Evaluating the Need for a New Clinical Decision Support Tool in Knee Arthroplasty Rehabilitation

Join us as this presenter discusses this poster live on May 25, 2021 | Track A at 1:00 PM Mountain

PRESENTER
JEREMY GRABER, PT, DPT
University of Colorado Anschutz Medical Campus
BACKGROUND
Physical therapists (PTs) use range of motion (ROM) and functional measures like the Timed Up and Go (TUG) to monitor patient recovery after total knee arthroplasty (TKA).1 We recently developed a clinical decision support tool which precisely predicts ROM and TUG recovery post-TKA; we believe this may augment PTs’ ability to monitor patient recovery. The purpose of this project was to assess PTs’ confidence and accuracy in monitoring post-TKA recovery prior to implementing our clinical decision support tool into practice.
SETTING/POPULATION
This project is part of a quality improvement collaboration between ATI Physical Therapy and the University of Colorado. All data collection occurred at two outpatient clinics in Greenville, SC.
METHODS
Eight PTs rated their confidence level in predicting ROM and TUG measurement in TKA rehabilitation; the survey was scored on a Likert scale ranging from 0 (not at all confident) to 3 (very confident). During standard rehabilitation, PTs regularly collected ROM, TUG, and other outcomes; these data were entered into a quality improvement database for all patients with TKA. At the first postoperative visit, PTs also estimated patients’ discharge knee flexion ROM and TUG values. We examined correlation (Pearson’s r) and agreement (Bland-Altman plots) between predicted and observed values for ROM and TUG.2 Observations recorded within 21 days of patients’ discharge date were eligible for inclusion. Patient records with an episode duration < 1 month were excluded to remove data associated with premature self-discharge.
RESULTS
Overall, PTs felt confident in their ability to predict patient outcomes. PTs reported feeling “confident” or “very confident” for both measures, except for two PTs who chose “somewhat confident” for TUG prediction. A total of 477 patient records were screened for inclusion in the accuracy assessment; only 25 were eligible for ROM assessment and 22 for TUG. The correlation between predicted and observed was moderate for ROM (r = 0.65) and weak for TUG (r = 0.29).3 The Bland-Altman limits of agreement were 0.8 + 16.3 degrees for knee flexion ROM and 0.3 + 4.2 seconds for TUG.
CONCLUSIONS
Although PTs rated themselves as confident in their ability to predict post-TKA ROM and TUG recovery, their accuracy suggests there is room for improvement. The Bland-Altman limits of agreement exceeded the minimal detectable change for both knee flexion ROM (6.4 – 7.1 degrees4) and TUG (2.49 seconds5). The observed correlations indicate PTs may benefit most from assistance monitoring TUG recovery. These results suggest our clinical decision support tool may provide PTs with a relative advantage-a key feature for disseminating innovations6-compared to standard practice. Our next steps will be to integrate the tool into the participating clinics to assess its effectiveness and implementation potential in outpatient TKA rehabilitation.
POSTER

Loader Loading...
EAD Logo Taking too long?

Reload Reload document
| Open Open in new tab

Posted in 2021 Poster Session, Clinical Decision Support & Technology Tools.