Dental insurance has been shown to be an important factor in people’s decisions to seek and use dental care services.^5–13 According to the 1989 National Health Interview Survey, or NHIS, while 70 percent of all people with dental insurance had at least one dental visit during 1989, only 50 percent of people without dental insurance reported at least one visit during the same period.⁵ Dental insurance also has been shown to affect dental provider treatment patterns.^12,13 For instance, Mueller and Monheit¹³ found that having coverage doubles the likelihood that a user will obtain bridgework and increases by 38 percent the probability that a user will receive a crown. According to Conrad and colleagues¹⁰ and Mueller and Monheit,¹³ while increased coverage among elderly patients did not result in uniform increased costs, it did result in a substitution in services provided. Since the effect of having insurance is to reduce the perceived price of care, it is not surprising that patients with coverage might substitute higher-cost treatment options in place of lower-cost, and perhaps less-preferred, alternatives.⁷

Dental insurance has had a profound effect on dentistry and dental care use. Insurance payments for dental care on behalf of enrollees are significant. During 1996, approximately 43 percent of all dental expenditures ($18 billion) were paid to dental providers from private insurance.¹⁴

We conducted this study to further examine the status of private dental insurance and the impact of private dental insurance coverage on the use of and expenditures for dental care services. To do this, we analyzed data from the 1996 Medical Expenditure Panel Survey, or MEPS.¹⁵

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
The U.S. Department of Health and Human Services has sponsored the administration of several national expenditure surveys since 1977. The 1996 MEPS is the third in a series of nationally representative health surveys of the U.S. community-based population sponsored by the Agency for Healthcare Research and Quality (formerly the Agency for Health Care Policy and Research and, before that, the National Center for Health Services Research). MEPS collects health care expenditure, use and payment source data, as well as socioeconomic, demographic and health insurance data similar to its previous surveys. It differs from the National Medical Expenditure Survey and the National Medical Care Expenditure Survey in that data on household respondents in each panel are collected for two consecutive years and the survey is fielded continuously (that is, a new panel is selected every day).¹⁵

The sample for the 1996 MEPS was 21,571 people in 10,500 households who participated in NHIS. To collect health expenditure and use data for 1996, field researchers interviewed each MEPS household in person three times over an 18-month period; the third round of interviews were conducted between February and May 1997.¹⁵ The combined full-year 1996 response rate of the MEPS sample through the third round was 70 percent.¹⁵

In this analysis, we focus on dental coverage and the effect of coverage on dental care use in 1996 for the U.S. noninstitutionalized population. Specifically, we provide national estimates for dental coverage status, percentage of population with a dental visit, number of visits per person for those who had a visit, and mean total expenditure for people with a visit by coverage for each of several socioeconomic and demographic categories during 1996.

We also conducted multivariate analyses to determine if the bivariate relationships shown in Table 1 would persist in a multivariate statistical model. We also assessed the relative impact of the socioeconomic and demographic variables on dental coverage status and utilization. Specifically, we estimated logistic regression models for dental coverage status (defined as = 1, if covered any time during the year; = 0, otherwise) and dental use (defined as = 1, if one or more dental visits during the year; = 0, otherwise). We estimated the impact of socioeconomic and demographic variables for age (18 years and younger, 19–44 years, 45–64 years and 65 years and older), sex, race/ethnicity (non-Hispanic black, Hispanic and white), family income (low income, middle income and high income), education and rural/urban status for the likelihood of having coverage. We then estimated the impact of coverage—controlling for age, sex, race/ethnicity, family income, education and rural/urban status—with two-part models for the likelihood of having a dental visit, the natural logarithm of the number of visits for those with a visit and the natural logarithm of expenditures for those with a visit.^16,17 Economic models of demand developed by Grossman¹⁸ and Phelps and Newhouse¹⁹ support the empirical specification of noncoverage variables in the model. To minimize the effect of nonprivate or public coverage, we truncated regression models to include people with incomes greater than 100 percent of the federal poverty level, or FPL.

