Application of Forcing Functions to Electronic Health Records Results in Improved Pain Control for Patients Undergoing Radiation Therapy for Bone Metastases

Abstract

Documentation of cancer pain is a metric used by national organizations including the American Society of Clinical Oncology and National Quality Foundation to measure quality of care. However, consistent documentation of pain remains sub-optimal in daily practice despite this imperative. The use of constraint systems, such as forcing functions, may improve physician compliance. This approach requires pain to be scored before the electronic health record (EHR) can be closed. We reviewed pain documentation rates at our institution, which recently installed changes to our EHR with successively more robust constraint systems. We hypothesized that these changes would lead to increase rates of pain documentation resulting in improved pain management for patients undergoing palliative radiation for painful bone metastases. We retrospectively compared patient on-treatment visit (OTV) pain scores from three different periods of documentation strategy: free-form (Aug-Oct 2014), pain score prompt (Jan-Mar 2015) and forcing function (May-June 2017). Patients were randomly selected from a 3-month period immediately before or after implementation of each new EHR design. The first and last OTV pain scores for patients with secondary bone metastases were selected from Jan to Dec 2016 (pain score prompt) and May to Dec 2017 (forcing function). Those with only one OTV were excluded. Pain score was documented as a numeric rating score (NRS). There were 150 patient OTVs studied for each documentation strategy (n = 450). The rate of documentation for free-form was 11% (95% CI: 7, 18); pain score prompts was 87% (95% CI: 81, 92); and forcing functions was 97% (95% CI: 93, 99). Documentation rates were different across the three groups (p < 0.001). The difference between each strategy was significant (p < 0.001 for free-form vs. prompt; p = 0.002 for prompt vs. forced function; p < 0.001 for free-form vs. forced function). Patients with secondary bone metastases (n = 72) had greater decreases in pain over the course of treatment using EHRs with forcing functions compared to pain score prompts (p = 0.026). The median improvement in NRS with use of forcing functions was -1.5 (IQR: -4 to 0), compared to pain score prompts, 0 (IQR: -1.25 to 0). Our results demonstrate that constraint systems applied to EHRs increase pain documentation rates. Forcing functions, being the most restrictive, have the highest rate of documentation. This improvement in pain documentation rates resulted in significant decreases in pain as compared to less restrictive documentation strategies for patients receiving palliative radiation therapy for painful bone metastases.