Patient-reported outcome questionnaires have become

Transcription

1 J Neurosurg Spine 16: , 2012 Minimum clinically important difference in pain, disability, and quality of life after neural decompression and fusion for same-level recurrent lumbar stenosis: understanding clinical versus statistical significance Clinical article Scott L. Parker, M.D., 1,2 Stephen K. Mendenhall, B.S., 1,2 David N. Shau, B.S., 1,2 Owoicho Adogwa, M.P.H., 1,2 William N. Anderson, Ph.D., 3 Clinton J. Devin, M.D., 1,2 and Matthew J. McGirt, M.D. 1,2 1 Department of Neurosurgery, Vanderbilt University Medical Center; 2 Vanderbilt University Spinal Column Surgical Quality and Outcomes Research Laboratory, Nashville, Tennessee; and 3 Lake Forest, California Object. Spine surgery outcome studies rely on patient-reported outcome (PRO) measurements to assess treatment effect, but the extent of improvement in the numerical scores of these questionnaires lacks a direct clinical meaning. Because of this, the concept of a minimum clinically important difference (MCID) has been used to measure the critical threshold needed to achieve clinically relevant treatment effectiveness. As utilization of spinal fusion has increased over the past decade, so has the incidence of same-level recurrent stenosis following index lumbar fusion, which commonly requires revision decompression and fusion. The MCID remains uninvestigated for any PROs in the setting of revision lumbar surgery for this pathology. Methods. In 53 consecutive patients undergoing revision surgery for same-level recurrent lumbar stenosis associated back and leg pain, PRO measures of back and leg pain were assessed preoperatively and 2 years postoperatively, using the visual analog scale for back pain (VAS-BP) and leg pain (VAS-LP), Oswestry Disability Index (ODI), Physical and Mental Component Summary categories of the12-item Short Form Health Survey (SF-12 PCS and MCS) for quality of life, Zung Depression Scale (ZDS), and EuroQol-5D health survey (EQ-5D). Four established anchor-based MCID calculation methods were used to calculate MCID (average change; minimum detectable change; change difference; and receiver operating characteristic curve analysis) for 2 separate anchors (health transition index of the SF-36 and the satisfaction index). Results. All patients were available for 2-year PRO assessment. Two years after surgery, a significant improvement was observed for all PROs assessed. The 4 MCID calculation methods generated a range of MCID values for each of the PROs (VAS-BP , VAS-LP , ODI , SF-12 PCS , SF-12 MCS , ZDS , and EQ-5D ). Each patient answered synchronously for the 2 anchors, suggesting both of these anchors are equally appropriate and valid for this patient population. Conclusions. The same-level recurrent stenosis surgery-specific MCID is highly variable based on calculation technique. The minimum detectable change approach is the most appropriate method for calculation of MCIDs in this population because it was the only method to reliably provide a threshold above the 95% confidence interval of the unimproved cohort (greater than the measurement error). Based on this method, the MCID thresholds following neural decompression and fusion for symptomatic same-level recurrent stenosis are 2.2 points for VAS-BP, 5.0 points for VAS-LP, 8.2 points for ODI, 2.5 points for SF-12 PCS, 10.1 points for SF-12 MCS, 4.9 points for ZDS, and 0.39 QALYs for EQ-5D. ( Key Words minimum clinically important difference failed back surgery syndrome lumbar spine Abbreviations used in this paper: EQ-5D = EuroQol-5D; MCID = minimum clinically important difference; MCS = Mental Component Summary; ODI = Oswestry Disability Index; PCS = Physical Component Summary; PRO = patient-reported outcome; QALY = quality-adjusted life year; ROC = receiver operating characteristic; SF-12, -36 = 12-, 36-Item Short Form Health Survey; VAS = visual analog scale; VAS-BP = VAS for back pain; VAS-LP = VAS for leg pain; ZDS = Zung Depression Scale. J Neurosurg: Spine / Volume 16 / May 2012 Patient-reported outcome questionnaires have become the standard measure for treatment effectiveness following spinal surgery. The VAS, 10 ODI, 8,9,22 SF-36, 27 and EQ-5D health survey 1,15,19 are the most commonly used PRO questionnaires for back and leg pain. A shortcoming is that the numerical scores of these outcome metrics alone lack a direct, clinically significant meaning. 471

