Text Messaging for Diabetes Prevention and Care in Argentina: Implementation Challenges and Lessons Learned (Preprint)

BACKGROUND In Argentina, diabetes is a growing public health concern, with a prevalence of 14% in 2024 and projections reaching 15.4% by 2050. In this context, a Diabetes Prevention and Care Program was implemented in low-income areas across three provinces. A key component of the program was a WhatsApp-based intervention aimed at promoting self-care, encouraging healthy behaviors, and supporting follow-up among people with diabetes, those at risk, and pregnant women. OBJECTIVE This study describes the implementation of a WhatsApp-based intervention within Argentina’s public health system, focusing on challenges encountered, implementation strategies used, and lessons learned across the three target populations METHODS The intervention was implemented in 40 primary care centers. Participants included adults with type 2 diabetes, people with moderate or high risk based on the FINDRISC score, and pregnant women. A set of 192 educational and reminder messages was developed and validated through expert input and community feedback. Messages were tailored to each target population and delivered via an automated system integrated with WhatsApp. We assessed implementation fidelity using Carroll’s framework, focusing on message delivery, participant responsiveness, and contextual barriers. RESULTS A total of 11,029 participants were enrolled in the study, of whom 9,983 (90.5%) had a valid mobile phone number registered in the system. Among these, 32.8% had a diagnosis of type 2 diabetes, 53.3% were identified as being at moderate or high risk based on the FINDRISC questionnaire, and 13.9% were pregnant women. Overall, 67.3% accepted the intervention. Acceptance was highest among those with diabetes (74.3%) and lowest among at-risk people (62.1%). Message adherence was high: 91.1% read the initial opt-in message required for enrollment, and 88.7% of participants received at least the minimum number of educational messages expected. The dropout rate was low (6.1%) but higher among pregnant participants (14.6%). Message delivery issues mostly included problems with WhatsApp on the mobile phones of participants. Technical challenges, including server overload, were addressed during implementation. CONCLUSIONS The text-messaging intervention was feasible and well-received, particularly among people with diabetes in public primary care settings. These findings support the potential scalability of WhatsApp-based intervention in low-resource settings. Future initiatives should prioritize integration with electronic clinical records, personalized content, and digital inclusion for underserved populations.