NSCH assessed race and ethnicity via self-reports. We restricted our analysis to non-Hispanic whites, non-Hispanic blacks and Hispanics. Other variables included in this study were age (3–17 years), sex, household poverty level (< 100 percent, 100–199 percent, 200–299 percent, 300–399 percent, 400 percent of the federal poverty level, which also takes into account household size), highest level of education of any household member (< high school, high school graduate, > high school), child’s country of birth (United States versus other), primary language spoken in the home (English versus other). Our analyses found that the last two items were relevant primarily to Hispanics and we did not include them in the logistic models described in Table 2. We assessed the role of dental insurance coverage from responses to one question about dental insurance ("Does the child have insurance that helps pay for any routine dental care including cleanings, X-rays and examinations?"). There were no other items on related determinants of dental care access.

View this table: [in this window] [in a new window]   TABLE 2 Odds ratio (OR) and 95% confidence intervals (CIs) for association between race/ethnicity and fair or poor oral health rating on the 2003 NSCH.*

– "How long has it been since the child last saw a dentist?" (question 1);

– "During the past 12 months, did the child see a dentist for any routine preventive dental care?" (question 2);

– "During the past 12 months, was there any time when the child needed routine preventive dental care?" (question 3).

From parents’ responses to these survey items, we created a categorical preventive-care variable indicating children who had received routine preventive care during the previous 12 months (answered "yes" to question 2 above); children who had had other dental visits (answered "no" to question 2, but had visited a dentist in the previous 12 months); children had not received routine preventive care but the parent thought it was needed (answered "yes" to question 3, but had not visited a dentist in the previous 12 months); and children who had not visited a dentist in the previous 12 months and the parent thought preventive care was not needed (answered "no" to question 3).

An additional question asked whether or not the child did "receive all the routine preventive dental care he or she needed" in the past 12 months.

We also created a dichotomous variable, "preventive-care attitude," that included one group of children who had received preventive dental care in the previous 12 months or whose parents thought that preventive care was needed, and a second group of children who had not received any preventive care during the previous 12 months and whose parents felt that no preventive care was necessary.

	STATISTICAL ANALYSIS

TOP ABSTRACT MATERIALS AND METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
To obtain population-based estimates, we weighted all of our analyses using NSCH sampling weights. Details of the weighting procedures can be found elsewhere.⁸ Briefly, the sampling weights accounted for multiple telephone lines within a household, nonresponse and the differential selection probabilities in each state. Furthermore, we calculated standard errors and 95 percent confidence intervals for all analyses that accounted for survey stratification.

The study results showed significant differences in parental ratings of their children’s oral health according to race/ethnicity.

To tabulate frequencies, we collapsed the upper and lower two categories of self-reported oral health status, resulting in a variable with three categories (excellent or very good, good, fair or poor). We calculated weighted proportions and 95 percent confidence intervals according to race/ethnicity. We used a ² test to test for differences between race/ethnicities. Using age- and sex-adjusted logistic regression models, we then evaluated the bivariate associations between the different variables and the oral health rating (fair or poor versus good or better) for each racial/ethnic group separately.

Furthermore, we analyzed the association between a fair or poor oral health rating (dependent variable) and race/ethnicity (independent variable) using weighted logistic regression models. To evaluate the extent to which the association between racial/ethnic group and oral health rating was explained by education and poverty level, dental insurance status and preventive-care attitude, we consecutively entered these prespecified variables into an age-and sex-adjusted model. Finally, we evaluated the association between preventive-care attitude (dependent variable) and race/ethnicity, education, poverty level and dental insurance status (independent variables) using weighted logistic regression. We also examined for possible interactions in our models (for example, between sex, education, insurance and poverty status), and we noted no meaningful interactions.

