Journal of Rehabilitation Medicine 51-3 | Page 16

162 M. Gustavsson et al. Table II. Participants’ characteristics Age, mean (SD) Women/men, n (%) Cohabitation, n (%) FAI pre-stroke (0–45), median Stroke severity, Barthel Index, n (%) Mild (50–100) Moderate/severe (< 49) Katz Extended Index of Independence in ADL pre-stroke, n (%) Dependent in both Independent in one Independent in both Intervention group (n  = 114) Control group (n  = 123) Inpatient rehabilitation (n  = 78) Home rehabilitation (n  = 36) Inpatient rehabilitation (n  = 51) Home rehabilitation (n  = 72) 72.6 (11.1) 38/40 (49/51) 40 (51) 29 74.6 (7.6) 13/23 (34/64) 21 (58) 31 69.3 (11.9) 17/34 (33/66) 32 (63) 32 70.2 (9.2) 27/45 (38/63) 41 (57) 35 53 (68) 32 (89) 41 (80) 66 (92) 25 (32) 4 (11) 10 (20) 6 (8) 2 (4) 1 (1) 19 (24) 12 (33) 11 (22) 20 (28) 52 (67) 19 (53) 38 (75) 51 (71) 5 (14) 7 (9) SD: standard deviation; FAI: Frenchay Activities Index; ADL: activities of daily living. The hypothesis was that client characteristics (age, sex, stroke severity and participation before stroke), rehabilitation context (inpatient or client’s home), and approach (enhanced client- centeredness or not) would differ between participants reporting a positive outcome and those reporting no positive outcome in participation in everyday life after stroke. Data from the inter- vention group and the control group were analysed separately. As a first step, univariable logistic regression analyses were performed to explore plausible associations in relation to the independent variables and a positive outcome regarding parti- cipation 12 months after inclusion. The independent variables were: age, sex, stroke severity according to the Barthel Index, rehabilitation context (inpatient medical-, inpatient geriatric-, or primary care home rehabilitation units) and participation in everyday life before stroke according to the FAI. Participants with moderate and severe stroke according to the Barthel Index were pooled and analysed as a single group due to a low number of participants with severe stroke. In addition, multivariable logistic regression analyses were performed, in which both the intervention and the control group were included. Since the results from the univariate analyses indicated a greater difference in outcome between home and the medical rehabilitation (reference in the univariate analyses) than between geriatric and medical rehabilitation and due to few participants from the medical rehabilitation unit, participants treated at medical and geriatric rehabilitation units were pooled to a single inpatient rehabilitation group. These analyses explored the relationship between context (inpatient or home rehabilitation) and approach (enhanced client-centred intervention or usual ADL intervention) of rehabi- litation and having a positive outcome in any of the 3 outcome measures. The p-value was 0.05. RESULTS Included in this study were 237 participants from the previous RCT; 123 from the intervention group and 114 from the control group. Of the 237 participants, 7 were missing data for SIS; 7 were missing data for FAI; and 1 was missing data for OGQ. The characteristics of the participants are presented in Table II. Data on the 3 outcome measures of participation at inclusion (0 or 3 months) and 12 months after inclusion are presented in Table III. Positive outcomes in participation in every- day life were found for: 208 (90%) participants using SIS, 154 (67%) using FAI, and 162 (69%) using OGQ. The univariable analyses for the 3 outcome measures (SIS, FAI and OGQ) are presented in Tables IVa–IVc. Among participants in the control group, there was a significant association between mild stroke and a po- sitive outcome using SIS domain 8 (OR 1.00 vs 0.25 for moderate/severe stroke) (see Table IVa). There was a significant association between mild stroke and a positive outcome using FAI for both the intervention group (OR 1.00 vs 0.36 for moderate/severe stroke) and the control group (OR 1.00 vs 0.12 for moderate/severe stroke) (see Table IVb). In addition, among participants in the control group there was a significant association between type of rehabilitation and a positive outcome using FAI (OR 3.20 for home rehabilitation vs 1.00 for Table III. Outcome measures at baseline and 12 months after inclusion Intervention group (n  = 114) SIS domain 8 participation, median FAI pre-stroke, median OGQ, mean Control group (n  = 123) Inpatient (n  = 78) Home-based (n  = 36) Inpatient (n  = 51) 0/3 months 12 months 0/3 months 12 months 0/3 months 12 months 0/3 months 12 months 25.9 30.5 8.3 66.4 19.5 6.4 25.9 31 7.2 68.1 20 4.9 24.1 32.5 9.2 58.3 21.5 5.2 75.0 25 3.8 SIS: Stroke Impact Scale 3.0; FAI: Frenchay Activities Index; OGQ: Occupational Gaps Questionnaire. www.medicaljournals.se/jrm Home-based (n  = 72) 30.9 33 5.6