2 S. L. Parker et al. Because of this, the concept of a minimum clinically important difference or MCID has been developed as a means to measure the critical threshold of improvement needed to achieve clinically meaningful treatment effectiveness. Treatment effects reaching the MCID threshold value imply clinical significance and justification for implementation into clinical practice. The MCID can be viewed as the smallest change in outcome measure that is important to patients. The most often used and validated methods to calculate MCID values are anchor-based approaches, which compare the change in PRO score following surgery to another measurement (anchor) of outcome such as perceived improvement after surgery or satisfaction with surgery. There have been a number of anchors and anchor-based calculation methods used in the literature to date, resulting in a wide variability in reported MCID values. 6,7,26 As a result of this variability, there has been no consensus on the best MCID calculation method, and definitive MCID values have yet to be established for the common PRO questionnaires used in spine surgery. It is important to appreciate that MCID values are specific to the PRO metric being used and the spinal pathology under study. Previous studies have attempted to determine the MCID of the VAS, ODI, and SF-36 in mixed spine surgery populations of various etiologies and surgical procedures. 5,6,12,29,30 Our group has previously assessed the MCID in patients undergoing transforaminal lumbar interbody fusion for spondylolisthesis-associated back and leg pain. To date, no studies have reported on MCID values for patients undergoing neural decompression and fusion for same-level recurrent lumbar stenosis; thus, we set out to determine the most appropriate MCID values for VAS, ODI, SF-12, ZDS, and EQ-5D for this patient population. Methods Patient Selection Fifty-three consecutive patients undergoing revision neural decompression (laminectomy/foraminotomy) and instrumented fusion for recurrent same-level lumbar stenosis at our institution between 2004 and 2008 were included in this study. The primary inclusion criteria were 1) prior laminectomy for lumbar stenosis; 2) MRI evidence of same-level stenosis; 3) presence of mechanical back and radicular leg pain; 4) age of years; and 5) failure of at least 6 months of conservative therapy. Patients were excluded if they had 1) an extraspinal cause of back pain; 2) infection, trauma, or neoplasia; 3) recurrent disc herniation after primary discectomy; or 4) an active workers compensation lawsuit. In all cases, posterior instrumented fusion was performed to treat to mechanical back pain from postlaminectomy instability or concern for destabilization. We received institutional review board approval for the present study from the Vanderbilt University Human Research Protection Program. Patient demographics, clinical presentation, indications for surgery, radiological studies, and operative variables were retrospectively reviewed for each case. Patient-assessed outcome measures were obtained via phone interview. All 4 surgeons participating in this study practiced similar postoperative treatment paradigms. In all cases, the surgeon encouraged discharge from the hospital beginning 48 hours after surgery, weaning of narcotics beginning 2 3 weeks after surgery, and return to work as soon as the patient felt capable. Pseudarthrosis or nonunion was defined as translation of greater than 3 mm and angulation of greater than 5 on flexion-extension radiographs or the presence of screw haloing or loosening on CT in the presence of corresponding mechanical back pain. 4 Patient-Reported Outcome Measures Institutional review board approval was obtained prior to contacting any patients for outcome assessment. Preoperative and 2-year outcomes after neural decompression and fusion for same-level recurrent lumbar stenosis were assessed by an independent investigator not involved with clinical care. Patient-reported outcome questionnaires included the VAS-BP, VAS-LP, 10,17 ODI, 8,9 SF-12 PCS and MCS scores, 11,27 ZDS, 24,31 and EQ-5D. 2,15,16 The EQ-5D was used for utility measurement because it has been validated and found to be responsive to low-back treatment. 