We obtained coverage from the Health Insurance Section of the Household Component of MEPS. Private dental coverage refers to coverage offered as a supplement to medical coverage or as coverage obtained from an unrelated and separate plan. Distinctions between these types of dental coverage offerings are not discernible with these data. Additionally, since some medical coverage plans may include dental benefits that were not reported by respondents, these estimates may understate somewhat the actual number of people with coverage.

To ensure sufficient numbers to produce reliable national estimates, we combined sociodemographic variable categories when necessary. We computed all estimates and statistics reported taking into account the complex sampling design of MEPS with the use of a software package (SUDAAN, release 6.40, Research Triangle Institute, Research Triangle Park, N.C.).²¹

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
There were 21,571 participants in the 1996 MEPS, representing approximately 268 million noninstitutionalized U.S. citizens. Of these, 51 percent were female (n = 11,282), 39 percent were nonwhite (n = 8,323), 58 percent were between the ages of 19 and 64 (n = 12,427), and 81 percent were in families with incomes above the FPL (n = 17,415).

Figure 1 shows the percentage of the population with coverage according to family income level. Figure 2 displays the percentage of the population with a dental visit for those with and without coverage by family income level. People in poor and low-income families were less likely (P < .05) to have dental coverage than were people in families with higher incomes. For those with coverage, people in poor and low-income families were the least likely (P > .05) to report a dental visit among all income groups. At each family income level, people without coverage were less likely (P > .05) to report a dental visit than were people with coverage.

View larger version (33K): [in this window] [in a new window]   Figure 1. U.S. population with private dental coverage by income, 1996. The population includes people in families with negative incomes. Poor: incomes below 100 percent of the federal poverty level, or FPL. Low income: incomes 100 to 199 percent of the FPL. Middle income: incomes 200 to 399 percent of the FPL. High income: incomes 400 percent of the FPL and higher. Source: Cohen.15

View larger version (40K): [in this window] [in a new window]   Figure 2. U.S. population with a dental visit, by income and private dental coverage, 1996. The population includes people in families with negative incomes. Poor: incomes below 100 percent of the federal poverty level, or FPL. Low income: incomes 100 to 199 percent of the FPL. Middle income: incomes 200 to 399 percent of the FPL. High income: incomes 400 percent of the FPL and higher. Source: Cohen.15

We noted differences in coverage rates for the race, income, age and education categories, but sex did not appear to affect the likelihood (P > .05) of coverage. Nonwhites, people in poorer families and the elderly (those 65 years of age and older) were less likely (P < .05) to have dental coverage than were whites, younger people or people in families with higher incomes. In addition, college graduates were more likely (P < .05) to have coverage than were high school graduates. Both college graduates and high-school graduates were more likely (P < .05) to have coverage than were people with less formal education (less than high-school education).

For each demographic and socioeconomic category, people with coverage were much more likely (P < .05) to have a dental visit than were people without coverage. The effect of coverage on visits per person and mean dental expenditures were mixed. For instance, coverage did not have an effect (P > .05) on visits per person in families at low, middle or high income levels and for children (18 years of age and younger), college graduates, Hispanics, people with limited formal education (less than high-school graduate) or people who resided in large metropolitan areas or non-metropolitan areas. For other demographic and socioeconomic categories, people with coverage reported more (P < .05) visits per year than did people without coverage. Coverage did not have an effect (P > .05) on mean dental expenditures for college graduates, non-Hispanic blacks, people at the poor family income level, or people who resided in nonmetropolitan areas. For other demographic and socioeconomic categories, people with coverage had higher (P < .05) mean total expenditures than did people without coverage.

