Validity and Reproducibility of the STarT Back Tool (Dutch Version) in Patients With Low Back Pain in Primary Care
Jasper D Bier1, Raymond W J G Ostelo2, Miranda L van Hooff3
1Department of General Practice, Erasmus MC, PO Box 2040, 3000 CA, Rotterdam, the Netherlands, and Fysiotherapie Fascinatio, Capelle aan den Ijssel, the Netherlands.
Objective:
The purpose of this study was to translate and to investigate the reliability and validity of the STarT Back screening tool (SBT) in the primary care setting among patients with nonspecific low back pain (LBP).
Design:
The SBT was formally translated into Dutch following a multistep approach for forward and backward translation. General practitioners and physical therapists included patients with LBP.
Methods:
Patients completed a baseline questionnaire and a follow-up at 3 days and 3 months. The construct validity was calculated with Pearson's correlation coefficient. The reproducibility was assessed using the quadratic weighted kappa and the specific agreement. Predictive validity was assessed using relative risk ratios for persisting disability at 3 months. Content validity was analyzed using floor and ceiling effects.
Results:
In total, 184 patients were included; 52.2% were categorized in the "low-risk" subgroup, 38.0% "medium-risk," and 9.8% "high-risk." For the construct validity we found, as expected, a moderate to high Pearson's correlation for questions 3 to 9 and a low correlation for questions 1 and 2 with their respective reference questionnaires. The reproducibility had a quadratic weighted kappa of 0.65 and the specific agreement of 82.4% for "low-risk," 53.3% for "medium-risk," and 33.3% for "high-risk." For the predictive validity for persisting disability we found a relative risk ratio for "medium-risk" of 1.8 (95% confidence interval [CI]: 1.0-3.1) and 2.7 (95% CI: 1.4-4.9) for "high-risk" compared with "low-risk." For the content validity, we found that no floor and ceiling effects were present.
Limitations:
There was a relatively small sample size for the retest reliability study. Patients were not compared between physical therapist and GP, as there were not enough patients in both groups. For practical reasons, the patients filled out the baseline questionnaire after receiving the first treatment/consultation; however, the questionnaire is intended to be filled in before the first consultation/treatment.
Conclusion:
The SBT has been successfully translated into Dutch. The psychometric analysis showed acceptable results and, therefore, the SBT is a valid screening tool for patients with LBP in Dutch primary care.