We performed all analyses using statistical software (STATA 7.0, StataCorp, College Station, Texas).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
This study included a total of 77,733 children 3 years or older for whom parental oral health ratings were available. Table 1 shows the distribution of basic demographic and socioeconomic data across racial/ethnic groups. The proportion of girls was similar across racial ethnic groups. The proportion of very young children (aged 3 to 5 years) was higher among Hispanics than among non-Hispanic whites and non-Hispanic blacks. Both non-Hispanic white children and non-Hispanic black children were born predominantly in the United States (> 97 percent) and their parents were primarily English speakers (> 98 percent), while 80.7 percent of Hispanic children were born in the United States. English was the primary language of only 40.2 percent of Hispanic parents.

We found major differences in education and family poverty level according to racial/ethnic groups, with Hispanics having the lowest educational and highest poverty levels and non-Hispanic whites having the highest educational and lowest poverty levels. Non-Hispanic blacks had the highest proportion of children with dental insurance coverage, followed by non-Hispanic whites and Hispanics. In the NSCH, insurance coverage included coverage by either private or public insurance programs, such as Medicaid.

As shown in Table 1, the study results showed significant differences in parental ratings of their children’s oral health according to race/ethnicity. These racial/ethnic differences were consistent among all age groups (data not shown). Among those who rated their children’s oral health as fair or poor, "cavities" (> 50 percent) and "crooked teeth" (> 30 percent) were the most prevalent specific problems reported, with no differences reported according to racial/ethnic groups. The prevalence of other specific dental problems reported was much lower (in almost all cases < 10 percent).

Pain was reported as a specific dental problem by 6.2 percent of Hispanics, 3.5 percent of non-Hispanic blacks and 3.1 percent of non-Hispanic whites. However, the absolute number of subjects reporting problems other than "cavities" or "crooked" teeth within racial/ethnic groups was low, so the resulting national prevalence estimates may not be reliable.

Frequency of dental visits. We found marked differences between racial/ethnic groups with regard to the frequency of dental visits, as measured by reports of "never been to a dentist" (Figure 1) and "no dental visits within past 12 months" (Figure 2). For example, Hispanic children were least likely to ever have seen a dentist (Figure 1) or to have had a dental visit in the previous year (Figure 2). Among children aged 12 to 17 years, 25.1 percent of Hispanics and 20.7 percent of non-Hispanic blacks had not seen a dentist within the previous 12 months, compared with 9.1 percent of non-Hispanic whites (Figure 2). We found similar racial/ethnic disparities for children aged 6 to 11 years, but the pattern was less clear among children aged 3 to 5 years.

View larger version (49K): [in this window] [in a new window]   Figure 1. Percentage of participants reporting that their children had "never been to a dentist," by race/ethnicity and age group. Bars represent 95 percent confidence intervals.

View larger version (49K): [in this window] [in a new window]   Figure 2. Percentage of participants reporting that their children had "no dental visits within past 12 months," by race/ethnicity and age group. Bars represent 95 percent confidence intervals.

View larger version (47K): [in this window] [in a new window]   Figure 3. Percentage of participants reporting that their children "did not receive all preventive care needed in past year," by race/ethnicity and age group. Bars represent 95 percent confidence intervals.

Furthermore, the study findings showed marked racial/ethnic differences with regard to parents’ attitudes about preventive care. Among parents of children aged 12 to 17 years, 20.4 percent of Hispanics compared with only 6.7 percent of non-Hispanic whites thought that routine preventive care was not necessary (Figure 4). The final logistic model (Table 3) on the predictors of parents’ preventive-care attitudes shows that being non-white remained significantly associated with a poorer or negative attitude (that is, children who did not receive any preventive care during the previous 12 months and whose parents felt that no preventive care was necessary).

View larger version (49K): [in this window] [in a new window]   Figure 4. Percentage of participants reporting that their children received "no preventive care in past year, and felt not needed," by race/ethnicity and age group. Bars represent 95 percent confidence intervals.

