00142075220J Public Health DentJ Public Health DentJournal of public health dentistry0022-40061752-732527701758801155710.1111/jphd.12180HHSPA1630639ArticlePrevalence of past-year dental visit among US adults, 1999–2010: comparison of trends and estimates between the Behavioral Risk Factor Surveillance System and three national surveysLinMeiMD, MPH, MSc1LiChien-HsunMS, MA2WeiLiangMS, MPH3NaavaalShillpaBDS, MPH, MS1Kolavic GrayShellieDMD, MPH4ManzMichael C.DDS, MPH, DrPH5BarkerLaurieMSPH1Division of Oral Health, Centers for Disease Control and Prevention, Atlanta, GA, USANorthrop Grumman Corp., Atlanta, GA, USADB Consulting Group, Inc., Atlanta, GA, USAIDONEOUS Consulting, Atlanta, GA, USASchool of Dentistry, University of Michigan, Ann Arbor, MI, USACorrespondence Mei Lin, Division of Oral Health, Centers for Disease Control and Prevention, 4770 Buford Highway, MailStop F80, Atlanta, GA 30341. Tel: 770-488-5109; Fax: 770-488-6080; mlin@cdc.gov. Shillpa Naavaal and Laurie Barker were with the Division of Oral Health at the Centers for Disease Control and Prevention. Chien-Hsun Li is with the Northrop Grumman Corp. Liang Wei is with the DB Consulting Group, Inc. Shellie Kolavic Gray is with the IDONEOUS Consulting. Michael Manz is with the University of Michigan, School of Dentistry.83202104102016320173132021772105114Objectives:

To compare estimated prevalence of past-year dental visit (PPYDV) among US adults aged ≥18 years from the Behavioral Risk Factor Surveillance System (BRFSS) to estimates from the Medical Expenditure Panel Survey (MEPS), National Health Interview Survey (NHIS), and National Health and Nutrition Examination Survey (NHANES).

Methods:

We estimated PPYDV adjusted for covariates (age, race/ethnicity, education level, poverty status, edentulism) using BRFSS, MEPS, and NHIS 1999–2010, and NHANES 1999–2004. We tested trend in overall PPYDV for BRFSS, MEPS, and NHIS from 1999–2010. For 2002 and 2010, we calculated absolute differences (AD) and 95% confidence intervals (CI) in PPYDV between BRFSS and each of the other surveys overall and among subpopulations defined by covariates. We pooled NHANES 1999–2004 data for comparison with BRFSS 2002.

Results:

From 1999 to 2010, BRFSS (68.5% vs. 67.5%), MEPS (43.5% vs. 39.7%), and NHIS (63.3% vs. 59.7%) showed small but significant decreases in overall PPYDV. In 2002, estimates for overall PPYDV were highest for BRFSS (70.0%) and lowest for MEPS (43.9%) with estimates for NHIS (61.5%) and NHANES (1999–2004: 58.1%) in between; the largest AD (26.2%, 95% CI: 25.0%–27.3%) was between BRFSS and MEPS. ADs were consistent in 2002 and 2010, overall and by covariates, except among edentate persons, where PPYDV estimates from BRFSS and NHIS were similar.

Conclusions:

Estimates of PPYDV from BRFSS were notably higher than estimates from MEPS, NHIS, or NHANES except among the edentate. Trends in PPYDV over time, however, were consistent across all surveys.

dental care usehealth surveyspopulation surveillanceadultsIntroduction

Since 1999, the prevalence of past-year dental visit (PPYDV) among adults has been a National Oral Health Surveillance System (NOHSS) indicator, with the Behavioral Risk Factor Surveillance System (BRFSS) as the official data source (1). This NOHSS indicator aligns with the national Healthy People (HP) 2020 objective to increase the proportion of the US population using the oral health care system in the past year, with the Medical Expenditure Panel Survey (MEPS) as the official data source (2). In practice, state progress in PPYDV has been monitored by comparing state-level estimates from BRFSS to HP targets based on MEPS, or previously, the National Health Interview Survey (NHIS) (3,4). An inherent assumption behind this approach is that estimates from BRFSS are similar to those from MEPS and NHIS, which may not hold for PPYDV.

National PPYDV estimates can also be obtained from the National Health and Nutrition Examination Survey (NHANES) (5). Stakeholders at the national, state, and local levels use PPYDV estimates from all four of these data sources to assess trends over time, evaluate impacts of programs and policies, and track progress toward targets for national objectives. The different emphases and designs of these surveys make them unique and complementary to one another, however, estimates of PPYDV from these surveys may not be directly comparable due to differences in survey design (6).

A comparison of PPYDV estimates in adults aged 25–59 years among MEPS, NHIS, and NHANES during 1971–1996 found that the lowest estimates were from MEPS and that trends over time in overall PPYDV estimates from these surveys were inconsistent (6). More recently, PPYDV estimates from BRFSS (e.g., 70.8% among ages ≥18 years, 2004 US median) (7) appear higher than estimates from MEPS (e.g., 43.0% among ages 21–64 years, 2004) (8), NHIS (e.g., 65.9% among ages 21–64 years, 2003) (9), and NHANES (e.g., 59.9% among ages 20–64 years, 1999–2004) (5). However, the practice of comparing BRFSS estimates to HP targets for PPYDV has continued, perhaps due in part to the lack of an accepted, comparable benchmark for estimates from BRFSS.

MEPS, NHIS, and NHANES are primarily used to monitor health at the national level. They are not designed to provide state-level estimates, although direct estimates may be obtained from MEPS and NHIS for some states with larger sample sizes, or through pooling multiple years of data (10–12). Although BRFSS is primarily used to provide state-level estimates, nationwide estimates can be obtained by pooling data for all states (13,14). Although each survey uses a different sampling design, all are limited to civilian, noninstitutionalized US residents.

