Reddy, Vishruth K. BA*; Parker, Scott L. MD*; Lockney, Dennis T. MD*; Patrawala, Samit A. MD*; Su, Pei-Fang PhD‡,§; Mericle, Robert A. MD¶
BACKGROUND: The Visual Analog Scale (VAS) and the Barrow Neurological Institute Pain Scale (BNI-PS) are 2 patient-reported outcome (PRO) tools frequently used to rate pain from trigeminal neuralgia (TN). Outcomes studies often use these patient-reported outcomes to assess treatment effectiveness, but it is unknown exactly what degree of change in the numerical scores constitutes the minimum clinically important difference (MCID). MCID remains uninvestigated for percutaneous stereotactic radiofrequency lesioning (RFL), a common surgical procedure for TN.
OBJECTIVE: To determine MCID values for the VAS and BNI-PS in patients undergoing RFL.
METHODS: Forty-three consecutive patients with TN who underwent RFL by a single surgeon were prospectively assessed with the VAS and BNI-PS preoperatively and 3 years postoperatively. Three anchors were used to assign each patient’s outcome: satisfaction, willingness to have the surgery again, and Health Transition Index. We then used 3 well-established, anchor-based methods to calculate MCID: average change, minimum detectable change, and change difference.
RESULTS: Patients experienced substantial improvement in both VAS (9.81 vs 3.35; P < .001) and BNI-PS (4.95 vs 2.44; P < .001) after RFL. The 3 MCID calculation methods generated a range of MCID values for each of the PROs (VAS, 4.13-8.20; BNI-PS, 1.03-3.30). The area under the receiver-operating characteristic curve was greater for BNI-PS compared with VAS for all 3 anchors, indicating that BNI-PS is probably better suited for calculating MCID.
CONCLUSION: RFL-specific MCID is variable on the basis of the calculation technique. With the use of the minimum detectable change calculation method with the Health Transition Index anchor, the minimum clinically important difference is 4.49 for VAS and 1.16 for BNI-PS after RFL for TN.
ABBREVIATIONS: AUC, area under the receiver-operating characteristic curve
BNI-PS, Barrow Neurological Institute Pain Scale
HTI, Health Transition Index
MCID, minimum clinically important difference
MDC, minimum detectable change
PRO, patient-reported outcome
RFL, percutaneous stereotactic radiofrequency lesioning
TN, trigeminal neuralgia
VAS, Visual Analog Scale
Patient-reported outcome (PRO) measures such as the Visual Analog Scale (VAS)1,2 and the Barrow Neurological Institute Pain Scale (BNI-PS)1,3,4 are frequently used to quantify pain improvement after various treatments for trigeminal neuralgia (TN) and therefore can help assess response to treatment. However, it is unknown exactly what degree of change in the numerical scores is necessary to be considered the minimum clinically important difference (MCID). We recently determined the MCID for these PROs specifically for patients who underwent microvascular decompression for classic TN.5 Because MCID calculations are specific for each treatment approach, it is necessary to determine the MCID specifically for each common surgical procedure for TN, including percutaneous stereotactic radiofrequency lesioning (RFL).
MCID can be thought of as the smallest change that is important to patients.6 Jaeschke et al7 first defined it as “the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient's management.”
Multiple methodological approaches have been described in previous studies to calculate MCID.8,9 Two of the most commonly used methods are the distribution-based and anchor-based calculation methods. The anchor-based approach remains the most frequently used and accepted method to calculate MCID.5,8-10 It compares the change in PRO score with some other measure of change, considered an anchor or external criterion.9 In this study, we used the following 3 anchors: satisfaction with surgery, willingness to have surgery again given the outcome experienced, and perceived improvement after surgery with the Health Transition Index (HTI) portion of the Short Form-36, which is an established quality-of-life measurement tool.11,12 To date, there is no consensus as to which anchors or anchor-based calculation techniques are superior for determining MCID values.
Although MCID values have been determined for patients undergoing various neurosurgical interventions,5,10-12 no studies have assessed MCID for patients undergoing RFL for TN. Therefore, our goals were to document the variability of MCID values obtained via common anchor-based calculations and to determine RFL-specific MCID values for VAS and BNI-PS in patients with TN.
