present study also adds to the noted paucity of highquality evaluation of interventions targeting pharmacists providing MA by using a mixed-methods approach including mystery clients. 20 The lack of significance of other measures may be study-designrelated-perhaps because the sample size of mystery clients was too small to detect differences. Additionally, mystery clients visited pharmacists before the end-line survey was conducted and only 1 month after the infographic was distributed. Thus, it is possible that pharmacists became more comfortable and familiar with the infographic over time and we may have seen similar results to the end-line survey had we collected mystery client data a few months later. It is also likely that more is needed than an infographic to change behaviours. Our infographic did not specifically motivate pharmacists to change behaviour, nor did we provide any training aimed at behaviour change. Perhaps training that included role-plays could enable pharmacists to translate this increase in knowledge into the actual provision of higher-quality information on MA to their clients.
Regardless, pharmacists welcomed the infographic and wanted this type of tool both for themselves and to give to their clients. In addition, it led to greater knowledge and increased sharing of information with clients visually. This suggests that expanding the provision of this infographic to pharmacists would be acceptable and could improve knowledge, and that carefully thinking through adapting this infographic (most likely with simpler language) for clients is key. This very simple intervention made measurable improvements in pharmacists' knowledge and could be quickly, easily, and economically scaled up, meeting pharmacists needs and, potentially, with additional time and edits, improving the quality of clients' experiences and information.

AUTHOR CONTRIBUTIONS
NDS designed the study, and led the data analysis and manuscript writing. BP aided in data analysis and manuscript writing. JP helped with manuscript writing and data coding, as well as training mystery clients.
MS helped with research-related activities and manuscript preparation. PD ran the data collection, data entry process, and manuscript preparation. AS helped co-design the study, oversaw research and data collection, and participated in manuscript preparation.
T A B L E 4 Intervention pharmacists' perspectives on the infographic.