RESUMEN
INTRODUCTION: Choric obstructive pulmonary disease (COPD) is the third mortality cause in the world, and the development of useful diagnostic tools is necessary to improve timely diagnostic rates in primary care settings. OBJECTIVE: To develop a web application displaying spirometric and clinical information - including respiratory symptoms and risk factors- to facilitate a COPD diagnosis. MATERIALS AND METHODS: In this cross-sectional study, an expert consensus was carried out with three specialists using the Delphi method to choose the relevant variables for COPD diagnosis. We developed a Python-based web application to diagnose COPD, displaying the clinical variables deemed relevant by the experts along the spirometric curve. RESULTS: Twenty-six clinical variables were included in the web application for the diagnosis of COPD. A fourth expert used the web application to classify a cohort of 695 patients who had undergone spirometry in a third-level centre and had answered at least one of five questionnaires for COPD screening. Out of the 695 subjects, 34% had COPD, according to the expert that diagnosed them using the web application. Only 42% of the patients in the COPD group had received a previous COPD diagnosis and 19% of the patients in the no COPD group had been misdiagnosed with the disease. CONCLUSION: We developed a web application that displays demographic and clinical information, as well as spirometric data, to facilitate the process of diagnosing COPD in primary care settings.
Introducción. La enfermedad pulmonar obstructiva crónica (EPOC) es la tercera causa de mortalidad en el mundo y es necesario el desarrollo de herramientas diagnósticas útiles para mejorar las tasas de diagnóstico oportuno en los entornos de atención primaria. Objetivo. Desarrollar una aplicación web que muestre la información clínica y de la espirometría incluyendo síntomas respiratorios y factores de riesgo para facilitar el diagnóstico de la EPOC. Materiales y métodos. En este estudio transversal se realizó un consenso de expertos con tres especialistas usando el método Delphi para elegir las variables relevantes para el diagnóstico de EPOC. Se desarrolló una aplicación web basada en Python que muestra la información clínica relevante según los expertos, junto con la curva y los datos de la espirometría para el diagnóstico de la EPOC. Resultados. Se incluyeron 26 variables clínicas para el diagnóstico de la EPOC. Un cuarto experto utilizó la aplicación web para clasificar una cohorte de 695 pacientes a los que se les había realizado una espirometría en un centro de tercer nivel y que habían contestado al menos uno de los cinco cuestionarios para la detección de la EPOC. De los 695 sujetos, el 34 % tenían EPOC según el experto que les diagnosticó usando la aplicación web. Sólo el 42 % de los pacientes del grupo con EPOC había recibido un diagnóstico previo de la enfermedad y el 19 % de los pacientes del grupo sin EPOC había sido diagnosticado erróneamente con la enfermedad. Conclusión. Se desarrolló una aplicación web que muestra información demográfica y clínica, así como datos espirométricos, para facilitar el proceso de diagnóstico de la EPOC en entornos de atención primaria.
Asunto(s)
Internet , Atención Primaria de Salud , Enfermedad Pulmonar Obstructiva Crónica , Espirometría , Enfermedad Pulmonar Obstructiva Crónica/diagnóstico , Enfermedad Pulmonar Obstructiva Crónica/fisiopatología , Humanos , Estudios Transversales , Masculino , Femenino , Persona de Mediana Edad , Anciano , Técnica DelphiRESUMEN
Background: Development of new tools in artificial intelligence has an outstanding performance in the recognition of multidimensional patterns, which is why they have proven to be useful in the diagnosis of Chronic Obstructive Pulmonary Disease (COPD). Methods: This was an observational analytical single-centre study in patients with spirometry performed in outpatient medical care. The segment that goes from the peak expiratory flow to the forced vital capacity was modelled with quadratic polynomials, the coefficients obtained were used to train and test neural networks in the task of classifying patients with COPD. Results: A total of 695 patient records were included in the analysis. The COPD group was significantly older than the No COPD group. The pre-bronchodilator (Pre BD) and post-bronchodilator (Post BD) spirometric curves were modelled with a quadratic polynomial, and the coefficients obtained were used to feed three neural networks (Pre BD, Post BD and all coefficients). The best neural network was the one that used the post-bronchodilator coefficients, which has an input layer of 3 neurons and three hidden layers with sigmoid activation function and two neurons in the output layer with softmax activation function. This system had an accuracy of 92.9% accuracy, a sensitivity of 88.2% and a specificity of 94.3% when assessed using expert judgment as the reference test. It also showed better performance than the current gold standard, especially in specificity and negative predictive value. Conclusion: Artificial Neural Networks fed with coefficients obtained from quadratic and cubic polynomials have interesting potential of emulating the clinical diagnostic process and can become an important aid in primary care to help diagnose COPD in an early stage.