Keith R, Wilson D, Gutierrez P. Acute and subacute rehabilitation for stroke: a comparison.

Keith R, Wilson D, Gutierrez P. Acute and subacute rehabilitation for stroke: a comparison. Arch Phys Med Rehabil. 1995 June; 76: 495-500.

Although I specified acute rehabilitation and SNFs in my clinical question, I was still interested in this study because it compared acute rehabilitation with subacute rehabilitation, which provides less intense therapy. In this study, the patient characteristics were similar for both groups (ie, no statistical difference was noted), and, more important, the characteristics of my patient were similar to the characteristics of the patients in this study with regard to sex, race, admission from an acute care hospital, and age. The authors reported limited variability between groups; however, their results did show a significant, albeit small, difference (P=.007) in the proportion of patients discharged to the community, with 71% of the acute rehabilitation group returning to a residential setting and 67% of the subacute rehabilitation group returning to the community.

The Functional Independence Measure (FIM) was used to measure functional outcome. Average FIM scores were not statistically different at admission between the 2 groups. At discharge, there was no statistical difference in total FIM scores between groups; however, the change in total functional gain was significantly different between groups, with the acute rehabilitation group having a greater improvement in total FIM scores. In addition, changes in the following individual FIM items were significantly higher for patients in acute rehabilitation: all areas of self-care, bowel and bladder management. toilet and tub/shower mobility, and walking/wheelchair locomotion. The authors performed stepwise regression analysis to determine the extent to which the independent variables (admission rIM score, age at admission, onset days to admission, days out of treatment, length of stay, therapy hours, and type of facility) predicted change in the FIM scores. Although this analysis did not show a relationship between a change in FIM score and facility type, according to Keith and colleagues, it did demonstrate that the other 6 variables--admission FIM score, age at admission, onset days to admission, days out of treatment, length of stay, and therapy hours--did predict change in FIM scores, suggesting that variations in intensity of treatment do influence outcomes. This study had a number of threats to internal validity. There was a large difference in number of subjects between groups (331 in the acute rehabilitation group compared with 97 in the subacute rehabilitation group at baseline). Subjects were recruited from only 2 facilities, which limits generalizability to other facilities. The retrospective design of the study may have contributed to subject selection bias. Authors were vague in reporting results, which made it difficult to accept their results and conclusions. They did not report P values in the text, and, although they reported P values in the tables, it was unclear which results were being statistically compared (ie, within-group versus between-group comparisons). Based on my interpretation of their reported results, however, rehabilitation services received at either type of rehabilitation facility appeared to foster discharge to the community for patients after a stroke, and, based on individual FIM items, higher functional outcomes were achieved when rehabilitation was received at an acute rehabilitation facility. Although this study had limitations, the results encouraged me to consider acute rehabilitation for my patient.