25 Mean values are presented ± SD. Anchors Two ad hoc anchors were used to demonstrate the anchor-based derivations of MCID. The first anchor used was derived from the health transition item of the SF-36. This item asks patients to rate how their current health compares with their health prior to surgery. The choices provided include worse, unchanged, slightly better, or markedly better. Patients answering slightly better or markedly better were classified as responders, whereas those answering unchanged or worse were classified as nonresponders. The second anchor used was derived by asking patients whether they were satisfied with the results of their surgery. Patients answering yes were classified as responders, whereas those answering no were classified as nonresponders. Anchor-Based Approaches We chose 4 previously reported anchor-based approaches to assess MCID following neural decompression and fusion for same-level recurrent lumbar stenosis. 6,7 Each anchor-based approach calculates the MCID based on unique definitions: 1) the average change approach defines the MCID as the average change score seen in the cohort defined to be responders; 2) the minimum-detectable-change approach defines the MCID as the smallest change that can be considered above the measurement error with a given level of confidence (often 95% CI); therefore, the MCID value is equal to the upper value of the 95% CI for the average change score seen in the cohort defined to be nonresponders; 3) the change difference approach defines the MCID as the difference of the average change score for responders and nonresponders; and 4) the ROC curve approach defines the MCID as the change value that provides the greatest sensitivity and/or specificity for a positive response. When used in the setting of MCID analyses, 472 J Neurosurg: Spine / Volume 16 / May 2012

3 Minimum clinically important difference sensitivity is defined as the proportion of patients who report improvement based on the external criterion and have a PRO score above the MCID threshold value. Conversely, specificity is defined as the proportion of patients who do not report improvement based on the external criterion and have a PRO score below the MCID threshold value. 7 The MCID value is defined as that which equally maximizes sensitivity and specificity. Additionally, the probability that scores will correctly discriminate between responders and nonresponders (accuracy) is depicted by the area under the ROC curve. This value ranges from 0.5 (discrimination is no better than pure chance) to 1.0 (all patients are able to be correctly discriminated). An area under the ROC curve of is considered adequate, whereas one of is considered excellent. 7 Results At our institution, 2100 index lumbar fusions were performed from 2004 through A total of 53 patients presenting with symptomatic same-level recurrent stenosis during this time were enrolled in the study. All patients presented with back and/or leg pain and radiographic evidence of central canal stenosis with or without evidence of lateral recess stenosis. All 53 patients underwent revision laminectomy/foraminotomy and posterior instrumented fusion in which pedicle screws were placed. No patients in the series were excluded or lost to follow-up, and all patients had a minimum follow-up of 2 years. The mean age of all patients was 56.3 ± 12.5 years. There were 35 women (66%) and 18 men (34%). Eight patients (15%) had diabetes, and 9 (17%) were either current or previous smokers. The mean body mass index was 29.8 ± 8.1. The mean number of years between prior lumbar fusion and revision arthrodesis was 4.0 ± 4.2 years (Table 1). At presentation, the mean VAS-BP and VAS-LP scores were 9.3 ± 1.0 and 9.5 ± 0.9, respectively. The mean