Quantifying Cumulative Postoperative Disability after Lumbar Spine Surgery with Spline-Based Modified Integrated Health State.
Academic Article
Overview
abstract
BACKGROUND CONTEXT: Postoperative burden following lumbar spine surgery plays a crucial role in determining surgical indication. Efforts to minimize surgical impact have been emphasized to improve patient selection and reduce unnecessary invasiveness. Although recent advances in spine surgery have led to improved outcomes at discrete time points, few studies have quantified the continuous and time-dependent burden experienced by patients throughout recovery. PURPOSE: To quantify cumulative postoperative disability using a continuously modeled recovery trajectory and to compare this metric across common lumbar procedures. STUDY DESIGN/SETTING: A retrospective analysis of a prospectively collected registry. PATIENT SAMPLE: Patients undergoing primary 1- to 4-level lumbar surgery between April 2017 and April 2024 in a single academic institution. OUTCOME MEASURES: Cumulative postoperative disability quantified with the modified Integrated Health State (mIHS), a novel continuous metric calculated as the area under the modeled Oswestry Disability Index recovery curve per week. METHODS: Recovery trajectories were modeled with multivariable mixed-effects regression using restricted cubic splines for postoperative day and an interaction term for surgical grades based on previously reported grading system (Grade 1 = decompression only; Grade 2 = single-approach fusion, 1-2 levels; Grade 3 = dual-approach fusion, 1 level; Grade 4 = dual-approach fusion, 2 levels; Grade 5 = dual-approach fusion, ≥3 levels). Differences in mIHS among grades were tested with one-way ANOVA followed by adjusted pairwise comparisons. Effect sizes were reported as eta squared (η²; small ≥ 0.01, medium ≥ 0.06, large ≥ 0.14) and Cohen's d (d; small ≥ 0.20, medium ≥ 0.50, large ≥ 0.80) to aid interpretation of differences among surgical grades. RESULTS: The multivariable model demonstrated a significant interaction between surgical grade and time (p < 0.001), indicating that recovery trajectories of ODI differed across surgical grades. The mIHS differed significantly among surgical grades at 10 weeks (Grade 1: 4.79 vs. Grade 2: 6.27 vs. Grade 3: 6.86 vs. Grade 4: 7.31 vs. Grade 5: 7.77; η2 = 0.38; p < 0.001), 20 weeks (η2 = 0.20, p < 0.001), 30 weeks (η2 = 0.11, p < 0.001), and 1 year (η2 = 0.03, p < 0.001). At 10 weeks, post hoc comparisons indicated stepwise increases in mIHS with higher surgical grades, with significant differences observed between Grade 1 vs. 2 (mean difference = 1.48 [95%CI 1.37 - 1.59]; d = 1.38; p < 0.001), Grade 2 vs. 3 (0.59 [95%CI 0.33 - 0.85]; d = 0.55; p < 0.001), and Grade 3 vs. 4 (0.45 [95%CI 0.02 - 0.87]; d = 0.42; p = 0.044). Significant differences persisted after 1 year between Grades 1 and 2 (mean difference = 0.40; p < 0.001; d = 0.25) and between Grades 2 and 4 (0.92; p = 0.003; d = 0.57). CONCLUSION: The mIHS, based on nonlinear mixed-effects modeling, effectively captured differences in cumulative postoperative disability among lumbar procedures, particularly within the first year. This novel approach may support comparative effectiveness research by quantifying total disability burden.