PATIENTS AND METHODS
We obtained approval from our institutional review board for this study. All patients underwent RFL by a single surgeon (R.A.M.). No patients had prior surgical treatment for TN. To be included, the patient had to have (1) a clinical diagnosis of TN based on presentation symptoms similar to those described by the International Headache Society classification,13 (2) failure of conservative and medical management defined as persistent breakthrough pain despite multiple antiepilepsy drugs in doses high enough to cause medication side effects, (3) pain reported as severe and significantly interfering with their activities of daily living despite maximum medical and nonsurgical treatments, and (4) any 1 or more of the following: poor candidacy for general anesthesia and/or suboccipital craniotomy, a concomitant diagnosis of multiple sclerosis, or patient preference to avoid craniotomy for microvascular decompression.
Surgical Technique: Percutaneous Stereotactic RFL
Patients were placed under methohexital anesthesia without intubation for all painful portions of the procedure. After standard surgical prepping, a stereotactic RFL stylet was advanced percutaneously approximately 3 cm lateral to the corner of the mouth and advanced up through the base of the skull into the gasserian ganglion. The stylet was then removed, and an electrode was placed within the gasserian ganglion. Patients were then awakened, and stimulation of the electrode was provided to achieve optimal positioning of the electrode. The radiofrequency generator was set to 70° for 90 seconds for the first radiofrequency lesion. After the lesion was completed, the patient was awakened and asked to confirm an absence of TN pain and at least 50% pinprick anesthesia. We assessed 50% pinprick anesthesia by first establishing the patient's baseline sensation on the contralateral, normal side of the face by defining a sharp stimulus (safety pin) as 10 of sharpness and defining a dull stimulus as 0 of sharpness. We then determined the patient's ability to detect the sharp stimulus on the operative side before and after the RFL procedure. Typically, the patient reported a 10 of sharpness before RFL, and the goal of the RFL procedure was to achieve approximately a 5 of sharpness. After this was achieved, the needle was removed, and patients were discharged home after 2 hours.
The VAS and BNI-PS are the 2 PRO measures used in this study. The VAS provides an estimate of pain intensity on a continuous scale, with a score of 0 representing no pain and a score of 10 representing worst pain.1 The BNI-PS rates pain on a scale of 1 to 5, incorporating degree of dependence on medications (Table 1).
Patients were prospectively enrolled in the study and available for evaluation at the 3-year follow-up. Preoperatively, PRO scores were obtained via face-to-face interviews. Approximately 3 years postoperatively, follow-up PRO scores were obtained via telephone interviews by an independent observer not involved in any aspect of patient care.
For derivations of MCID values, we used 3 anchors that have been used in multiple previous studies.5,9,10,12 They were satisfaction with surgery, willingness to have surgery again given the outcome experienced, and perceived improvement after surgery using the HTI portion of the Short Form-36, which is an established quality-of-life measurement tool.11,12 The satisfaction with surgery anchor functioned by asking patients whether they were satisfied with the results of their surgery. Those answering “yes” were labeled responders, and those answering “no” were labeled nonresponders. The willingness to have surgery again anchor functioned by asking patients if they would have the surgery again based on their outcome. Those answering “yes” were labeled responders, and those who responded “no” were labeled nonresponders. The perceived improvement after surgery anchor was derived from the HTI portion of the Short Form-36 quality-of-life health survey, which rates patients' health state postoperatively compared with their health state preoperatively. The choices included worse, unchanged, slightly better, significantly better, or completely better. Patients answering “significantly better” or “completely better” were labeled responders, and patients answering “worse,” “unchanged,” or “slightly better” were labeled nonresponders.
