Artificial Intelligence Symposium (THAIS)

Artificial Intelligence Symposium (THAIS). Parc Taulí. Sabadell

1. The challenge of efficiency: Primary Care. Are we ready? Josep Vidal Alaball (MD, PhD , MPH) Family Physician Head of Central Catalonia Research and Innovation Unit in Primary Care

2. Amara’s Law We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.

3. AI in Primary Care Conviction Desperation “It’s not so much about what jobs will we do, but how will we do our jobs.” “Everyone will do their job differently, working with machines over the next 20 years.” Economist Andrew Charlton

4. AI in Primary Care

5. Real examples of AI in Primary Care Retinography Dermatology Radiology Voice assistance

6. 2017 ! Validation of a Deep Learning Algorithm for Diabetic Retinopathy . Romero-Aroca P, Verges -Puig R, de la Torre J, Valls A, Relaño-Barambio N, Puig D, Baget-Bernaldiz M. Telemed J E Health . 2020 Aug;26(8):1001-1009. doi : 10.1089/tmj.2019.0137. Epub 2019 Nov 4. PMID: 31682189. [ Institut de Investigació Sanitària Pere Virgili (IISPV) de Reus] Retinography

7. Retinography Vidal-Alaball J, Arocas Bonache A, Solé-Casals J, Royo Fibla D, Marin-Gomez FX, Distéfano LN, Boixadera A, Casado-García Á , García-Domínguez M, Inés A, Heras J, Zapata MA Clinical validation of an artificial intelligence algorithms for the detection of different central-involved retinal pathologies and glaucoma by using a non-mydriatic camera; Heliyon 2024; under review A total of 1652 eyes from 871 patients [1652 images] were analyzed, Sensitivity/Specificity : 86.8%/95.6% for detecting DR 94.9%/94.3% for detecting age-related macular degeneration (AMD) 82.7%/92.4% for detecting glaucoma tous optic neuropathy (GON) 87.0%/87.5% for detecting epiretinal membrane 89.7%/98.0% for detecting nevus

8. Dermatology 44 Conditions Benign and malignant melanocytic pathology
Benign tumor pathology and malignancy
Inflammatory pathology
Infectious pathology
Genital pathology

9. GENERAL OBJECTIVES SPECIFIC OBJECTIVES To carry out a prospective validation of a machine learning (ML) model as a diagnostic support tool for cutaneous pathology in Primary Care. To evaluate the diagnostic accuracy and efficacy of AA models in a real clinical settings . Comparing their diagnostic capacity with that of Primary Care doctors and dermatologists. To detect which skin pathologies are not included i n the current model that contains 44 diagnoses. To know the level of satisfaction of Primary Care professionals with the use of the artificial intelligence model. DESIGN AND METHODOLOGY Prospective, multicenter observational pilot study. N= 100 patients . Study population: > 18 years old who consult for a skin lesion in the Primary Care of Central Catalonia. Escalé-Besa A, Fuster -Casanovas A, Börve A, Yélamos O, Fustà -Novell X, Esquius Rafat M, Marin-Gomez FX, Vidal-Alaball J. Using Artificial Intelligence as a Diagnostic Decision Support Tool in Skin Disease: Protocol for an Observational Prospective Cohort Study . JMIR Res Protoc . 2022 Aug 31;11(8):e37531. doi : 10.2196/37531. PMID: 36044249; PMCID: PMC9475422. Dermatology

10. Dermatology 36 pathologies in total 12 pathologies not included in the used version of the AI model ( Autoderm ®)

11. Radiology Risk of repeating the same biases that science has made so far: gender bias Accuracy: 0.95 - Specificity: 0.98 - Sensibility: 0.48

12. Voice Assistance – Generative AI

13. Conclusions We are “ready” – desperate “In House” development/validation EFFICIENT– bias Generative IA huge potential – EFFICIENT (time)

14. Josep Vidal i Alaball [email protected] https://www.slideshare.net/jvalaball