Jacek A. Kopec

The Quebec Back Pain Disability Scale. Measurement properties.

Study Design. The Quebec Back Pain Disability Scale is a 20-item self-administered instrument designed to assess the level of functional disability in individuals with back pain. The scale was administered as part of a larger questionaire to a group of 242 back pain patients. Follow-up data were o btained after several days and after 2 to 6 months. Objectives. The goal of this study was to determine whether the Quebec scale is a reliable, vaild, and responsive measure of disability in back pain, and to compare it with other disability scales. Summary of Background Data. A number of functional disability scales for back pain are being used, but their conceptual validity is uncertain, unlike most published instruments,the Quebec scale was constructed using a conceptual approach to disability assessment and empirical methods of item development, analysis, and selection. Methods. The authors calculated test-retest and internal consistency coefficients, evaluated construct validity of the scale, and tested its responsiveness against a global index of change. Direct comparisons with the Roland, Oswestry, and SF-36 scales were carried out. Results. Test-retest reliability was 0.92, and Cronbachs alpha coefficient was 0.96. The scale correlated as expected with other measures of disability, pain, medical history, and utilization variables, work-related variables, and socio-demographic characteristics. Significant changes in disability over time, and differences in change scores between patients that were expected to differ in the direction of change, were found. Conclusions. The Quebec scale can be recommended as an outcome measure in clinical trials, and for monitoring the progress of individual patients participating in treatment or rehabilitation programs.

A Comparative Review of Four Preference-Weighted Measures of Health-related Quality of Life

Four generic, preference-weighted, health-related quality of life (HRQL) questionnaires (the Quality of Well-Being, the Health Utilities Index, the EQ-5D, and the SF-6D) are reviewed. Although all of these questionnaires are designed to measure the same concept, each uses a different model of health, a different method of deriving preferences, and a different scoring formula. Head-to-head comparisons of preference-weighted questionnaires are limited. By considering several hypothetical health states, we found that utility scores for equivalent states can vary substantially, depending on the measure used. Clinicians and researchers applying preference-weighted HRQL questionnaires should be aware of such differences and exercise caution when interpreting the results.

Functional disability scales for back pain.

Functional disability scales for back pain are reviewed for content and measurement properties, concentrating on five widely used questionnaires. Current methodological issues in functional assessment also are discussed.

The use of the Tobit model for analyzing measures of health status.

Self-reported health status is often measured using psychometric or utility indices that provide a score intended to summarize an individuals health. Measurements of health status can be subject to a ceiling effect. Frequently, researchers want to examine relationships between determinants of health and measures of health status. Regression methods that ignore the presence of a ceiling effect, or of censoring in the health status measurements can produce biased coefficient estimates. The Tobit regression model is a frequently used tool for modeling censored variables in econometrics research. The authors carried out a Monte-Carlo simulation study to contrast the performance of the Tobit model for censored data with that of ordinary least squares (OLS) regression. It was demonstrated that in the presence of a ceiling effect, if the conditional distribution of the measure of health status had uniform variance, then the coefficient estimates from the Tobit model have superior performance compared with estimates from OLS regression. However, if the conditional distribution had non-uniform variance, then the Tobit model performed at least as poorly as the OLS model.

Predictors of back pain in a general population cohort.

Study Design. The study used longitudinal data from the first and second cycles (1994–1995 and 1996–1997) of the Canadian National Population Health Survey. Objective. Our objective was to derive prediction models for back pain in the general male and female household populations. Summary of Background Data. Little is known about the predictors of back pain in the general population. Most previous studies focused on specific occupational groups and used a cross-sectional or case-control design. Methods. The study cohort consisted of all respondents aged 18+ years who reported no back problems in the 1994–1995 National Population Health Survey cycle (N = 11,063). Potential predictors of chronic back pain were classified into nine groups and entered into stepwise logistic regression models. Bootstrap methods were used to derive the final models and assess their predictive power. Results. The overall incidence of back pain was 44.7 per 1,000 person-years and was higher in women (47.0 per 1,000 person-years) compared with men (42.2 per 1,000 person-years). In men, significant predictors of back pain were age (peak effect in 45–64 years), height, self-rated health, usual pattern of activity (especially heavy work), yard work or gardening (negative association), and general chronic stress. In women, significant factors were self-reported restrictions in activity, being diagnosed with arthritis, personal stress, and history of psychological trauma in childhood or adolescence. Conclusions. Overall health and psychosocial factors are important predictors of back pain in both men and women. Other risk factors differ between the two sexes.