MCID Anchor-Based Calculations
Three previously reported anchor-based calculation methods to determine MCID values were chosen for this study, as used by multiple other authors.5,9,10,12 They included average change, minimum detectable change (MDC), and change difference. With the average change calculation method, MCID equals the average difference in pain preoperatively vs postoperatively for those patients labeled responders. With the MDC calculation method, MCID equals the smallest change that can be considered above the measurement error, calculated here with the 95% confidence level. Thus, MCID equals the upper value of the 95% confidence interval for the average change score in the group of patients labeled nonresponders. With the change difference calculation method, MCID equals the difference in the average change scores for the groups of patients labeled responders and nonresponders. To determine the probability that these MCID values would accurately delineate between responders and nonresponders, we calculated the area under the receiver-operating characteristic curve (AUC). The AUC values can range from 0.5, where delineation equals that of pure chance, to 1.0, where all patients are correctly delineated.9,10 An area of 0.7 to 0.8 is considered acceptable, whereas an area of 0.8 to 0.9 is considered excellent.9
Forty-three consecutive patients (29 women, 14 men) undergoing RFL were enrolled in the study, and all were available for evaluation at the 3-year follow-up. The mean follow-up time between PRO measurements was 3.01 years (95% confidence interval, 2.53-3.50), with a median follow-up time of 2.95 years. No patient received subsequent surgical treatment for TN in the 3-year follow-up period. The mean ± SD age was 68.9 ± 11.5 years. At the time of presentation, VAS and BNI-PS were 9.81 ± 0.50 and 4.95 ± 0.21, respectively. Three years postoperatively, each of the outcome measures assessed demonstrated substantial improvement in our patient population. Therefore, the mean ± SD change (improvement) scores for VAS and BNI-PS were 6.47 ± 4.19 and 2.51 ± 1.53, respectively (Figure 1). For the HTI anchor, 6 patients (14%) assessed themselves as worse, 2 (5%) as unchanged, 5 (12%) as slightly better, 14 (33%) as significantly better, and 16 (37%) as completely better. For the satisfaction with surgery anchor, 32 patients (74%) answered “satisfied,” whereas the remaining 11 (26%) answered “not satisfied.” For the surgery again anchor, 33 patients (77%) patients answered “yes” to willingness to have surgery again, and the remaining 10 (23%) answered “no.”
Depending on calculation method, the MCID threshold ranged from 4.49 to 8.20 for the HTI anchor, 4.82 to 7.81 for the satisfaction with surgery anchor, and 4.13 to 7.76 for the surgery again anchor (Table 2). For all anchors, the smallest threshold was derived from the MDC approach and the largest from the average change approach.
The MCID threshold ranged from 1.16 to 3.30 for the HTI anchor, 1.03 to 3.19 for the satisfaction with surgery anchor, and 1.12 to 3.09 for the surgery again anchor (Table 2). For all anchors, the smallest threshold was derived from the MDC approach and the largest from the average change approach.
Comparison of Anchor and MCID Calculation
The MCID values varied depending on calculation methods (Table 2). The AUC was slightly greater for the HTI anchor compared with the satisfaction and surgery again anchors when the VAS was used (Figure 2). The AUC was also slightly greater for the HTI anchor compared with the satisfaction and surgery again anchors when the BNI-PS was used (Figure 2). For the VAS, the average change approach consistently produced the largest MCID value, whereas the MDC approach consistently produced the smallest MCID value. For the BNI-PS, the average change approach consistently produced the largest MCID value, and the MDC approach consistently produced the smallest MCID value.
In our cohort, the MDC method appeared to be the most appropriate calculation technique because it provided a threshold value above the 95% confidence interval of the unimproved cohort (greater than the measurement error) and corresponded to patient perception of importance of change. With the use of the MDC method along with the HTI anchor, which consistently demonstrated the highest AUC values, the MCID was 4.49 for VAS and 1.16 for BNI-PS.
The average change MCID calculation reflects the average difference in pain preoperatively vs postoperatively in responders. Patients with TN who are considering surgery generally have preoperative pain rated at 9 or 10 on the VAS because the pain from TN is often referred to as the worst pain known to mankind.14 The pain relief achieved in our cohort was very high, with most patients improving to a ≤ 3 on the VAS after RFL. This calculates to an average change MCID of 8.20, 7.81, or 7.76, depending on the anchor used. Certainly, with such a dramatic improvement in pain, this is an overestimate of the MCID.
The MDC represents the upper value of the 95% confidence interval for the average change score calculated in the group of patients labeled nonresponders. In our cohort, when the MDC calculation technique was used, the MCID for the VAS was 4.49, 4.82, or 4.13, depending on the anchor used. Correspondingly, the change difference calculation technique, which represents the difference in pain improvement between responders and nonresponders, approximated MCID for the VAS as 5.74, 5.27, or 5.56, depending on the anchor used. This represents a reduction in pain of approximately 50% and results in substantially lower MCID values than those determined via the average change technique.