BRFSS, an annual telephone survey of adults, was established in 1984 specifically to provide state-level data for chronic diseases, health-related behaviors, and health care use. By 1993, all 50 states and DC participated in the core questionnaire, supporting nationwide estimates from pooled data (15). To address the lack of comparable state-level oral health data, the questionnaires for 1999, and even-numbered years since 2002 have included a question about time since last dental visit (1).

MEPS has been conducted and its data released annually since 1996. MEPS provides national estimates of health care use and expenditures (16). MEPS is the only survey of these four to use a panel design that includes five rounds of in-person interviews over 2 years (16).

NHIS, an in-person household survey conducted continuously since 1957 has provided the longest time series of national data on a wide range of health indicators and supports annual estimates within many subpopulations (12). NHIS has a central role in the design of federal household surveys, such as providing the sampling frame for MEPS (12).

NHANES, first established in the early 1960s and conducted continuously since 1999, combines in-person interviews with laboratory tests and standardized physical examinations, including an oral health examination (17). Released every two years, it is a uniquely rich data source for monitoring oral health status within the overall health context. Because of the intensive nature of NHANES, a much smaller sample size than NHIS and MEPS is used.

To our knowledge, no studies have provided comparisons of PPYDV estimates from BRFSS with those from MEPS, NHIS, and NHANES, overall and within subpopulations, in one report. Because BRFSS is the only survey of these four to support direct PPYDV estimates for all states, we compare national trends and estimates from BRFSS with those from MEPS, NHIS, and NHANES during 1999–2010. Findings may inform development and use of aligned state and national oral health indicators.

MethodsData sources

Our analysis of publicly released, de-identified data sets was exempt from IRB and human subjects research review; however, data collection for all four surveys was reviewed by the appropriate institutional or ethics review boards and either approved or found to be exempt from review.

BRFSS

We pooled BRFSS data for all states and DC to create nationwide estimates for PPYDV for 1999, 2002, 2004, 2006, 2008, and 2010 (1). We did not include data after 2010 because BRFSS changed the sampling frame and weighting methods from 2011 and forward (18).

MEPS

We estimated PPYDV for each year during 1999–2010 from the MEPS-household component (MEPS-HC) (16). The MEPS-HC sample is drawn from the previous year’s NHIS sample households; participants in each panel are asked about dental visits in five rounds of in-person interviews over 2 years (16).

NHIS

We estimated PPYDV for each year during 1999–2010 from the NHIS Sample Adult data. NHIS Household and Family components collect demographic and limited health data on all members of each NHIS household and family. From each NHIS family, a sample adult is randomly selected to complete the Sample Adult core questionnaires, on which the question on time since last dental visit is included (12). We obtained covariates from the NHIS Household and Family components.

NHANES

In 1999–2010, the comparable time period with BRFSS, the NHANES oral health questionnaire included a dental visit question in 1999–2004, but not in 2005–2010 (17). We pooled NHANES data from the 1999–2000, 2001–2002, and 2003–2004 cycles to estimate PPYDV during 1999–2004 (17) for comparison with the BRFSS 2002 estimates. The NHANES analytic guidelines recommend pooling four or more years of data to produce stable estimates for subpopulations (10). Pooling 1999–2004 data placed 2002 closer to the midpoint than pooling 4 years of data (1999–2002 or 2001– 2004).

Study population

We limited our analyses to adults aged ≥18 years with information on PPYDV. Because the NOHSS indicator and HP2020 objective for PPYDV (1,2) are not limited to the dentate population, our study population included dentate and edentulous adults.

Outcome

The outcome was PPYDV based on self-reported visit to any type of dental care provider. For each survey, we reviewed questions and responses representing PPYDV, along with interview mode, which could affect the response or dictate the questionnaire structure (Table 1). BRFSS, NHIS, and NHANES participants were asked how long it had been since their last dental visit. The questions and response choices for each survey did not change from 1999–2010 except for minor changes in response options for BRFSS since 2002. BRFSS used a telephone interview; the other three surveys used in-person interviews. MEPS participants were asked about dental visits in each of the five interview rounds conducted over two years: since January 1 for the first round; since the previous round for the second through fourth rounds; and between the fourth round and December 31 for the fifth round. From MEPS, we estimated the prevalence of having ≥1 dental visit during each calendar year, using data from all panels with interview rounds during the calendar year (16).

Covariates

We selected socio-demographic factors for which disparities in dental care use have been reported (6,19–21). We also included dentate status because difference in PPYDV between socio-demographic groups may be due in part to differences in prevalence of edentulism (21).

Socio-demographic factors included age group (18–44, 45–64, and 65+ years), race/ethnicity (non–Hispanic White (NHW), non–Hispanic Black (NHB), Hispanic), education level (less than high school, high school graduate, greater than high school), and poverty status (Poor: <100% Federal Poverty Level (FPL), Near-poor: 100–199% FPL, Non–poor: ≥200% FPL). Adults reporting multiple races or race/ethnicities other than NHW, NHB or Hispanic were included in the analyses, however, we do not report separate estimates for these groups. The NHANES 1999–2004 estimates for the Hispanic population are representative only of the Mexican American population (10).

Edentulism (yes, no) was based on self-report of having lost all natural teeth for BRFSS, MEPS, and NHIS. For NHANES, edentulism was defined as having no natural teeth recorded in any of the 32 tooth spaces from the oral health examination. Adults lacking examination data were coded as dentate status unknown and retained in the denominator for the analyses.

Statistical analyses

We estimated crude and adjusted PPYDV overall for BRFSS, MEPS, and NHIS from 1999–2010 and for NHANES 1999–2004. To estimate adjusted PPYDV, we used predicted marginal probabilities from binomial multiple regression models that simultaneously adjusted for all covariates (22).

