Measuring clinically relevant improvement after lumbar spine surgery: is it time for something new?

Abstract

BACKGROUND CONTEXT

Minimum clinically important difference (MCID) for patient-reported outcome measures is commonly used to assess clinical improvement. However, recent literature suggests that an absolute point-change may not be an effective or reliable marker of response to treatment for patients with low or high baseline patient-reported outcome scores. The multitude of established MCIDs also makes it difficult to compare outcomes across studies and different spine surgery procedures.

PURPOSE

To determine whether a 30% reduction from baseline in disability and pain is a valid method for determining clinical improvement after lumbar spine surgery.

STUDY DESIGN

Retrospective analysis of prospective data from a national spine registry, the Quality Outcomes Database.

PATIENT SAMPLE

There were 23,280 participants undergoing elective lumbar spine surgery for degenerative disease who completed a baseline and follow-up assessment at 12 months.

OUTCOME MEASURES

Patient-reported disability (Oswestry Disability Index [ODI]), back and leg pain (11-point Numeric Rating Scale [NRS]), and satisfaction (NASS scale).

METHODS

Patients completed baseline and a 12-month postoperative assessment to evaluate the outcomes of disability, pain, and satisfaction. The change in ODI and NRS pain scores was categorized as met (≥30%) or not met (<30%) percent reduction MCID. The 30% reduction from baseline was compared with a wide range of well-established absolute point-change MCID values. The relationship between 30% reduction and absolute change values and satisfaction were primarily compared using receiver operating characteristic (ROC) curves, area under the curve (AUROC), and logistic regression analyses. Analyses were conducted for overall scores and for disability and pain severity categories and by surgical procedure.

RESULTS

Thirty percent reduction in ODI and back and leg pain predicted satisfaction with more accuracy than absolute point-change values for the total population and across all procedure categories (p<.001), except for when compared with the highest absolute point-change threshold for leg pain (3.5-point reduction). The largest AUROC differences, in favor of 30% reduction, were found for the lowest disability (ODI 0–20%: 21.8%) and bed-bound disability (ODI 81%–100%: 13.9%) categories. For pain, there was a 3.4%–12.4% and 1.3%–9.8% AUROC difference for no/mild back and leg pain (NRS 0–4), respectively, in favor of a 30% reduction threshold.

CONCLUSIONS

A 30% reduction MCID either outperformed or was similar to absolute point-change MCID values. Results indicate that a 30% reduction (baseline to 12 months after surgery) in disability and pain is a valid method for determining clinically relevant improvement in a broad spine surgery population. Furthermore, a 30% reduction was most accurate for patients in the lowest and highest disability and lowest pain severity categories. A 30% reduction MCID allows for a standard cut-off for disability and pain that can be used to compare outcomes across various spine surgery procedures.

Introduction

Lumbar spine surgery is one of the most prevalent and fastest-growing medical interventions in the United States. The rates of elective lumbar spine surgery have increased 300% in recent decades [1], [2], [3]. In 2011, spinal fusions cost the United States health-care system a total of $12.8 billion, more than any other procedure type [4]. Due to the prevalence, incidence, and cost of these procedures along with the shift toward value-based care, accurate, and efficient assessment of patient benefit from lumbar spine surgery is imperative.

Minimum clinically important difference (MCID) [5] represents the smallest, clinically relevant change in a patient-reported outcome (PRO) score such as the Oswestry Disability Index (ODI) [6,7] and numeric rating scales (NRS) for back pain and leg pain [8]. MCID thresholds in spine surgery are used to determine whether interventions benefit patients [9], [10], [11], [12], [13] and to develop predictive calculators to facilitate informed medical decision making [14], [15], [16]. The most widely used MCID values in lumbar spine surgery for ODI, NRS back pain, and NRS leg pain are 12.8, 1.2, and 1.6 points, respectively [17]. These values were determined by Copay et al. in patients with diverse lumbar pathologies undergoing various surgical procedures [17]. Other studies have defined MCID point-change values for a wide range of lumbar surgical subpopulations, including transforaminal lumbar interbody fusion for degenerative lumbar spondylolisthesis [18], neural decompression and fusion for same level recurrent lumbar stenosis [19], and revision fusion for symptomatic pseudarthrosis [20].

Although MCID is an accepted classification strategy for clinical effectiveness, the literature suggests that an absolute change from baseline, as opposed to relative or proportional, may not be an effective or reliable marker of response to treatment for patients with low or high baseline PRO scores [21,22]. Studies in the chronic pain literature that compare absolute change measures to thresholds of clinical relevance defined by percent reduction from baseline show improved performance with the percent reduction measurements [23]. Furthermore, Ostelo et al. developed recommendations on back pain outcome measures based on literature review, expert panel, and a workshop and concluded that a 30% change from baseline can be considered clinically meaningful improvement [24]. To date, one study in the spine surgery literature has calculated percent improvement thresholds for various PROs [25]. However, these values were developed with the goal of identifying thresholds for substantial clinical benefit, which is a higher change threshold than MCID, and from a single center study.

The primary objective of this study was to determine whether a percent reduction from baseline in disability and back and leg pain is a valid method for determining MCID after lumbar spine surgery. We hypothesized that a 30% reduction would more accurately represent clinical improvement than established MCIDs that utilize absolute point-change values. Percent reduction from baseline may be a more clinically relevant method for identifying response to treatment among spinal surgery populations by accounting for baseline PRO score.