preoperative ODI and EQ-5D scores were 36.0 ± 6.0 and 0.18 ± 0.22, respectively. The mean preoperative PCS and MCS scores on the SF-12 were 25.1 ± 5.8 and 34.9 ± 12.8, respectively, and the mean self-rating ZDS score was 49.9 ± In all cases, patients underwent a second decompression at the level(s) of their index surgery. Posterior instrumented fusion was performed in all cases due to mechanical back pain due to postlaminectomy instability or TABLE 1: Summary of baseline characteristics Baseline Characteristic No. of Cases or Mean (%) no. of patients 53 mean age (yrs) 56.3 ± 12.5 no. of males 18 (34) mean body mass index 29.8 ± 8.1 diabetes 8 (15) smoking history 9 (17) bone morphogenetic protein 8 (15) mean hospital stay (days) 4.4 ± 1.6 mean yrs btwn index op & reop 4.0 ± 4.2 J Neurosurg: Spine / Volume 16 / May 2012 concern for destabilization during more extensive revision laminectomy or facetectomy. 20,21 All surgeries were performed with open technique and used similar pedicle screw systems. The mean duration of the hospital stay was 4.4 ± 1.6 days. One dural tear (1.8%) occurred and was repaired intraoperatively without complication. One patient (1.8%) experienced an incisional CSF leak perioperatively. There were 2 cases (3.7%) of surgical site infection, both treated with intravenous antibiotics. One patient (1.8%) had a malpositioned pedicle screw and required a second operation in the perioperative period. Hardware failure and radiculopathy were absent in all cases. Two years after revision, no patients experienced significant persistence of back pain in conjunction with radiographic evidence of nonunion (Fig. 1); however, it is possible that longer follow-up may reveal delayed pseudarthrosis in a subset of patients. Patient-Reported Outcome Scores at 2 Years At 2 years postoperatively, each of the outcome measures assessed had significantly improved in the population as a whole. The mean improvements in VAS-BP and -LP scores were 4.3 ± 3.1 (p < 0.001) and 6.1 ± 3.3 (p < 0.001), respectively. The mean improvements in ODI and EQ-5D were 14.3 ± 12.4 (p < 0.001) and 0.41 ± 0.34 (p < 0.001), respectively. The mean improvements in SF-12 PCS, SF-12 MCS, and ZDS scores were 7.2 ± 11.2 (p < 0.001), 12.6 ± 13.0 (p < 0.001), and 12.4 ± 16.8 (p < 0.001), respectively (Fig. 2). For both the health transition item and satisfaction anchors, 34 patients (64%) were classified as responders and 19 (36%) as nonresponders. Each patient responded synchronously to each anchor question. There was a significant difference in 2-year change scores for responders compared with nonresponders. The mean improvement in VAS-BP and -LP scores in responders versus nonresponders was 6.0 versus 1.3 (p < 0.001) and 7.5 versus 3.6 (p < 0.001), respectively. The mean improvement in ODI and EQ-5D scores for responders versus nonresponders was 19.9 versus 4.3 (p < 0.001) and 0.52 versus 0.23 (p = 0.003), respectively. The mean improvement in SF-12 PCS, SF-12 MCS, and ZDS scores for responders versus nonresponders was 11.5 versus (p < 0.001), 15.9 versus 6.7 (p = 0.003), and 18.6 versus 1.3 (p < 0.001), respectively (Fig. 3). Visual Analog Scale for Low-Back Pain The MCID threshold ranged from 2.2 to 6.0 for the The smallest threshold was derived from the minimumdetectable-change approach, whereas the average-change approach yielded the largest threshold. The area under the curve for each anchor was 0.93 (Table 2). Visual Analog Scale for Leg Pain The MCID threshold ranged from 3.9 to 7.5 for the The smallest threshold was derived from the changedifference approach, whereas the average-change approach yielded the largest threshold. The area under the curve for each anchor was 0.83 (Table 2). 473

