Dental Care Utilization among Veterans by Smoking Status

Background Given the critical importance of dental care utilization among veterans and the overall health consequences of tobacco use in all populations, the purpose of this research is to examine smoking as a risk factor for poor dental care utilization among United States Veterans. Methods A secondary data analysis of cross-sectional data from the National Survey of Veterans was conducted. The primary outcome was dental care utilization (Yes, No). Frequency, chi-square analyses, and multivariate logistic regression statistical tests were performed while adjusting for confounding factors. Results There were 6,308 veterans in the study. Veterans who were current smokers were less likely to have dental care utilization within the previous six months than former smokers or never smokers. In unadjusted logistic regression analysis, current smokers had an odds ratio of 2.83 [95% CI: 2.36, 3.40] as compared with never smokers. The adjusted odds ratio for current smoking on dental care utilization was 1.71 [95% CI: 1.40, 2.09] as compared with never smoking. Conclusions Since veterans who smoked are less likely to have dental care utilization within the previous six months, they are at higher risk for later diagnosis of dental problems. Veterans who smoke should be specifically targeted with interventions to ensure frequent dental visits, so future problems may be averted or managed early in their development.


Introduction
Keeping veterans healthy is a priority for the Veterans Health Administration. Smoking is a known risk factor for poor health. In the past, the military had a history and culture supporting smoking. Many veterans began smoking during their service. Although 57% of the U.S. population reports never smoking, only 32% of veterans are never smokers [1,2]. However, the percentage of veterans and nonveterans who are current smokers is similar (approximately 20%) [2]. ere were 19.9 million veterans in the United States in 2017 [1]. Of these veterans, there were 14.8 million under age 65 years, 2.9 million aged 65-69 years, and 2.2 million who were aged 70-74 years; 46.2% were members of the Army (9.2 million) and 81.9% were White (16.3 million) [1]. A total of 6.07 million veterans (31.1%) received care from the Veterans Health Administration (VHA) in 2017 [1]. Veteran healthcare expenditures are categorized in priorities from 1 to 8 with several subsections within each priority. e average healthcare expediture in 2014 varied from $1,111 in priority 8 (subsection G) to $11,702 in priority 1, whereas the average healthcare expenditure for nonveterans was $1,068 [1].
A modifiable behavior, such as smoking, is a risk factor for chronic conditions including metabolic disorders, cardiovascular diseases, and poor oral health [3]. In a systematic review, researchers found a dose-response relationship between smoking and tooth loss [4]. Smoking was identified as an independent risk factor for the development and severity of periodontal disease and tooth loss by other researchers [5][6][7]. Most periodontal disease cases in the U.S. are associated with smoking [8,9]. Smoking is implicated as a component cause in over three-quarters of oral cancers [10], significantly increasing the risk of oral cancer among smokers [11,12].
Periodic, regular, and professional dental visits help dentists to prevent, control, delay, or treat dental problems [13]. In a longitudinal study of over 5,000 participants, associations between preventive dental care service utilization and fewer tooth losses were observed [14]. Participants who reported smoking or diabetes had lower frequency of tooth loss with biannual dental visits compared with participants who smoked and only had annual visits. Additionally, epidemiologists suggested a link between less dental care service utilization and advanced stage of oral cancer at diagnosis [15,16]. e main nationally representative data sources for dental care utilization are the Medical Expenditure Panel Survey (MEPS), the National Health and Nutrition Examination Survey (NHANES), and the National Health Interview Survey (NHIS). While estimates of dental care utilization differ among these data sources [17,18], they remain useful, especially for national trends. Adult dental utilization had been declining from 2003 to 2012 due to a decline in the number of adults with private dental benefits; however, it was stabilized in 2013-2014 [19]. e stabilization may be the result of provisions of the Affordable Care Act in which Medicaid eligibility was expanded and included dental benefits to adults in several states [19].
Within national data sets, there is a lack of information on U.S. veterans. A U.S. veteran is any person who served on active duty in the U.S. Armed Forces (Army, Navy, Air Force, Marine Corps, and Coast Guard). e researchers of existent studies have consistently identified dental care as a critical need for veterans. For example, researchers in a Massachusetts county identified oral health as the greatest unmet need among veterans [20]. Researchers conducting a veterans' needs assessment in California showed dental care among their top three needs [21]. Researchers in a Connecticut county also identified dental care as a significant healthcare concern among veterans [22]. In a study conducted using data from the Veterans Health Administration, researchers concluded that half had periodontitis [13]. Additionally, there are metropolitan/ rural differences with rural veterans being less likely to utilize dental services and more likely to have poorer oral health outcomes [23].
Dental emergencies have a negative effect on military operations, should someone be recalled to duty [24]. e Department of Defense identified regular dental visits as critical to individual medical readiness. Given the critical importance of dental care utilization among veterans and the overall health consequences of tobacco use among veterans, the purpose of this population-based analysis is to examine smoking as a risk factor for poor dental care utilization among United States veterans.

