This study consisted of three in-person meetings held at month 1 (T0), month 3 (T3), and month 6 (T6). An online questionnaire was administered at the beginning of each meeting, focusing on the knowledge and practices of HCPs regarding medication adherence. The study targeted HCPs involved in the management of dyslipidemic patients from two centers: one in Mexico and one in Thailand.

Meetings in Mexico were held between November 2023 and May 2024, while those in Thailand took place between May and November 2024. The meetings included the administration of online questionnaires, lectures, and resources on medication adherence. They were led by the ambassadors Prof. Makarawate in Thailand, and Dr. Araiza Garaygordobil in Mexico. The first meeting last 1.5 h. Participants first completed the questionnaire to assess baseline knowledge and use of resources related to medication adherence. This was followed by a lecture on medication adherence, as well as access to some videos and articles for deeper understanding. Participants also received explanations and access to two medication adherence tools: Insight, and my a:care app.

The second and third meetings were each 45 min long. Participants completed the same questionnaire, followed by a refresher lecture and Q&A session to discuss their learnings and challenges in practice and with adherence tools. The third meeting included a final session to review learnings and discuss difficulties encountered during the program.

In addition to the lectures, participants had access to three resources throughout the study: (1) Online articles and videos on improving adherence through behavioral science and technology; (2) Access to Insight tool, an adaptative instrument measuring key patient-level risk factors for adherence problems in chronic disease, based on four key pillar: Social (patient’s perception of support and impact of treatment on their social role), Psychological (self-concept, reactance and discounting of future benefits), Usage (practical and control elements) and Rational (perceptions of treatment benefits, illness severity, and perceived risk for adverse outcomes) (14); and (3) Access to the my a:care app, a mHealth application designed to address medication non-adherence using established behavioral methods. HCP could recommend the app to patients to encourage and monitor adherence. The app link adherence to self-care and provides motivational messages, health insights, challenges, and pill reminders.

The questionnaire used in this study was developed specifically for the purpose of evaluating healthcare professionals’ knowledge, attitudes, and practices regarding medication adherence in dyslipidemia management. Items development was based on a review of the existing adherence literature addressing HCPs behavior and educational programs, as well as based on the authors expertise. Items assessing the learning of the training content delivered in the program were also added. Finally, authors from Thailand and Mexico also assess the questionnaire for relevance and cultural appropriateness. Meetings and emails discussions were carried out until final validation of the questionnaire. The questionnaire was then developed on the SurveyMonkey^® Platform, and diffused during the centers meeting via a QR code. Each participant had to acknowledge the privacy policy and sign a consent form before accessing the questionnaire, which took approximately 15 min to complete. The final questionnaire consisted of three parts: (1) Profiling, to collect information on respondents such as gender, time spent working in the center, occupation, number of patients seen per month, time spent per consultation and adherence-related habits; (2) Educational, to evaluate respondents’ knowledge of non-adherence and its evolution over time; and (3) Impact on clinical practice, to assess respondents’ practices and changes over time. Each item was answered using a Likert-scale to facilitate the analysis and observe changes over time.

The first questionnaire (T0) included the profiling and clinical practice section. Since the participants remained the same throughout the study, the profiling section was not repeated at T3 and T6. The T0 questionnaire was administered at the start of the project, before any interventions, to evaluate baseline knowledge and practices. The second questionnaire (T3) included a training assessment section to evaluate what participants had learned, as well as the clinical practice section, and was administered 3 months after the start of the project, before the refresher session. The third questionnaire (T6) was identical to the second, and was administered 3 months after the second meeting, before the final discussion session. All responses were closed-end and self-reported, reflecting the perceptions of the respondents. Questionnaires were translated in Spanish and Thai to facilitate the understanding of the questions. The English questionnaire and privacy policies are available as Supplementary materials.

