Differences in estimates are likely the result of differences in methodology and do not diminish their importance.¹¹ Moreover, noted associations between important socioeconomic and demographic variables appear to remain consistent across the surveys.¹¹ Therefore, although these surveys may produce differing estimates of utilization, analyses of data from these surveys have resulted in a number of important reports that have helped to characterize the dental utilization patterns of most Americans.^10–15

For the most part, these analyses have focused on utilization estimates for a particular survey period or year. Fewer studies have examined utilization patterns over time. However, a few exceptions are notable. For instance, in an analysis of expenditure data, Moeller and Levy¹² estimated that between 1977 and 1987, the percentage of the population with at least one dental visit remained almost stable (41.1 percent in 1977 compared with 42.4 percent in 1987). In addition, they found that the per capita use declined from 3.2 to 2.9 visits for people with a visit during the same period.

Brown and Lazar¹⁵ analyzed data from the NHIS, and reported that the overall rate of dental utilization for Americans increased from 42 percent between July 1963 and June 1964 to more than 57 percent during 1989. White and colleagues,¹⁶ in a report of the U.S. Public Health Service, noted that the proportion of Americans with at least one dental visit per year increased modestly from 55 percent in 1983 to just over 57 percent in 1989. Eklund and colleagues¹⁷ studied insured Americans and reported that the overall dental utilization by this group increased from 60 percent in 1980 to about 70 percent in 1995. These studies provide important information about the changing utilization patterns of Americans over time.

This study provides a comparative analysis of utilization using 1977, 1987 and 1996 expenditure data.

The purpose of this article is to augment these utilization studies with a detailed comparative analysis of data available from a 20-year period. Specifically, this study provides a bivariate and multivariate comparative analysis of utilization using 1977, 1987 and 1996 expenditure data.

	METHODS

TOP ABSTRACT METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Since 1977, the DHHS has sponsored several national expenditure surveys.^3–5 The National Medical Care Expenditure Survey, or NMCES, sponsored by the National Center for Health Services Research, or NCHSR, provided detailed national estimates of expenditures for various types of health care in 1977. NMCES was a panel study of approximately 14,000 households and 38,815 people and included data in regard to 1977 dental care expenditures, demographic and socioeconomic status, and insurance payments. Data were collected in five rounds, with a combined response rate of 82 percent.³

The 1987 National Medical Expenditure Survey, or NMES, sponsored by the Agency for Healthcare Research and Quality, or AHRQ (formerly NCHSR and the Agency for Health Care Policy and Research, or AHCPR), provided detailed national estimates of expenditures for various types of health care in 1987. Specifically, NMES provided detailed national estimates of health expenditures, utilization of services, sources of payment and insurance coverage for the U.S. civilian population during the period from Jan. 1, 1987, through Dec. 31, 1987. The NMES surveyed approximately 14,000 households and 34,459 people and oversampled population groups of particular policy interest, including the elderly, people with limitations in activities of daily living, blacks, Hispanics and the poor. All survey components were designed to provide statistically unbiased estimates that were representative of the civilian, noninstitutionalized U.S. population. Data were gathered in five rounds of interviews during an 18-month period in 1987 and 1988; the combined response rate was 79.7 percent.⁴ The 1987 NMES data were released for public use in October 1992.⁴

The 1996 MEPS is the third in a series of nationally representative health surveys of the U.S. community-based population sponsored by AHRQ. The MEPS collects health care expenditure, utilization and payment source data, along with socioeconomic, demographic and health insurance data similar to those in its predecessor surveys. It differs from the NMES and NMCES in that data on household respondents in each panel were collected for two consecutive years (as opposed to one year for the other surveys) and the ongoing survey is fielded continuously (that is, a new panel is selected every year).

The target for the 1996 MEPS was a sample of 10,500 NHIS households and 21,571 people. To collect health expenditure and use data for 1996, interviewers questioned people in each MEPS household in person three times during an approximate 18-month period, with the third round conducted between February and May of 1997. The combined (that is, averaged) full-year 1996 response rate for the MEPS sample through the third round was 70 percent.⁵

The focus of our analysis is on use of dental services and changes in use for the civilian, community-based U.S. population during 1977, 1987 and 1996. Specifically, we provide national estimates of dental visits for each of several socioeconomic and demographic categories during these three years.