4 S. L. Parker et al. Fig. 1. A and B: Sagittal and axial preoperative MR images revealing severe central canal stenosis and bilateral neural foraminal narrowing at L4 5 with Grade I spondylolisthesis of L-4 on L-5. The patient had previously undergone L3 5 laminectomies. C E: Postoperative dynamic upright, extension, and flexion radiographs demonstrating no pathological movement, suggesting solid fusion. Oswestry Disability Index The MCID threshold ranged from 8.2 to 19.9 for the The smallest threshold was derived from the minimumdetectable-change approach, whereas the average-change approach yielded the largest threshold. The area under the curve for each anchor was 0.88 (Table 2). SF-12 PCS Score The MCID threshold ranged from 2.5 to 12.1 for the The smallest threshold was derived from the minimum-detectable-change approach, whereas the change-difference approach yielded the largest threshold. The area under the curve for each anchor was 0.87 (Table 2). SF-12 MCS Score The MCID threshold ranged from 7.0 to 15.9 for the The smallest threshold was derived from the ROC curve approach, whereas the average-change approach yielded the largest threshold. The area under the curve for each anchor was 0.68 (Table 2). Self-Reporting ZDS The MCID threshold ranged from 3.0 to 18.6 for the The smallest threshold was derived from the ROC curve approach, whereas the average-change approach yielded the largest threshold. The area under the curve for each anchor was 0.83 (Table 2). EuroQol-5D The MCID threshold ranged from 0.29 to 0.52 for the health transition item and satisfaction with surgery anchors. The smallest threshold was derived from the change-difference approach, whereas the average-change approach yielded the largest threshold. The area under the curve for each anchor was 0.75 (Table 2). Comparison of Anchor and MCID Calculation Method Unlike previous studies that have shown variation in response to multiple anchors, 19 each patient responded synchronously for the 2 anchors in the present study. As a result, there was no difference in the MCID threshold generated with the health transition item anchor compared with the satisfaction with surgery anchor. The area under the ROC curve ranged from 0.68 to 0.93 for the 7 PRO metrics assessed in this study, indicating that these 2 anchors provide an appropriate means to delineate responders from nonresponders in this patient population. The averagechange approach produced the largest MCID threshold for the majority of PRO metrics assessed, whereas the minimum-detectable-change, change-difference, and ROC curve approaches provided the smallest MCID threshold (Table 2). The minimum-detectable-change approach appears to generate the most appropriate MCID threshold in this patient population, as it was the only method to consistently provide a threshold above the 95% CI of the unimproved cohort (greater than the measurement error). Based on the minimum-detectable-change method with either the health transition item or satisfaction with surgery anchor, the 474 J Neurosurg: Spine / Volume 16 / May 2012

5 Minimum clinically important difference Fig. 2. Baseline and 2-year scores obtained in patients undergoing revision neural decompression and instrumented fusion for same-level recurrent stenosis associated back and leg pain. Each PRO measure showed significant improvement at 2 years postoperatively. The overall mean change scores were as follows: VAS-BP 4.3 ± 3.1 (p < 0.001), VAS-LP 6.1 ± 3.3 (p < 0.001), ODI 14.3 ± 12.4 (p < 0.001), SF-12 PCS 7.2 ± 11.2 (p < 0.001), SF-12 MCS 12.6 ± 13.0 (p < 0.001), ZDS 12.4 ± 16.8 (p < 0.001), and EQ-5D 0.18 ± 0.19 (p < 0.001). MCID thresholds following neural decompression and fusion for symptomatic same-level recurrent stenosis are 2.2 points for VAS-BP, 5.0 points for VAS-LP, 8.2 points for ODI, 2.5 points for SF-12 PCS, 10.1 points for SF-12 MCS, 4.9 points for ZDS, and 0.39 QALYs for EQ-5D. Discussion In this study, we set out to determine the variability of MCID thresholds using various calculation techniques in a homogeneous patient population and to define an optimal MCID value for 7 commonly used PRO questionnaires in patients undergoing neural decompression and instrumented fusion for symptomatic same-level recurrent stenosis. Each of the 4 anchor-based approaches used to calculate an MCID value resulted in a significant range of thresholds: for VAS-BP, for VAS-LP, 8.2 J Neurosurg: Spine / Volume 16 / May for ODI, for SF-12 PCS, for SF-12 MCS, for ZDS, and for EQ-5D. The largest MCID threshold was most often generated by the