Enhancing Medication Risk Communication in Developing Countries: A Mixed Method Survey Among Doctors and Pharmacists in Malaysia


 Background: Medication risk communication is essential to ensure the safe use of medicines. However, very few nations worldwide have established effective risk communication systems. To date, the effectiveness of risk communication among healthcare professionals in Malaysia has never been evaluated. Our study aimed to (i) evaluate doctors’ and pharmacists’ awareness of risk communication methods; (ii) identify factors predicting the usefulness of these methods; and (iii) compare respondents’ preferences for risk communication to outline suggestions for enhancement. Methods: We conducted a nationwide cross-sectional survey covering four commonly used risk communications, namely a national drug bulletin, safety alerts, Direct Healthcare Professional Communication letters (DHPCs), and educational materials. Multiple logistic regression analysis was used to determine the association between independent variables and the usefulness of risk communication. We performed qualitative analysis to gain insights on respondents’ perspectives on risk communication.Results: Of the 1146 responses received, 650 were from pharmacists (56.7%). Among the four methods surveyed, 71.5% of respondents were aware of educational materials, while awareness of the other three methods ranged from 20.7% to 53.9%. Pharmacists had higher awareness of all four methods compared to doctors. Private sector respondents were more aware of DHPCs compared to those from the public sector. The strongest predictors for finding risk communication useful were being a pharmacist [odds ratio (OR) = 18.2; 95% CI (10.98-30.07); p < 0.001], having ≥30 years’ work experience [OR= 4.9; 95% CI 1.98-12.08; p < 0.001], and working in the pharmaceutical industry [OR= 4.6; 95% CI (1.08-19.72); p = 0.039]. Both doctors and pharmacists preferred risk communication in the English-language and electronic format. However, other preferences differed between the professions and sectors. Analysis of free-text comments revealed five core themes to guide risk communication enhancement strategies.Conclusions: Risk communication awareness differed between public and private sector doctors and pharmacists depending on communication source. Integrating our findings with the theory of effective communication, we provide suggestions for developing strategic plans on enhancing risk communication. Public-private sector collaboration is key in ensuring risk communication effectiveness.


Study design
We conducted a cross-sectional, multicentre, self-administered web-based survey involving doctors and pharmacists across Malaysia from March to June 2021.

Study population
The Malaysian healthcare system is divided into the government-funded public sector and the pay-forservice private sector. [15,16] The public healthcare sector, estimated to serve approximately 65% of the population, comprises hospitals and health clinics where services are heavily subsidised. Meanwhile, the private sector consists of a network of hospitals, specialist or general practitioner clinics, and retail pharmacies. The types and brands of medicines used in these 2 sectors usually vary because of differences in funding. [15] The study population comprises all registered doctors and pharmacists in the Malaysian public and private healthcare sectors. As of June 2020, there were 71,041 doctors (51,912 serving with the Ministry of Health-MOH) and 19,341 pharmacists (11,616 serving with the MOH) registered in Malaysia. [17] Inclusion and exclusion criteria All fully registered doctors and pharmacists who gave their consent to participate in the online survey were included. Doctors and pharmacists undergoing pre-registration training (house o cers and provisionally registered pharmacists) and those not currently practising in Malaysia were excluded.

Sample size
Based on previous surveys involving doctors and pharmacists in Malaysia, the calculated minimum sample size required would be 383, with a 5% margin of error and con dence interval of 95%. [18,19] However, based on the documented response rate of about 35% for similar studies, the calculated sample size needed for this study was 1100.

