Evaluation of a Clinical Decision Support System for Dermatology in a Remote Area: Insights from Martinique

Alice Callens¹Moustapha Dramé¹Julia Dugardin¹Emilie Baubion¹Adrien Perrier¹Romain Blaizot²Nicola Briand¹Emmanuelle Amazan¹Gladys Ferrati-Fidelin¹Ferdaous Khammouma¹Nicolas Olivier¹Melissa Heleine¹Aziz Baccouche¹Raymond Helenon³Aurélie Martin⁴Claire Jacquin⁵Anne-Laure MESSAGIER^1*

• ¹University Hospital of Martinique, Fort-de-France, Martinique
• ²Service de Dermatologie, Hôpital de Cayenne, Cayenne, French Guiana
• ³Private dermatology practice, Fort de France, Martinique
• ⁴Centre Hospitalier Universitaire de Nîmes, Nîmes, Languedoc-Roussillon, France
• ⁵Private dermatology practice, Nouméa, New Caledonia

Background: Dermatology faces workforce shortages in Martinique, a French Overseas Department, where general practitioners are often the first point of care. The lack of dermatology CDSS tools adapted to tropical and Francophone contexts, especially across diverse phototypes, underscores the need for innovative digital solutions. Objective: The Research Objective (RO) of this study was to evaluate a newly developed dermatology CDSS ("Dermagic") in Martinique, focusing on its reliability, usefulness, and effectiveness across diverse phototypes. To guide this evaluation, we formulated three Research Questions (RQ): What evidence supports acceptance and relevance among physicians? Which barriers and underused features may guide future improvements? How adaptable is the tool to other Francophone or tropical regions? Methods: A cross-sectional study was conducted from February 7 to March 7, 2024, using a 21-item questionnaire aligned with HONcode and Netscoring criteria. The survey was sent to 117 physicians; 64 responses were analyzed. Reliability, usefulness, and satisfaction were assessed, including the System Usability Scale (SUS). Subgroup analyses were performed using exact Mid-P tests to explore variability of use across specialties. Results: Respondents (response rate 55%) highlighted the tool's quick accessibility (98.4%), improved prescription facilitation (100%), ergonomic design (90.6%), and adaptability to local dermatology needs, including diverse phototypes (95.3%). Regular use was reported by 62% of physicians, with 20% using the tool daily and 42% weekly. Subgroup analysis revealed higher use of the Medical Coding module among dermatologists compared to other non-generalist specialists (p = 0.016), while no significant differences were observed between dermatologists and general practitioners. A high System Usability Scale (SUS) score of 87.7/100 indicated excellent usability, and 96% of users reported being satisfied. Conclusion: This CDSS is a reliable, user-friendly tool that supports dermatological practice in Martinique, particularly addressing the management of diverse phototypes and the needs of local doctors and patients. Limitations include the modest sample size and self-reported nature of responses. Future directions include cross-regional evaluations, integration of AI for diagnostic support, and independent validation studies to strengthen its scalability and impact.