Measuring functional outcomes in persons with back pain: A review of back-specific questionnaires

It is now widely recognized that patients’ perspectives are essential in making medical decisions and judging the results of treatment. Patient-based outcome measures are usually classified as generic or disease specific. Both types of instruments have been used in back pain research during the past two decades. Generic measures include global ratings of health status, as well as multidimensional measures of health-related quality of life (HRQOL), such as the Sickness Impact Profile (SIP), SF-36 Health Survey, Nottingham Health Profile (NHP), or shorter instruments. Generic measures are applicable to patients across different types of conditions. The hallmark of a disease-specific measure is the attribution of symptoms and functional limitations to a specific disease or condition. Unlike items in a generic measure, items in a disease-specific measure assess only those aspects of health that tend to be affected by the disease. The goal is to achieve high relevance and responsiveness of the scales without undue burden to the patient. This review is limited to validated functional questionnaires that were designed by selecting items relevant to back pain. Generic measures of HRQOL and pain measures are discussed elsewhere in this issue of Spine. It should be noted that some general pain and disability instruments have been tested in patients with low back pain. Furthermore, a large number of other questionnaires and rating schemes have been used to evaluate outcomes in patients with back pain. However, most of these questionnaires have been developed ad hoc for a single study, their psychometric properties have not been reported, and comparisons with other scales have not been made.

A comparison of four indirect methods of assessing utility values in rheumatoid arthritis.

Objectives:Utility scores can be assessed indirectly using preference-based instruments and used as weightings for quality-adjusted life years in economic analyses. It is not clear whether available instruments yield similar results or what domains of health are contributing to the overall score in a sample of patients with rheumatoid arthritis (RA). Subjects:Our study included 313 individuals with rheumatologist-confirmed RA. Measures:A self-completed survey that permitted scoring of 4 indirect utility instruments (the Health Utilities Index Mark 2 and 3 (HUI-2 and HUI-3), the EuroQoL (EQ-5D), and the Short Form 6D (SF-6D) was the basis of our study. Results:Mean (standard deviation) global utility scores were 0.63 (0.24) for the SF-6D, 0.66 (0.13) for the EQ-5D, 0.71 (0.19) for the HUI-2, and 0.53 (0.29) for the HUI-3 (P = 0.02 by repeated-measures analysis of variance). The intraclass correlation across all the indices was 0.67 (95% confidence interval 0.62–0.71). Bland-Altman plots revealed that agreement among instruments was poor at lower utility values. In this elderly RA sample, all of the global utilities mostly measured functional ability and pain. Conclusions:There are significant differences in utilities obtained from different indirect methods. Agreement among the instruments was moderate but poorer at lower utilities. It is unlikely that these utility values, if used as the weightings for quality-adjusted life years, would result in comparable estimates.

Bias in case-control studies. A review.

consistent terminology has not evolved. Different authors have used different terms when referring to essentially the same type of bias. For example, Kleinbaum et al1 and Rothman2 classify biases into selection bias, information bias and confounding. Sackett offered a more detailed classification, with five major types of sampling bias (eg, admission rate bias, membership bias) and four types of measurement bias (eg, diagnostic suspicion bias, recall bias).3 Feinstein distinguishes, inter alia, among susceptibility bias, performance bias, transfer bias and detection bias.4 Lasts Dictionary of epidemiology5 gives definitions of 26 biases but fails to mention many of the terms proposed by other authors. It is not our objective to provide a comprehensive list of biases, or to propose a new, improved taxonomy. Our purpose is to review the most common threats to validity in the design of case-control studies?commonly regarded as being particularly susceptible to bias?and to discuss some useful strategies in dealing with these problems. Certain issues will not be addressed, notably methods of controlling confounding, and the statistical analysis of case control studies.

Patient-Reported Outcomes in Sentinel Node–Negative Adjuvant Breast Cancer Patients Receiving Sentinel-Node Biopsy or Axillary Dissection: National Surgical Adjuvant Breast and Bowel Project Phase III Protocol B-32

PURPOSE Sentinel lymph node resection (SNR) may reduce morbidity while providing the same clinical utility as conventional axillary dissection (AD). National Surgical Adjuvant Breast and Bowel Project (NSABP) B-32 is a randomized phase III trial comparing SNR immediately followed by AD (SNAD) to SNR and subsequent AD if SN is positive. We report the definitive patient-reported outcomes (PRO) comparisons. PATIENTS AND METHODS Eligible patients had clinically node-negative, operable invasive breast cancer. The PRO substudy included all SN-negative participants enrolled May 2001 to February 2004 at community institutions in the United States (n = 749; 78% age > or = 50; 87% clinical tumor size < or = 2.0 cm; 84% lumpectomy; 87% white). They completed questionnaires presurgery, 1 and 2 to 3 weeks postoperatively, and every 6 months through year 3. Arm symptoms, arm use avoidance, activity limitations, and quality of life (QOL) were compared with intent-to-treat two-sample t-tests and repeated measures analyses. RESULTS Arm symptoms were significantly more bothersome for SNAD compared with SNR patients at 6 months (mean, 4.8 v 3.0; P < .001) and at 12 months (3.6 v 2.5; P = .006). Longitudinally, SNAD patients were more likely to experience ipsilateral arm and breast symptoms, restricted work and social activity, and impaired QOL (P < or = .002 all items). From 12 to 36 months, fewer than 15% of either SNAD or SNR patients reported moderate or greater severity of any given symptom or activity limitation. CONCLUSION Arm morbidity was greater with SNAD than with SNR. Despite considerable fears about complications from AD for breast cancer, this study demonstrates that initial problems with either surgery resolve over time.

Association of Biomarkers With Pre-Radiographically Defined and Radiographically Defined Knee Osteoarthritis in a Population-Based Study

OBJECTIVE To evaluate 10 biomarkers in magnetic resonance imaging (MRI)-determined, pre-radiographically defined osteoarthritis (pre-ROA) and radiographically defined OA (ROA) in a population-based cohort of subjects with symptomatic knee pain. METHODS Two hundred one white subjects with knee pain, ages 40-79 years, were classified into OA subgroups according to MRI-based cartilage (MRC) scores (range 0-4) and Kellgren/Lawrence (K/L) grades of radiographic severity (range 0-4): no OA (MRC score 0, K/L grade<2), pre-ROA (MRC score>or=1, K/L grade<2), or ROA (MRC score>or=1, K/L grade>or=2). Urine and serum samples were assessed for levels of the following biomarkers: urinary biomarkers C-telopeptide of type II collagen (uCTX-II), type II and types I and II collagen cleavage neoepitopes (uC2C and uC1,2C, respectively), and N-telopeptide of type I collagen, and serum biomarkers sC1,2C, sC2C, C-propeptide of type II procollagen (sCPII), chondroitin sulfate 846 epitope, cartilage oligomeric matrix protein, and hyaluronic acid. Multicategory logistic regression was performed to evaluate the association of OA subgroup with individual biomarker levels and biomarker ratios, adjusted for age, sex, and body mass index. RESULTS The risk of ROA versus no OA increased with increasing levels of uCTX-II (odds ratio [OR] 3.12, 95% confidence interval [95% CI] 1.35-7.21), uC2C (OR 2.13, 95% CI 1.04-4.37), and uC1,2C (OR 2.07, 95% CI 1.06-4.04), and was reduced in association with high levels of sCPII (OR 0.53, 95% CI 0.30-0.94). The risk of pre-ROA versus no OA increased with increasing levels of uC2C (OR 2.06, 95% CI 1.05-4.01) and uC1,2C (OR 2.06, 95% CI 1.12-3.77). The ratios of type II collagen degradation markers to collagen synthesis markers were better than individual biomarkers at differentiating the OA subgroups, e.g., the ratio of [uCTX-II][uC1,2C] to sCPII was associated with a risk of ROA versus no OA of 3.47 (95% CI 1.34-9.03) and a risk of pre-ROA versus no OA of 2.56 (95% CI 1.03-6.40). CONCLUSION Different cartilage degradation markers are associated with pre-ROA than are associated with ROA, indicating that their use as diagnostic markers depends on the stage of OA. Biomarker ratios contrasting cartilage degradation with cartilage synthesis are better able to differentiate OA stages compared with levels of the individual markers.