Real-world Implementation of an eHealth System Based on Artificial Intelligence Designed to Predict and Reduce Emergency Department Visits by Older Adults: Pragmatic Trial

Abstract

Background Frail older people use emergency services extensively, and digital systems that monitor health remotely could be useful in reducing these visits by earlier detection of worsening health conditions. Objective We aimed to implement a system that produces alerts when the machine learning algorithm identifies a short-term risk for an emergency department (ED) visit and examine health interventions delivered after these alerts and users’ experience. This study highlights the feasibility of the general system and its performance in reducing ED visits. It also evaluates the accuracy of alerts’ prediction. Methods An uncontrolled multicenter trial was conducted in community-dwelling older adults receiving assistance from home aides (HAs). We implemented an eHealth system that produces an alert for a high risk of ED visits. After each home visit, the HAs completed a questionnaire on participants’ functional status, using a smartphone app, and the information was processed in real time by a previously developed machine learning algorithm that identifies patients at risk of an ED visit within 14 days. In case of risk, the eHealth system alerted a coordinating nurse who could then inform the family carer and the patient’s nurses or general practitioner. The primary outcomes were the rate of ED visits and the number of deaths after alert-triggered health interventions (ATHIs) and users’ experience with the eHealth system; the secondary outcome was the accuracy of the eHealth system in predicting ED visits. Results We included 206 patients (mean age 85, SD 8 years; 161/206, 78% women) who received aid from 109 HAs, and the mean follow-up period was 10 months. The HAs monitored 2656 visits, which resulted in 405 alerts. Two ED visits were recorded following 131 alerts with an ATHI (2/131, 1.5%), whereas 36 ED visits were recorded following 274 alerts that did not result in an ATHI (36/274, 13.4%), corresponding to an odds ratio of 0.10 (95% IC 0.02-0.43; P<.001). Five patients died during the study. All had alerts, 4 did not have an ATHI and were hospitalized, and 1 had an ATHI (P=.04). In terms of overall usability, the digital system was easy to use for 90% (98/109) of HAs, and response time was acceptable for 89% (98/109) of them. Conclusions The eHealth system has been successfully implemented, was appreciated by users, and produced relevant alerts. ATHIs were associated with a lower rate of ED visits, suggesting that the eHealth system might be effective in lowering the number of ED visits in this population. Trial Registration clinicaltrials.gov NCT05221697; https://clinicaltrials.gov/ct2/show/NCT05221697.

Digital home care interventions and quality of primary care for older adults: a scoping review

“…In the period corresponding to the years 2020 to 2023, marked by the COVID-19 pandemic, 20 documents were published equivalent to 29.41% of the sample-among which 14 of the evidence was collected or analyzed during the COVID-19 pandemic [ 36 , 74 , 75 , 83 – 85 , 87 , 89 , 92 , 93 , 96 , 100 – 102 ], thus being influenced by the scenario of the health crisis. In the other 6 documents, data collection took place before the pandemic, and the results were not related to pandemic outcomes [ 78 , 79 , 81 , 82 , 86 , 97 ].…”

“…The United States [ 35 , 38 – 43 , 46 , 47 , 49 , 50 , 52 , 53 , 55 , 58 , 61 – 63 , 67 , 69 , 73 , 80 , 83 , 85 , 87 , 89 , 94 , 101 ] stood out for representing 28 (41.18%) of the studies, followed by Sweden [ 54 , 60 , 65 , 71 , 78 , 79 , 84 , 88 ] with 8 (11.76%) and Canada [ 37 , 64 , 72 , 93 , 95 , 100 ] which was identified in 6 (8.82%) of the studies, Netherlands [ 66 , 70 , 98 , 99 ] with 4 publications (5.88%), Brazil [ 74 , 91 , 96 ] and Norway [ 77 , 81 , 90 ] with 4.41% each. Moreover, data from France [ 45 , 75 ], Germany [ 57 , 97 ], and New Zealand [ 56 , 86 ] were mapped at 2.94% each. The countries with the fewest publications, corresponding to 1 document in each country, were Italy [ 76 ], Spain [ 44 ], Mexico [ 48 ], Portugal [ 82 ], Scotland [ 51 ], Australia [ 68 ], Finland [ 91 ], South Korea [ 55 ], Hong Kong [ 92 ], China [ …”

How Might We Have Known? Using Administrative Data to Predict 30-Day Hospital Readmission in Clients Receiving Home Care Services from 2018 to 2021

“…30 This study provides insight into factors that influence the risk of 30-day hospital readmission, some of which may be modifiable. These highlight an opportunity to leverage existing, routinely-collected administrative data to identify those at greatest risk 31 and allow supportive interventions (eg, social support and additional home care services) to be targeted appropriately. Results can be used to drive evidence-informed interventions to interrupt the traumatic and expensive cycle of hospital readmissions experienced by a subset of older home care clients.…”

Inteligência Artificial, Microcredenciais E O Processo De Enfermagem: Potencialidades Para a Aprendizagem Baseada Em Competências

“…Dentre as inúmeras aplicabilidades da IA, está o sistema eHealth, cujo objetivo é identificar situações de alto risco para idosos (Belmin et al, 2022). Este sistema baseia-se num registo de informações que é alimentado por uma aplicação para telemóvel (App), em situações de risco é emitido um alerta para o enfermeiro coordenador, este por sua vez entra em contacto com o doente ou familiar para procurar atendimento em uma unidade de urgência (Belmin et al, 2022).…”