Estudio comparativo de herramientas de inteligencia artificial para el diagnóstico y orientación diagnóstica basada en síntomas

Abstract

Introduction: Entelai Doc is an AI-driven healthcare system that allows patients to resolve medical queries based on their symptoms. The aim of our study is to compare the diagnostic accuracy of Entelai Doc with another symptom checker tool with regulatory approval.

Materials and Methods: We used clinical cases published in the medical literature, involving patients over 16 years of age and excluding pregnant women. Cases were selected based on the most frequent reasons for medical consultation. The performance of Entelai Doc (Aenti SRL, Argentina) was analyzed in terms of diagnostic effectiveness, accurate severity assessment, appropriate case referral, and the number of questions asked. Each variable was statistically compared with Ada (Company Name, Germany).

Results: A total of 68 cases were evaluated using Entelai Doc and ADA. The effectiveness in reaching the definitive diagnosis was 45.5% for both tools. However, when considering the top 5 diagnoses, Entelai Doc showed an effectiveness of 98.5% vs. 72% for ADA (p<0.0001). The effectiveness in case referral and determination of appropriate urgency level was high for both tools (98.5% vs. 95.6%). Nevertheless, the medical approach based on the definitive diagnosis (100% vs. 71%; p<0.0001) and top 5 diagnoses (100% vs. 81%; p<<0.0001) was superior for Entelai Doc. The median number of questions asked by Entelai Doc was 18 vs. 31 for ADA (p<0.0001).

Conclusion: Entelai Doc and Ada adequately address clinical cases of the most common reasons for medical consultation. Entelai Doc demonstrated higher effectiveness in guiding the medical approach according to the diagnosis. AI-driven symptom checker tools offer a high degree of precision in referrals, with the next step being a performance comparison in a clinical setting alongside medical professionals.