We tested linear trend over time in overall PPYDV from 1999–2010 for BRFSS, MEPS, and NHIS through binomial regression models, controlling for other covariates. Linear trend was present if the ß coefficient of the survey year variable was not equal to zero (increasing trend if ß >0; decreasing trend if ß <0); statistical significance of the trend was assessed with a t-test.

Analyses stratified by the covariates were conducted for BRFSS, MEPS, and NHIS using data from 2002 and 2010 and for NHANES using data from 1999–2004. PPYDV estimates among subpopulations were adjusted for all other covariates.

We calculated absolute differences (AD) in adjusted PPYDV between BRFSS and each of other three surveys. We provided 95% confidence intervals (CI) for AD using the Z distribution. Z-tests were used to test hypotheses of AD greater than zero.

We describe statistically significant results at the P≤0.05 level, unless otherwise stated. All analyses were conducted using SAS-callable SUDAAN version 11 (23), accounted for the complex, multistage sampling designs, and were weighted for disproportionate sampling, nonresponse and post–stratification to the US adult population. For NHANES estimates, we used the interview weights because adults who did not participate in the examination were retained in the denominators. Trend analyses incorporated the variance structures described in the NHIS and MEPS analysis guidelines to account for statistical dependence between samples drawn within the same design period (e.g., 1995–2005, 2006–2010) and geographic area (16,24).

Results

The sample size for each year was largest for BRFSS (range: 153,669–441,607), followed by NHIS (range: 21,361–32,792), MEPS (range: 16,748–24,925), and NHANES 1999–2004 (16,997).

Trends in overall adjusted PPYDV from BRFSS, MEPS, and NHIS during 1999–2010 were generally parallel with small declines (P<0.001) (Figure 1 and Table 2).

Crude and adjusted PPYDV were similar overall and within most subpopulations, except among older adults aged ≥65 years, Hispanics, adults with less than high-school education, and poor adults. For example, the crude estimates were lower than adjusted estimates among older adults from BRFSS (2002: 64.6% vs. 72.7%), MEPS (2002: 42.3% vs. 50.9%), NHIS (2002: 54.8% vs. 66.3%), and NHANES (1999–2004: 53.5% vs. 62.9%). We present only adjusted estimates to reduce the chance that differences between estimates were due to confounding (Tables 3 and 4).

In 2002, the highest overall adjusted PPYDV estimate was from BRFSS (70.0%), followed by NHIS (61.5%), NHANES 1999–2004 (58.1%), and MEPS (43.9%) (Table 3). Among most subpopulations, with the exception of edentulous adults, the highest estimates were from BRFSS and the lowest from MEPS. The 2010 overall adjusted estimates were ordered similarly: BRFSS (67.5%), NHIS (59.7%), and MEPS (39.7%) (Table 4). The AD in overall adjusted PPYDV between BRFSS and the other three surveys was largest for MEPS (2002: AD = 26.2%, 95% CI: 25.0%–27.3%), followed by NHANES (1999–2004: AD = 12.0%, 95% CI: 10.3%–13.6%) and NHIS (2002: AD = 8.5%, 95% CI: 7.8%–9.3%) (Table 5). ADs in 2010 were similar to those in 2002 among most subpopulations (Table 5).

In the edentate subpopulation, the BRFSS estimate (27.5%, 2002) was higher than those from MEPS (19.8%, 2002) and NHANES (19.8%, 1999–2004). However, we did not find differences in adjusted PPYDV between the BRFSS and NHIS estimates (26.7%, 2002) (AD = 0.9%, 95% CI: −1.5%–3.2%) (Tables 3 and 5).

Discussion

We found that MEPS yielded the lowest PPYDV estimate among adults relative to estimates from NHIS and NHANES, consistent with Macek et al.’s finding (6). However, we found generally parallel trends in adjusted overall estimates of PPYDV during 1999–2010 for BRFSS, MEPS, and NHIS, in contrast to Macek et al. (6), who found inconsistent trends in overall PPYDV estimates among adults aged 25–59 years among MEPS and its precursors (1977 National Medical Care Expenditure Survey, 1987 National Medical Expenditure Survey, 1996 MEPS), NHIS (1986, 1989, 1993) and NHANES (1971–1975, 1988–1994). These inconsistencies perhaps were due in part to differing and limited number of time periods with available data at the time of their study. Our trend analyses were based on 12 years of NHIS and MEPS data and 6 years of BRFSS data, included older adults and controlled for multiple socio-demographic factors and edentulism.

We found that BRFSS produced substantially higher PPYDV estimates than MEPS, NHIS, and NHANES for the time periods and most subpopulations we examined. Overall, BRFSS estimates differed from NHIS, NHANES, and MEPS by about 8, 12, and 26 percentage points, respectively (Figure 1). One exception was among the edentate, for whom PPYDV estimates from BRFSS were higher than from MEPS and NHANES, but not different from NHIS. There is no clear explanation for these differences among edentate persons.

To our knowledge, no previous studies included BRFSS in such comparisons. The differences in survey design, interview mode (telephone vs. in-person) and question (Table 1) may explain the relative order of PPYDV estimates among the surveys. For example, our finding of the highest PPYDV estimate among adults from BRFSS versus the lowest estimate from MEPS parallels Romaire et al.’s finding of the highest PPYDV estimate among children from the National Survey of Children’s Health (NSCH) versus the lowest estimate from MEPS (20). Like BRFSS, NSCH is a telephone survey designed to provide state and national data, although for children through parent interviews.

Although all four surveys collect only self-reported dental visit data, Macek et al. and Romaire et al. (6,20) speculated that MEPS might provide a more accurate estimate of PPYDV than NHIS, NHANES, and NSCH because the shorter recall period of the MEPS panel design may improve recall accuracy. This feature affirms the rationale for use of MEPS as the official data source for the HP objective on dental care use (6). BRFSS, NHIS, and NHANES – each with recall time of a year or longer – could be subject to greater recall error than MEPS. However, it isn’t clear whether recall error from a longer recall period would lead to systematic overestimation or underestimation of PPYDV (6,20,25).