	STATISTICAL ANALYSIS

TOP ABSTRACT METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
We conducted multivariate analyses to determine if the bivariate relationships in Table 1 would persist in a pooled, multivariate statistical model, and to assess the relative impact of the socioeconomic and demographic variables on dental utilization during the 20-year period. To accomplish this analysis, we pooled the data from the three survey sources and subjected them to multivariate statistical analyses. Regression models were estimated with the survey data pooled across the three survey periods for the use rate (defined as 1 if one or more dental visits occurred during the year or 0 if no dental visits occurred) and for the natural logarithm of the number of dental visits during the year for people with at least one visit. The impact of socioeconomic and demographic variables (that is, age, sex, race/ethnicity, family income, education and employment status) on the use rate and on the number of annual dental visits then was estimated with these pooled models.

All reported estimates and statistics were computed by taking into account the complex sampling design of NMCES, NMES and MEPS, with the use of a statistical software package (SUDAAN).¹⁸

	RESULTS

TOP ABSTRACT METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
There were 38,815 participants in the 1977 NMCES, representing 212,097,954 noninstitutionalized U.S. civilians. Of these, 52 percent were female (n = 20,100), 13 percent (n = 5,097) were black and 56 percent (n = 21,615) were between the ages of 19 and 64 years.

A total of 34,459 people participated in the 1987 NMES, representing 239,392,856 noninstitutionalized U.S. civilians. Of these, 53 percent were female (n = 18,376), 21 percent (n = 7,406) were black and 54 percent (n = 18,696) were between the ages of 19 and 64 years. A total of 21,571 people participated in the 1996 MEPS, representing 268,130,477 noninstitutionalized U.S. civilians. Of these, 52 percent were female (n = 11,282), 13 percent (n = 2,907) were black and 58 percent (n = 12,424) were between the ages of 19 and 64 years.

Table 1 shows that slightly more than 40 percent of the U.S. population had at least one dental visit during 1977, 1987 and 1996. Overall, no significant difference (P > .05) was observed in the rate of utilization during this 20-year period. On the other hand, for people with a dental visit, the number of visits per person per year decreased (P < .05) from 1977 to 1987 and from 1987 to 1996.

Elderly respondents and young children were more likely (P < .05) to report having had a dental visit in 1996 than in 1987 or in 1977. In 1996, 41 percent of seniors and 21 percent of young children had at least one visit, while slightly less than 30 percent of seniors and less than 14 percent of young children reported having had at least one visit during 1977. As expected, poorer people and people with less education were less likely (P < .05) to report having had a dental visit than people with more income or more education during each of these periods. Female and employed participants were more likely (P < .05) to report having had a dental visit than were their unemployed and male counterparts during 1977, 1987 and 1996.

Tables 2 and 3 are based on a single regression analysis of the probability of having had at least one dental visit during the year and the annual number of dental visits per person given use. Each analysis in Tables 2 and 3 is based on cross-sectional data that are pooled for the years 1977, 1987 and 1996. Table 2 includes data from 1977 and Table 3 includes data from 1987 and 1996, which are provided to illustrate the changes from 1977. (The left three columns report the logistic estimation results for the probability of visiting the dentist at least once during the year and the right three columns report the ordinary-least-squares results for the natural logarithm of the number of annual dental visits given use during the year.)

As shown in Table 1, dental use rates for children between 6 and 18 years of age were the highest of any age group in each of the three years studied. The positive-valued and significant changes in the coefficients for the 0- to 5-year-old group and the 65-years-and-older group suggest that the gap in use rate between the 6- to 18-year-olds and these two age groups tended to narrow slightly between 1977 and 1996. However, the coefficient change estimate for the 19- to 44-year-old age group in 1996 suggests a widening of the gap in the use rate between children aged 6 to 18 years and adults aged 19 to 44 years between 1977 and 1996.

For participants with at least one dental visit during the year, seniors and adults aged 45 to 64 years appeared to visit a dentist as often as did children aged 6 to 18 years. However, evidence suggests that dental visits for seniors were increasing relative to visits for children aged 6 to 18 years during the 20-year period of this study. Children younger than 6 years of age and people 19 to 44 years of age had fewer dental visits than people aged 6 to 18 years. In addition, the results indicate a widening of the gap in use rates and number of dental visits (for people with a visit) between people in the 19- to 44-year-old category and people in the 6- to 18-year-old category by the end of the 20-year period (1996).