Age- and sex-adjusted analyses show strong associations between socioeconomic indicators and self-reports of fair or poor oral health among all racial/ethnic groups (Table 4). The associations between educational levels and poverty levels of families and a parental report of fair or poor oral health in their children were strongest among Hispanics.

Dental insurance coverage was associated with better oral health ratings among all racial/ethnic groups, with the strongest association among Hispanics. Among Hispanics, U.S. nativity and English as the primary language were strongly associated with better oral health ratings (Table 4).

Parents whose children did not receive preventive care but who thought it was necessary reported significantly worse oral health in all racial/ethnic groups compared with parents of children who received routine preventive care in the previous 12 months (Table 4). Parents who thought that routine preventive care was not necessary did not report worse oral health than did parents of children who did receive routine preventive care. Parents who reported that their children had not received all needed preventive dental care in the previous 12 months were more than five times as likely to report that their children had fair or poor oral health, with no significant differences found among racial/ethnic groups (Table 4).

Compared with non-Hispanic whites, non-Hispanic blacks had about twice the odds and Hispanics had more than four times the odds of reporting fair or poor oral health independent of their children’s age and sex (Table 2). Further adjustments for poverty level and education markedly attenuated this association. However, additional adjustments for dental insurance and attitude toward preventive dental care did not further attenuate the observed associations (Table 2). Even after such adjustments, we found that the odds of reporting fair or poor oral health were more than twice as high among Hispanics compared with non-Hispanic whites.

The National Survey of Children’s Health data show significant differences in parents’ ratings of children’s oral health according to race/ethnicity.

We should point out that Hispanic parents were more likely than non-Hispanic white parents to report that preventive dental care was not necessary even if their children had not received any preventive dental care in the preceding year. However, this attitude regarding preventive dental care was strongly associated with lower education, higher poverty level and lack of dental insurance. After adjusting for these socioeconomic variables, we found that the association of attitude toward preventive dental care with race/ethnicity was attenuated markedly (Table 3), although both Hispanics and non-Hispanic blacks still were significantly more likely to have a poorer preventive-care attitude (that is, children who did not receive any preventive care during the previous 12 months and whose parents felt that no preventive care was necessary) than were non-Hispanic whites.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Racial/ethnic disparities. The NSCH data show significant differences in parents’ ratings of children’s oral health according to race/ethnicity, with Hispanics much more likely to report fair or poor oral health. Using data from the NSCH, Liu and colleagues¹⁰ described in detail the racial/ethnic disparities in dental care and insurance coverage. It is important to note that their results mirror the pattern of disparities in clinical oral health status reported by other researchers.^13,14 Hispanic families are by far the most economically disadvantaged group represented within the NSCH data, with greater proportions reporting incomes below the federal poverty level, less than a high school education and a primary language other than English (Table 1).

Disparities in reported oral health status can be explained, at least partially, by these measures of socioeconomic status (Table 2). However, even after adjustment for education and poverty level, Hispanic parents were twice as likely as non-Hispanic white parents to report their children’s oral health as being fair or poor (Table 2). We should note that poverty level and education measured at one point in time are inexact measures of socioeconomic status, and it is plausible that other dimensions of socioeconomic status may explain the remaining association between race/ethnicity and parental reports of their children’s oral health status.

Overall, these results are consistent with those of previous studies that found that parents from minority groups and those with lower education and greater poverty levels were more likely to report that their children had poor oral health.^15–17

Parental reports of children’s oral health. Childhood is a critical time to initiate prevention and health promotion efforts. Children must rely on their caregivers for health-related decision making and access to medical and dental care.¹⁸ Thus, parents’ perceptions of their children’s oral health status and health care needs are vitally important. Our research focuses on parents’ perceptions of their children’s oral health status. Parental perceptions appear to underreport clinically determined oral health needs,^15,17 while such parental perceptions may be influenced by socioeconomic and racial/ethnic characteristics.^15–17