eoretical Framework.
e Andersen Expanded Behavioral Model is a theoretical model which is used to explain general healthcare utilization and the risk factors that influence healthcare utilization. It was chosen as the theoretical framework for this study on dental care utilization. Andersen proposed categories of risk factors which influence healthcare utilization. ese are predisposing factors which are generally immutable factors (such as sex, age, and race/ethnicity), enabling factors (such as marital status, education, insurance, and income), need factors (such as health status), and personal health/behavioral practices (such as smoking, drinking, physical activity, and obesity) [25]. Based upon the healthcare utilization model, dental utilization is hypothesized to have similar factors. Smoking, as a personal health/behavioral practice, is being hypothesized as a major contributor to poor dental utilization and fewer dental visits. e Andersen model was used to select the other explanatory or potentially confounding variables in the study (predisposing factors such as sex, age, and race/ ethnicity were included; enabling factors such as income, marital status, education, and insurance were included; and the need factor of self-reported health was included). It should be noted that due to (1) data availability, (2) the specific questions posed in the data source, and (3) degrees of freedom, all explanatory or potentially confounding variables could not be included in the final analysis. e rationale for this study is that previous dental utilization studies concerning smoking focused on other groups. For example, U.S. youths in grades 9-12 were studied with results showing an association with smoking and dental visits in the previous 12 months with an adjusted odds ratio of 0.75 [26]. Also, a study of children aged 10-17 years living in a household with smokers less likely to have a dental visit within the previous year was reported [27]. It is important to determine if a similar pattern exists for veterans [28].

Statistical Methodology.
e analyses included frequency analyses and weighted percentage estimates of the sample. Bivariate analyses (chi-square analyses) were conducted to determine associations between smoking status, sex, race/ethnicity, age, marital status, education, income, insurance, and self-reported health status with dental care utilization. e level of significance selected, a priori, was 0.05. Unadjusted logistic regression analysis and adjusted logistic regression analyses were conducted on dental care utilization by smoking status. e adjusted logistic regression modeling was conducted controlling for sex, race/ethnicity, age, marital status, education, income, insurance, health selfperception, and oral health perception. Due to the number of participants not reporting their type of address and the difficulty in interpreting rurality from the manner in which the question was posed, the type of address was not included in the logistic regression analyses. Due to small sample size of veterans who were females and the small sample size of veterans who were not White, two sensitivity analyses were conducted excluding race as a covariate in one of the analyses and sex as a covariate in the other. All statistical analyses incorporated the weightings and accounted for the complex design of NSV2010. Analyses were conducted with SAS 9.3 (Carey, NC, USA). All independent variables were entered in a single step in the logistic regression model. Variables were selected and included in the model based on their importance to the dental visits shown in the literature.

