Asthma and chronic obstructive pulmonary disease (COPD) are recognized as common health problems worldwide. High rate of sub-diagnoses or miss-diagnoses on our country is commonly observed due to different diagnostic criteria and low availability of gold standard diagnoses techniques such as spirometry. Objective: To develop an artificial intelligence (AI) model to standardize the diagnoses of asthma and COPD.

Methods: Subjects signed ICF previous to enter the study. We included 4223 subjects with asthma and 3830 with COPD during 2019 -2021. Subject demographics were collected for all subjects. Using clinical variables in addition to spirometric variables, we develop an artificial intelligence model. A logistic regression algorithm was used to train the IA models for asthma and COPD respectively

Results: The IA algorithm identified 2 scorecards scaling from 0 to 250 and 300 to 620 points for asthma and COPD, respectively. Four groups were identified based on the risk of developing asthma and COPD. A total of 14.95% subjects were diagnosed with asthma, and most patients (61.22%) had a score ranging from 176 to 220. COPD was diagnosed in 3.10% with 41.38% having a score of 540 to 620. The sensitivity, specificity, positive, and negative predictive values for asthma and COPD were 64%, 72%, 30%, 92%, and 83%, 77%, 12%, and 99%, respectively

Conclusion: The AI model effectively predicted the risk of developing asthma and COPD. This is a exploratory approach target to facilitate diagnoses of asthma and COPD in a primary care setting. Further research is needed.