The effect of community dialogues and sensitization on patient reporting of adverse events in rural Uganda: Uncontrolled before-after study

Background Patients experiencing adverse drug events (ADE) in many developing countries are in the best position to report these events to the authorities but need to be empowered to do so. Systematic evaluation of community engagement and patient support especially in rural areas would provide evidence for a program to monitor potential harm from medicines. The aim of this study was to assess the effects of a community dialogue and sensitization (CDS) program on the knowledge, attitude and practises of community members for reporting ADE. Methods This an uncontrolled before-after study was conducted in two eastern Ugandan districts between September 2016 and August 2017 Results After implementation of the community dialogue and sensitization (CDS) program, there was an overall 20% (95% CI:16% to 25%) increase in knowledge about ADE in the community compared to before the program began. Awareness levels increased by 50% (95% CI: 37% to 63%) among those with little or no education and by41% (95% CI: 31% to 52%) among young people (15–24 years). Furthermore, 5% (95% CI: 3% to 7%) more respondents recognized the need for reporting ADEs compared to before the program. Finally, there was a significant increase of 115% (95% CI:137% to 217%) in respondent recognition and reporting of ADEs compared to the beginning of the CDS program. Overall, this community found the CDS program acceptable and proposed aspects that could be improved for future use. Conclusion Our evaluation showed that the CDS program increased knowledge and improved attitudes by catalyzing discussions among community members and healthcare professionals on health issues and monitoring safety of medicines compared to before the program. Successful implementation of the program depends on holistic health systems strengthening and adaptation to the community’s way of life.

157 Assuming a baseline proportion with an acceptable level ρ of 0.5, and testing at the 0.05 158 level, a sample size of 389 for each survey was determined to give 80 per cent power to 159 detect a change of at least ten per cent of the primary outcome. To allow adjustment for 160 confounders, non-response and design effect, we doubled the sample size, to obtain the 161 required sample size. Sampling involved a single-stage household sampling. In each of the 65 162 villages in the IMHDSS surveillance area, the study team sampled an equal number of 163 households using a simple random sampling approach with the help of community leaders.
164 For the purpose of the surveys, a household was defined as a group of people who routinely 165 lived and ate together. One person per selected household was interviewed. The target for 166 interviews was the person best placed to answer questions about the household's health in the 167 community members' questionnaire (Additional file # 1). All community based health 168 facilities in the intervention area were included in the survey. At least two healthcare workers 169 from each health facility were randomly selected and interviewed using a specific healthcare 170 provider questionnaire (Additional file # 2). The facilities considered were both private and 171 public owned and these included hospitals, health centre VI-II, pharmacies, drug shops and 174 Data collection 175 A quantitative assessment of target communities' knowledge, attitudes and practices of 176 patient ADE reporting was done using identical questions before and after the CDS 177 intervention.

178
179 Interviews were carried out by local field researchers, using a pre-tested structured and 180 validated questionnaire translated in the local language (Lusoga). The field researchers were 181 instructed to read out the survey questions exactly as rendered on the questionnaire.
182 Instructions for field researchers with regard to whether the question required a single 183 response or whether multiple responses were possible and answer options were provided in 184 the questionnaire in English. Answer options were not read out aloud and field researchers 185 were instructed not to suggest answers to the respondent. An 'other' category was provided 186 for most questions and field researchers were instructed to note down respondents' answers if 187 they could not clearly assign the answer to an existing answer category.

188
189 All field team members had previous experience of conducting or supervising field research.
190 Field researchers and supervisors attended a two-day training course covering data collection 191 tools, field procedures and interview techniques. Field supervisors received an additional 192 days' training focusing on supervision of field teams as well as the sampling process.
193 Following the pre-test, a half-day training session was conducted to discuss challenges 194 identified during the pre-test. The training materials were prepared by MUCHAP and NDA 195 and the training was conducted by MUCHAP, who were responsible for coordination and 196 supervision during the field work. Field supervisors were tasked with monitoring the quality 197 of the data collected and seeking clarification from the field researchers where necessary. At 198 the end of each day, they were responsible for conducting a feedback meeting with their 12 226 they would report any ADE if encountered. The secondary outcome of this study was to 227 establish the best ways to engage the community and this was included as part of the 228 questionnaire during the surveys. 237 This is true though the number in baseline were more than those in end-line survey. Table 2 238 summarizes survey respondents' profile in terms of age, education, religion and occupation.