Data from the Third National Health and Nutrition Examination Survey show that among children aged 2 to 18 years, 22.5 percent had untreated decay, while only 18.0 percent of parents reported a need for dental treatment.¹⁵ A survey conducted among Maryland schoolchildren comparing clinical examinations in children with parental reports of oral health status found that only 34 percent of parents whose children had untreated decay were aware of it.¹⁷ However, although the sensitivity of parental reports regarding children’s oral health may be low, the specificity of such reports appears to be high.¹⁷

Factors other than active disease, such as functional ability and cosmetic appearance, also may affect both clinically determined treatment needs and self-ratings of oral health.¹³ Because the rating of overall oral health does not discriminate between true dental pathology and esthetic concerns, we conducted a sensitivity analysis limited to those children for whom fair or poor oral health ratings were based on unequivocally nonesthetic problems (such as pain, cavities, broken teeth), which yielded similar results (data not shown).

The relationship between perceived needs, clinical needs and dental care utilization is complex and warrants further exploration; therefore, the results of our study must be interpreted with caution. However, although parents may not be fully cognizant of their children’s dental treatment needs, those who do perceive that their children need care are more likely to seek it. Furthermore, dental care utilization among children in all age, racial and ethnic groups has been shown to be inadequate.¹⁵

Preventive-care attitudes. Previous research has shown that low-income children aged 5 through 11 years are more likely to have had dental visits if their mother reported that she felt dental care was important and perceived that her child needed dental care.¹⁹ In our study, we attempted to explore parental attitudes toward prevention as a possible determinant of racial/ethnic disparities. In our analyses, we categorized parents as having a negative prevention attitude if they indicated that preventive dental visits were not needed for their children and had not taken their children for such visits in the previous year.

We found that Hispanic parents were more likely to have this negative attitude than were non-Hispanic white or non-Hispanic black parents. However, within each racial/ethnic group, parents who had a positive prevention attitude and parents who did not were equally likely to report their children’s oral health as fair or poor (Table 4). Hence, our results do not suggest that preventive-care attitude, as measured in this study, is a significant determinant of racial/ethnic disparities in parental reports of their children’s oral health (Table 2).

It may be that some parents do not consider routine preventive care necessary unless they consider their children’s oral health as fair or poor. Nevertheless, a negative prevention attitude was strongly associated with lower socioeconomic status. In fact, differences in preventive-care attitude between racial/ethnic groups were greatly attenuated after we controlled for socioeconomic indicators (Table 3). However, routine preventive care, which was defined in the survey as "check-ups, screenings, and sealants," does not include any restorative or emergency dental treatment. It is possible that the parent’s attitude toward prevention may differ from his or her attitude toward treatment of diseases, and the latter may be more important with regard to racial/ethnic disparities in children’s oral health.

	CONCLUSION

TOP ABSTRACT MATERIALS AND METHODS STATISTICAL ANALYSIS RESULTS DISCUSSION CONCLUSION REFERENCES

 
The results of our study show that, on the basis of parental reports, there are significant racial/ethnic disparities in the oral health of U.S. children, with both Hispanic children and non-Hispanic black children being worse off than white children. In addition, Hispanic children also tend to be worse off than non-Hispanic black children, on the basis of parental reports of their children’s oral health status. Although poverty level and education explain a large part of the observed racial/ethnic disparities, significant disparities exist independent of such socioeconomic variables. In particular, even after accounting for such factors, we found that Hispanic children remained significantly more likely than white or black children to have received a parental report of fair or poor oral health (Table 2).

The remaining contributors to such oral health disparities may be related to provider-based factors or to other patient-based factors not assessed by the NSCH. It is important to point out that although we found that self-reported attitude toward preventive care differed between white parents and the other two groups (Table 3), we found no evidence (comparison of model 3 with model 4 in Table 2) that self-reported attitude toward preventive care is a significant determinant of racial/ethnic disparities in children’s oral health status.