An Assessment of a Short Composite Questionnaire Designed for Use in an Interventional Spine Pain Management Setting

Transcription

1 Original Research An Assessment of a Short Composite Questionnaire Designed for Use in an Interventional Spine Pain Management Setting Robert Burnham, MSc, MD, FRCPC, Gordon Stanford, DC, Lori Gray, PhD Objective: To evaluate the reliability, validity, responsiveness, and practicality of a short composite questionnaire designed for use in an interventional spine pain management setting. Setting: A rural interventional spine pain management practice. Patients: Sixty subjects undergoing spinal intervention procedures (injection or neurotomy). Methods: The Pain Disability Quality of Life Questionnaire-Spine (PDQQ-S) assesses spine pain quality, related disability, and life satisfaction/quality with the use of 2 questions for each domain. To assess the PDQQ-S, patients completed 4 questionnaires (McGill Pain Questionnaire [MPQ], Oswestry Disability Questionnaire [ODQ], Assessment of Quality of Life [AQoL], and PDQQ-S). Patients completed the questionnaires twice before their procedure and once after their procedure. Main Outcome Measures: Test-retest reliability was assessed by calculating the Pearson correlation coefficient for the 2 preprocedure questionnaire completions. Validity was assessed by calculating the Pearson correlation of the pain quality, disability, and life quality/satisfaction question scores of the PDQQ-S with the MPQ, ODQ, and AQoL, respectively. Questionnaire responsiveness was explored by calculating responsiveness ratio scores. Practicality was assessed by recording the time required to complete each questionnaire and the ease with which it was completed. Results: Test-retest reliability correlation scores were as follows: PDQQ-S 0.73; MPQ 0.88; ODQ 0.89; and AQoL Acceptable correlations existed between the pain, disability, and life quality/satisfaction domains of the PDQQ-S and the MPQ (.50), ODQ (.71), and AQoL (.43), respectively. Responsiveness ratio scores for the PDQQ-S, MPQ, ODQ, and AQoL were 4.4, 1.9, 2.3, and 0.7, respectively. Compared with the other questionnaires, the PDQQ-S required approximately one fifth the time to complete and was significantly easier to understand. Conclusions: The PDQQ-S demonstrates adequate reliability and validity and superior responsiveness and practicality in persons with low back pain undergoing interventional spine pain management procedures. PM R 2012;4: INTRODUCTION Self-report questionnaires commonly are used to assess the severity of spine pain and associated disability [1]. Questionnaires are available for each of these domains; however, because of the length and complexity of some questionnaires, the burden on the participant and administrator makes them impractical to use, particularly in a busy clinical setting. Composite questionnaires that are used to measure several pain-related domains simultaneously are less common [2,3]. An instrument that is easy for the patient to understand, can be administered quickly, provides reliable and valid composite information about the patient s pain syndrome, and is sensitive to treatment-related changes would be useful, particularly in an interventional pain management practice setting. Recently, a 4-item questionnaire called the Pain Disability Questionnaire-Spine (PDQ-S), which is used to measure the domains of pain quality and disability, was specifically designed for that purpose. Via psychometric assessment, researchers found that it is an efficient, user-friendly tool with adequate reliability and validity and excellent responsive- R.B. Central Alberta Pain & Rehabilitation Institute, Lacombe, and Division of Physical Medicine & Rehabilitation, University of Alberta, Edmonton, Alberta, Canada. Address correspondence to: R.B., 1, 6220 Highway 2A, Lacombe, Alberta, Canada T4L-2G5; e- mail: rburnham@telusplanet.net Disclosure: nothing to disclose G.S. Central Alberta Pain & Rehabilitation Institute, Lacombe, Alberta, Canada Disclosure: nothing to disclose L.G. Canadian University College, Lacombe, Alberta, Canada Disclosure: nothing to disclose Peer reviewers and all others who control content have no relevant financial relationships to disclose Disclosure Key can be found on the Table of Contents and at Submitted for publication July 22, 2011; accepted March 24, PM&R 2012 by the American Academy of Physical Medicine and Rehabilitation /12/$36.00 Vol. 4, , June 2012 Printed in U.S.A

2 414 Burnham et al SHORT QUESTIONNAIRE FOR SPINAL PAIN INTERVENTION ness to intervention in subjects with low back pain [4]. However, the important domain of pain-related disruption of life quality/satisfaction was not assessed with the PDQ-S. To address this deficiency, we incorporated 2 additional life quality/satisfaction questions to the PDQ-S, making it the Pain Disability Quality of Life Questionnaire Spine (PDQQ- S). The purpose of this study was to evaluate the reliability, validity, responsiveness, and practicality of the PDQQ-S. METHODS The PDQQ-S was developed by the staff of the Central Alberta Pain and Rehabilitation Institute after a review of the literature and team discussion regarding the measurement of chronic spine pain and its associated effects. Team members representing the disciplines of medicine, physical therapy, nursing, and psychology selected, adapted, and