Early prediction of outcome after stroke is becoming increasingly important, since most patients are discharged from hospital stroke units within several days after stroke.
1) To determine the accuracy of physical therapists (PTs) prediction at hospital stroke unit regarding upper limb (UL) function; 2) to develop a Computational Prediction Model (CPM) and 3) to compare the accuracy of PTs’ and CPMs' prediction. Secondary objectives were to explore the impact of timing on the accuracy of PTs and CPMs' predictions and to investigate the direction of the difference between predicted and observed outcome. Finally, we investigated if the accuracy of PTs' prediction was affected by their experience in stroke rehabilitation.
Prospective cohort study.
PTs made predictions at T72h and Tdischarge about UL function after 6 months in 3 categories, derived from the Action Research Arm Test. At the same time, clinical variables were measured to derive a CPM. The accuracy of PTs and CPMs’ predictions was evaluated by calculating Spearman rank correlation coefficients (rs) between predicted and observed outcome.
131 patients and 20 PTs participated. For T72h, rs between predicted and observed outcome was 0.63 for PTs’ predictions, and 0.75 for the CPM, which is significantly higher (P<0.01). For Tdischarge, rs for PTs’ predictions significantly (P<0.01) improved to 0.75, the rs for CPMs’ predictions improved slightly, although not significantly, to 0.76.
PTs administered a test battery every 3 days, which might enhance accuracy of prediction.
Accuracy of PTs’ predictions at T72h is lower than that of the CPM. At Tdischarge, PTs and CPMs’ predictions are about equally accurate.
Source: http://ptjournal.apta.org/cgi/content/short/ptj.20120112v1?rss=1
Walang komento:
Mag-post ng isang Komento