Questionnaire
We adapted the English-language survey questions from the Strengthening Collaboration for Operating Pharmacovigilance in Europe (SCOPE) Joint Action Work Package 6: Healthcare Professional Survey, [20] with some modi cations based on the literature review. The adapted questionnaire contained six domains with 33 items related to medication risk communication.
Our questionnaire covered four risk communication methods used in Malaysia, namely a national drug bulletin called the Malaysian Adverse Drug Reactions Advisory Committee (MADRAC) bulletin, NPRA safety alerts, DHPCs, and educational materials. Table 1 [21,22] shows a comparison of these four methods. For each method, respondents were asked to respond on (a) their awareness of the method[a], (b) how useful they nd the method [b], and (c) how likely they would take the action recommended in the communication [c].
We assessed the preferences of respondents regarding risk communication and perceptions on the factors that may affect their response to the risk communication through eight questions. The questionnaire included ve optional free-text boxes allowing respondents to leave comments on each of the four methods and suggestions for improving risk communication.
For assessment of face and content validity of the questionnaire, we distributed the adapted questionnaire to a group of six experts in medication risk communication or pharmacovigilance. The experts were asked to rate their judgment on the relevance of each item to the measured domain using a 4-point Likert scale. We calculated the content validity index (CVI) [see Additional le 1] to allow objective assessment of content validity, using a CVI cut-off score of at least 0.83 as evaluation was carried out by six experts. [23][24][25][26] A scale-level CVI averaging (S-CVI/Avg) score of 0.97 was obtained, and we modi ed the questionnaire according to comments from the expert panel. Informed consent was obtained from respondents on the rst page of the questionnaire. The respondents participated voluntarily and were not given any form of remuneration. We disseminated reminder emails at 3 and 8-weeks following the initial questionnaire distribution.

Quantitative analysis
We analysed the data using IBM SPSS version 26 (SPSS Inc., Chicago, Illinois). Categorical data were presented as absolute numbers or percentages, while means and standard deviations (SD) were determined for the continuous numerical variables. We used the Pearson Chi-squared test of independence or independent t-test to compare the demographic data as well as preferences of doctors and pharmacists on risk communication. Level of signi cance (α) was set at 0.05.
We performed multiple logistic regression analyses to determine the association of seven independent variables (gender, ethnicity, designation, work setting, work experience, strata, and training experience) with the usefulness of risk communication methods in Malaysia. The dependent variable "Overall, do you think NPRA risk communication is useful?" is a dichotomous measure coded 1= Yes and 0= No. Univariate logistic regression was performed to identify variables to be included in the model, based on a p < 0.25 signi cance level to ensure identi cation of variables known to be important. [27] The best t model was selected using the Hosmer-Lemeshow test and classi cation table.

Qualitative analysis
We analysed responses obtained from free-text comments using manual thematic analysis, according to the six-phase procedure outlined by Braun and Clarke (2019).
[28] Two researchers (RP, ZA) examined the data before independently assigning initial codes. The codes were reviewed and grouped together through concensus among the two members to identify common themes. We then reviewed these themes before generating a thematic map for the entire dataset. Coding and resulting theme generation were veri ed by the supervisory team. Footnote: [a] Four answer options were given [1 = Yes, I have seen and read it, 2 = Yes, I have seen but have never read it, 3 = No, I have heard of but never seen it, 4 = No, I have never heard of it].

Results
Socio-demographic characteristics A total of 1146 healthcare professionals completed the survey, comprising 650 pharmacists (56.7%) and 496 doctors (43.3%). [29] 1 compares respondents' awareness of the four risk communication methods surveyed, how useful they nd each method, and how often they take the recommended risk minimization actions. Overall, 71.5% of respondents had seen and read educational materials, [29] while only 20.7% were aware of DHPCs. Fig. 1(a) shows that pharmacists had signi cantly higher awareness of all four methods compared to doctors. Over 70% of pharmacists had seen and read the MADRAC Bulletin and NPRA Safety Alerts, as opposed to less than 26% of doctors. Over 60% of doctors had never heard of the MADRAC Bulletin and DHPCs.
More than 80% of respondents found each of the four methods useful or very useful, with the highest usefulness reported for educational materials (90.2%), as shown in Fig. 1(b). Respondents received educational materials most often from pharmaceutical companies (56.1%), compared to from NPRA (28.2%) or other MOH sources (40.8%). Fig. 1(c) shows that DHPCs were the most likely risk communication method to encourage respondents to take the recommended risk minimization actions, with 67.2% "often" or "always" taking the recommended action, followed by NPRA Safety Alerts (61.9%), and the MADRAC Bulletin (47.9%).
Comparison of awareness between public and private sectors Predictors of risk communication usefulness     Fig. 3, ve core themes on risk communication enhancement emerged from analysis of the comments and are supported by illustrative quotations from the participants.

