Uncertainty in the Relationship Between Sagittal Alignment and Patient-Reported Outcomes

BACKGROUND: Previous studies have reported correlations and precise quantitative relationships between sagittal alignment and health-related quality-of-life (HRQOL) scores. These studies have not reported the extent of uncertainty in these relationships. OBJECTIVE: To explore the uncertainty in the overall relationships between sagittal alignment and HRQOL and in the predictions of individual patient pain and disability. METHODS: A retrospective analysis of all new adult patients with long-cassette radiographs and complete outcomes questionnaires presenting to the senior author from 2012 to 2014 was performed. Univariable maximum a posteriori linear regression analyses using Bayesian methods were performed. High-density probability intervals for mean regression relationships and for individual values were calculated using minimally informative prior distributions. RESULTS: A total of 134 patients satisfied inclusion criteria and were included. For Oswestry Disability Index (ODI) vs pelvic incidence-lumbar lordosis (LL), the 90% high-density probability interval ranged from -0.04 to 0.23, indicating that both the magnitude and direction of the relationship were uncertain. For both ODI vs sagittal vertical axis and ODI vs LL, there was uncertainty in the magnitude of the slope. Wide regions of uncertainty were also seen for predicting individual patient scores. CONCLUSION: We report the previously unpublished degree of uncertainty in the mean quantitative relationships between radiographic sagittal alignment and patient-reported outcomes and in individual patient outcomes scores. Based on these results, establishing treatment thresholds or predicting an individual's outcome is unreliable. Further research efforts should be focused on developing multilevel hierarchical models incorporating parameter uncertainty and heterogeneous effects.