Table 2 shows the results of the logistic regression analyses for the probability of having private dental care coverage (second and third columns) and for the probability of having at least one dental visit during the year (fourth and fifth columns). Table 2 also shows the ordinary least squares, or OLS, regression results for the natural logarithm of dental visits for people with at least one visit (sixth and seventh columns) and for the natural logarithm of expenditures for people with expenditures (eighth and ninth columns) during 1996. The explanatory variables in the coverage equation in the second and third columns include age, sex, race and ethnicity, family income level, education and rural/urban status. The other three equations in Table 2 contain the same explanatory variables, as well as a coverage status variable added to each of these models. We restricted Table 2 to people in families with incomes of 100 percent of the FPL or more, while we made estimations in Table 1 without this restriction.

In the logistic equation for the probability of a dental visit in the fourth and fifth columns of Table 2, we found that people with dental coverage were more likely (P < .05) to have a dental visit than were people without coverage. This result, which we found in Table 1 without controlling for other factors, is confirmed to hold after controls. In addition, people aged 45 to 64 years and 65 years and older were more likely (P < .05) to have had a dental visit than were people 18 years of age and younger. Females and college graduates were more likely to have had a dental visit than were males and people with limited formal education (less than high-school graduate). On the other hand, people aged 19 to 44 years, non-Hispanic blacks, Hispanics and people in families at middle or lower income levels were less likely (P < .05) to have a dental visit than were people 18 years of age and younger, whites and people in families with high income. Rural/urban status did not add to the precision of this model.

The sixth and seventh columns in Table 2 show the OLS regression for the natural logarithm of the number of dental visits during 1996 for people with at least one visit during the MEPS survey year. Controlling for other factors, people with dental coverage who visited the dentist at least once (P < .05) during the year had more dental visits during the year than did people who were dental service users without coverage. People who visited the dentist at least once during the year and who were aged 65 years and older had more dental visits than did dental users aged 18 years and younger. Women and people residing in large metropolitan areas (P < .05) who visited the dentist at least once during the year had more dental visits than did men and people residing in non-metropolitan areas who used dental services. On the other hand, people who visited the dentist at least once during the year, were aged 19 to 44 years, were non-Hispanic black or Hispanic, and in families with low incomes had fewer dental visits than did dental users aged 18 years and younger, whites and people in families with high incomes.

The eighth and ninth columns of Table 2 show the OLS regression estimates for the natural logarithm of dental expenditures for people with dental service expenditures during 1996. Controlling for other factors, this equation confirms the Table 1 result that among people who incurred dental expenses during the year those with dental coverage had higher dental expenditures than did those without coverage. In addition, dental users in the age groups 45 to 64 years and 65 years and older (P < .05) spent more on dental services than did people in the group aged 18 years and younger. People residing in metropolitan areas (P < .05) had higher dental expenditures than did people residing in nonmetropolitan areas. On the other hand, non-Hispanic blacks, Hispanics and people in families with low incomes had lower dental expenditures than did whites and people in families with high incomes.

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
While these data and analyses are useful, they do have limitations. For instance, self-reporting of data is less accurate than data collected by observation or by dental record abstraction, and analyses of data from different survey sources historically have resulted in varying national estimates.²² In addition, the specification of the dental coverage variable is a function of a report of coverage, a report of payment for dental care by a third party or both. Also, adverse selection may overstate the effect of dental coverage on use and expenditures should people needing extensive services self-select into coverage. This effect, however, may be mitigated in the MEPS data to the extent that most dental coverage is made available through employer-sponsored plans that offer little, if any, choice of plans. Finally, individual coverage plans may vary considerably in their degree of benefit generosity. Findings from previous studies have shown that the demand for dental services is related positively to the generosity of dental coverage.^12,13

On the other hand, these data are useful and comprehensive and provide estimates that are nationally representative. As such, MEPS data are unique and provide important information from which dental visits and expenditures can be compared and analyzed in the context of dental care coverage. While the specification for dental coverage has limitations, its formulation is based on previously used and accepted methods, provides an acceptable nationally representative measure of dental care coverage, and makes possible valuable analyses and comparisons.⁸

Several studies^9–13 have shown that the presence of private dental insurance coverage is positively associated with having a dental visit. The current study is one of the first to simultaneously describe the association between presence of private dental insurance and three important health practice measures: the probability of at least one dental visit during the year, the number of dental visits during the last year and expenditures. By looking at these three health practice measures side by side, we are able to describe the interrelation between dental care use and expenditures, across selected demographic and socioeconomic status, or SES, variables.

