321 Using PROs as Time-Series to Establish a Risk-Informed “Value” Categorization for Patients Before Surgery for Lumbar Spinal Stenosis

INTRODUCTION: Patient-reported outcome measures (PROMs) are pivotal in patient-centered care, offering insights into patients' perceptions of their health and quality of life (QOL). Their application in lumbar spinal stenosis (LSS) surgery outcomes reveals significant heterogeneity over time, underscoring the need for nuanced analytical methods that account for time-dependent improvement of QOL following surgery. METHODS: This single-institution retrospective cohort study investigated LSS patients undergoing index surgery from 2/2015 to 09/2023 with at least 2 pre-operative and 2 post-operative PROMIS physical health (PROMIS-PH) scores. The time-weighted average of the difference in PROMIS-PH over the “baseline” was plotted against the standard deviation of PROMIS-PH over the postoperative period. Four groups were defined: high-mean, high-variance (HMHV), low-mean, high-variance (LMHV), high-mean, low-variance (HMLV), and low-mean, low-variance (LMLV). Predictive modeling was used to explore the relationship between preoperative characteristics, and pre-operative MV categorization with postoperative MV categorizations. McNemar’s tests compared our dependent samples across 4 categorizations from pre-op to post-op. RESULTS: The study assessed 1,190 unique patients with 11,681 PROMIS-PH scores. The distribution of postoperative MV categorization varied significantly from preoperative MV categorization (p<0.001). Patients with LMLV pre-operatively were more likely to show LMLV post-operatively (p<0.001). Patients with HMHV pre-operatively were most likely to benefit from surgery and show HMLV postoperatively (p<0.001). CONCLUSIONS: The novel application of economic principles and time series analysis to PROMs data offers a sophisticated approach to understanding the heterogeneity in LSS surgery outcomes over time and can aid in pre-operative prediction of surgical benefit. Patients declining pre-operatively were more likely to continue worsening postoperatively, while those with volatile but mildly improving QOL pre-operatively achieved more stable long-term responses to surgery postoperatively.