developed questions they perceived as being particularly relevant yet parsimonious in the assessment of chronic pain in patients undergoing therapeutic, minimally invasive spine interventions. The patient s score for each question was to reflect their average experience of the previous week relative to their index pain. Each question was designed to be answered with the use of a 10-cm Visual Analogue Scale (VAS) estimate of severity. A composite score then could be calculated by summing the VAS scores of the individual questions. The PDQQ-S consists of 2 questions relating to the quality of the patient s pain (ie, pain severity and frequency), 2 questions relating to disability (each patient selected 2 everyday life activities that were limited as a result of their spine pain, were of particular importance to them, and that they would like improved or restored), and 2 questions exploring life quality/satisfaction. The PDQQ-S is shown in Figure 1. The question relating to pain intensity is in common use in various established questionnaires [5]. The questions relating to disability were adapted from the Patient-Specific Functional Scale [6]. The question relating to the level of satisfaction in living with the current degree of pain was adapted from item 3 of the proposed Core Outcome Measure Questionnaire recommended for clinical care [7]. The questions relating to pain frequency and effect on quality of life were original. The index pain on the diagram of the PDQQ-S was defined by the physician after the history was obtained and the physical examination was conducted. The interventional suite nurse confirmed the index pain with patients and helped them define their 2 relevant disabilities. Subjects were recruited from patients attending the Central Alberta Pain and Rehabilitation Institute, a rural multidisciplinary chronic pain management clinic. When it was determined that a lumbopelvic spinal intervention procedure (ie, a facet, sacroiliac joint, or epidural corticosteroid injection, a radiofrequency neurotomy of the facet or sacroiliac joints, or a TransDiscal Biacuplasty [Baylis Medical, Montreal, Quebec, Canada]) was clinically indicated, patients were invited to participate in the study. All patients accepted the invitation, and each signed an informed consent form before they participated. The consent form indicated that the usefulness of several questionnaires was being assessed but did not identify the PDQQ-S as the target questionnaire. Subjects were not remunerated for their participation. All procedures were performed by the same practitioner (R.B.). Four questionnaires were used in the study: the PDQQ-S, the adjective checklist portion of the short-form McGill Pain Questionnaire (MPQ) [8], the Oswestry Disability Questionnaire (ODQ) [9], and the Assessment of Quality of Life Questionnaire (AQoL) [10]. The latter 3 comparison questionnaires were chosen because they have established reliability and validity and are in common use [3,11]. Questionnaires were completed the day of the procedure and at a designated time after the procedure (ie, 2 weeks after a corticosteroid injection and 6-8 weeks after radiofrequency neurotomy or TransDiscal biacuplasty). A limited subset of subjects also completed the questionnaire 1 month before they underwent their procedure. Specifically, 13 subjects completed the 1-month preprocedure PDQQ-S and AQoL, and 22 completed the MPQ and ODQ. Reliability (test-retest stability) was assessed by calculating repeated measures analysis of variance (ANOVA) and the Pearson correlation coefficients for the 2 pretreatment administrations of each questionnaire. For the PDQQ-S, the same analyses also were performed for each domain (ie, pain, disability, and life quality/satisfaction) of the questionnaire. A reliability coefficient of 0.7 was considered evidence of adequate test-retest reliability [12]. Construct validity was explored by assessing the correlation between the relevant questions from the PDQQ-S and the other established questionnaires measuring a similar construct. Specifically, the summed scores of the 2 questions relating to the subjects pain quality (pain intensity and frequency) were correlated with the MPQ. The summed scores of the 2 questions relating to disability were correlated with the ODQ, and the summed scores of the 2 questions relating to life quality/satisfaction were correlated with the AQoL. Evidence of convergent validity exists when the correlations fall in the mid range ( ) [13,14]. Responsiveness that is, the ability of an instrument to detect clinically important changes over time was explored for each questionnaire and for each domain of the PDQQ-S by calculating a responsiveness ratio (RR). The RR evaluates intervention-related change over time while considering the between-subject variability in within-subject changes in stable subjects according to the following formula: mean change in the pre post-treatment score divided by the square root of twice the mean square error of the 2 pretreatment measurements [15].To be of significance, RR values should be 1.96 [12]. Practicality was assessed by determining how long it took to complete each questionnaire (in minutes) and the ease with which the subjects understood each questionnaire (a VAS score for the statement understanding and completing this form was: 0 extremely easy; 10 extremely hard).

