Validity of self-reported periodontal questions in a New Zealand cohort

Abstract

Objective

This study aims to assess the validity of four self-reported questions for measuring periodontitis in a birth cohort.

Methods

Full-mouth periodontal examinations (three sites/tooth) were undertaken at age 38 in a complete birth cohort born in 1972/1973 in New Zealand. Four self-reported periodontal screening questions were included (“Do you think you have gum disease”; “Has a dental professional ever told you that you have lost bone around your teeth”; “Have you ever had scaling, root planing, surgery, or other treatment for gum disease” and “Have you ever had any teeth that have become loose by themselves without some injury”), and the sensitivity and specificity of those self-reported items were calculated for individual questions and using a multivariable binary logistic regression model. Generalised linear models were used to compare relative risks for periodontitis and smoking, using the (a) clinical measures and (b) self-reported questions.

Results

Among the 895 who had periodontal examinations, the prevalence of periodontitis was 43.7, 22.8 and 12.0 %, respectively, for one or more sites with ≥4, ≥5 and ≥6 mm clinical attachment loss (AL). The specificity of the four self-reported questions was high (82–94 %), but the sensitivity was low for all, except the question: “Do you think you have gum disease”. The four questions’ highest combined sensitivity + specificity value was 1.33 for one or more sites with ≥4 mm AL, with the area under the receiver operating characteristic (ROC) curve being greatest for one or more sites with ≥6 mm AL, at 0.84. For the smoking–periodontitis association, the estimates of relative risk for periodontitis among smokers were as follows: (a) 1.81, 2.88 and 5.79, respectively, clinically determined to have one or more sites with ≥4, ≥5 and ≥6 mm AL and (b) 2.19, 2.17, 1.23 and 1.89, respectively, for the four self-reported questions.

Conclusion

The four self-reported periodontal screening questions performed adequately in identifying clinically determined periodontal disease, and they showed moderate validity when used together as a set. However, the strength of the association between smoking and periodontitis was underestimated when they were used instead of clinically determined periodontal disease.

Clinical relevance

These findings suggest that clinical examinations remain to be the desired approach for periodontal surveys, but where resource constraints preclude those, self-reported methods can provide useful information; after all, some periodontal information is better than none at all.