Older adults recovering from illness or functional decline did just as well when part of their hospital-based rehabilitation was shifted to their own homes, according to a new multicenter study from the Netherlands. The research, published in European Geriatric Medicine, compared a structured program called Better@Home, in which a shortened inpatient stay is followed by rehabilitation delivered at home, with conventional inpatient geriatric rehabilitation. On the study’s primary measure, independence in activities of daily living, the two approaches produced statistically indistinguishable results at discharge.
The Better@Home study was a prospective cohort study with a matched historical control group, carried out across eight geriatric rehabilitation facilities in the Netherlands embedded in three university networks for the care of older people. In total, 350 participants were included: 110 in the Better@Home group and 240 in the control group. The control group consisted of patients who had received conventional inpatient geriatric rehabilitation at the same facilities between January 2022 and June 2023, admitted at least six months before the new program was implemented to minimize contamination bias. Patients who had already received home-based rehabilitation were excluded from the historical control group.
The Better@Home program was developed in co-creation with rehabilitation professionals and rests on five core components: replacing part of inpatient rehabilitation with home-based rehabilitation, focusing on participation goals, using eHealth to support rehabilitation, promoting patients’ self-management skills through reablement strategies, and close collaboration among all relevant care partners. The medical service profiles of the program and conventional inpatient rehabilitation were otherwise identical. During the home-based phase, treatment was delivered through home-based, outpatient, or remote sessions, with patients otherwise residing at home. The duration and intensity of treatment were individualized according to multidisciplinary goals and patient progress, with no fixed duration or fixed number of therapist sessions prescribed.
The primary outcome was independence in activities of daily living, assessed with the Barthel Index, a ten-item measure scored from 0 to 20 in the modified version used, where higher scores indicate greater independence. For the historical control group, data came from electronic patient files, which meant the Barthel Index was the only comparative outcome available at both admission and discharge for both cohorts. In two of the participating facilities, scores from a different instrument, the Utrecht Scale for Evaluation of Rehabilitation, were converted to Barthel Index values on a per-item level. The index was measured within two weeks following admission and discharge, and interviews were performed by a trained researcher or research assistant.
Matching between the groups was conducted on Barthel Index scores at admission, with strata constructed per diagnosis group in each facility and frequency matching performed using an online randomization tool. Baseline characteristics were similar between the groups. The analysis used two complementary statistical models: a Cox regression model to identify factors associated with rehabilitation duration, and a linear mixed-effects model to evaluate the treatment effect on the Barthel Index, allowing adjustment for relevant confounders and inclusion of participants with missing follow-up scores under a missing-at-random assumption.
The central finding was clear. The adjusted mean difference in Barthel Index score at discharge, comparing the Better@Home group with the control group, was 0.48 points, with a 95 percent confidence interval running from -0.32 to 1.28 and a p-value of 0.24. This difference was not statistically significant, and it was also well below the minimal important change for the Barthel Index, which the authors cite as 3.1 points. In other words, the observed gap between the two models of care was both statistically negligible and clinically small.
The rehabilitation trajectories themselves differed substantially in shape. Mean inpatient rehabilitation duration was 33.1 days in the Better@Home group versus 37.6 days in the control group, a difference of 4.5 days that was not statistically significant. However, mean total rehabilitation duration, including the home-based phase, was 64.8 days in the Better@Home group compared with 37.6 days in the control group, a difference of 27.3 days that was highly significant. Within the Better@Home group, the mean completed duration of home-based rehabilitation was 36.3 days and included a mean of 9.7 treatment sessions per patient. Almost all patients, 97 percent, received treatment sessions at home, 76 percent received remote sessions, and 14 percent received outpatient sessions. The Cox regression model showed that the control group had a significantly shorter rehabilitation duration compared with the Better@Home group, although this difference diminished over time, and the authors note that sensitivity analyses indicated the results were affected by how censored observations were handled, warranting cautious interpretation.
The study’s authors are careful about the limits of what can be concluded. The non-randomized design means residual confounding could not be fully ruled out despite multivariable adjustment and frequency matching. Eligibility for the Better@Home program could not be retrospectively established for the remaining control patients, so comparability between groups regarding suitability for home-based rehabilitation cannot be assured, although matching per facility within each diagnosis group mitigated this risk. The validity of the mixed-model results depends on the plausibility of the missing-at-random assumption; the proportion of missing data was relatively small for this frail population, 15 percent at discharge, and did not differ significantly between groups. The structure and definition of rehabilitation completion also differed between groups, since the control group’s completion signified discharge from the inpatient facility whereas the Better@Home group’s completion followed the full inpatient and home-based phases, meaning the comparison reflects different measurement time points. The authors note that most functional gains in post-acute populations occur during the early inpatient period, and that home-based rehabilitation focuses primarily on transferring functional gains into everyday activities rather than continued gains in independence.
Another important consideration is a potential ceiling effect of the Barthel Index itself. At discharge, 35 percent of participants scored 19 or 20 on the index, the highest levels of independence the scale can record. The authors suggest this ceiling may have reduced the instrument’s sensitivity to detect between-group differences among patients reaching higher levels of daily living independence. They also point out that the Barthel Index, while useful for screening and rapid assessment of basic dependency, may be less sensitive to subtle functional changes in relatively high-functioning patients or in home-based settings, and that improvements in more complex domains relevant to home-based rehabilitation are better assessed by instrumental activities of daily living or participation measures.
The researchers conclude that a structured home-based geriatric rehabilitation program replacing part of inpatient rehabilitation can achieve comparable levels of independence in activities of daily living to conventional inpatient care, while emphasizing that the findings currently provide insufficient evidence to recommend widespread implementation. Data from ongoing economic, feasibility, and follow-up evaluations of the Better@Home study are still required to comprehensively assess the cost-effectiveness and feasibility of the program and to inform evidence-based implementation strategies. Future studies, they argue, should include broader outcome domains such as instrumental activities of daily living and participation measures, and the authors suggest that introducing a standard core set of outcome measures could facilitate such efforts. With aging populations increasing demand for geriatric rehabilitation and workforce shortages and budgetary pressures straining the system, the prospect of delivering part of rehabilitation in patients’ own homes, without sacrificing recovery of daily living independence, represents a finding of considerable practical interest for the future organization of care for older adults.
Subject of Research: Home-based geriatric rehabilitation after shortened inpatient rehabilitation and independence in activities of daily living
Article Title: Independence in activities of daily living for home-based geriatric rehabilitation after shortened inpatient rehabilitation: the “Better@Home” center prospective cohort study with a matched historical control group
Article References: Gamble, C. J., Vaz, M., de Waal, M. W. M., Groen, W. G., Pol, M. C., van Dam van Isselt, E. F., Innocenti, F., Zwakhalen, S. M. G., & van Haastregt, J. C. M. (2026). Independence in activities of daily living for home-based geriatric rehabilitation after shortened inpatient rehabilitation: the “Better@Home” center prospective cohort study with a matched historical control group. European Geriatric Medicine. https://doi.org/10.1007/s41999-026-01628-2
Image Credits: AI Generated
DOI: 10.1007/s41999-026-01628-2
Keywords: geriatric rehabilitation, home-based rehabilitation, activities of daily living, Barthel Index, Better@Home, older adults, inpatient rehabilitation, Netherlands, cohort study, functional independence, discharge outcomes, reablement
News Source: Beatrice Stafford. (October 27, 2026). Home-Based Geriatric Rehabilitation Matches Inpatient Care for Daily Living Independence. Scienmag.
