Artificial intelligence empowered digital platform to support patients with heart failure (AZIMUTH platform)

Abstract Background Despite continuous efforts, Heart Failure (HF) is the leading cause of hospital readmission in developed countries. Since this tendency worsens prognosis of patients and increases health care costs, it raises concerns about the sustainability of the overburdened traditional model of care. Contributing factors to readmission include inadequate treatment of risk factors and low adherence to medications. Purpose The AZIMUTH study aimed at evaluating the feasibility, patient acceptance and perceived value of a certified medical device class IIa, smartphone app-based model of care, designed to improve quality of care and clinical outcomes for HF patients. Methods The AZIMUTH-I multicentric trial enrolled 185 patients with a diagnosis of chronic HF. At enrollment, patients are supplied with the mobile application and IoT devices. Data were accessible through a clinical dashboard meant for the healthcare professionals to remotely monitor the individual patients. During the observational period of 12 months, the patients were asked to daily report their feelings and symptoms, to share vital parameters, to record medications and to answer questionnaires regarding quality of life (KCCQ) and medication adherence (MMAS-8). They were also allowed to digitally send medical reports and to address questions to a mutidisciplinary clinical staff. The user acceptance was recorded using the User Experience Questionnaire (UEQ-S) and the Friends and Family Test (FFT). Results Our ehealth-empowered approach was tested in a cohort of 185 chronic heart failure patients (mean age 74 ± 12 years, mean LVEF 47 ± 13 %, mean NYHA class 2 ± 0.5). After a median follow-up of 11 months, all patients completed the study. The engagement was very high: 75% of opened the app at least once a day, showing a remarkable willingness to share additional data, consisting in 8231 "patient-reported outcome measures" (PROM). The average adherence to mandatory tasks was 74.8% whereas the adherence to medication reporting was 84.2%. The median result of the UEQ was "excellent" and the FFT revealed an overall experience between "positive" and "very positive". A significant decrease in the rate of re-hospitalization was observed at 12 months when compared with the same period before the implementation of the AZIMUTH model of care (1±0.2 vs 3±0.3 p<0.01) and was accompanied with a significant increase in the perceived quality of life (KCCQ overall score 71.2±20.7vs78.5±23.5; clinical score 77.4±20.1vs82.6±21.5 p<0.05). Conclusion This trial proved that our ehealth-integrated HF care is feasibile, user-friendly, scalable with high adherence and perceived benefits. The preliminary clinical findings and the high number of interactions suggest that the adoption of similar digital platform may not only improve the quality of life but also generate savings in health expenditures. Further study is eagerly awaited to evaluate the clinical efficacy of this new model of care.