The two selected centers were hospitals with no previous experience in adherence programs. Each project ambassador selected one center, based on the following criteria: at least 15 HCPs enrolled (no maximum), all from the same hospital or clinic (can be from different units) managing patients with dyslipidemia (cardiologists, nephrologists, endocrinologists, internists, nurses, pharmacists), able to attend to follow-up meetings every 3 months for 6 months, and open to implementing new practices in their daily consultations. Email and personal invitations were shared within the two centers, and HCPs willing to participates could register by contacting their respective ambassadors. All participants agreed to and signed a consent form before starting the program. Responses were collected only in aggregate and anonymized. The platform complied with the RGPD policies for data protection.

In the Thailand center, a small remuneration was offered to the participants to encourage their participation.

Descriptive and statistical analysis were performed. Given the pilot design of this study and the small number of participants, a 90% confidence interval (CI) was used to balance the limited sample size with the need to maintain reasonable statistical power. Using a 90% CI reduces the risk of false negatives and enhances sensitivity to detect potential trends in practice changes.

A Chi-squared test was conducted to evaluate significant differences, and for some questions, a mean test (Student’s test) was also performed to maximize the identification of potential trends. Analyses were performed using COSI v4.1 software. Some participants missed one of the follow-up meetings (T3 or T6), leading to incomplete response profiles. No imputation was performed; analyses used all available data at each timepoint, and results are presented as percentages to facilitate comparison despite varying denominators.

In Thailand, 27 HCPs were enrolled, between May and November 2024. All participants (100%) were nurses, most of whom dedicated 100% of their working time to the center and conducted more than 41 dyslipidemia consultations per month. On average, they saw 54 patients per month, with consultations lasting 15 to 20 min. Two participants reported using methods to assess the risk of non-adherence prior to the training (health behavior assessment forms).

In Mexico, 15 HCPs were enrolled, between November 2023 and May 2024. 60% of participants were cardiologists, 40% were general practitioners, and 80% were men. Most participants worked full-time at the study center and conducted more than 41 dyslipidemia consultations per month, for an average number of patients seen per month of 62, with average consultation durations of 16–30 min. Four participants reported prior use of methods to assess non-adherence risk (online cardiovascular risk scales, or direct questions on adherence, side effects, accessibility, and dosage of the treatment).

This section evaluated participants’ learning resulting from the training and self-study of educational materials. Overall, participants assimilated definitions from WHO and OECD reports (1, 4), although recalling the phases of adherence and phases of treatment adherence remained challenging. Both groups (Thailand and Mexico) showed increased correct responses regarding stakeholders’ involvement, with high accuracy rate by the end of the program. Nearly all participants believed by the end of the program that behavioral science can support adherence (14/15 in Mexico, and 27/27 in Thailand).

This study has several strengths. It provides an overview of how a structured training program was implemented within healthcare centers and its impact on clinical practice. The multicenter design, involving two countries from distinct socio-cultural backgrounds, strengthens the generalizability of the findings. Importantly, the training content and assessment tools were standardized across both sites, enhancing consistency and reinforcing the real-world applicability of this pilot study. The longitudinal design further allowed for the monitoring of potential changes in practices over time. However, some limitations must be acknowledged. First, the relatively small number of participants (<30) limits the statistical power and requires cautious interpretation of the results. Second, the absence of a control group makes it difficult to directly attribute observed changes to the training intervention alone. Third, declarative (self-reported) data may be subject to reporting biases, and potential selection bias cannot be excluded, as participants may have been more motivated or receptive to change. Additionally, the Hawthorne effect—behavior modification due to being observed—may have influenced responses. Finally, in the Thai center, only nurses participated due to the local care structure, which may have introduced some bias in responses to some questions.

This program may be extended to other centers, led by an ambassador within each center, using the same training content, program structure, and questionnaire, to confirm these results. Future studies should aim to build on these preliminary findings by including larger sample sizes and extending follow-up periods to better evaluate the long-term impact and sustainability of training-induced changes. Interventional trials could also help determine the effectiveness of specific components of the training. Further research should also explore whether these changes in HCPs practice translate into improved patient outcomes. In addition, studies are needed to assess how this type of personalized program could be developed in larger scale and implemented in general practices in diverse healthcare settings.