Background Several questionnaires on neck pain are usually assumed to be unidimensional as their raw scores are considered interval measures. However, the assumption of unidimensionality requires the questionnaire data to satisfy certain criteria when tested with modern psychometric approaches, such as Item Response Theory. The Neck Bournemouth Questionnaire (NBQ) has already been cross-culturally adapted and validated in Italian on a population of subjects with chronic neck pain (Geri et al., 2014). We examined this scale by means of Rasch analysis an Item Response Theory model in order to provide a disease-specific questionnaire whose scores can be assumed as interval variables and, consequently, legitimately their change scores as meaningful.
Objectives Validation of the Italian NBQ with Rasch analysis.
Methods A sample of 161 subjects with chronic neck pain was assessed with the NBQ. The structure of the scale was preliminarily tested with both exploratory and confirmatory factor analyses. Rasch analysis was performed analysing the following features: the goodness of fit of the data to the model, thresholds ordering, unidimensionality, local independence of the items and Differential Item Functioning (DIF) were examined to consider the NBQ an interval scale. Further, the Person Separation Index and the Mean Location of Persons were used, respectively, to assess the internal consistency and the targeting of the scale.
Results Both the exploratory and confirmatory factor analyses revealed a two factors structure of the NBQ. Factor 1 (item 1, 2, 3, 6) fitted the model (χ2=10.65, df 8, p=0.22) after removal of item 7, which had low communalities on exploratory factor analysis and further showed disordered thresholds (Figure). Item 2 presented a marginal uniform Differential Item Functioning across educational level. The Person Separation Index was 0.80 and the Mean Location of Persons was 0.48 (SD=1.02). Factor 2 (Item 4 and 5) fitted the model (χ2=3.86, df 4, p=0.42). The Person Separation Index and the Mean Location of Persons of factor 2 were, respectively, 0.77 and -0.71 (SD=1.57).
Conclusions The NBQ is a useful tool for the assessment of patients with chronic neck pain. The NBQ showed a two factors structure, each factor fitted the Rasch model and therefore works as a unidimensional measure. Using the interval scores of the NBQ may provide meaningful change scores to establish the effectiveness of treatments for chronic neck pain in clinical trials.
Geri T, Signori A, Gianola S, Rossettini G, Grenat G, Checchia G, et al. Cross-cultural adaptation and validation of the Neck Bournemouth Questionnaire in the Italian population. Quality of Life Research. 2014. Epub ahead of print 23 September 2014.
Disclosure of Interest None declared