3 PM&R Vol. 4, Iss. 6, Figure 1. The Pain Disability Quality of Life Questionnaire-Spine instrument.

4 416 Burnham et al SHORT QUESTIONNAIRE FOR SPINAL PAIN INTERVENTION RESULTS Sixty subjects participated in the study. Sixty-seven percent were women, and the average age was 60 (12.4) years. Of the 60 subjects, 12 underwent corticosteroid injection (intra-articular facet/sacroiliac joint or epidural), 38 underwent a lumbar facet joint radiofrequency neurotomy, 1 underwent a sacroiliac joint radiofrequency neurotomy, and 9 underwent TransDiscal Biacuplasty. To assess reliability, repeated-measures ANOVA were performed for each questionnaire, comparing the scores for 1 month before and the day of treatment. No significant differences were found (PDQQ-S, F.01; MPQ, F.2; ODQ, F 2.0; AQoL, F 1.1). The test-retest correlations of each questionnaire exceeded 0.70, suggesting acceptable reliability (Table 1). Repeated-measures ANOVA of each of the pain, disability, and life quality/satisfaction domains of the PDQQ-S showed no significant test-retest differences (pain, F.3; disability, F 3.1; life quality/satisfaction, F 1.7). Test-retest correlations for the pain, disability, and life quality/satisfaction domain questions of the PDQQ-S also were.70 (Table 2). Correlations between the pain domain questions of the PDQQ-S and the MPQ, the disability domain questions of the PDQQ-S and the ODQ, and the life quality/satisfaction domain questions of the PDQQ-S and the AQoL ranged between 0.4 and 0.8, suggesting acceptable convergent validity (Table 2). RR values for each of the 4 questionnaires are summarized in Table 1, and RR values for the pain, disability, and life quality/satisfaction domains of the PDQQ-S are summarized Table 1. A comparison of the reliability, responsiveness, and practicality of 4 questionnaires Domain Analysis Reliability, r (95% CI) PDQQ-S 0.73 ( ) MPQ 0.88 ( ) ODQ 0.89 ( ) AQoL 0.83 ( ) Responsiveness, RR PDQQ-S 4.4 MPQ 1.9 ODQ 2.3 AQoL 0.7 Practicality: time to complete questionnaire, min, mean (SD) PDQQ-S 0.9 (0.9) MPQ 5.7 (7.1) ODQ 6.2 (4.3) AQoL 6.7 (4.3) Practicality: questionnaire understandability, VAS, mean (SD) PDQQ-S 0.3 (0.8) MPQ 0.8 (2.1) ODQ 1.2 (1.9) AQoL 2.0 (2.2) CI confidence interval; PDQQ-S Pain Disability Quality of Life Questionnaire Spine; MPQ McGill Pain Questionnaire; ODQ Oswestry Disability Questionnaire; AQoL Assessment of Quality of Life questionnaire; RR responsiveness ratio; SD standard deviation; VAS Visual Analogue Scale. Table 2. Reliability, validity, responsiveness, and practicality of the PDQQ-S Domain Analysis Reliability, r (95% CI) Pain questions.72 ( ) Disability questions.80 ( ) Quality of life questions.76 ( ) Composite questionnaire.73 ( ) Validity, r (95% CI) Pain questions: MPQ.50 ( ) Disability questions: ODQ.71 ( ) Quality of life questions: AQOL 0.43 (.21 to.61) Responsiveness, RR Pain questions 2.9 Disability questions 3.6 Quality of life questions 5.1 Composite questionnaire 4.4 Practicality, mean (SD) Time to complete, min.9 (.9) Questionnaire understandability.3 (.8) PDQQ-S Pain Disability Quality of Life Questionnaire Spine; CI confidence interval; MPQ McGill Pain Questionnaire; ODQ Oswestry Disability Questionnaire; AQoL Assessment of Quality of Life Questionnaire; RR responsiveness ratio; SD standard deviation. in Table 2. The life quality/satisfaction domains of the PDQQ-S are negatively correlated with the AQoL because they are measured on opposite scales. Specifically, a high score on the PDQQ-S reflects a high level of quality of life disruption, whereas a high score on the AQoL reflects a high quality of life. Only the PDQQ-S and ODQ were significant, and the RR of the PDQQ-S was superior. Each domain of the PDQQ-S showed acceptable responsiveness. In terms of practicality, ANOVA confirmed that the PDQQ-S could be completed significantly more quickly (F 12.1; P.0001) and was easier to understand (F 3.3; P.04) than the other questionnaires (Table 2). DISCUSSION The primary purpose of this study was to evaluate the reliability, validity, responsiveness, and practicality of the PDQQ-S. Its suitability for use in an interventional spine pain management setting was of particular interest. The PDQQ-S demonstrated acceptable reliability and construct