Female and male participants. As shown in Table 1, female participants were more likely to see a dentist during the year than male participants, and among both female and male participants who saw a dentist during the year, female participants had more frequent visits than male participants. Table 2 shows there was no evidence that these gender gaps changed during the 20-year period.

Evidence suggests that dental visits for seniors were increasing relative to visits for children aged 6 to 18 years during this study.

Race/ethnicity. We pooled blacks, Hispanics and people in all other racial/ethnic groups except whites into one nonwhite group for comparison with whites in these multivariate analyses. People in the non-white group were less likely to have visited a dentist during the year than were whites, and nonwhites with at least one visit saw their dentist fewer times during the year than did white users of dental services. By 1996, a significant narrowing in the use rate gap was observed between whites and nonwhites, but the gap in the number of dental visits for users between whites and nonwhites showed no clear signs of changing during this 20-year period.

Income. As shown in Table 1, people in low-income families with incomes no greater than 200 percent of the poverty threshold had lower dental use rates than people in middle- and high-income families. The data also showed that the gap in use rates between lower- and higher-income people widened during the 20-year period. Additionally, while the annual number of dental visits for users did not appear to differ according to income class at the beginning of the study period (1977), a gap between lower- and higher-income people did appear by the middle of the period (1987) and continued to widen by the end of the study period (1996).

Education. Both the dental use rate and the annual number of dental visits for dental users increased with a person’s educational level (or with the parent’s educational level for children younger than 18 years of age) (Table 1). In addition, the gap in the dental use rate between people with a high-school or college education and those with less education increased by the middle of the 20-year period (1987), but by the end of the 20-year period (1996), the gap had returned to the level observed at the beginning of the study period (1977). The gap in annual dental visits for dental users between people without a high-school education and those with at least a high-school education remained stable during the 20-year period.

Employment status. As shown in Table 1, people who were employed at any time during the study year (or children younger than 16 years of age with at least one parent who was employed at any time during the year) were more likely to see a dentist in 1977 than were people who were not employed at all during the year. However, this employment gap declined by 1987 and disappeared by 1996. On the contrary, employed people visited the dentist less frequently than did people without employment, and this relationship remained stable during the entire 20-year study period.

	DISCUSSION

TOP ABSTRACT METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Although these data and analyses are useful, they do have limitations. For instance, self-reporting of data is less accurate than data collection by observation or by dental record abstraction, which potentially limits the usefulness of these data. In addition, while survey design and methodology for each of the three surveys are similar, they are not identical.

On the other hand, these data are useful and comparable and provide estimates that are nationally representative. The NMCES, NMES and MEPS data are unique, provide important information and establish a mechanism for comparing and analyzing dental visits in a new and important approach during a 20-year period.

Multivariate analysis. Generally, the results of the multivariate analysis shown in Tables 2 and 3 are consistent with the bivariate relationships shown in Table 1. For example, the 11-percentage-point increase in the use rate for people aged 65 years and older and the 7-percentage-point increase in the use rate for children younger than 6 years of age during the 20-year period (Table 1) are consistent with the use rate trends for this period shown in Tables 2 and 3. We observed a narrowing of the gap in the dental use rates between people 6 to 18 years of age and children younger than 6 years of age during the 20-year period. This estimated effect, after other socioeconomic and demographic influences on the use rate are controlled for, is consistent with an increase in the dental use rate for children younger than 6 years of age during this period. The estimated narrowing of the gap in the dental use rate between people aged 6 to 18 years and people aged 65 years and older during the 20-year period also is consistent with an increase in the dental use rate for seniors during the same period.

The authors observed a narrowing of the gap in the dental use rates between people 6 to 18 years of age and children younger than 6 years of age.

These data provide practitioners, oral health advocates and policy-makers with new and important baseline information with which dental visits can be compared and analyzed over a 20-year period. For instance, by understanding these analyses and their results, dental professionals will be better positioned to provide care and meet the oral health needs of older adults and young children whose use rates have increased during this study period.

Trends. The multivariate analysis enables us to gain additional insights into trends shown by the bivariate relationships in Table 1. For example, in each of the three years shown in Table 1, employed people had a higher use rate than their unemployed counterparts in the population. However, as shown in Table 2, we see that after other influences on the dental use rate are controlled for, this association between employment status and the dental use rate at the beginning of the 20-year period (1977) weakens in 1987 and vanishes by 1996.