Awareness and publicity
The most commonly recurring theme was to increase awareness of the risk communication methods especially among doctors and the private healthcare sector. "Never heard of NPRA, should consider promotion of this service" [respondent 179 (R179), doctor, 38 years]. "Reach out more to community pharmacy or general practitioner settings" (R021, pharmacist, 34 years). "I think the NPRA needs to be more present in platforms commonly accessed by doctors" (R120, doctor, 40 years). Respondents suggested increasing the presence of NPRA in social media. Many respondents were keen for virtual continuing medical education (CME) sessions on medication safety updates to be held regularly.

Attractive and concise content
Majority of respondents stated a preference for more concise, evidence-based, and original content. One respondent stated that the MADRAC Bulletin and Safety Alerts contained "mainly cut and paste" information (R030, doctor, 57 years). "Make the risk communication simple and succinct" (R022, pharmacist, 57 years). Many suggested the use of infographics, with a standardized layout to make all risk communication from the national regulatory authority easily recognizable.

Dissemination
Respondents preferred more frequent risk communication, with repetition of important messages. "NPRA needs to send same messages repeatedly so that we are aware" (R096, doctor, 52 years). Various suggestions were provided to improve the communication outreach, such as collaboration with professional bodies, dissemination through trusted sources including pharmacists or the administrative heads of healthcare departments and sending brief safety alerts via mobile phone with links to further details.

Technological advances
Many respondents were keen for a mobile phone application to receive medication risk communication and suggested an "app with automatic messages on drug safety that can be tailored according to specialty area" (R029, doctor, 37 years). Some speci c comments were also received regarding improvements required on the regulatory agency website. Of the four available risk communication methods assessed, doctors and pharmacists had the highest awareness of educational materials. This is probably because these materials are often made available through multiple sources for a prolonged period, [22] resulting in increased awareness. About half of the respondents were aware of the MADRAC Bulletin and NPRA Safety Alerts which are directly distributed by the national regulatory authority, most of those aware being pharmacists from the public healthcare sector. Awareness was lowest for DHPCs, which are solely distributed by pharmaceutical companies to users of the medicinal product mentioned in the letter. Our ndings indicate that the methods of risk communication dissemination impact the awareness levels.
Almost 60% of respondents had never heard of DHPCs, although these communications have been widely used worldwide since 2012. [14] However, compared to healthcare professionals from the public sector, those from the private sector had a higher level of awareness of DHPCs. This is possibly because those from the private sector are more often directly involved in purchasing medication from the pharmaceutical companies. Internationally, studies have shown DHPCs to be ineffective due to several factors. DHPCs have been viewed as a defensive practice to transfer responsibility from manufacturers to prescribers, or a form of advertising as it is signed and distributed by the pharmaceutical companies. [14] Surprisingly, our study found that DHPCs were the most effective risk communication method in prompting respondents to take the recommended risk minimization actions. An earlier study also showed DHPCs to be more effective than drug bulletins. [30] Thus, changes to the DHPC dissemination methods are needed to fully utilize this promising communication channel. Besides distribution by the pharmaceutical companies, the DHPCs could be made accessible online through regulatory authority websites to ensure coverage of both public and private sectors. A recent change made in Denmark could be adapted by other countries, whereby DHPCs are still signed by the industry but are disseminated by the Danish Medicines Agency to increase levels of trust. [14] The national regulatory authority, Ministry of Health, and professional bodies emerged as the preferred sources of risk communication in this study, ahead of international regulatory agencies. This concurs with a previous study which reported that Asian health o cials preferred guidance from in-country sources. [31] Trust in the source is a major factor in the success of risk communication, with low trust resulting in decreased uptake. [32] When organizations are not transparent in communicating risks, their credibility is reduced. [33] Our study shows that pharmacists are signi cantly more aware of medication risk communication compared to doctors, most likely because the Malaysian regulatory authority is part of the Ministry of Health Pharmacy Services Programme. A previous study on awareness of medication error reporting also revealed higher awareness among pharmacists.
[18] Pharmacists are a trusted source of medication safety information for other healthcare professionals, patients and the public. [12,34] As pharmacists are ideally placed across the country in various healthcare sectors, they can play a key role in increasing risk communication outreach to urban and rural areas, other healthcare professionals, policymakers, the pharmaceutical industry, and the general community. Regulatory authorities should conduct regular training for pharmacists on risk communication methods, latest safety issues and effective dissemination. Eventually, these trained pharmacists could serve as the trainers in their respective facilities.
Respondents' preferences on risk communication as reported in this study are in agreement with the seven key aspects of effective communication, namely a trusted sender, relevant context and content, clear and concise wording, repetition of the message, use of multiple channels to reach target recipients, and suitable for capability of the audience [35,36]. Therefore strategies to enhance risk communication should be planned based on these preferences.
It is encouraging to note that respondents who have received training in medication safety had a signi cantly higher awareness of NPRA risk communication.

