Table 5.
Trial and recruitment characteristics that significantly affected randomisation rate for included stroke rehabilitation RCTs
| Trial feature | Kruskal–Wallis | P | RCTs | Subgrouping (median) | |||||
|---|---|---|---|---|---|---|---|---|---|
| Randomisation rate | Trial characteristic | Stroke survivor living context | X2(3) = 10.11 | 0.018 | 239 | Community (48%)* | |||
| General hospital (38%) | |||||||||
| Stroke-specific environment (27%) | |||||||||
| Stage of rehabilitation | X2(5) = 16.64 | 0.002 | 292 | Acute (23%) | |||||
| Acute–subacute (25%) | |||||||||
| Subacute (29%) | |||||||||
| Subacute–chronic (26%) | |||||||||
| Chronic (48%)* | |||||||||
| Recruitment characteristic | Recruitment strategy | X2(2) = 10.34 | 0.006 | 167 | Screening stroke survivors in the community (49%)* | ||||
| Screening admissions (22%) | |||||||||
| Screening inpatients (35%) | |||||||||
| Number of recruiters | X2(2) = 6.06 | 0.048 | 133 | Single recruiter (29%) | |||||
| Two or three recruiters (40%)* | |||||||||
| Four or more recruiters (21%) |
* Best recruitment condition
Kruskal–Wallis = appropriate statistics for Kruskal–Wallis test, p = significance level, RCTs = number of RCTs contributing to this finding, X2 = chi squared test