Study Design and Data Source.
is study received an acknowledgment as nonhuman subject research (protocol 1511920072) from the researchers' academic institutional review board. A cross-sectional study design was used involving secondary data analyses of previously collected public data from the National Survey of Veterans, 2010 (NSV2010) [28]. e primary aim of the NSV2010 data source was to survey veterans with the purpose of planning and allocating resources for veterans [28]. e NSV2010 was the sixth, and most recent, comprehensive nationwide survey for the Department of Veteran Affairs [28].
e NSV2010 researchers used a cross-sectional research design for their study of noninstitutionalized veterans [28]. ey used two sampling approaches: an address-based sampling (ABS) and a list-based sampling approach [28]. Potential addressbased participants received a notification, a screening letter, and, if eligible, were sampled for either the mailed or web-based survey. Potential list-based participants received personal contact letters [28].
ere were 10,972 surveys mailed and 8,710 surveys returned (66.7% response rate). e surveys were weighted for probability of selection/ nonresponse [28]. Data were poststratified to known population totals to represent the noninstitutional veteran population [28].

Study Sample.
is study included veterans, aged 20 years and above, from the NSV2010 study who had complete data on the dental visit, smoking status, sex, race, age, marital status, income, insurance adequacy, and selfreported health status. e final study sample included the data from 6,308 veterans.

Key Dependent Variable.
e key variable was dental care utilization within the previous six months (Yes, No). Information for this variable was gathered from the NSV2010 question "In the last six months, have you had any dental care or visited a dentist?" [28]. is was the only dental utilization question posed on the NSV2010 questionnaire.

Key Independent Variable.
e key independent variable was smoking status. Participants were asked in the NSV2010 if they had smoked at least 100 cigarettes in their lifetime.
e question had dichotomous response categories of "Yes" (to indicate current or former smokers) or "No" (to indicate never smokers). A follow-up question was posed to respondents who endorsed "Yes" to the previous question. e veteran was asked if he or she smoked cigarettes now with "Yes" being used to identify a current smoker and "No" being used to identify a former smoker. Both questions were used to create the smoking variable (current smoker, past smoker, and nonsmoker) for this current study.

Other Variables.
Other variables are known to be associated with access and utilization of dental services and/ or be potential explanatory or confounding variables. Of these potential variables, the following were available to be included in the analyses: sex (categorized as male or female); race/ethnicity (categorized as White or Other due to the small number of minorities in the survey); age in years (50 years and above, 35 years to less than 50 years , and 20 years to less than 35 years as the 2010 survey did not have participants who were born before 1991; that is, it did not include participants ages 18 years to less than 20 years); marital status (categorized as married or not married); education (categorized as high school graduate or less, some college, college degree or more); income (categorized as less than $20,000, $20,000 to less than $30,000, or $30,000 and above); insurance (categorized as adequate, or inadequate: based on the question "my family has a health insurance plan that adequately covers me and my family" [25] in which the responses "completely agree, agree, and neither agree nor disagree" were categorized as adequate and "disagree and completely disagree" were categorized as inadequate); health self-perception (categorized as excellent/very good/good or fair/poor based on the question, "In general, would you say your health is ... excellent; very good; good; fair; poor?"); oral health perception (categorized as excellent/very good/good, or fair/poor: based on the question, "How would you rate the health of your teeth and gums? Would you say it is excellent; very good; good; fair; poor?"); and mailing address (categorized as rural route, street, U.S.PO box/other box, or missing based on the question, "At which of the following types of addresses does your household receive mail? A street address with a house or building number, an address with a rural route number, a U.S. Post Office box, a commercial mailbox establishment)" [25].