average-change approach, whereas the minimum-detectable-change, change-difference, and ROC curve approaches provided the smallest MCID threshold values. All 4 of the calculation methods assessed are currently used in the literature, and the variability reported here highlights the need to standardize the most appropriate calculation method for MCID. We also compared 2 anchors: 1) general health assessment in the form of the health transition index of the SF- 36 health survey, and 2) satisfaction with surgery. General health assessments have traditionally been used as anchors and have been shown to be sensitive to change. 12 They are also widely used for their convenience, as they are part of the commonly used SF-36. In the current study, there 475

6 S. L. Parker et al. Fig. 3. There was a significant difference in change score between responders and nonresponders for each of the PRO metrics assessed. Mean improvement in VAS-BP, VAS-LP, ODI, SF-12 PCS, SF-12 MCS, ZDS, and EQ-5D scores for responders versus nonresponders was 6.0 versus 1.3 (p < 0.001), 7.5 versus 3.6 (p < 0.001), 19.9 versus 4.3 (p < 0.001), 11.5 versus (p < 0.001), 15.9 versus 6.7 (p = 0.003), 18.6 versus 1.3 (p < 0.001), and 0.52 versus 0.23 (p = 0.003), respectively. was no variability between the 2 anchors. Each patient answered synchronously for the 2 anchors, and as a result, both of these anchors appeared equally appropriate and valid for this patient population. To date, no consensus has been reached as to the superior method with which to calculate an MCID. Calculation methods that have been previously supported in the literature include standard error of measurement, 14,29,30 one-half standard deviation, 3,18,28 effect size, 13,23 minimum detectable change, 6,19 and ROC curve. 5 Copay and coworkers 6 systematically compared calculation methods and their resulting values within the same patient population. Similar to the results of the present study, a wide variability in MCID values was observed depending on the MCID calculation method that was used. Based on the fact that an appropriate MCID value should 1) be greater than the measurement error of patients defined to be nonresponders, and 2) correspond to the patient s perception of an important amount TABLE 2: Summary of the 4 MCID calculations generating a range of values for each of the PROs assessed* Anchor-Based Approach Calculated MCID Value VAS-BP VAS-LP ODI SF-12 PCS SF-12 MCS ZDS EQ-5D average change minimum detectable change (95% CI) change difference ROC curve (area under the curve) 4.0 (0.93) 6.0 (0.83) 9.0 (0.88) 4.4 (0.87) 7.0 (0.68) 3.0 (0.83) 0.43 (0.75) * The health transition index of the SF-36 was used as one anchor and compared to a satisfaction-based anchor. All patients responded synchronously to each anchor, suggesting that the health transition index and satisfaction anchors are each equally appropriate in the population. The ROC curve analysis approach may be the most appropriate MCID calculation because it generated a threshold of improvement that was statistically greater than chance error from unimproved patients (minimum detectable change), was most consistent with the difference in change scores between responders and nonresponders (change difference), and predicted patient-reported improvement with equal specificity and sensitivity. 476 J Neurosurg: Spine / Volume 16 / May 2012

7 Minimum clinically important difference of change, each study concludes that the minimum-detectable-change approach is the superior calculation method for determining an MCID threshold in patients undergoing various lumbar spine surgeries. In a prior study, our group defined MCID 2 years following transforaminal lumbar interbody fusion in patients with low-grade degenerative lumbar spondylolisthesis. 