Limitations
This study has several limitations. First, the use of non-probability sampling is not ideal but was unavoidable, as a list of complete sampling frame could not be obtained because of data protection issues. However, this limitation was minimized by obtaining a large sample size and responses from all states in Malaysia, including both urban and rural areas. The responses from doctors were lower than pharmacists, though their number is three times larger than pharmacists. While this was expected based on doctor response rates in other studies, it may increase the chances that data from doctors who responded are not representative of the population. Second, the use of an online survey platform may have excluded non-technologically savvy groups, such as those in areas without internet connection.
Third, the entire survey was conducted while the COVID-19 pandemic was raging in Malaysia, therefore there may be a bias in terms of the groups who were likely to participate as many clinicians would have been very occupied with the pandemic response. Finally, our study is subject to the known biases of cross-sectional, self-reporting survey methodology including recall bias, answering tendencies, and misunderstanding on terminology. However we reduced this limitation by validating the questionnaire with experts and piloting it among the target population. We also added qualitative research methods to provide a deeper understanding of the respondents' preferences on risk communication. This will help guide the development of a strategic plan for enhancing medication risk communication.

Future research
Effective risk communication is a multi-stage process. The initial steps are for the message to be sent and received, as have been evaluated in this study. This paper could be used to reach out and collaborate with other stakeholders. The constructive comments and suggestions provided by the respondents have been categorized into different target areas. These could serve as a guide for focus group discussions involving all stakeholders to develop a speci c, evidence-based, achievable and sustainable national strategic plan for the enhancement of risk communication.
Moving forward, we need to ensure the message is understood and prompts action or change in KKM/NIHSEC/ P20-2263). All methods were carried out in accordance with relevant guidelines and regulations. We performed the study is in accordance with the principles of the Declaration of Helsinki. Participation in this research was entirely voluntary. We provided all respondents with information on the study aims, procedures, risks, bene ts, and protection for individual privacy. Written informed consent was obtained from all individual participants included in the study.

Consent for publication
Not applicable Availability of data and materials The analysis results generated in the current study are presented in this published article. The research datasets will not be publicly available because they contain sensitive material, identifying participant information. The datasets generated for this study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This research did not receive funding from any agencies in the public, commercial, or not-for-pro t sectors.
Authors' contributions R.P. and Z.A. conceived and designed this study. R.P. developed the questionnaire, collected, analysed, and interpreted the data, and drafted the manuscript. Z.A. and A.K. reviewed the manuscript for important intellectual content. Z.A. supervised the data analysis and interpretation, critically revised the manuscript and provided nal approval of the version to be submitted. All authors read and approved the nal manuscript for publication. Thematic map of healthcare professionals' perspectives on medication risk communication

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.