Results
ere were 6,308 veterans in this study sample. Most were male (91.6%) and White (86.9%). Almost two-thirds (64.5%) were current or past smokers and 35.4% reported never smoking.
ere were 55.3% who reported having a dental visit within the previous six months. Details are presented in Table 1.
e Rao-Scott chi-square test results for dental visits within the previous six months and smoking and other variables are presented in Table 2. Veterans who were current smokers were less likely to have dental care utilization within the previous six months than former or never smokers. Veterans who reported a dental care utilization in the previous six months were more likely to be white, older (50 years of age or older), married, have a college degree or higher, have an annual income of $30,000 International Journal of Dentistry or higher, and have adequate insurance and were more likely to report excellent/very good/good general and oral health perception. e unadjusted odds ratio (OR) and adjusted odds ratio (AOR) with their confidence intervals (CIs) of smoking on no report of a dental care utilization within the previous six months are presented in Table 3. In unadjusted logistic regression analysis, current smokers were more likely to not have had dental care utilization as nonsmokers (OR 2.81; 95% CI: 2.34, 3.38; P < 0.0001). After controlling for other risk factors and potential confounders, the odds ratio of a current smoker not having dental care utilization in the previous six months as compared with nonsmokers remained statistically significant (AOR 1.50; 95% CI: 1.22, 1.84; P < 0.0001). e OR and AOR for the comparison of former smokers with never smokers failed to reach significance. Veterans who did not have dental care utilization within the previous six months were more likely to be other than white (AOR 1.37; 95% CI: 1.08, 1.74); not married as compared with married (AOR 1.24; 95% CI: 1.06, 1.47); educated through high school or less as compared with being educated to college or more (AOR 2.41; 95% CI: 2.05, 2.83); and to have a family income less than $20,000 as compared with $30,000 or above (AOR 2.69; 95% CI: 2.14, 3.38). Other results are presented in Table 3. ree sensitivity analyses to determine covariate effects were conducted. e results were similar to the results presented in Table 3. In the first sensitivity analysis, race/ethnicity was not included among the covariates. In the second analysis, sex was not included among the covariates. In the third sensitivity analysis, self-reported health was not included among the covariates. ere were significant associations between smoking categories and education categories in subgroup analysis where the reference categories were never smoking and college degree or beyond. e only exception was that the association between former smoker and college degree or beyond failed to reach statistical significance.

Discussion
In a large nationally representative sample of U.S. veterans, we found that veterans who were current smokers were more likely not to have dental care utilization within the previous six months as compared with never smoking veterans [AOR, 1.50; 95% CI: 1.22,1.84]. Only 37.0% of veterans who were current smokers reported having dental care utilization within the previous six months compared with 62.3% of never smoking veterans who reported having a dental care utilization within the previous six months (P < 0.0001). ese results are consistent with the Andersen Expanded Behavioral Model for overall healthcare utilization that explains healthcare utilization as resulting from several factors, among which are personal health/behavioral practices such as smoking. In this study, the authors hypothesized that smoking, as a personal health/behavioral practice, is a factor in poor dental utilization and fewer dental visits for veterans based upon previous studies of other groups.
While there is a lack of similar studies among veterans, this study's results corroborate findings from research studies not specific to veterans. For example, researchers of a study of 15,250 U.S. adults, using MEPS 2000 data, reported that current smokers were less likely to have had a dental visit within the previous year than nonsmokers [AOR 0.78; 95% CI: 0.69, 0.88] [29]. Our study's results also support the findings from another study of 2,119 U.S. adults that showed that long-term smokers were less likely to have had dental care utilization in the previous year than never smokers [AOR: 0.69; 95% CI: 0.48, 0.99] [12].
Bloom et al. [30] reported that current smokers were more likely to delay dental visits. In a study using Behavioral  (1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008), never smokers reported a higher median annual dental visit as compared with former smokers (3% difference) [31]. Current smokers were also more likely to visit only in case of dental emergency [9]. However, some researchers reported results dissimilar to this study. In a Japanese study, current smokers had more dental visits within the previous year as compared with nonsmokers (P � 0.0003) [32]. Using 2014 BRFSS data, researchers found that smokers who were attempting to quit were more likely to have a recent dental visit [33].
More research is needed to understand the smokingdental care utilization relationship. Statistically significant reasons for low dental care utilization include dental anxiety and financial barriers [34]. Specifically for U.S. veterans, dental coverage eligibility requirements are stringent [35]. Any needed dental care is provided to service-connected dental disability, to disabilities which are 100% disabling, or to veterans who were former prisoners of war [35]. However, exclusions apply for veterans who do not meet those criteria [35]. e implication is that access to needed dental care is limited.
e Office of the Actuary, the Department of Veterans Affairs has projected a steeper growth in the female, minority, and younger age veterans between 2010 and 2040 [36]. Given a new profile in the coming decades, researchers will need to reassess veterans' oral health needs.
e Department of Veterans Affairs policy makers may International Journal of Dentistry want to incorporate features of successful dental service models across the country that proactively supports Veterans [37].
Dentists and dental healthcare professionals are in the position and have the ability to help with tobacco cessation [38] as many people have routine dental utilization. However, access to this specific population is problematic due to the lower dental utilization. Although not a focus of this study, higher education increased the likelihood of veterans attending dental visits and higher income increased the likelihood of veterans attending dental visits. erefore, it is particularly important to target dental public health messages concerning the importance of dental visits to the groups who would most benefit from that knowledge and encourage them to have routine dental care.

