Early predictors of functional independence 2 years after spinal cord injury

To determine: (1) how well factors measured at admission to an acute care facility predict functional independence measure (FIM) scores, use of personal care assistance, and wheelchair ownership 2 years after traumatic spinal cord injury (SCI); (2) the extent that factors measured during inpatient stay add to these predictions; and (3) if FIM scores differ through use of assistance and wheelchair ownership 2 years after SCI.

Prospective, longitudinal.

Tertiary care acute, rehabilitation hospitals and home settings.

One hundred sixty SCI admissions.

FIM, use of personal care assistance (yes/no), and wheelchair ownership (manual/electric/none) 2 years after SCI.

Year 2 FIM scores were highly correlated (> or = .68) to the ASIA admission and discharge light touch, pin prick, and motor scores. Admission neurological status and age accounted for 65% of year 2 FIM score variance. Adding hospital events and the discharge ASIA motor score increased prediction to 76% of the variance. A separate regression model using only year 2 neurological scores and age accounted for 73% of the total FIM variance. Discriminant function analysis indicated 86% correct classification regarding use of personal care assistance and 88% correct classification of wheelchair ownership. Using a separate cross-validation sample, overall classification accuracy for assistance was 80% and wheelchair ownership 67%. FIM scores were significantly lower in assistance users (78 +/- 24) than nonusers (120 +/- 8) and were significantly different between wheelchair ownership groups: manual (103 +/- 21), electric (61 +/- 15), and none (125 +/- 2).

Late disability can be predicted using early impairment measures. The FIM prediction from variables measured during the early treatment phase was as good as prediction based on concurrent measures.