validity and superior responsiveness and practicality compared with established questionnaires measuring similar constructs. The PDQQ-S evaluates pain quality, pain-related disability, and life quality/satisfaction. These domains constitute 3 of the 5 cardinal constructs of the minimum outcome core set recommended for the evaluation of patients with low back pain [2,7]. The other 2 recommended constructs include generic health status and work disability, neither of which are likely to change quickly in response to an interventional spinal injection or neurotomy. Compressing, filtering, and distilling already validated questionnaires risks compromising the validity and usefulness of the subsequent questionnaire. The development of the PDQQ-S involved constructing a questionnaire that was

5 PM&R Vol. 4, Iss. 6, useful in a setting in which a very brief pain assessment instrument was desired but also allowed the evaluation of multiple pain-related domains. Thus the PDQQ-S was not developed to be an in-depth general questionnaire that comprehensively assesses the 3 domains it examines. For example, the PDQQ-S was not intended for use with many different clinical pain syndromes, nor was it intended to be used to aid in the differential diagnosis among pain syndromes. As a result, many of the items on questionnaires that have been designed to have a discriminative capacity, such as the descriptor items on the short-form of the MPQ, were deemed unnecessary. However, because these items were not included, the PDQQ-S measures pain intensity and frequency but does not measure pain affect and quality. In examining the domain of disability, the reasoning used in item selection was as follows. Both instruments recommended by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials, the Roland-Morris Disability Questionnaire, and the ODQ for measuring clinically important differences in physical functioning in patients with low back pain [16] have been criticized regarding comprehensiveness in terms of how well they represent the wide range of activities that can be affected by lower back pain [17]. Given that the ODQ has only 10 items, it is quite likely that areas of functioning important to the patient are not assessed by the questionnaire. To address this issue and yet keep the PDQQ-S brief enough for use in its intended setting, it was decided to have the patients identify, with the aid of the interventional suite nurse, the 2 activities they believed were most relevant in terms of the functional impact of their index pain. As for life quality/satisfaction, similar to Dworkin et al [16], who recommended that single-item global ratings of improvement or treatment satisfaction be used as core outcome measures, it was decided that it was important to afford patients the opportunity to provide an overall integrated assessment of the effect of the intervention procedure on both their satisfaction with and quality of life. However, it is important to note that because the items used in the PDQQ-S give a compressed assessment of pain, disability, and quality of life/satisfaction, it is not appropriate for this instrument to be used in situations in which one wants an in-depth assessment of these domains. Should a comprehensive evaluation of these domains be desired, it would be more appropriate to use one of the more established instruments currently available that have an accumulation of evidence attesting to their reliability, validity, and usefulness in regard to this use. Two existing pain-related composite questionnaires also may be suitable in an interventional spine pain management practice context. The Bournemouth and the modified Von Korff questionnaires both possess qualities of reliability, validity, and practicality [18,19]. However, differences exist in the time frame of the assessment, the number and representation of the domains being assessed, the scale of assessment, and their psychometric characteristics. Like the Bournemouth questionnaire, we chose to measure the patient s pain-related experience during the previous week. The modified Von Korff questionnaire measures the pain-related experience during the previous month. The original Von Korff questionnaire measures the pain-related experience during previous 