Predicting disability and institutionalization in elderly people: how strong is the evidence for the predictive value of the short physical performance battery?

Gawel et al. conducted a review to investigate the ability of the short physical performance battery (SPPB) to predict long term disability or future institutionalization. The authors concluded that the SPPB is a good predictor of long term disability or institutionalization in community-dwelling elderly people in the US aged 65 years and older. The systematic search they conducted was guided by the PRISMA statement to minimize the risk of faulty reporting and to improve the transparency of the review. After the extensive search and selection process, five articles were found eligible for inclusion by both independent reviewers. The reviewers used the Newcastle–Ottowa Scale to rate the quality of these studies to ensure that only studies with high enough quality were included. The above mentioned efforts contributed positively to the methodological quality of the review. Unfortunately, the evidence that Gawel et al. base their conclusions on might be considered rather weak. Three included studies investigated the predictive value of the SPPB on disability and two included studies investigated the predictive value of the SPPB on institutionalization. One of the three studies concerning disability reported an odds ratio that was not significant. Therefore, this study does not support the conclusion of the reviewers. Furthermore, one of the two studies concerning institutionalization included women aged 49–65 years old at baseline. The study had a follow-up of 3 years, meaning that most of the participants had not reached the age of 65 years or older by the end of the study. Therefore, this study also does not support the conclusion of the review. When taking these findings into consideration, only two studies concerning disability and one study concerning institutionalization remain that support the conclusions of the reviewers. These three remaining studies were all conducted by Guralnik et al. and were based on the same cohort study, namely the Established Populations for Epidemiologic Research in the Elderly cohort. So, the conclusions that Gawel et al. draw regarding the predictive value of the SPPB are merely based on the data from one cohort study. Therefore, in our opinion the conclusion that ‘the SPPB is a good predictor of long term disability or institutionalization in community-dwelling elderly people aged 65 years and older living in the US’ could be considered an overestimation of the actual evidence that is reported in the review. Besides that, the two studies by Guralnik et al. regarding disability investigated the predictive value of the SPPB in elderly people who were free of disability at baseline. So, the results of these studies do not necessarily support the conclusion that the SPPB is a predictor of long term disability, but rather provide evidence for the predictive value of the SPPB for the onset of disability. Previous reviews suggest that other indicators might be stronger predictors of disability or institutionalization in community-dwelling elderly people compared to the SPPB. A recent review by Vermeulen et al. revealed that a decrease in individual physical functions predicts the onset and increase of disability in community-dwelling elderly people. The results from this systematic literature review suggest that decreased gait speed, decreased physical activity and weight loss are stronger predictors of disability compared to lower extremity function. A review by Stuck et al. showed that lower extremity function is, among others, a predictor of disability development in elderly people. However, most studies regarding lower extremity function that were included in the two reviews mentioned above used a single performance measure instead of the SPPB to assess lower extremity function. Furthermore, a review by Gaugler et al. about predictors of institutionalization included 77 reports from 12 different cohort studies. A meta-analysis revealed that three or more dependencies in activities of daily living, cognitive impairment, and prior nursing home use were the strongest predictors of nursing home admission. A review by Luppa et al. concluded that strong evidence exists that the following variables are predictors of institutionalization: age, self-rated health status, functional impairment, cognitive impairment, prior nursing home placement, and number of prescriptions. Correspondence to: J Vermeulen, Department of Health Services Research, Faculty of Health Medicine and Life Sciences, Maastricht University, PO Box 616, 6200 MD Maastricht, The Netherlands. Email: j.vermeulen@maastrichtuniversity.nl Vermeulen and Neyens Commentaries