Affiliation:
1. Department of Neurosurgery, The Johns Hopkins University School of Medicine, Baltimore, Maryland;
2. Department of Neurosurgery, Vanderbilt University Medical Center, Nashville, Tennessee; and
3. Independent Statistician, Lake Forest, California
Abstract
Object
Outcome studies for spine surgery rely on patient-reported outcomes (PROs) to assess treatment effects. Commonly used health-related quality-of-life questionnaires include the following scales: back pain and leg pain visual analog scale (BP-VAS and LP-VAS); the Oswestry Disability Index (ODI); and the EuroQol-5D health survey (EQ-5D). A shortcoming of these questionnaires is that their numerical scores lack a direct meaning or clinical significance. Because of this, the concept of the minimum clinically important difference (MCID) has been put forth as a measure for the critical threshold needed to achieve treatment effectiveness. By this measure, treatment effects reaching the MCID threshold value imply clinical significance and justification for implementation into clinical practice.
Methods
In 45 consecutive patients undergoing transforaminal lumbar interbody fusion (TLIF) for low-grade degenerative lumbar spondylolisthesis-associated back and leg pain, PRO questionnaires measuring BP-VAS, LPVAS, ODI, and EQ-5D were administered preoperatively and at 2 years postoperatively, and 2-year change scores were calculated. Four established anchor-based MCID calculation methods were used to calculate MCID, as follows: 1) average change; 2) minimum detectable change (MDC); 3) change difference; and 4) receiver operating characteristic curve analysis for two separate anchors (the health transition index [HTI] of the 36-Item Short Form Health Survey [SF-36], and the satisfaction index).
Results
All patients were available at the 2-year follow-up. The 2-year improvements in BP-VAS, LP-VAS, ODI, and EQ-5D scores were 4.3 ± 2.9, 3.8 ± 3.4, 19.5 ± 11.3, and 0.43 ± 0.44, respectively (mean ± SD). The 4 MCID calculation methods generated a range of MCID values for each of the PROs (BP-VAS, 2.1–5.3; LP-VAS, 2.1–4.7; ODI, 11–22.9; and EQ-5D, 0.15–0.54). The mean area under the curve (AUC) for the receiver operating characteristic curve from the 4 PRO-specific calculations was greater for the HTI versus satisfaction anchor (HTI [AUC 0.73] vs satisfaction [AUC 0.69]), suggesting HTI as a more accurate anchor.
Conclusions
The TLIF-specific MCID is highly variable based on calculation technique. The MDC approach with the SF-36 HTI anchor appears to be most appropriate for calculating MCID because it provided a threshold above the 95% CI of the unimproved cohort (greater than the measurement error), was closest to the mean change score reported by improved and satisfied patients, and was least affected by the choice of anchor. Based on the MDC method with HTI anchor, MCID scores following TLIF are 2.1 points for BP-VAS, 2.8 points for LP-VAS, 14.9 points for ODI, and 0.46 quality-adjusted life years for EQ-5D.
Publisher
Journal of Neurosurgery Publishing Group (JNSPG)