6 months [20]. Because the duration of therapeutic effect from local anesthetic/corticosteroid injection intervention can be short, the 1-week assessment is more relevant in an interventional spine care setting. The modified Von Korff questionnaire measures the domains of disability and pain (3 questions for each domain), whereas the Bournemouth questionnaire assesses 3 domains (pain, disability, and emotional impact). However, these 3 domains are not equally represented in the questionnaire. Two questions relate to pain, 2 relate to emotional impact, and 3 relate to disability. The PDQQ-S measures 3 domains (pain, disability, and life quality/satisfaction), each with equal representation. The effect of intervention on the construct of life quality/satisfaction may be more rapid than on the emotional impact constructs of depression and anxiety included in the modified Von Korff Questionnaire. The Bournemouth and modified Von Korff questionnaires both use an 11-point numerical rating scale, whereas the PDQQ-S uses a 10-cm long VAS. We acknowledge that the VAS adds an administrative burden of having to measure the subject s responses with a ruler. However, given the short duration between the administration of the pretreatment and posttreatment questionnaires (weeks), it is possible that patients could remember their pretreatment numerical rating score, which may bias their post-treatment score. This recall may be less likely to happen when using the VAS. Neither the Bournemouth nor the modified Von Korff questionnaire has been assessed in terms of responsiveness. Good responsiveness is a key psychometric characteristic of a suitable instrument for interventional pain management when the effect of the intervention is the outcome of interest. The PDQQ-S was shown in our study to have superior responsiveness to the effects of interventional procedures in comparison with the other questionnaires we evaluated. Future research comparing the psychometric properties of these 3 questionnaires in an interventional spine pain management setting is encouraged. Our study has limitations. First, the number of subjects is relatively low, particularly for the reliability assessment. Of the 4 questionnaires, the reliability correlation was weakest for the PDQQ-S. Nonetheless, a test-retest reliability score of 0.7 is generally considered acceptable, and all questionnaires met this requirement. That finding is notable because 2 factors would tend to weaken the PDQQ-S correlation and confidence interval. Specifically, only 13 subjects completed the PDQQ-S both 1 month before and the day of the procedure, and the PDQQ-S includes only 6 items compared with 10 to 15 in the other questionnaires. A second limitation relates to generalizability. The PDQQ-S was developed specifically for use in an interventional spine pain management setting. This study has demonstrated favorable psychometric characteristics in patients undergoing lumbar spine and sacroiliac joint procedures. Its usefulness in patients undergoing

6 418 Burnham et al SHORT QUESTIONNAIRE FOR SPINAL PAIN INTERVENTION thoracic or cervical spine procedures is yet to be established. Finally, the complete administrative burden of the PDQQ-S was underestimated in the study. Specifically, the time required to explain the concept of index pain was not recorded, nor was the time required to measure the 6 VAS scores recorded. We assert, however, that the concept of index pain is relevant to the administration of any questionnaire measuring the impact of localized pain and that the scoring time of the 6 VAS questions is negligible. A third limitation is that, because of its brevity, the PDQQ-S is very restricted in terms of the domain of content that it examines in regard to patient outcome after the lumbopelvic spinal interventions used. As a result, aspects of patients pain, disability, and quality of life/satisfaction that have been identified in the literature as being important outcome variables that can be examined by using more comprehensive established questionnaires were not measured, nor were changes in these areas resulting from the interventions that were used assessed. The PDQQ-S therefore may not be as sensitive in detecting improvements after treatment as a test