German Translation and Cross-Cultural Adaptation of the STarT Back Screening Tool

Abstract

Background: Although evidence based treatment approaches for acute low back pain are available, the prevention of persistent disabling symptoms remains a challenge. Subgroup targeted treatment using adequate screening tools may be a key component for the development of new treatment concepts and is demonstrating promising early evidence. The Keele STarT Back Screening Tool is a practical instrument, developed to stratify patients with back pain according to their risk of persistent disabling symptoms. The aim of this study was to translate and cross-culturally adapt the STarT tool into German (STarT-G) and to investigate its psychometric properties.

Methods: The translation was performed according to internationally accepted guidelines and pretested to assess face validity among patients. Psychometric testing was then performed within a cross-sectional cohort of adult patients attending physiotherapy practices for back pain. Patients completed a booklet containing STarT-G and 5 reference standard questionnaires. Measurement properties of the STarT-G were explored including construct validity, floor and ceiling effects, and discriminative abilities.

Results: The pretests (n=25) showed good face validity including strong comprehension and acceptability of the STarT-G with only item 5 (fear avoidance) manifesting some ambiguities. The questionnaires were sent to 74 and completed by 50 patients (68%) of whom mean age was 46 (SD 14.5) years and 52% were male. Spearman's rank correlations for construct validity ranged from 0.35 to 0.56. AUCs for discriminative ability ranged from 0.79 to 0.91. Neither floor nor ceiling effects were observed. There were 28 (57%) participants defined as low risk, 17 (35%) as medium risk, and 4 (8%) as high risk.

Conclusion: STarT-G is linguistically valid for German speaking countries. For the selected population, the correlations indicate acceptable validity and AUC showed satisfying discrimination. Data for psychometric properties have to be confirmed in a large scale study with a representative sample.

Fig 1. Distribution STarT-G totalscore.

Fig 2. Distribution STarT-G psychosocial subscore.

Fig 3. STarT-G totalscore against RMDQ Score.

RMDQ: Roland Morris Disability Questionnaire.

Fig 4. STarT-G psychosocial subscore against PCS Score.

PCS: Pain Catastrophizing Scale.

Fig 5. Receiver operating characteristic curve for STarT-G totalscore against RMDQ cases.

AUC 0.79 (95% CI 0.67–0.92); RMDQ: Roland Morris Disability Questionnaire.

Fig 6. Receiver operating characteristic curve for STarT-G totalscore against composite reference standard cases.

AUC 0.81 (95% CI 0.65–0.98).

Fig 7. Receiver operating characteristic curve for STarT-G psychosocial subscore against composite reference standard cases.

AUC 0.88 (95% CI 0.71–1.00).

Fig 8. Receiver operating characteristic curve for risk group against composite reference standard cases.

AUC 0.91 (95% CI 0.77–1.00).