Seungwon Jeong, Researcher, PhD, Yusuke Inoue, Researcher, PhD,
Katsunori Kondo, MD, PhD, Daisuke Matsumoto, RPT, Master of Health Science,
Nariaki Shiraishi, RPT, PhD
Jpn J Compr Rehabil Sci 5: 19-25, 2014
Objective: To develop formulas for predicting
Functional Independence Measure (FIM) at the time
of discharge from an acute or convalescent hospital
ward using multicenter data.
Methods: Data from 4,311 acute patients (22 hospitals)
and 1,941 convalescent patients (24 hospitals) were
divided into two groups (calculation group and
verification group). Multiple regression analysis was
performed to develop formulas for predicting
discharge FIM and test their validity with data from
the verification group.
Results: The formula derived for predicting discharge
FIM for acute patients was 85.04 + (-0.53 × age) +
(12.06 × subarachnoid hemorrhage) + (-7.90 ×
complication present) + (-0.70 × number of days from
onset of stroke until admission) + (1.24 × admission
GCS) + (-1.08 × admission NIHSS) + (-4.15 ×
modified Rankin Scale score before stroke) + (0.30 ×
admission motor FIM) + (1.03 × admission cognitive
FIM), with R2 = 0.78. The formula derived for
predicting discharge FIM for convalescent patients
was 33.04 + (-0.34 × age) + (-3.88 × complication
present) + (-0.11 × number of days from onset of
stroke until admission) + (2.44 × admission GCS) +
(-1.68 × modified Rankin Scale score before stroke) +
(0.53 × admission motor FIM) + (1.25 × admission
cognitive FIM) (R2 = 0.66).
Conclusion: Using a large multicenter database, we
developed separate formulas for predicting FIM at
discharge from an acute ward and from a convalescent
ward with proven external validity.
Key words: prediction of FIM at hospital discharge, stroke patients, acute rehabilitation, convalescent rehabilitation, rehabilitation patient data bank