Background/Purpose: Variation in clinical care of patients post total knee replacement (TKR) is well-known. Experts suggest that electronic health records (EHR) can be used to decrease unwarranted practice variation and to provide real world evidence for comparative effectiveness research (CER). However close examination of EHRs shows that they rarely include adequate intervention detail to replicate or replace randomized controlled trials. Using content recommendations from PT experts we designed a web-based data capture system to collect detailed, quantifiable intervention data from physical therapists (PTs) treating patients following TKR. PTs were instructed to enter every intervention provided at every visit using drop-down menus. If unable to describe their intervention using the menus, PTs provided explanations of the interventions in a Clinical Notes text box. We deployed the system to 8 outpatient physical therapy practices and 33 sites in 3 different US states. The purpose of this study was to evaluate the system’s ability to capture detailed quantifiable intervention data from multiple clinical sites and practitioners by describing the use of the Clinical Notes text box to identify interventions.

Methods: 81 PTs entered intervention data for 153 patients over a total of 2442 visits. All clinical notes were reviewed by at least two trained physical therapy reviewers to determine if the interventions in the clinical notes were: 1) additional information about interventions already identified by the PT in the menus; 2) additional interventions that were available but not identified in the menu by the PT; or 3) additional interventions not represented in the menus. We identified the frequency of visits when interventions were entered into the clinical texts but not entered using the menus. We determined the number of PTs who included interventions in the clinical notes but not in the menus. We also determined what interventions listed in the text box were unavailable in the menus.

Results: 46 PTs (57 %) used the Clinical Notes text box to include interventions they had not chosen from the menus. Clinical Notes from 424 visits (17%) included interventions which had not been identified in the menus.  The most common reason (148 visits) for including interventions in the text box but not in the menu was because the therapist exceeded the maximum allowable number (10) of strengthening exercises. In 257 visits (11%) the PT listed interventions in the text box that had been available in the menus. 19 visits (0.008%) included interventions in the text box that could not have been identified in the menus. PTs reported documentation was quick and easy

Conclusion: It is possible to use a structured, menu-driven data capture system to collect detailed and quantifiable intervention data across multiple practice sites and among multiple PTs in patients post TKR. While many PTs used text to describe their interventions, better training of the PTs and minor changes to the capture system will minimize the need for text in data capture, making detailed and quantifiable descriptions of interventions and dosages straight forward across treatment sites and practitioners. This will allow large pragmatic CER and may lead to decreased practice variation.

« Back to ACR Convergence 2020

ACR Meeting Abstracts - https://acrabstracts.org/abstract/a-web-based-data-capture-system-can-successfully-collect-detailed-and-quantifiable-physical-therapy-intervention-data-post-total-knee-replacement/