19 Just as in the current study, the minimum-detectablechange method was determined to be the superior calculation technique. The MCID thresholds 2 years following transforaminal lumbar interbody fusion for degenerative lumbar spondylolisthesis were 2.1 and 2.8 points, respectively, for VAS-BP and VAS-LP, 14.9 points for ODI, and 0.46 QALYs for EQ-5D. It is essential to appreciate the importance and utility of determining MCID thresholds for homogeneous patient population undergoing a specific surgical procedure. Each patient population will have unique emotional and physical stressors, different perceptions of pain, and varying degrees of improvement viewed as significant. These differences can be elicited with the use of patient reported outcome questionnaires. The revision fusion patient population described in the current study is different than a population presenting with de novo back pain and seeking primary fusion; therefore, differences in satisfaction, a patient s intrinsic view of a meaningful improvement, and subsequent MCID thresholds should not be unexpected. While the power of this study did not allow characterization of differences between responders and nonresponders, this important point should be analyzed in future studies as it could help identify patients that may be the most appropriate surgical candidates. Conclusions Same-level recurrent stenosis specific MCID thresholds are variable based on calculation technique. The health transition item and satisfaction with surgery anchors appear equally appropriate in this patient population. The minimum-detectable-change approach is the most appropriate method for calculation of MCID in this population because it was the only method to reliably provide a threshold above the 95% CI of the unimproved cohort (greater than the measurement error). Based on this method with either the health transition item or satisfaction with surgery anchor, the MCID thresholds following neural decompression and fusion for symptomatic same-level recurrent stenosis are 2.2 and 5.0 points, respectively, for VAS-BP and VAS-LP, 8.2 points for ODI, 2.5 points for SF-12 PCS, 10.1 points for SF-12 MCS, 4.9 points for ZDS, and 0.39 QALYs for EQ-5D. Disclosure The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: Parker, McGirt. Acquisition of data: Parker, Mendenhall, Shau, Adogwa. Analysis and interpretation of data: Parker, McGirt. Drafting the article: Parker. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of J Neurosurg: Spine / Volume 16 / May 2012 the manuscript on behalf of all authors: Parker. Statistical analysis: Anderson. References 1. Adogwa O, Parker SL, Shau D, Mendelhall SK, Cheng J, Aaronson O, et al: Long-term outcomes of revision fusion for lumbar pseudarthrosis. Clinical article. J Neurosurg Spine 15: , Badia X, Diaz-Prieto A, Gorriz MT, Herdman M, Torrado H, Farrero E, et al: Using the EuroQol-5D to measure changes in quality of life 12 months after discharge from an intensive care unit. Intensive Care Med 27: , Beaton DE: Simple as possible? Or too simple? Possible limits to the universality of the one half standard deviation. Med Care 41: , Burkus JK, Dorchak JD, Sanders DL: Radiographic assessment of interbody fusion using recombinant human bone morphogenetic protein type 2. Spine (Phila Pa 1976) 28: , Carreon LY, Glassman SD, Campbell MJ, Anderson PA: Neck Disability Index, short form-36 physical component summary, and pain scales for neck and arm pain: the minimum clinically important difference and substantial clinical benefit after cervical spine fusion. Spine J 10: , Copay AG, Glassman SD, Subach BR, Berven S, Schuler TC, Carreon LY: Minimum clinically important difference in lumbar spine surgery patients: a choice of methods using the Oswestry Disability Index, Medical Outcomes Study questionnaire Short Form 36, and pain scales. Spine J 8: , Copay AG, Subach BR, Glassman SD, Polly DW Jr, Schuler TC: Understanding the minimum clinically important difference: a review of concepts and methods. Spine J 7: , Fairbank JC, Couper J, Davies JB, O Brien JP: The Oswestry low back pain disability questionnaire. Physiotherapy 66: , Fairbank JC, Pynsent PB: The Oswestry Disability Index. Spine (Phila Pa 1976) 25: , Gallagher EJ, Liebman M, Bijur PE: Prospective validation of clinically important changes in pain severity measured on a visual analog scale. Ann Emerg Med 38: , Gandek B, Ware JE, Aaronson NK, Apolone G, Bjorner JB, Brazier JE, et al: Cross-validation of item selection and