Strengths and Limitations.
is study has several strengths. Researchers used a large, nationally representative sample of U.S. veterans in which weights were applied to maintain population estimates in the data analyses. e survey questions made it possible to determine participants who were current smokers, former smokers, and who never smoked. e authors were able to indicate the independent effect of smoking, while controlling for other potentially confounding variables. e results that veterans who smoke are less likely to have dental care utilization within the previous six months may be helpful to dental care providers and policy makers to create focused interventions. e veteran health is a major concern and access to quality care is challenging. ere have been many changes since the NSV2010 was completed. e NSV2010 was the most current of the NSV surveys to use.
ere is a need for continued surveillance in this important group.
As a cross-sectional study, the results cannot be interpreted as causal. In addition, the data on both key variables, dental care utilization and smoking status, were selfreported. Since smoking is socially unacceptable, participants may have underreported their smoking status. e dental care utilization data included an aggregation of all dental services. Since the researchers could not distinguish among the services, the assessment of the frequency of specific dental visits could not be determined. e dental utilization question also had a 6-month time frame imposed upon it. Health self-perception was used as specific comorbidities that were not available in the original dataset; however, in sensitivity analysis excluding overall health, the results remained similar to those reported. Additionally, other covariates (metropolitan versus rural status, number of missing teeth) would have strengthened the study but also were not available. However, in the primary dataset, the purpose was not to capture co-morbidities, so questions about hypertension, diabetes, cardiovascular disease, obesity, etc. were not posed to the participants.

Conclusion
Based upon the Andersen eoretical Model, in this study of 6,308 veterans, those who smoked were more likely to not have dental care utilization within the previous six months, indicating that personal health/behavioral factors influence not only general health but also oral health. ese findings have implications for oral conditions. Since smokers are less likely to have routine dental visits, they are at higher risk for late diagnosis of their dental problems or other oral conditions such as potentially malignant and malignant lesions [39]. Veterans who smoke should be specifically targeted with promotional messages regarding oral healthcare utilization.

Data Availability
Previously reported 2010 National Survey of Veterans data were used to support this study and are available at https:// catalog.data.gov/dataset/national-survey-of-veterans-baea8. is dataset is cited at relevant places within the text as Reference [28].

Additional Points
Practical Implications. Population subgroups often have factors which influence healthcare more so than in the population at large. e association of smoking and dental visits is such an example in the veteran subgroup. Disclosure e content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. e funders had no role in the study design, data collection, analysis, decision to publish, or preparation of the manuscript.

Conflicts of Interest
e authors have no conflicts of interest to declare.

Authors' Contributions
R. Constance Wiener developed the concept, conducted the data analyses, and wrote the first draft. Ruchi Bhandari, Alcinda Shockey, and Christopher Waters reviewed the data analyses, contributed to the writing and editing of the drafts, and approved the final version of the manuscript.