Abstract
<title>Abstract</title> <p>Background Artificial intelligence (AI) is increasingly integrated into modern healthcare, serving as a critical clinical supportive tool. However, geographical and linguistic variations necessitate local independent validation of these digital diagnostic applications, particularly within the Middle East. This study aimed to evaluate the diagnostic and therapeutic efficacy of intelligent symptom checkers against established physician clinical diagnoses. Methods This retrospective, cross-sectional study analyzed 100 diverse clinical cases collected from five community pharmacies in Damascus, Syria. Patient-reported symptoms and clinical histories were systematically evaluated using two AI applications in the local healthcare contex: Ada ("Check your Health") and Symptomate. Software-generated diagnostic outputs and therapeutic recommendations were recorded and statistically compared against the physician's prescription (the gold standard) using Cohen’s Kappa, Chi-Square, and McNemar’s tests. Results Ada demonstrated superior diagnostic accuracy with an 82% absolute match and 17% correlation rate compared to Symptomate (68% match, 25% correlation). For therapeutic regimens, Ada showed a higher concordance rate (49%) than Symptomate (37%). Inferential statistics confirmed a significant difference in diagnostic proficiency (McNemar /X^2 = 5.14, P = 0.023) and treatment profiles (X^2 = 7.56, P = 0.0059). Symptomate exhibited a higher risk-averse threshold, deferring to clinical consultation in 53% of cases compared to Ada's 43%. Conclusions Both applications demonstrate robust alignment with human clinical reasoning, confirming their value as clinical support tools. Ada proved more algorithmically consistent across medical specialties, while both tools appropriately emphasize physician reliance for complex and emergency clinical triage.</p>