Predicting Follow-Up Functional Outcomes in Outpatient Rehabilitation.

Baker JG, Fiedler RC, Ottenbacher K J, et al (School of Medicine and Biomedical Sciences, State University of New York at Buffalo, Buffalo, NY; School of Allied Health Sciences, University of Texas Medical Branch at Galveston, Tex; Rehabilitation Institute of Chicago, Northwestern University School of Medicine, Chicago, Ill), Am J Phys Med Rehabil. 1998;77:202-212.

Payer pressure to limit the utilization of outpatient rehabilitation has resulted in increasing interest in functional outcomes. Credible outcomes data require the identification and measurement of factors directly related to functional status and the use of valid statistical analyses. This study addresses functional status at the beginning of treatment, at the end of treatment, and at an average of 6 months after rehabilitation.

The subjects in this study were patients with musculoskeletal complaints who were being treated at an outpatient rehabilitation clinic in New York. The 44 patients who completed the follow-up questionnaire comprised the sample for the study. The subjects ranged in age from 22 to 72 years. Sixty-one percent were female; 22% were covered by workers' compensation. They received therapy for low back pain (60%), neck and arm pain (23%), other pain (5%), and other problems (12%).

The authors used the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) and the Medical Rehabilitation Follow Along (MRFA(TM)) as scales in the data analyses. The SF-36 is a short form of a measurement tool that assesses general health status and includes scales measuring physical functioning, physical role functioning, bodily pain, general health perceptions, social functioning, vitality, emotional role functioning, and mental health. The MRFA instrument used in this study consists of 26 items that assess functional outcomes in physical functioning, the experience of pain, and affective well-being.

The primary data analysis technique was logistic regression. Independent variables included in the regression analyses were age, sex, presenting problem, workers' compensation coverage, functioning at the initial assessment, number of visits, length of program, and program intensity.

According to the researchers, the results of the logistic regression analyses indicate that this approach is a viable method of studying long-term outcomes of outpatients at the time of follow-up. The authors concluded that logistic regression equations demonstrated their usefulness in studying predictors of dichotomous outcomes. The authors described other variables, such as whether an individual has returned to work at follow-up, that could have been used in the analyses. In addition, the authors described the limitations of this study, including the use of broad, self-reported problems rather than using more specific measurements of function. The authors believed that there may be an opportunity to develop a software program that could help clinicians modify variables such as intensity to better predict successful outcomes.

Mitchell Tannenbaum, PT Naperville, Ill