Given the increase in annual dental care utilization estimated by other surveys, such as the NHIS series,^6–8 it may be surprising that Moeller and Levy¹² observed a trend toward no increase in utilization when comparing 1987 NMES data with data collected comparably in 1977, and we continued this observation through 1996 in the present study. Interestingly, this observation was not found to be consistent across all age groups. For example, dental care use increased among people at the extremes of life (that is, people younger than age 6 years and people 65 years of age and older).

People aged 65 years and older. The increased use of dental services by people aged 65 years and older may reflect increased retention of natural teeth. The percentage of people aged 65 to 74 years without teeth decreased from about 46 percent in 1977 to about 29 percent in 1996.^9,10,19 One-half of the people aged 65 years and older in a 1989 survey reported that having "no teeth" was their reason for not visiting a dentist.⁷ This marked decrease in edentulousness, along with the retention of more teeth by dentate seniors, may be responsible for the increasing use of dental care to a level similar to that of 19- to 44-year-olds by 1996.

Children younger than age 6 years. In contrast, the increased use of dental services by children younger than age 6 years cannot be explained by a significant trend in health status. If anything, dental caries experience has decreased during the study period. The increased proportion of children younger than age 6 years who visit a dentist may reflect an increasing recognition of the importance of the primary dentition, a decreased tolerance of untreated decay or an increased number of dentists who are willing to treat young children. It is interesting that the trend toward decreased visits per person for those with a visit is consistent across all age groups, indicating that people who use dental care are more likely to have fewer visits per year on a per capita basis than they did in 1977. Either fewer services are needed or requested or dentists have been able to meet patients’ needs with fewer visits.

On the other hand, the decreased number of visits per person across all age groups and across groups characterized by other demographic variables could represent decreased disease among people seeking care, but further insight is not possible from our data. Other studies have reported that an increasing proportion of dental services are for routine diagnostic and preventive services.¹⁷ These findings are consistent with recent recommendations for preventive services without regard to dental risk status.²⁰

Untreated dental caries. Approximately 30 percent of people of all ages have untreated dental caries (NHANES III, 1988–1994, unpublished data analyzed by the Division of Oral Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, 2000). Although the decreased number of visits by users of dental care could reflect decreased restorative needs among these people, the fact that about half of the people aged 6 to 64 years in the NMCES, NMES and MEPS did not use dental care, and one-third of the U.S. population has unrestored dental caries indicates that there are many people who could benefit from dental care but are not receiving it.

After controlling for other socioeconomic and demographic influences, we found that the gap in dental use rates between whites and nonwhites narrowed significantly by the end of the 20-year period, but the racial gap in regard to the number of dental visits between white and nonwhite users of dental services persisted throughout this period. This suggests that policy efforts to close these gaps have been only partially successful. In addition, any policy efforts during the past 20 years to close the gaps in dental use rates and visits for users of dental services between low-income and high-income people appear to have been unsuccessful. In fact, the use rate gaps between these income groups widened during this period. Gaps in dental visits that were not evident at the beginning of the study period did, in fact, emerge by 1996.

A more extensive model of dental utilization would have to be established to better understand why these changes occurred during the 20-year period. For instance, why did the elderly and youngest children have increased dental use rates during this period relative to those of older children? Future research into the role of dental health insurance and other potentially relevant influences on dental care not included in our model may further explain these results.

	CONCLUSION

TOP ABSTRACT METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
The data presented above are unique and comparable and establish a mechanism by which dental visits can be compared during a 20-year period. While aggregate utilization rates generally were stable during this period, some differences within socioeconomic and demographic groups are notable. For example, the use rate increased during the 20-year period for people 65 years of age and older and for children younger than 6 years of age. By understanding these analyses and their results, U.S. dentists will be better positioned to provide treatment and meet the oral health care needs of all Americans.

View larger version (133K): [in this window] [in a new window]   Dr. Manski is a professor, Department of Oral Health Care Delivery, Dental School, University of Maryland, and a visiting scholar, Center for Cost and Financing Studies, Agency for Healthcare Research and Quality, Rockville, Md. Address reprint requests to Dr. Manski, University of Maryland, 666 W. Baltimore St., Baltimore, Md. 21201, e-mail "Manski@Dental. umaryland.edu".

View larger version (128K): [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.

View larger version (129K): [in this window] [in a new window]   Dr. Maas is director, Division of Oral Health, Centers for Disease Control and Prevention, Atlanta.