1. P139. Clinically relevant percent reduction (CRPR): a new definition of clinically significant change for lumbar spine surgery.
- Author
-
Archer, Kristin R., Asher, Anthony M., Pennings, Jacquelyn S., Khan, Inamullah, Sivaganesan, Ahilan, Bydon, Mohamad, Devin, Clinton J., and Asher, Anthony L.
- Subjects
- *
LUMBAR vertebrae , *DISABILITIES , *SPINAL surgery , *LEG pain , *DISCECTOMY , *LUMBAR pain , *RECEIVER operating characteristic curves - Abstract
Minimal clinically important difference (MCID) represents the smallest, clinically relevant change in a patient-reported outcome (PRO) score. However, the literature suggests that an absolute change from baseline may not be a reliable marker of response to treatment for patients with a low or high baseline PRO score. An alternative to MCID is a threshold of clinical relevance defined by percent reduction from baseline PRO score. The purpose of this study was to determine whether a clinically relevant percent reduction (CRPR) of 30% in disability and pain scores is a valid method for determining clinical improvement at 12 months after lumbar spine surgery. Retrospective evaluation of prospectively collected data from a national surgical spine registry, the Quality Outcomes Database (QOD). A total of 23,280 participants undergoing elective surgery for a lumbar degenerative condition (primary stenosis, disc herniation, spondylolisthesis (Grade I) and symptomatic mechanical disc collapse or revision surgery for recurrent same-level disc herniation, pseudarthrosis and adjacent segment disease). Outcomes were disability (ODI), back and leg pain (NRS-BP/LP), and satisfaction (NASS scale). Participants completed a preoperative assessment and follow-up assessment at 1-year. The change in ODI and NRS-BP/LP scores were categorized as met CRPR (percent change ≥30%) or not met CRPR (percent change < 30%) and by achieving MCID or not achieving MCID. MCID was based on established values of 12.8 for ODI, 1.2 for NRS-BP, and 1.6 for NRS-LP. The relationship between each categorization (30% CRPR and MCID) and satisfaction were compared with logistic regression models and receiver operating characteristic (ROC) curves. Analyses were conducted in the entire sample and by surgical procedure (microdiscectomy, decompression, decompression with fusion). In addition, the ability of CRPR and MCID to predict satisfaction was examined by severity of disability and pain. Significance was set at p <.05. Results from the satisfaction prediction models found that the OR for 30% ODI CRPR was 11.1 (10.2 to 12.1) and for ODI MCID was 7.8 (7.2 to 8.4), while 30% NRS-BP and NRS-LP CRPR was 7.9 (7.3 to 8.5) and 6.7 (6.2 to 7.2) compared to MCID ORs of 5.5 (5.1 to 5.9) and 5.6 (5.2 to 6.0) (p <.001). A CRPR threshold of 30% reduction in ODI and NRS-BP/LP predicted satisfaction with 76.9%, 73.7% and 70.9% accuracy, respectively, whereas MCID thresholds predicted satisfaction with 73.5%, 69.5% and 68.2% accuracy. Accuracy was similar across surgical procedure categories. The largest difference in predictive power for 30% CRPR compared to MCID was found for patients with minimal disability (ODI 0-20%: 75.6% vs 65.1%) and bed-bound disability (ODI 81-100%: 74.8% vs 60.9%). For pain, there was no difference in predictive power for patients with moderate pain (NRS 5-6), however there was a 3.9% and 1.7% difference for no/low back and leg pain (NRS 0-4) and a 3.8% and 2.6% difference for severe back and leg pain (NRS 7-10) in favor of a 30% CRPR threshold. CRPR may be a more clinically relevant method for identifying response to treatment since it can be applied to broad spinal surgery populations and takes into account the baseline PRO score. Furthermore, a 30% CRPR appears to outperform established MCID thresholds for patients with either low or high preoperative disability and pain scores. This abstract does not discuss or include any applicable devices or drugs. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF