The utility of the STarT back screening tool in a population with chronic low back pain: A prospective study

Abstract

Study design: Cross sectional (Study 1) and prospective (Study 2). Background: Chronic low back pain (LBP) is problematic with significant personal, social, and economic impact. The need to screen for indicators of poor prognosis and/or stratify patients with LBP based on risk is highlighted in the literature. The STarT Back Screening Tool (SBST) is one tool that has been developed to meet this need. There is a growing body of evidence for the psychometric properties of the SBST and for the tool’s predictive and discriminative ability, particularly in populations with LBP of variable episode duration. Studies have shown that the SBST’s performance has differed depending on the population in which it was evaluated. Clinical setting, cultural context, and LBP episode duration have all influenced the tool’s performance. There are a lack of studies evaluating the utility of the SBST in a population exclusively with chronic LBP. While the SBST is commonly used in Australia, the tool is yet to be evaluated in an Australian context. Aims: This study aimed to determine the utility of the SBST in a population exclusively with chronic non-specific LBP in Australia. The aim of the cross sectional study was to profile the SBST risk subgroups with respect to pain, disability, and psychological measures. The aim of the prospective study was to evaluate the predictive and discriminative ability of the SBST for pain intensity, self-reported LBP related disability, and global self-perceived change at one year follow-up.Methods: Participants with dominant axial non-specific LBP of at least three months duration were recruited (n=290). At baseline participants completed the SBST and data were collected on demographic information, pain intensity (11-point numerical rating scale [NRS]), disability (Roland Morris Disability Questionnaire [RMDQ]), depression/anxiety/stress (Depression Anxiety Depression Scale), fear avoidance beliefs (Fear Avoidance Beliefs Questionnaire [FABQ]), catastrophising (Pain Catastrophising Scale), perceived risk of pain persistence (11-point NRS), self-efficacy (Pain Self-efficacy Questionnaire), and chronic pain acceptance 5 (Chronic Pain Acceptance Questionnaire). Baseline SBST risk subgroup differences were examined using a one-way analysis of variance (ANOVA), Kruskal-Wallis test, or Chi Squared test depending on the data type and distribution. Follow-up data were collected after one year (n=264). The follow-up measures were pain intensity (NRS), disability (RMDQ), and global perceived change (7-point global rating of change scale). The follow-up measures were dichotomised into recovered and not recovered for analysis. The proportion of participants who were not recovered with respect to each of the follow-up measure was calculated at a cohort level and by baseline SBST risk subgroup. The predictive ability of the SBST was evaluated with risk ratios (RR) using the low risk group as the reference category. In order to evaluate the discriminative ability of the SBST, receiver operating characteristic (ROC) curves were constructed and the area under the curve (AUC) was calculated for both the SBST total score and psychological subscale score for each of the follow-up measures.Likelihood ratios, sensitivity, specificity, and the diagnostic odds ratio (DOR) for the low risk group versus the medium/high risk group and the low/medium risk group versus the high risk group were calculated for the significant AUC values. Statistical significance was considered to be at the p<0.05 for all analyses. Results: At baseline, the SBST categorised 82 participants (28.3%) as low risk, 116 (40.0%) as medium risk, and 92 (31.7%) as high risk. There were significant differences between the SBST risk subgroups for pain intensity (p≤0.001), disability (p≤0.001), and for the scores on the psychological questionnaires (p≤0.001 except FABQ work subscale p=0.010). The SBST risk subgroups demonstrated increasingly higher levels of pain, disability, and negative psychological affect and cognitions as subgroup categorisation increased from low to medium to high risk. Non-recovery rates at the one year follow-up were 76.1% (n=201) for pain, 31.4% (n=83) for disability, and 44.5% (n=117) for global perceived change. The predictive ability of the SBST was the strongest for disability. The medium risk group (RR 2.30 [95% confidence interval (CI) 1.28, 4.10], p=0.003) and the high risk group (RR 2.86 [1.60, 5.11], p≤0.001) had a meaningfully greater relative risk of being considered disabled compared to the low risk group. The medium risk group (RR 1.25 [1.04, 1.51], p=0.013) and the high risk group (RR 1.26 [1.03, 1.52], p=0.020) had a statistically greater risk of non-recovery with respect to pain but this relative risk was unlikely to be of any clinical significance.Conclusions: The results of this study provide valuable information to clinicians on the usefulness and limitations of using the SBST in patients with chronic LBP in an Australian setting. The SBST should be applied with caution for patients presenting with chronic LBP. The SBST has value as a substitute for the administration of multiple full-length, unidimensional questionnaires to initially screen for high levels of pain, disability, and negative psychological affect and cognitions in clinical practice. Therefore the tool can alert the clinician to the need for further assessment. The SBST has moderate predictive and acceptable discriminative ability for future disability. However, the SBST’s predictive and discriminative ability was relatively weak for future pain and the tool was unable to identify those participants who perceived themselves as improved versus not improved at the one year follow-up.