December 2007

Aims To evaluate the reliability and construct validity of the Dental Neglect Scale (DNS) and to estimate the level of dental neglect in the adult Norwegian population. Methods A questionnaire containing socio-demographics, oral health attitude variables, self-reported service use and a translated version of the original DNS was tested in two samples: 1) a convenience sample of University employees (n=263) and 2) a proportionate random sample (n=2000) drawn from the national population register (age 16-79 years). The reliability assessment of the instrument was by internal consistency (Cronbach’s alpha) and factor analysis (principal component) (n=1309). The test-retest comparisons (n=108) were analyzed by Spearman’s rho for the sum-scores, and kappa statistics for single items. Logistic regression analyses were used to evaluate the construct validity of the DNS. Results The Cronbach’s alpha coefficient for the overall construct of DNS was 0.67 (n=173) and 0.57 (n=1301). Only one factor was extracted, explaining a total of 36% of the scale variance (n=1301). Cohen’s kappa for the test-retest comparisons ranged from 0.21-0.79 (n=108), and Spearman’s rho for the test-retest sum-scores was 0.60. Higher neglect scores were positively related to a negative opinion about own dental health (OR=3.3), last dental appointment because of pain or other problems (OR= 2.3), less than 20 teeth (OR=2.2), drinking soft drinks with sugar every day (OR=2.1), non-regular dental service use (OR=2.2) and using floss or toothpicks seldom or never (OR=1.6). The prevalence of high dental neglect was 20%. Conclusions The analyses indicated construct validity for the Dental Neglect Scale but low reliability for some of its items. One fifth of this representative sample of Norwegian adults reported a high level of dental neglect. Key words: Dental neglect, reliability, validity