Learning From Those Who ‘Don’t Know:’ Exploring Self-Reported Oral Health

Objectives: Self-report data can extend the reach of oral health research, but ‘don’t know’ responses can threaten validity. Such participants may be excluded, which reduces sample size and may induce selection bias if ‘don’t know’ respondents are systematically different from those included. We aimed to compare characteristics of those responding with ‘don’t know,’ compared to ‘yes’ or ‘no,’ to assess the potential for systematic differences.
Methods: We used data from three questionnaire-based observational studies. Pregnancy Study Online (PRESTO) (N=7,343), the Black Women’s Health Study (BWHS) (N=483), and the National Health and Nutrition Examination Survey (NHANES) (N=6,980). We evaluated responses to single questionnaire items meant to determine the presence of ‘gum disease’ within each of the three cohorts (i.e., “Has a dentist or dental hygienist ever told you that you have periodontitis or gum disease?”). We compared sample characteristics, including self-reported general oral health, engagement with oral healthcare, and demographic characteristics across each response category (yes/no/don’t know).
Results: There were systematic differences between those who reported ‘don’t know’ and those who reported ‘yes’ or ‘no.’ Those differences were related to markers of socioeconomic status (SES), self-rated oral health, and health-related behaviors including healthcare engagement. In the specific case of oral healthcare use, 55% of ‘don’t know’ respondents in PRESTO had visited a dentist in the past year, compared with 72% of ‘yes’ and 69% of ‘no.’ In BWHS 55% of ‘don’t know’ respondents had visited a dentist compared with 84% of ‘yes’ and ‘81%’ of ‘no, and in NHANES, 32% of ‘don’t know’ respondents had visited a dentist compared with 49% of ‘yes’ and 62% of ‘no.’
Conclusions: Those who responded ‘don’t know’ had distinctly different patterns of oral health engagement and SES predictors of oral health. This systematic difference warrants development of better self-reported measures and/or use of methods to address threats to validity.