Functional mobility measures in older adults after hip fracture.

To the Editor: We submitted an Invited Editorial to accompany the article by Dr. Mendelsohn and colleagues, published in the October issue of the American Journal of Physical Medicine & Rehabilitation, as it relates to the comparison of hip fracture outcomes in one facility in Ontario, Canada, with those of inpatient rehabilitation facilities in the United States. Whether acute care after hip fracture should occur in an inpatient rehabilitation facility or in a skilled nursing facility is currently under scrutiny by the Centers for Medicare and Medicaid Services as part of the general concern related to enforcement of the 75% rule, which requires that a facility’s patient mix include 75% from ten listed conditions. Hip fracture is among the ten conditions. Further information may be obtained at the following Web site: http://cms.hhs.gov/providers/irfpps The reason for this Letter to the Editor is that we feel there is a need to clarify some of the impressions given in the article. In essence, the objective, design, methodology, and conclusions of this article are to determine the relationships between measures of functional mobility (Timed-Up-and-Go [TUG], Self-Paced Walking [SPW], and the Berg Balance Scale [BBS]) and functional status (FIM instrument) and two combinations of the FIM component motor items. Admission and discharge tests were performed on 20 rehabilitation inpatients after hip fracture. Pearson r correlations were computed for all combinations of the tests to determine the relationships between scores at admission and at discharge and the change in the scores. The conclusions are that the FIM scores do not correlate well with the other functional scores; therefore, the FIM instrument may not be a specific measure of functional mobility. With respect to the objective and the conclusions, each of the scales (FIM, BBS, TUG, and SPW) has a specific role in the analysis of patients with disability who may have limitations in ambulation. No one test is a substitute for another. Yet, the caution expressed by the authors that the FIM instrument may not be a specific measure of functional mobility is inappropriate because the FIM instrument is not intended to be a specific measure of functional mobility. The FIM instrument was designed and has been validated as a generic measure of disability, from the perspective of independence vs. dependence in basic daily activities. The correlations found in the article (with r approaching 0.5) are in line with the expectations from the literature, for related yet distinct domains of functioning (i.e., structural or physiologic impairment on one hand and integrated whole body activity on the other). A high correlation should not be expected between the FIM instrument and other tests of function. In fact: a) When the patient no longer needs significant assistance from another person, then independence has been achieved. Measurement of how well an independent person performs is achieved through other tools, similar to those discussed in this article. b) The FIM instrument explores what the patient does “most of the time” in his/her actual environment; the TUG, SPW, and BBS are tests of performance under controlled (i.e., laboratory) conditions. What is missing is proof that the results of these tests can be generalized to everyday activities. c) The FIM items, expressed as motor and cognition tasks, together represent quantitative levels of a shared construct (independence in daily life). The cumulative score is the valid measure as it relates to the need for assistance from another person. The FIM instrument is not intended to be a measure of subdomains (e.g., locomotion or transfer mobility) independently of the other items. With respect to design and methodology, correlation was probably underestimated. In fact: a) The reliability of the FIM raters is unknown. Evidence is not presented to support adherence to the standard FIM rating criteria and there is no evidence that the raters passed a basic mastery test. b) The measurements are intrinsically nonconcomitant, because the FIM instrument describes what the patient does most of the time, whereas the TUG, SPW, and BBS are tested within a specified time span. c) It would seem probable that only patients with FIM locomotion independence ratings of 6 or 7, and possibly 5, would be capable of completing the TUG and SPW tests by themselves (FIM ratings 1–4 indicate dependence on another person). Then, variance of the FIM ratings was reduced (so that covariance, and hence correlation, was hindered). Also, it is questionable—contrary to the authors’ suggestion—that TUG and SPW have no constraints on their range. Time and forward speed are bound between 0 and infinity, so that the assumption of a symmetrical, normal distribution (particularly for changes) may not be true. d) Computing correlation coefficients to determine a linear relationship assumes that the variables are normally distributed. However, this is hardly supported by the data (see, for instance, the scatterplot in Fig. 7, algebraically giving r 0.5; P 0.05) . More appropriately, FIM,