The inclusion in MEPS of detailed follow-up questions on each dental visit could protect against overestimation of PPYDV due to social desirability bias (6), and lead to underestimation of PPYDV, if participants underreport visits, because they know that additional follow-up questions regarding that visit will be asked (6,26).

Different modes of data collection might also explain the difference in PPYDV estimates. During this time period, BRFSS used a landline-based telephone interview whereas MEPS, NHIS, and NHANES used an in-person interview (25,27). Landline-based telephone interview surveys may have a bias toward higher-income, older, non–minority respondents due to telephone noncoverage than in-person household interviews, although BRFSS minimized the impact of noncoverage with special weights (27). Congruent with our findings, Romaire et al. (20) reported much higher PPYDV estimates from NSCH, also a landline-based telephone interview survey, than from the face-to-face interview surveys, MEPS, NHIS, and NHANES. However, the difference in estimates between these telephone and face-to-face interview surveys can vary by indicator. Previous studies reported that BRFSS, NHIS, and NHANES provided similar estimates for several measures, such as smoking and diabetes (13,14).

A few limitations of our study should be noted. Our study used self-reported time since last dental visit. One study examined the accuracy of self-reported dental visit in 6 months against dental record among adults aged ≥45 years in north Florida and agreement ranged from 84 to 91% (28). No studies of internal validity were available to confirm which of the four surveys produced more accurate PPYDV estimates, although MEPS estimates may be subject to fewer sources of bias than BRFSS, NHANES, or NHIS. The relative order and size of PPYDV estimates between BRFSS and the other surveys may not apply to other indicators. Data for categorizing subpopulations by Hispanic ethnicity and edentate status differed by survey, thus between-survey PPYDV comparisons for these groups should be interpreted with caution.

It is perhaps most important to note that BRFSS has made major methodological changes since 2011, including adding cell phone interviews and adopting an advanced weighting method to improve the survey’s representativeness(18). However, three recent reports of PPYDV among adults aged 65+ years in 2012 indicate that estimates from BRFSS (66.0%) (1) remained higher than estimates from NHIS (61.8%) (29) and substantially higher than estimates from MEPS (42%) (30). When three or more years of data are available from BRFSS (e.g., 2012, 2014, and 2016), a study is needed to determine if trends from the surveys remain approximately parallel and the magnitude of differences in PPYDV estimates between BRFSS and the other surveys changes over time, especially among young adults and minority groups who were more likely to be cell phone only users than their counterparts (18). Despite the limitations, our study highlights differences that existed among these surveys in the last decade, establishes a good point of reference for future study, and highlights the need for a comparable benchmark for state estimates of PPYDV.

Our study findings indicated that it may not be appropriate to compare state estimates for PPYDV from BRFSS to national estimates from MEPS, NHIS, or NHANES, such as targets for HP2020 objectives derived from MEPS – due to substantial differences in PPYDV estimates from BRFSS as compared to MEPS, NHIS, and NHANES. For example, if compared with the HP2020 target of 49% on use of dental care in the past year developed from MEPS (2), 1999–2010 BRFSS estimates (Figure 1) were far beyond the target, giving a falsely optimistic view of dental care use at the state level during that time period, in which MEPS estimates (Figure 1) fell short of the target.

Recognizing these challenges, we suggest some practical strategies on how BRFSS could be used to monitor state-level data of dental care use in alignment with the HP2020 objective rather than compared directly (and inappropriately) to the national target. HP objectives reflect issues of national importance. HP provides priority areas, objectives and measurable indicators that states, as well as local areas and communities, can use to guide their efforts in improving the health of their populations. However, states do not necessarily monitor their progress against national targets such as a HP2020 target or other estimates from the official HP data source, especially when the official data source was not designed to support state-level estimates.

Although PPYDV estimates from BRFSS were higher than those from the other three surveys, the direction and size of the differences was relatively constant during 1999–2010 and the trends over time were consistent from the four surveys. Thus, BRFSS could be used to monitor trends over time and sociodemographic disparities in PPYDV.

Rather than monitoring progress relative to national estimates or targets developed from MEPS, NHANES, or NHIS, states may consider using BRFSS to monitor trends in and set state targets for PPYDV. State developed targets could better reflect the impact of state-specific public health program and policy efforts on dental care utilization. Availability of both US and state-level estimates from BRFSS allows comparisons among state estimates and between a state and the US estimate from BRFSS (7). It would also support ongoing comparison of PPYDV estimates from BRFSS to estimates from other national surveys, which contain more detailed data to explore main drivers of trends and provide context for more careful interpretation of the BRFSS estimates.

In conclusion, our study found substantially higher PPYDV estimates among adults from the state-based BRFSS than from national surveys, MEPS, NHIS, and NHANES, even after adjustment for socio-demographic factors and edentulism. The lowest estimates were from MEPS, which were more than 20 percentage points below BRFSS estimates. Despite these differences, BRFSS, MEPS, and NHIS all displayed small decreasing trends in overall PPYDV estimates from 1999–2010. Between-survey comparisons in PPYDV estimates for 2002 and 2010 were similar overall and in most subpopulations. Given the magnitude of the differences in PPYDV estimates between BRFSS and these three national surveys, it may not be appropriate to compare BRFSS estimates of PPYDV with estimates or targets derived from MEPS, NHIS, or NHANES. However, BRFSS could be used to monitor trends in PPYDV over time, and state-level targets for PPYDV could be established using state or nationwide estimates from BRFSS.

Acknowledgment

We acknowledge excellent suggestions provided by Drs. Susan Griffin, Valerie Robison, and Barbara Gooch with the Division of Oral Health at the Centers for Disease Control and Prevention.

Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

ReferencesCenters for Disease Control and Prevention. Oral Health Data. [cited 2016 Apr 29]. Available from: http://www.cdc.gov/oralhealthdata/.US Department of Health and Human Services. Healthy People 2020. 2010. [cited 2014 Aug 13]. Available from: http://www.healthypeople.gov/2020/default.aspx.XuF, TownM, BalluzLS, BartoliWP, MurphyW, ChowdhuryPP, GarvinWS, PierannunziC, ZhongY, SalandySW, JonesCK, CrawfordCA. Surveillance for certain health behaviors among States and selected local areas – United States, 2010. MMWR Surveill Summ. 2013;62(1):1–247.Centers for Disease Control and Prevention. Healthy People2000 progress review – oral health. [cited 2016 May 11]. Available from: http://www.cdc.gov/nchs/healthy_people/hp2000/reviews/oral_health.htm.DyeBA, TanS, SmithV, LewisBG, BarkerLK, Thornton-EvansG, EkePI, Beltran-AguilarED, HorowitzAM, LiCH, Trends in oral health status: United States, 1988–1994 and 1999–2004. Vital Health Stat. 2007;11(248):1–92.MacekMD, ManskiRJ, VargasCM, MoellerJF. Comparing oral health care utilization estimates in the United States across three nationally representative surveys. Health Serv Res. 2002;37(2):499–521.12036005Centers for Disease Control and Prevention. Behavioral Risk Factor Surveillance System prevalence and trends data. [cited 2014 June 20]. Available from: http://www.cdc.gov/brfss/brfssprevalence/.ManskiRJ, BrownE. Dental use, expenses, private dental coverage, and changes, 1996 and 2004. Chartbook #17. [cited 2015 Jul 9]. Available from: http://meps.ahrq.gov/mepsweb/data_files/publications/cb17/cb17.shtml.WallTP, VujicicM, NassehK. Recent trends in the utilization of dental care in the United States. J Dent Educ. 2012;76(8): 1020–7.22855587JohnsonCL, Paulose-RamR, OgdenCL. National Health and Nutrition Examination Survey: analytic guidelines, 1999–2010. Natl Center Health Stat Vital Health Stat. 2013; 2(161).Ezzati-RiceTM, RohdeF, GreenblattJ. Sample design of the Medical Expenditure Panel Survey household component, 1998–2007. Methodology Report No. 22. 2008. [cited 2015 Aug 7]. Available from: http://www.meps.ahrq.gov/mepsweb/data_files/publications/mr22/mr22.pdf.Centers for Disease Control and Prevention. 2010 National Health Interview Survey (NHIS) public use data release – NHIS survey description. 2011. [cited 2014 Aug 13]. Available from:LiC, BalluzLS, FordES, OkoroCA, ZhaoG, PierannunziC.A comparison of prevalence estimates for selected health indicators and chronic diseases or conditions from the Behavioral Risk Factor Surveillance System, the National Health Interview Survey, and the National Health and Nutrition Examination Survey, 2007–2008. Prev Med. 2012; 54(6):381–7.22521996FahimiM, LinkM, MokdadA, SchwartzDA, LevyP. Tracking chronic disease and risk behavior prevalence as survey participation declines: statistics from the Behavioral Risk Factor Surveillance System and other national surveys. Prev Chronic Dis. 2008;5(3):A80.18558030Centers for Disease Control and Prevention. About the Behavioral Risk Factor Surveillance System (BRFSS). [cited 2014 June 20]. Available from: http://www.cdc.gov/brfss/about/about_brfss.htm.Agency for Healthcare Research and Quality. Medical Expenditure Panel Survey household component-135B: 2010 dental visits. 2012. [cited 2014 June 20]. Available from: http://meps.ahrq.gov/data_stats/download_data/pufs/h135b/h135bdoc.pdf.Centers for Disease Control and Prevention. National Health and Nutrition Examination Survey – questionnaires, datasets, and related documentation. [cited 2014 Apr 2]. Available from: http://www.cdc.gov/nchs/nhanes/nhanes_questionnaires.htm.Centers for Disease Control and Prevention. Frequently asked questions about changes to the Behavioral Risk Factor Surveillance System. [cited 2016 Apr 29]. Available from: http://www.cdc.gov/surveillancepractice/reports/brfss/brfss_faqs.html.US Department of Health and Human Services. Oral Health in America: A Report of the Surgeon General. 2000. [cited 2014 July 9]. Available from: http://silk.nih.gov/public/hck1ocv.@www.surgeon.fullrpt.pdf.RomaireMA, BellJF, HuebnerCE. Variations in children’s dental service use based on four national health surveys. Pediatrics. 2012;130(5):e1182–9.23071211LiKY, WongMC, LamKF, SchwarzE. Age, period, and cohort analysis of regular dental care behavior and edentulism: a marginal approach. BMC Oral Health. 2011;11:9.21410991BielerGS, BrownGG, WilliamsRL, BroganDJ. Estimating model-adjusted risks, risk differences, and risk ratios from complex survey data. Am J Epidemiol. 2010;171(5):618–23.20133516Research Triangle Institute. SUDAAN 11.0.0. Research Triangle Park, NC: Research Triangle Institute; 2012.Centers for Disease Control and Prevention. National Health Interview Survey – questionnaires, datasets, and related documentation. [cited 2013 Nov 5]. Available from: http://www.cdc.gov/nchs/nhis/nhis_questionnaires.htm.BhandariA, WagnerT. Self-reported utilization of health care services: improving measurement and accuracy. Med Care Res Rev. 2006;63(2):217–35.16595412LoftusEF, SmithKD, Klinger MR FiedlerJ. Memory mismemory for health events. In: TanurJM, editor. Questions about questions inquiries into cognitive bases of surveys. New York: Russell Sage Foundation; 1992.Centers for Disease Control and Prevention. Comparability of data: BRFSS 2010. [cited 2014 Aug 18]. Available from: http://www.cdc.gov/brfss/annual_data/2010/compare_10.pdf.GilbertGH, RoseJS, SheltonBJ. A prospective study of the validity of data on self-reported dental visits. Community Dent Oral Epidemiol. 2002;30(5):352–62.12236826National Center for Health Statistics. Health, United States,2013. 2014. [cited 2016 May 19]. Available from: http://www.cdc.gov/nchs/data/hus/hus13.pdf.NassehK, VujicicM. Dental care utilization rate continues to increase among children, holds steady among working age adults and the elderly. Health Policy Institute Research Brief. American Dental Association. 2015. [cited 2016 May 19]. Available from: http://www.ada.org/~/media/ADA/Science%20and%20Research/HPI/Files/HPIBrief_1015_1.ashx.

