Validation of PROMIS score as a measure of postoperative success following decompressive surgery for Chiari I malformation

Introduction: Evaluating postoperative outcomes in patients with Chiari I malformation is often challenging and has become an area of investigation. Many metrics have been previously studied, however, few have been subjected to comprehensive validation or are easily implemented. The Patient-Reported Outcomes Measurement Information System (PROMIS) Physical Health (PH) and Mental Health (MH) metrics are a National Institutes of Health (NIH)-funded, publicly available, and a highly reliable measure of patient-reported health status. While commonly used and validated in spine and stroke patient populations, these metrics have yet to be studied in adult Chiari I malformation patients undergoing surgical decompression. We aim to investigate the validity of PROMIS-PH and PROMIS-MH in tracking post-operative success following decompressive surgery for Chiari I malformation.

Methods: Adult patients undergoing decompressive surgery with duraplasty for Chiari I malformation at our institution from 2009-2023 were identified. Prospectively administered pre- and postoperative PROMIS-PH and MH scores were of interest, from which the postoperative score improvement was calculated. The current standard for discerning surgical success, postoperative physician gestalt, was retrospectively assigned to patients as “improved” or “unimproved” following review of the clinical documentation at a time point coinciding with the postoperative PROMIS score data collection. Univariable logistic regression and receiver operative characteristic (ROC) analyses were performed to validate postoperative PROMIS improvement as a metric for discriminating clinical success following decompression.

Results: We identified 113 patients with average age of 38.5 years and mean follow-up time of 14.4 months. Seventy-four (65.5%) patients experienced clinical improvement according to postoperative physician gestalt assessment. Postoperative PROMIS-PH score improvement was superior in the “improved” patients (improved: 8.5, unimproved: -2.5), p<.001. Logistic regression demonstrated PROMIS-PH to be an independent predictor of clinical improvement, with OR of 1.44 (95% CI: 1.27-1.70). ROC area under the curve was 0.852 with mean clinically important difference of 3.0, corresponding to a sensitivity and specificity of 0.80 and 0.90, respectively. PROMIS-MH was not a satisfactory predictor of surgical success with an ROC area under the curve of 0.740.

Conclusion : Postoperative PROMIS-PH score improvement serves as a responsive and valid metric for trending postoperative outcomes in the Chiari I population.