Abstract
<sec> <title>BACKGROUND</title> <p>Remote monitoring reduces hospitalisations and mortality in selected patients with heart failure, but access among very old, frail patients in routine practice is poorly described. Non-enrolment may encompass distinct situations—including death, clinical ineligibility, referral to another pathway, patient refusal, and practical barriers—that should not be interpreted as a homogeneous outcome.</p> </sec> <sec> <title>OBJECTIVE</title> <p>To classify reasons for non-enrolment in a hospital-based telemonitoring pathway into meaningful categories, quantify non-enrolment among patients for whom enrolment was clinically relevant, and identify its determinants among these eligible patients.</p> </sec> <sec> <title>METHODS</title> <p>We conducted a retrospective, single-centre cohort study of 1386 patients hospitalised for heart failure in a cardiogeriatrics unit (TRIAGE-HF cohort; 16 December 2024 to 8 June 2026). Among 1063 patients not already telemonitored before admission, we classified the 451 not enrolled at discharge into six categories: death before a decision was possible, clinical/programme ineligibility, referral elsewhere, patient/family refusal, a potentially addressable practical or organisational barrier (including nursing home residence as a blanket exclusion), or uncertain eligibility. Patients with death, ineligibility, or referral elsewhere were excluded from the primary analysis. The primary outcome was non-enrolment among the 859 remaining eligible patients (612 newly enrolled, 247 not enrolled), assessed using a multivariable logistic regression model with a priori selected candidate variables; age was categorised to account for a detected non-linear association.</p> </sec> <sec> <title>RESULTS</title> <p>Among the 451 non-enrolled patients, 66 (14.6%) had died before a decision was possible, 84 (18.6%) were ineligible, 54 (12.0%) were referred elsewhere, 54 (12.0%) declined participation, 186 (41.2%) faced a practical/organisational barrier (132 with nursing home residence as the recorded reason), and 7 (1.6%) were undocumented. Among the 859 eligible patients, 247 (28.8%) were not enrolled. In the multivariable model (n=847, 240 events; age categorised because of a non-linear association), nursing home residence at admission was the strongest determinant (adjusted odds ratio [aOR] 30.00, 95% CI 15.01-59.93), followed by social isolation (aOR 3.46), depressive syndrome (aOR 1.73), and severe cognitive impairment (aOR 1.66); lower BMI and lower ADL score were also associated with non-enrolment. Age showed a non-monotonic pattern, with higher odds under 80 years (aOR 1.93) than at 80-89 years (reference) and no difference at 90 years or older. LVEF, NT-proBNP, atrial fibrillation/flutter, and Charlson index were not independently associated with non-enrolment (apparent AUC 0.830). Among 164 nursing home residents cohort-wide, only 34 (20.7%) were ever enrolled.</p> </sec> <sec> <title>CONCLUSIONS</title> <p>Once death, ineligibility, and referral elsewhere are separated out, and preference is distinguished from structural barriers, non-enrolment affects roughly three in ten eligible patients and is driven by nursing home residence, social isolation, depressive syndrome, and cognitive impairment rather than cardiac severity. Because most of this gap reflects potentially addressable practical or organisational barriers rather than documented patient or family refusal, adaptive, frailty-aware enrolment pathways for institutionalised patients warrant evaluation.</p> </sec>