Adjusted prevalence (%)a of having a past-year dental visit (PPYDV) among adults aged 18 years and older, 1999–2010 BRFSSb, MEPS and NHIS and NHANES 1999–2004c.

aPrevalence adjusted for age, race/ethnicity, education level, poverty status, and edentulism except for adjusted prevalence from MEPS 1999, which did not adjust for edentulism due to edentulism not available for MEPS 1999.

bDental visit question was included in the BRFSS emerging core questionnaire in 1999 and rotating core questionnaire in even years since 2002.

cNHANES 1999–2004 was used for comparison with BRFSS 2002; dental visit question not available from NHANES 2005–2010.

BRFSS, Behavioral Risk Factor Surveillance System; MEPS, Medical Expenditure Panel Survey; NHIS, National Health Interview Survey; NHANES, National Health and Nutrition Examination Survey. [Color figure can be viewed at wileyonlinelibrary.com]

Questions Used to Estimate Prevalence of Past-Year Dental Visit, BRFSS, MEPS, NHIS, and NHANES 1999–2010

Survey
BRFSS*MEPS †NHISNHANES ‡
Interview mode
Telephone interviewIn-person interview, 5 roundsIn-person interviewIn-person interview
Years question(s) asked
1999, 2002, 2004, 2006, 2008, 20101999–20101999–20101999–2004
Single or multiple questions used
Single question usedMultiple questions usedSingle question usedSingle question used
Question wording
How long has it been since you last visited a dentist or a dental clinic for any reason? Include visits to dental specialists, such as orthodontists{Since (START DATE)/ Between (START DATE) and (END DATE)}, did (PERSON) see or talk to any type of dental care provider, such as the types listed on this card, for dental care or a dental check-up?About how long has it been since you last saw or talked to a dentist? Include all types of dentists, such as orthodontists, oral surgeons, and all other dental specialists, as well as dental hygienists.About how long has it been since {you/SP} last visited a dentist? Include all types of dentists, such as, orthodontists, oral surgeons, and all other dental specialists, as well as dental hygienists.
{Since (START DATE)/ Between (START DATE) and (END DATE)}, did (PERSON) see or talk to any other type of dental care provider, such as the types listed on this card (other than what you’ve already told me about)?
What type of dental care provider did (PERSON) see during this visit?
What did (PERSON) have done during this visit?
Responses used to determine past year dental visit
Within the past year (anytime less than 12 months ago) (2002, 2004, 2006, 2008, 2010)Reported dental visit at any round of interview during the calendar year6 months or less, or more than 6 months, but not more than 1 year ago6 months or less, or more than 6 months, but not more than 1 year ago
Within the past year (1 to 12 months ago) (1999)

Dental visit question was included in the BRFSS emerging core questionnaire in 1999 and rotating core questionnaire in even years since 2002

Dental care providers in the MEPS question refer to general dentist, dental hygienist, dental technician, dental surgeon, orthodontist, endodontist, periodontist, or other dental specialist

Question and response listed were for NHANES 1999–2004; Dental visit question was not available in NHANES 2005–2010

BRFSS, Behavioral Risk Factor Surveillance System; MEPS, Medical Expenditure Panel Survey; NHIS, National Health Interview Survey; NHANES, National Health and Nutrition Examination Survey

Trends* in Prevalence (%) of Having a Past-Year Dental Visit (PPYDV) Among Adults Aged 18 Years and Older, from 1999 to 2010 BRFSS, MEPS, and NHIS

TrendYearsß coefficient of survey year†P-value†
BRFSS1999–2010−0.00235<0.001
MEPS‡2000–2010−0.00620<0.001
MEPS‡1999–2010−0.00477<0.001
NHIS1999–2010−0.00597<0.001

BRFSS, Behavioral Risk Factor Surveillance System; MEPS, Medical Expenditure Panel Survey; NHIS, National Health Interview Survey.

Trends in PPYDV adjusted for age, race/ethnicity, education level, poverty status, and edentulism.

Linear trend was present if the ß coefficient of the survey year variable was not equal to zero (increasing trend if ß > 0; decreasing trend if ß < 0); statistical significance of the trend was assessed with a t-test.

Because edentulism was not available in MEPS 1999, trend in PPYDV from MEPS 2000–2010 adjusted for all the covariates including edentulism, and trend in PPYDV from MEPS 1999–2010 adjusted for all the covariates excluding edentulism.

