Validity of a questionnaire in estimating restorative treatment need among young adults

There are no verified anamnestic sets of questions for assessing restorative treatment need. Finnish conscripts responded to a computer-based questionnaire on oral health during their oral health screening in 2011. This study compared the outcomes of the screening and the questionnaire concerning restorative treatment need with the aim of finding and validating a set of questions with the best association between the two methods.

The study group comprised 8,566 conscripts. Of 50 original questions, 22 questions were chosen for closer analysis. The generalised linear mixed model was used to calculate the OR values (95% CI) for each of these questions, using restorative treatment need at individual level as the response variable. Questions with the best association (Q1-Q10) were selected for the final set; the inclusion criterion was p < 0.05. The area under curve (AUC) value was calculated for the sum function of these 10 questions.

Among the final set of 10 questions, the OR values varied between 1.12 and 4.61. The AUC value was 0.75. By increasing the number of positive responses to 8 questions, the odds for restorative treatment need were OR 69.27 and increased to infinity with 10 questions.

A selected set of questions together with clinical screening, or even alone, can be a valid instrument for screening people for restorative treatment need. The method is particularly useful in large populations. This statistical method might identify appropriate sets of questions for different contexts.