Both the MDC and the change difference calculation techniques yielded MCID values that were more representative of our patients' expectations for a successful result. When asked preoperatively what the minimum improvement in VAS would warrant proceeding with the RFL surgery, many of the patients in our cohort stated at least a 50% improvement in pain. Our results are consistent with assertions from Copay et al9 and Parker et al10 that a truly sound MCID value should correspond to patients' perceptions of the importance of the change. Because the MDC method provided a threshold value above the 95% confidence interval of the unimproved cohort (greater than the measurement error) and corresponded to patient perception of importance of change, it appeared to be the most appropriate calculation method for our cohort.
The AUC is a calculation that weighs the sensitivity and specificity of each of the anchors. In the present study, the anchor that consistently demonstrated the highest AUC value was the HTI anchor. Additionally, the AUC was greater for the BNI-PS compared with the VAS for all 3 anchors, indicating that BNI-PS is probably better suited for calculating MCID.
This MCID analysis has several limitations. The most important is that MCID calculations are probably influenced by baseline pain severity, and in our study, the baseline pain was very high in every patient. It is difficult to calculate MCID if not enough patients have poorer outcomes. Because few patients in this cohort reported no pain relief after surgery, the responders and nonresponders were not balanced. An additional limitation is in the selection of anchors, which have been variable in the literature, as well as our small sample size of 43 patients.
RFL-specific MCID is variable depending on calculation technique. Depending on the MDC calculation method with the HTI anchor, the MCID is 4.49 for VAS and 1.16 for BNI-PS after RFL for TN.
Mr Reddy acknowledges support from National Institutes of Health grant T35 HL090555 and Public Health Service Award T32 GM07347 from the National Institute of General Medical Studies for the Vanderbilt Medical Scientist Training Program. The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.
1. Chen HI, Lee JY. The measurement of pain in patients with trigeminal neuralgia. Clin Neurosurg. 2010;57:129–133.
2. Tang CT, Chang SD, Tseng KY, Liu MY, Ju DT. CyberKnife stereotactic radiosurgical rhizotomy for refractory trigeminal neuralgia. J Clin Neurosci. 2011;18(11):1449–1453.
3. Park SH, Hwang SK. Outcomes of Gamma Knife radiosurgery for trigeminal neuralgia after a minimum 3-year follow-up. J Clin Neurosci. 2011;18(5):645–648.
4. Rogers CL, Shetter AG, Fiedler JA, Smith KA, Han PP, Speiser BL. Gamma Knife radiosurgery for trigeminal neuralgia: the initial experience of the Barrow Neurological Institute. Int J Radiat Oncol Biol Phys. 2000;47(4):1013–1019.
5. Reddy VK, Parker SL, Patrawala SA, Lockney DT, Su P-F, Mericle RA. Microvascular decompression for classic trigeminal neuralgia: determination of minimum clinically important difference in pain improvement for patient reported outcomes. Neurosurgery. 2013;72(5):749–754.
6. Stratford PW, Binkley JM, Riddle DL, Guyatt GH. Sensitivity to change of the Roland-Morris Back Pain Questionnaire: part 1. Phys Ther. 1998;78(11):1186–1196.
7. Jaeschke R, Singer J, Guyatt GH. Measurement of health status: ascertaining the minimal clinically important difference. Control Clin Trials. 1989;10(4):407–415.
8. Wright AA, Cook CE, Baxter GD, Dockerty JD, Abbott JH. A comparison of 3 methodological approaches to defining major clinically important improvement of 4 performance measures in patients with hip osteoarthritis. J Orthop Sports Phys Ther. 2011;41(5):319–327.
9. 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. 2007;7(5):541–546.
10. Parker SL, Adogwa O, Paul AR, et al.. Utility of minimum clinically important difference in assessing pain, disability, and health state after transforaminal lumbar interbody fusion for degenerative lumbar spondylolisthesis. J Neurosurg Spine. 2011;14(5):598–604.
11. 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. 2010;10(6):469–474.
12. 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. 2008;8(6):968–974.
13. Headache Classification Subcommittee of the International Headache Society. The International Classification of Headache Disorders: 2nd edition. Cephalalgia. 2004;24(suppl 1):9–160.
14. Eller JL, Raslan AM, Burchiel KJ. Trigeminal neuralgia: definition and classification. Neurosurg Focus. 2005;18(5):E3.
Comparison of treatment success is one of the most challenging aspects of the surgical trigeminal neuralgia literature. The authors provide a novel quantitative approach to an age-old subjective problem. With this in place, more interesting questions about technique and patient selection can be addressed.
Brian Harris Kopell
New York, New York