Adjusted Prevalence (%)* and 95% Confidence Interval (CI) of Having a Past-Year Dental Visit Among Adults Aged 18 Years and Older, 2002 BRFSS, MEPS and NHIS, and NHANES 1999–2004

BRFSS 2002
MEPS 2002
NHIS 2002
NHANES 1999–2004
Sample sizeAdjusted prevalence A1(%)† (95% CI)Sample sizeAdjusted prevalence B1(%)† (95% CI)Sample sizeAdjusted prevalence C1(%)† (95% CI)Sample sizeAdjusted prevalence D1(%)† (95% CI)
Total236,45770.0 (69.7, 70.4)26,36443.9 (42.8, 44.9)30,50261.5 (60.8, 62.2)16,99758.1 (56.5, 59.8)
Age (years)
 18–44104,42167.9 (67.4, 68.4)13,77239.5 (38.1, 40.9)15,48359.0 (58.1, 60.0)8,39356.8 (54.8, 58.8)
 45–6483,35370.7 (70.1, 71.3)8,26845.9 (44.5, 47.3)9,30863.0 (61.9, 64.1)4,29561.1 (58.6, 63.7)
 65+48,68372.7 (71.8, 73.5)4,32450.9 (48.8, 53.1)5,71166.3 (64.8, 67.7)4,30962.9 (59.7, 66.2)
Race/ethnicity
 NHW†188,86670.9 (70.6, 71.3)15,44747.5 (46.2, 48.9)20,14463.0 (62.1, 63.8)8,28361.1 (58.8, 63.4)
 NHB†17,26065.2 (64.0, 66.5)3,63028.9 (26.9, 31.1)4,06854.8 (53.1, 56.7)3,44353.3 (50.7, 56.1)
 Hispanic†13,34065.9 (64.2, 67.6)5,78930.0 (28.1, 32.1)5,18357.9 (56.0, 59.8)3,99748.5 (44.0, 53.6)
Education level
 Less than high school25,16754.0 (52.5, 55.5)6,93430.5 (28.7, 32.4)5,81543.7 (42.1, 45.4)5,85444.5 (41.1, 48.1)
 High school graduate74,76467.2 (66.5, 67.8)12,71940.9 (39.6, 42.2)8,70556.4 (55.2, 57.5)4,16954.4 (52.1, 56.9)
 More than high school136,03773.4 (72.9, 73.9)6,59552.6 (51.3, 54.0)15,69768.2 (67.4, 69.1)6,93466.1 (64.1, 68.1)
Poverty status
 Poor†18,83957.8 (56.2, 59.6)3,83731.1 (29.0, 33.3)4,71851.9 (49.7, 54.3)3,14751.3 (48.0, 54.8)
 Near-poor†48,48161.9 (61.1, 62.7)5,51733.6 (31.7, 35.6)6,05152.5 (50.9, 54.0)4,09248.6 (46.1, 51.3)
 Non-poor†135,63573.5 (73.0, 73.9)17,01046.4 (45.1, 47.7)19,73365.1 (64.3, 65.9)8,09863.0 (61.0, 65.1)
Edentulism
 Yes19,51727.5 (26.3, 28.8)2,59119.8 (18.0, 21.9)2,85826.7 (24.7, 28.8)1,33119.8 (17.3, 22.7)
 No213,52872.1 (71.7, 72.5)23,36246.1 (44.9, 47.2)27,61364.6 (63.9, 65.4)13,28362.0 (60.3, 63.8)

BRFSS, Behavioral Risk Factor Surveillance System; MEPS, Medical Expenditure Panel Survey; NHIS, National Health Interview Survey; NHANES, National Health and Nutrition Examination Survey.

Prevalence adjusted for age, race/ethnicity, education level, poverty status, and edentulism.

A1, B1, C1, and D1: adjusted prevalence from 2002 BRFSS, 2002 MEPS, 2002 NHIS, and NHANES 1999–2004, respectively; NHW: Non–Hispanic White; NHB: Non–Hispanic Black; Hispanic: Mexican-American only for NHANES 1999–2004; Poor: <100% Federal Poverty Level (FPL); Near-poor: 100–199% FPL; Non–poor: ≥200% FPL.

Adjusted Prevalence (%)* and 95% Confidence Interval (CI) of Having a Past-Year Dental Visit Among Adults Aged 18 Years and Older, 2010 BRFSS, MEPS, and NHIS

BRFSS 2010
MEPS 2010
NHIS 2010
Sample sizeAdjusted prevalence A2(%)† (95% CI)Sample sizeAdjusted prevalence B2(%)† (95% CI)Sample sizeAdjusted prevalence C2(%)† (95% CI)
Total441,60767.5 (67.3, 67.8)22,55939.7 (38.5, 40.8)26,72559.7 (58.9, 60.5)
Age (years)
 18–44103,85766.4(65.9, 66.9)11,14435.9 (34.2, 37.6)12,40257.4 (56.4, 58.5)
 45–64189,43768.6 (68.3, 69.0)7,69542.1 (40.4, 43.8)8,97261.1 (59.9, 62.3)
 65+148,31372.2 (71.8, 72.6)3,72048.0 (45.7, 50.4)5,35164.4 (62.7, 66.1)
Race/ethnicity
 NHW†349,29469.4 (69.1, 69.8)10,64143.6 (42.1, 45.1)15,35661.4 (60.4, 62.4)
 NHB†34,63463.8 (62.8, 64.9)4,25731.6 (29.5, 33.9)4,41955.8 (54.0, 57.6)
 Hispanic27,87566.0 (64.8, 67.3)5,46229.3 (26.8, 32.1)5,07355.4 (53.6, 57.2)
Education level
 Less than high school40,90553.8 (52.5, 55.1)5,01128.6 (26.4, 30.9)4,56544.7 (42.7, 46.9)
 High school graduate131,35863.1 (62.5, 63.7)10,79035.2 (33.7, 36.8)7,05053.1 (51.7, 54.5)
 More than high school267,85471.9 (71.6, 72.3)6,56749.4 (47.9, 50.8)15,00065.7 (64.7, 66.7)
Poverty status
 Poor†43,02549.6 (48.5, 50.7)4,02028.6 (26.1, 31.3)4,78045.8 (43.9, 47.8)
 Near-Poor†78,53758.0 (57.2, 58.8)4,89531.0 (29.1, 33.0)5,70448.4 (46.8, 50.1)
 Non-Poor†257,96874.7 (74.3, 75.1)13,64443.6 (42.1, 45.1)16,24166.1 (65.1, 67.0)
Edentulism
 Yes38,17228.2 (27.3, 29.2)1,67620.6 (16.8, 25.2)2,29430.1 (27.8, 32.5)
 No395,29670.0 (69.7, 70.3)20,41242.2 (40.9, 43.6)24,42362.6 (61.7, 63.5)

BRFSS, Behavioral Risk Factor Surveillance System; MEPS, Medical Expenditure Panel Survey; NHIS, National Health Interview Survey.