battery comprising 3 questionnaires that were developed specifically to comprehensively measure 1 of these 3 areas. In conclusion, this study suggests that the PDQQ-S is an efficient, user-friendly composite questionnaire of pain, disability, and life quality/satisfaction that demonstrates adequate reliability and validity and superior responsiveness to interventionrelated change in persons with low back pain. It is well suited for interventional spine pain management practice. ACKNOWLEDGMENTS We acknowledge Greg Turgeon and Dr. Michael Pappachan for their assistance with data collection. REFERENCES 1. Bombardier C. Outcome assessments in the evaluation of treatment of spinal disorders: Summary and general recommendations. Spine 2000; 25: Ferrer M, Pellise F, Escudero O, et al. Validation of a minimum outcome core set in the evaluation of patients with back pain. Spine 2006;31: Rocchi M, Sisti D, Benedetti P, et al. Critical comparison of nine different self-administered questionnaires for the evaluation of disability caused by low back pain. Eura Medicophys 2005;41: Burnham R, Stanford G, Gray L. An assessment of the reliability, validity, responsiveness and practicality of a short composite pain questionnaire (abstract). Pain Med 2009;10: Jensen M, Karoly P. Self-report scales and procedures for assessing pain in adults. In: Handbook of Pain Assessment. New York, NY: The Guilford Press; 2011, pages Stratford P, Gill C, Westaway M, Binkley J. Assessing disability and change of individual patients: A report of a patient specific measure. Physiother Can 1995;47: Deyo R, Battie M, Beurskens A, et al. Outcome measures for low back pain research: A proposal for standardized use. Spine 1998;23: Melzack R. The short-form McGill Pain Questionnaire. Pain 1987;30: Fairbank J, Davies J, Couper J, O Brien J. The Oswestry low back pain disability questionnaire. Physiotherapy 1980;66: Hawthorne G, Richardson J, Osborne R. The Assessment of Quality of Life (AQoL) instrument: A psychometric measure of health-related quality of life. Qual Life Res 1999;8: Burckhardt C, Jones K. Adult measures of pain: The McGill Pain Questionnaire (MPQ), Rheumatoid Arthritis Pain Scale (RAPS), Short-Form McGill Pain Questionnaire (SF-MPQ), Verbal Descriptive Scale (VDS), Visual Analog Scale (VAS), and West Haven-Yale Multidisciplinary Pain Inventory (WHYMPI). Arthritis Rheum 2003;49:S96-S Terwee C, Bot S, de Boer M, et al. Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol 2007;60: Stinson J, Kavanagh T, Yamada J, et al. Systematic review of the psychometric properties, interpretability and feasibility of self-report pain intensity measures for use in clinical trials in children and adolescents. Pain 2006;125: Streiner D, Norman R. Health Measurement Scales [electronic resource]: A Practical Guide to Their Development and Use. New York, NY: Oxford University Press; 2008, Guyatt G, Walter S, Norman G. Measuring change over time: Assessing the usefulness of evaluative instruments. J Chron Dis 1987;40: Dworkin R, Turk D, Wyrwich K, et al. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMPAACT recommendations. J Pain 2008;9: Watson PJ. Clinical Assessment of low back pain. In: Turk DC, Melzack R, eds. Handbook of Pain Assessment. New York, NY: Guilford Press; 2011, pages Bolton J, Breen A. The Bournemouth Questionnaire: A short-form comprehensive outcome measure. I. Psychometric properties in back pain patients. J Manipul Physiol Ther 1999;22: Underwood J, Barnett A, Vickers M. Evaluation of two time-specific back pain outcome measures. Spine 1999;24: Von Korff M, Ormel J, Keefe F, Dworkin S. Grading the severity of chronic pain. Pain 1992;50: This CME activity is designated for 1.0 AMA PRA Category 1 Credit and can be completed online at me.aapmr.org. Log on to go to Lifelong Learning (CME) and select Journal-based CME from the drop down menu. This activity is FREE to AAPM&R members and $25 for non-members. CME Question The Pain Disability Quality of Life Questionnaire-Spine (PDQQ-S) is not suggested to be an efficient, user-friendly composite questionnaire for intervention-related change of: a. pain intensity and frequency b. pain-related disability c. satisfaction with and quality of life d. generic health status Answer online at me.aapmr.org