Study Design and Population
NHANES is cross-sectional survey of the non-institutionalized US civilians to collect data through a combination of laboratory assessments, self-reported questionnaires, and clinical exams. It is conducted bi-annually, however, the COVID-19 pandemic prevented field operations in March 2020, which resulted in the incomplete data collection for the 2019–2020 cycle, making the collected data not nationally representative. To address this issue, the data collected from 2019 until March 2020 were combined with data from the NHANES 2017–2018 cycle to create a nationally representative sample of NHANES 2017-March 2020 pre-pandemic data.[13]
The unweighted response rate of the examined sample was 47% for the 2017–March 2020 cycle. We included all participants one year or older with the complete overall health exam status with at least one natural tooth, exclusive of third molars, and answered the questions about their last dental appointment. The total sample size included 12,944 participants, and all participants provided written informed consent prior to study participation. The study was approved by the ethical review boards of the National Centre for Health Statistics (approval protocol numbers: 2011–17 and 2018–01).[14]
Clinical assessment of oral diseases
Trained and calibrated dental professionals conducted all the clinical examinations to assess the oral health status. We described teeth with untreated coronal caries as any dental cavity in the crown of a tooth that was both active and untreated, excluding third molars. For adults aged 18 years or older, we defined untreated root caries as any carious lesion located below the cementoenamel junction and above the gingival margin of teeth with gum recession, excluding third molars. We categorized missing teeth as teeth that had been lost due to caries or periodontal disease. Additionally, we determined the number of teeth that had coronal caries and the number of missing teeth.
Utilization of Dental Care and demographic factors
The primary predictor was the time since the last dental appointment. Based on the question, “When did you last visit a dentist?”. We categorized participants into three groups: if they had a dental appointment within a year; if more than a year, but within three years; if their last appointment to the dentist was more than 3 years ago or never.
We further categorized participants, based on the type of appointment, into routine dental care attendees or urgent appointments. Using their answers to the question “What was the main reason you last visited the dentist?”, participants were considered routine dental care attendees if they answered “Went in on own for check-up, examination, or cleaning” or “Was called in by the dentist for check-up, examination, or cleaning” or “Went for treatment of a condition that dentist discovered at earlier checkup or examination”, while urgent attendees were identified if they answered "Something was wrong, bothering or hurting me". We assigned individuals who have never been to a dentist as urgent attendees.
In our analyses, we took into account sociodemographic confounders, including age, gender, race/ethnicity, family income based on federal poverty level, and education level. The age variable was divided into seven groups (1–5, 6–11, 12–19, 20–34, 35–49, 50–64, 65+). Gender was either male or female. The race/ethnicity variable was divided into five groups, including Non-Hispanic Asian, non-Hispanic Black, non-Hispanic White, Mexican American/Hispanic, and Other, which included multi-racial groups. Family income was divided into four groups based on the ratio of family income to the federal poverty level (FPL). Finally, education level was divided into five groups: younger than 20 years old (education not reported), less than high school, completed high school/GED, some college or AA degree, and college graduate or above.
Statistical analysis plan
We described first the demographic distribution of our study population. We used a chi-square test to assess the distribution of these characteristics by the time since their last dental appointment. We reported the prevalence of having teeth with untreated coronal caries, teeth with root caries, missing any tooth, as well as the mean number of teeth with untreated coronal caries and mean number of missing teeth. National weighted estimates were reported, with the corresponding 95% confidence intervals (95%CI) and stratified by the reason of dental appointments. Taylor linearization methods were used in the survey procedures for standard error estimations, with the publicly provided masked variance pseudoprimary and masked variance pseudostratum sampling units.
Logistic regressions were used for the binary dental outcomes: presence of untreated coronal caries, presence of untreated root caries, and presence of missing teeth. Poisson regressions were used to assess the mean ratio for the count outcomes: number of teeth with untreated coronal caries and number of missing teeth. Simple logistic/Poisson regressions were run first to assess the crude estimates of the oral health outcomes by the time since last dental appointment, stratified by the reason of the appointment (routine or urgent). Then, we adjusted for demographic characteristics in the final multiple regression models with the interaction between time since the last dental appointment and reason of the appointment. Alpha was set at 0.05, and all analyses carried out using Stata 17.0 (StataCorp).