Six studies (n=approximately 4,000, sample size 193 to 1,854) of five screening tools were included in the review. QUADAS scores ranged from 12 to 14 (one study scored 14). Blinding of assessors was generally unreported. No study validated any instrument independently of its development setting.
The end-point for functional status determination by reference standard varied across the studies (one to six months, where reported). Area under the ROC curve was reported for four of the five screening tools (four studies, one study per tool). Area under curve was 0.64 and 0.66 for the TRST tool (30 day and 120 day end-point), 0.65 for the HARP tool (discharge end-point), 0.71 for the ISAR tool (six month end-point) and 0.73 for the SHERPA tool (three month end-point). One study reported only sensitivity (88%) and specificity (54%) for a low cut-off, graded 0 (low risk) out of 4.
One study reported the reliability of the ISAR instrument: concordance correlation coefficient for test-retest reliability was 0.73.
The paper also included a discussion of generalisability and clinical utility, but no supporting data were available.