Several studies^9–13 have shown that the presence of private dental insurance coverage is positively associated with having a dental visit. MEPS data supported these findings; however, they also demonstrated that among people with a dental visit in the last year, private dental insurance coverage is associated with a greater number of dental visits per year and higher expenditures, when we controlled for selected demographic and SES factors. These findings are important, because only 51.4 percent of Americans have private coverage.

Our investigation also demonstrated that key demographic and SES variables were associated with dental care use, independent of private dental insurance coverage. For example, people with lower SES, as measured by education level and income level, were significantly less likely to have visited a dentist in the last year than were people with higher education levels or income levels. In addition, females were more likely to have visited a dentist than were males, and whites were more likely to have visited a dentist than were non-Hispanic blacks or Hispanics, when we controlled for insurance coverage and other covariates.

Among those who had visited the dentist in the last year, MEPS also showed differences in the number of visits and level of expenditures across demographic and SES variables, independent of private dental insurance coverage. For example, people at a low income level made fewer visits to the dentist and had lower expenditures than did people at a high income level, regardless of insurance coverage. In addition, whites had more dental visits and higher expenditures than did non-Hispanic blacks or Hispanics. Finally, although females had more dental visits than did males, there were no differences in expenditures between the two sexes.

Age-specific and rural- or urban-specific associations were more complex. Compared with people aged 18 years and younger, adults aged 19 to 44 years were less likely to have visited a dentist in the last year, and people 45 years and older were more likely to have visited a dentist in the last year, when we controlled for private dental insurance coverage. Among people with a dental visit in the last year and compared with people aged 18 years and younger, adults aged 19 to 44 years had fewer visits, whereas adults aged 65 years and older had more visits, independent of insurance coverage. By contrast, adults aged 45 years and older had higher expenditures than did people aged 18 years and younger. Regarding rural- or urban-specific associations, we found no differences in the probability of a dental visit in the last year across categories, when we controlled for private dental insurance coverage. People residing in large metropolitan areas, however, had more dental visits than did those residing in nonmetropolitan areas, and people residing in small metropolitan and large metropolitan areas had higher expenditures than did those residing in non-metropolitan areas.

Despite differences across income levels, we found no statistically significant differences in number of visits or expenditures across education levels. These findings imply that the financial elements of SES play a more influential role in dental care use and costs than do the educational components of SES, at least for those with a dental visit in the last year. Future studies could investigate these possibilities in greater detail.

	CONCLUSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Although presence of private dental insurance coverage is strongly associated with receipt of care, it is clear from this investigation that demographic and SES factors also must play a role. Programs designed to improve dental care use or control expenditures must take into consideration each of these determinants to be successful. Private dental insurance coverage is important, but it is only one piece of a complex puzzle.

View larger version (131K): [in this window] [in a new window]   Dr. Manski is a professor, Department of Oral Health Care Delivery, Dental School, University of Maryland, Baltimore, and a senior scholar, Center for Cost and Financing Studies, Agency for Healthcare Research and Quality, Rockville, Md. Address reprint requests to Dr. Manski at Department of Oral Health Care Delivery, Dental School, University of Maryland, 666 West Baltimore St., Baltimore, Md. 21201, e-mail "Manski{at}Dental.umaryland.edu".

View larger version (137K): [in this window] [in a new window]   Dr. Macek is an assistant professor, Department of Oral Health Care Delivery Dental School, University of Maryland, Baltimore.

View larger version (124K): [in this window] [in a new window]   Dr. Moeller is a health economist, Center for Cost and Financing Studies, Agency for Healthcare Research and Quality, Rockville, Md.