Three measures of physical rehabilitation effectiveness in elderly patients: a prospective, longitudinal, comparative analysis

BACKGROUND: Rehabilitation success is measured by instruments that assess performance of activities of daily living. Guidelines on the use and choice of these instruments are lacking. The present study aimed to analyse prognostic indicators of physical rehabilitation effectiveness in elderly patients according to three rehabilitation impact indices. METHODS: Prospective, longitudinal study in a post-acute care unit. The study included rehabilitation-eligible deconditioned elderly in-patients prospectively admitted to post-acute care (n = 685, aged 83.2 ± 8.3 years, mean length of stay 15 ± 9.2 days). Data Collection: Premorbid health status variables (PHSV): age, sex, comorbidity (Charlson index), medical history (heart failure, pulmonary disease, cerebrovascular disease, dementia), previous living situation and pre-admission functional status (premorbid Lawton and Barthel indices). Admission health status variables (AHSV): main diagnoses, referral source, physical (Barthel-adm) and cognitive function (Pfeiffer test), undernutrition and dysphagia. Outcome Measures: Absolute functional gain (AFG, admission-to-discharge Barthel change), relative functional gain (RFG, achieved percentage of potential gain) and rehabilitation efficiency index (REI, AFG over length of stay). Univariate analysis considered these parameters, along with PHSV and AHSV. Multivariate logistic regression analysis was performed for AFG ≥20, RFG ≥35 % and REI ≥ 0.50. RESULTS: Greater AFG was associated with 14 variables, 8 PHSV (57.1 %) and 6 AHSV (42.8 %); greater RFG with 9 variables, 3 PHSV (33.3 %) and 6 AHSV (66.6 %); and REI with 9 variables, 4 PHSV (44.4 %) and 5 AHSV (55.5 %). Mean AFG value was 34.5 ± 15.8 in patients who achieved complete recovery (RFG 100 %, n = 189, 27.5 %) and 35.3 ± 15.0 (p = 0.593) in the remaining patients (n = 311, 45.4 %). In multivariate analysis, only Barthel-adm was related to all three rehabilitation impact indices. CONCLUSIONS: Both premorbid and acute-process variables have a greater impact on AFG and REI, compared to RFG. Although AFG gives information about the degree of reduction in dependence, it does not provide clinical information about post-rehabilitation functional status (mean AFG values did not differ between patients with and without complete recovery). A future implication for evaluating rehabilitation effectiveness in elderly patients is to recommend RFG corrected by premorbid Barthel score, which is less affected by previous health conditions, as the optimum method to assess the degree to which maximum potential improvement was achieved. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1186/s12877-015-0138-5) contains supplementary material, which is available to authorized users.