Remote telemonitoring supported by an artificial intelligence algorithm for the early detection of heart failure and COPD exacerbations

Background: The progressive aging of the world population has increased the number of patients with multiple comorbidities leading to frequent hospitalizations and an increase in healthcare expenditure. The use of a telemonitoring system supported by an artificial intelligence algorithm could represent an instrument to support patients, reducing these outcomes. Aim of the study: Evaluate the effectiveness of the remote multi-parameter monitoring system supported by the artificial intelligence algorithm BPCO Media in early detection of COPD and heart failure exacerbations. Study design: Single-centre, prospective study conducted in Policlinico Campus Bio-medico of Rome. 40 patients have been enrolled: 22 patients affected by heart failure, 18 patients affected by COPD. Patients enrolled were telemonitored using an artificial intelligence algorithm. In the training phase, the algorithm defines patient “normality pattern” measuring the interaction between the parameters of heart failure and oxyhaemoglobin saturation. In case of significant deviations from the measurement range, an alarm signal was detected and patients were contacted to evaluate any new onset symptoms. The follow-up lasted 14 months. Results: In the study 40 patients have been enrolled: 22 patients affected by heart failure (mean (SD) age 78.6 [+-9.9] years, 54.5% were women; COPD patients were 18 (mean (SD) age 72.8 [+- 9.59], mostly females (61%). Patients were independent in their daily activities: the mean ADL score was 5.82 (SD 0.5) and IADL mean was 7.55 (SD 0.8) in heart failure group, the mean ADL score was 5.89 (SD 0.47) and IADL mean was 7.78 (SD 0.65) in patients affected by COPD. Patients were also affected by other chronic diseases: hypertension, chronic kidney disease, atrial fibrillation, dyslipidemia, diabetes, depression, anemia, benign prostatic hyperplasia. During the remote telemonitoring study supported by artificial intelligence algorithm, patients affected by heart failure had 2 exacerbations and 2 hospitalizations compared with the 22 exacerbations and hospitalizations occurred in the previous year, with a 91% reduction of the events (IC 95%: 61-98%, p value 0.008). During the trial, COPD patients reported 1 exacerbation and 1 hospitalization compared with the 15 exacerbations and 15 hospitalizations of the previous year, with a 93 % reduction of the outcomes (IC 95%: 44%-99% p value 0.001). Conclusions: the results of the study on the efficacy of remote telemonitoring supported by an artificial intelligence algorithm in COPD and heart failure patients are encouraging but these findings need to be confirmed in larger studies in the future.