WSO - April 2024 - 433

MacDonald et al.
433
Ethics
The use of the data in this project is authorized under section
45 of Ontario's Personal Health Information Protection
Act (PHIPA) and does not require review by a Research
Ethics Board.
Statistical methods
RI was stratified into quartiles for baseline and crude outcomes
analyses to aide in visualizing the data. The association
between RI (treated as a continuous variable) and the
outcomes of interest were analyzed using regression models
with restricted cubic splines. The total mean RI was
used as the reference. Analyses were adjusted for age, sex
(male vs. female), CCI (low vs. high), income quintile, living
alone prior to admission (yes vs. no), preadmission living
setting (home, assisted living, or other), rural (yes vs.
no), treated on a stroke unit at any time during their acute
inpatient stay (yes vs. no), acute care LOS (days), total
rehabilitation admission FIM, and rehabilitation institution.
Post hoc analysis included calculating the RI above which
there was diminishing incremental benefit in FIM change.
The number and proportion of patients receiving that RI or
greater was calculated. Furthermore, a propensity score
(PS) weighting analysis was conducted using stabilized
average treatment effect weights to fit each outcome model,
with the exposure of RI dichotomized into above or below
this cutoff. We performed balance diagnostics of the baseline
covariates to ensure the weighted standardized differences
were <0.1. The continuous covariates were modeled
using restricted cubic splines when estimating the PS. All
analyses were carried out using SAS version 9.4.
Results
Participants
Of 15,558 individuals admitted to inpatient stroke rehabilitation
within 72 h of discharge from acute care, 12,770 met
the inclusion criteria and were analyzed. Reasons for exclusion
included age <19 or >100 years (n = 14), rehabilitation
LOS <3 days or missing/unknown (n = 1088), admitted
from LTC or admission setting missing/unknown (n = 92),
acute stroke 5 years prior to the index date (n = 492), missing
FIM information (n = 281), missing/unknown discharge
setting (n = 25), RI<1st or >99th percentile (n = 156), discharge
to acute care (n = 541), and missing RI (n = 99).
Individuals' mean age was 72.6 years (SD = 13.4), 5877
(46.0%) were female, and 11,191 (87.6%) had an ischemic
stroke. Table 1 summarizes the individuals' clinical and
socio-demographic characteristics.
RI ranged from 5 to 162 min/day with an overall
median (Quartile 1-Quartile 3 (Q1-Q3) of 72 (54-94)
min/day. There were statistically significant differences in
the characteristics of individuals in each RI quartile.
Individuals who received greater RI were male (p trend
⩽ 0.0001), younger (p trend ⩽ 0.0001), more likely to be
treated on an acute stroke unit (p trend ⩽ 0.0001), have
longer acute LOS (p trend ⩽ 0.0001), live in a rural community
(p trend ⩽ 0.0001), be admitted from home
(p trend = 0.0001), not live alone (p trend = 0.0098), and
have moderately severe strokes (admission FIM scores:
40-80; p trend ⩽ 0.0001).
Total FIM change
As shown in Table 2, the mean total FIM change increased
as RI increased (mean FIM change 21.52 (SD = 14.73) for
quartile 1, compared with 31.52 (SD = 14.99) for quartile 4;
p ⩽ 0.0001). Figure 1 demonstrates the positive association
between RI and total FIM change remained after adjusting
for baseline factors. Despite the association, there was little
incremental improvement in FIM change after reaching a
RI of 95 min/day.
Motor and cognitive FIM change
Both motor and cognitive FIM change increased with
increasing RI (mean motor FIM change 19.20 (SD = 13.31)
for quartile 1 compared with 27.29 (SD = 13.50) for quartile
4; p ⩽ 0.0001 and mean cognitive FIM change 2.32
(SD = 3.51) for quartile 1 compared with 4.22 (SD = 3.92)
for quartile 4; p ⩽ 0.0001). In the adjusted analysis, after a
RI of 95 min/day there was minimal change in motor FIM
change; however, cognitive FIM change continued to demonstrate
improvement (Figures 2 and 3).
Rehabilitation effectiveness
Rehabilitation effectiveness increased as RI increased
(mean rehabilitation effectiveness: 44.70 (SD = 26.64) for
quartile 1 compared with 56.55 (SD = 21.27) for quartile 4;
p < 0.0001). There was minimal increase in rehabilitation
effectiveness after a RI of 85 min/day (Figure 4).
90-day home time
The amount of home time in the first 90 days poststroke
increased as RI increased. However, this association was
strongest for individuals receiving 35 min of rehabilitation
or less a day; individuals receiving a RI of 35 min/day or
less had the shortest amount of time at home (Figure 5).
Discharge location
Lower RI was strongly associated with lower probability of
discharge back to preadmission setting, most significantly
when RI was ⩽35 min/day (Figure 6). A similar picture was
seen regarding discharge to LTC, in which individuals who
received higher RI (⩾95 min/day) were significantly less
likely to be discharged to LTC (Figure 7).
International Journal of Stroke, 19(4)
WSO - April 2024

Table of Contents for the Digital Edition of WSO - April 2024