scoring for the SF-12 Health Survey in nine countries: results from the IQOLA Project. J Clin Epidemiol 51: , Hägg O, Fritzell P, Nordwall A: The clinical importance of changes in outcome scores after treatment for chronic low back pain. Eur Spine J 12:12 20, Hays RD, Woolley JM: The concept of clinically meaningful difference in health-related quality-of-life research. How meaningful is it? Pharmacoeconomics 18: , Jaeschke R, Singer J, Guyatt GH: Measurement of health status. Ascertaining the minimal clinically important difference. Control Clin Trials 10: , Jansson KA, Németh G, Granath F, Jönsson B, Blomqvist P: Health-related quality of life (EQ-5D) before and one year after surgery for lumbar spinal stenosis. J Bone Joint Surg Br 91: , Johnson JA, Coons SJ, Ergo A, Szava-Kovats G: Valuation of EuroQOL (EQ-5D) health states in an adult US sample. Pharmacoeconomics 13: , Kelly AM: The minimum clinically significant difference in visual analogue scale pain score does not differ with severity of pain. Emerg Med J 18: , Norman GR, Sloan JA, Wyrwich KW: Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care 41: , Parker SL, Adogwa O, Paul AR, Anderson WN, Aaronson O, Cheng JS, et al: Utility of minimum clinically important difference in assessing pain, disability, and health state after transfo- 477

8 S. L. Parker et al. raminal lumbar interbody fusion for degenerative lumbar spondylolisthesis. Clinical article. J Neurosurg Spine 14: , Resnick DK, Choudhri TF, Dailey AT, Groff MW, Khoo L, Matz PG, et al: Guidelines for the performance of fusion procedures for degenerative disease of the lumbar spine. Part 9: fusion in patients with stenosis and spondylolisthesis. J Neurosurg Spine 2: , Resnick DK, Choudhri TF, Dailey AT, Groff MW, Khoo L, Matz PG, et al: Guidelines for the performance of fusion procedures for degenerative disease of the lumbar spine. Part 10: fusion following decompression in patients with stenosis without spondylolisthesis. J Neurosurg Spine 2: , Roland M, Fairbank J: The Roland-Morris Disability Questionnaire and the Oswestry Disability Questionnaire. Spine (Phila Pa 1976) 25: , 2000 (Erratum in Spine (Phila Pa 1976) 26:847, 2001) 23. Samsa G, Edelman D, Rothman ML, Williams GR, Lipscomb J, Matchar D: Determining clinically important differences in health status measures: a general approach with illustration to the Health Utilities Index Mark II. Pharmacoeconomics 15: , Thurber S, Snow M, Honts CR: The Zung Self-Rating Depression Scale: convergent validity and diagnostic discrimination. Assessment 9: , Tosteson AN, Lurie JD, Tosteson TD, Skinner JS, Herkowitz H, Albert T, et al: Surgical treatment of spinal stenosis with and without degenerative spondylolisthesis: cost-effectiveness after 2 years. Ann Intern Med 149: , van der Roer N, Ostelo RW, Bekkering GE, van Tulder MW, de Vet HC: Minimal clinically important change for pain intensity, functional status, and general health status in patients with nonspecific low back pain. Spine (Phila Pa 1976) 31: , Ware JE Jr: SF-36 health survey update. Spine (Phila Pa 1976) 25: , Wright JG: Interpreting health-related quality of life scores: the simple rule of seven may not be so simple. Med Care 41: , Wyrwich KW, Nienaber NA, Tierney WM, Wolinsky FD: Linking clinical relevance and statistical significance in evaluating intra-individual changes in health-related quality of life. Med Care 37: , Wyrwich KW, Tierney WM, Wolinsky FD: Further evidence supporting an SEM-based criterion for identifying meaningful intra-individual changes in health-related quality of life. J Clin Epidemiol 52: , Zung WW, Richards CB, Short MJ: Self-rating depression scale in an outpatient clinic. Further validation of the SDS. Arch Gen Psychiatry 13: , 1965 Manuscript submitted September 26, Accepted January 16, Please include this information when citing this paper: published online February 10, 2012; DOI: / SPINE Address correspondence to: Scott L. Parker, M.D., Department of Neurosurgery, Vanderbilt University Medical Center, 4347 Village at Vanderbilt, Nashville, Tennessee scott. parker@vanderbilt.edu. 478 J Neurosurg: Spine / Volume 16 / May 2012