Prevalence adjusted for age, race/ethnicity, education level, poverty status, and edentulism; the dental visit question was not available in NHANES 2005–2010.

A2, B2, and C2: adjusted prevalence from 2010 BRFSS, MEPS, and NHIS, respectively; NHW: Non–Hispanic White; NHB: Non–Hispanic Black; Poor: <100% Federal Poverty Level (FPL); Near-poor: 100–199% FPL; Non–poor: ≥200% FPL.

Absolute Difference (AD) (%) and 95% Confidence Interval (CI) in Adjusted Prevalence of Having a Past-Year Dental Visit Among Adults Aged 18 Years and Older, Between BRFSS and MEPS, NHIS and NHANES*

2002 BRFSS–2002 MEPS2002 BRFSS–2002 NHIS2002 BRFSS–1999–2004 NHANES2010 BRFSS–2010 MEPS2010 BRFSS–2010 NHIS
AD (95% CI) AD = A1-B1†AD (95% CI) AD = A1-C1 †AD (95% CI) AD = A1-D1 †AD (95% CI) AD = A2-B2†AD (95% CI) AD = A2-C2†
Total26.2 (25.0, 27.3)8.5 (7.8, 9.3)12.0 (10.3, 13.6)27.9 (26.7, 29.1)7.8 (7.0, 8.7)
Age (years)
 18–4428.4 (26.9, 29.9)8.9 (7.8, 9.9)11.1 (9.1, 13.1)30.5 (28.8, 32.3)8.9 (7.7, 10.1)
 45–6424.8 (23.3, 26.3)7.7 (6.5, 8.9)9.6 (7.0, 12.1)26.6 (24.9, 28.3)7.5 (6.3, 8.8)
 65+21.8 (19.4, 24.1)6.4 (4.7, 8.1)9.8 (6.5, 13.1)24.2 (21.8, 26.6)7.8 (6.1, 9.6)
Race/ethnicity
 NHW†23.4 (22.0, 24.8)8.0 (7.0, 8.9)9.9 (7.6, 12.2)25.9 (24.3, 27.4)8.1 (7.0, 9.1)
 NHB†36.3 (33.9, 38.7)10.4 (8.2, 12.5)11.9 (9.0, 14.8)32.2 (29.8, 34.7)8.1 (6.0, 10.2)
 Hispanic†35.9 (33.2, 38.5)8.0 (5.4, 10.5)17.4 (12.4, 22.3)36.7 (33.8, 39.6)10.6 (8.4, 12.8)
Education level
 Less than high school23.5 (21.1, 25.8)10.3 (8.1, 12.4)9.5 (5.8, 13.2)25.2 (22.6, 27.8)9.0 (6.5, 11.5)
 High school graduate26.3 (24.8, 27.7)10.8 (9.5, 12.1)12.7 (10.3, 15.2)27.9 (26.3, 29.5)10.0 (8.5, 11.6)
 More than high school20.8 (19.3, 22.2)5.2 (4.2, 6.1)7.3 (5.3, 9.3)22.6 (21.1, 24.0)6.2 (5.2, 7.3)
Poverty status
 Poor†26.8 (24.0, 29.5)5.9 (3.1, 8.8)6.6 (2.9, 10.3)21.0 (18.2, 23.9)3.8 (1.6, 6.0)
 Near-poor†28.3 (26.2, 30.4)9.4 (7.7, 11.2)13.3 (10.6, 15.9)27.0 (24.9, 29.1)9.6 (7.7, 11.4)
 Non-poor†27.1 (25.7, 28.5)8.4 (7.4, 9.3)10.5 (8.4, 12.5)31.1 (29.6, 32.7)8.6 (7.6, 9.7)
Edentulism
 Yes7.7 (5.4, 10.0)0.9 (−1.5, 3.2)7.7 (4.8, 10.6)7.6 (3.4, 11.9)−1.9 (−4.4, 0.7)
 No26.0 (24.8, 27.3)7.5 (6.7, 8.3)10.1 (8.3, 11.8)27.8 (26.5, 29.2)7.4 (6.5, 8.4)

BRFSS, Behavioral Risk Factor Surveillance System; MEPS, Medical Expenditure Panel Survey; NHIS, National Health Interview Survey; NHANES, National Health and Nutrition Examination Survey.

Prevalence adjusted for age, race/ethnicity, education level, poverty status, and edentulism; the dental visit question was not available in NHANES 2005–2010; all ADs in adjusted prevalence between BRFSS and each of other three surveys MEPS, NHIS, and NHANES were statistically significant (P ≤ 0.05) except for AD between BRFSS and NHIS among edentate adults.

A1, B1, C1, and D1: adjusted prevalence from 2002 BRFSS, 2002 MEPS, 2002 NHIS, and NHANES 1999–2004 respectively; A2, B2, and C2: adjusted prevalence from 2010 BRFSS, MEPS, and NHIS, respectively; NHW: Non–Hispanic White; NHB: Non–Hispanic Black; Hispanic: Mexican-American only for NHANES 1999–2004; Poor: <100% Federal Poverty Level (FPL); Near-poor: 100–199% FPL; Non–poor: ≥200% FPL.