The use of dental care services is complex, as it involves insurance coverage and affordability, geographic accessibility of providers, provider availability and participation in various types of insurance plans, provider interest in treating certain subpopulations, valuation of health care services and perceived need for health care services.² An additional factor that may influence the use of dental services is the presence of a "dental home," which has been described by Nowak and Casamassimo³ as an accessible, family-centered, continuous, comprehensive, coordinated, compassionate and culturally competent source for dental care. Understanding of the sociodemographic variables that predict patients’ use of a dental home might help reduce the observed disparities in oral health among minority groups.

One characteristic that may be linked to who seeks regular dental care is patients’ level of trust in the health care system and health care providers. Several authors have posited that mistrust of medicine and research among African-Americans and other minority groups may be rooted in experiences extending back to slavery and continuing to the present day.^4–6 One view is that blacks’ cultural mistrust and suspicion developed in response to racism and mistreatment in medical research and clinical settings by the larger American society.^7,8 For example, Thomas and Quinn⁹ stated that the strategies used to recruit and retain participants in the Tuskegee Syphilis Study of untreated syphilis in black men were quite similar to those being advocated for HIV/AIDS prevention programs today. While trust affects almost every aspect of the doctor-patient interaction, from personal disclosure to cooperation in treatment,¹⁰ culture strongly influences how patients experience health, illness and medical care.¹¹

Using a community-based representative sample, we tested the hypothesis that sociodemographic psychosocial variables—trust, education, income, race/ethnicity and perceived health status—predict having a regular dental home.

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Subjects. We obtained data for this study from a 2001 telephone-based survey of 1,005 low-income people (403 men and 602 women) aged 18 years and older that was conducted to assess their knowledge of and attitudes about preventive health care. We randomly selected the subjects from two metropolitan counties in Florida: Miami-Dade (n = 518) and Duval (n = 487). Complete data on the major analytic variables were available for 962 (95.7 percent) respondents. On the basis of responses to questions about race and Hispanic origin, we classified 128 respondents as Hispanic (of any race), 515 as non-Hispanic blacks (blacks), 255 as non-Hispanic whites (whites) and 64 as other unspecified race/ethnicity. The mean age of the subjects was 43.9 years, and the median education level was 13 years.

Instrument development. The 70-item survey instrument consisted of items that assessed trust in the health care system, concerns about AIDS, attitudes about and willingness to participate in cancer screenings, perceived racial discrimination in the health care system, medical and dental care attendance patterns and various demographic information. The questions were drawn from multiple sources including prior surveys,^12,13 a previously published study¹⁴ and items created for this survey.

A panel of experts reviewed the questionnaire. Then we pilot tested the questionnaire among a convenience sample of adults, modified it and had it reviewed by a panel of telephone survey experts with the University of Florida Bureau of Economic and Business Research. We administered the resulting survey in September and October 2001 from 9 a.m. to 9 p.m. Monday through Friday, Saturday morning and afternoon, and Sunday afternoon and evening.

Telephone survey and sampling scheme. We used a random digit dialing, or RDD, telephone-sampling database (GENESYS, Marketing Systems Group, Fort Washington, Pa.) to select the sample for this study. The database, which is updated quarterly, contains telephone banks (area code plus prefix plus first two digits of suffix) that have at least one residential number listed in the white pages. The telephone banks we targeted represented low-to-moderate income households according to U.S. census tract data. The telephone banks, which are geo-coded to census tracts, permit links to the corresponding census and Current Population Survey data for those census tracts. This enabled us to isolate telephone banks in a particular county where we expected to find a minimum proportion of black households.

The primary objective of the sample design was to ensure that we could compare white and black respondents. The sample design we selected met this need, as it was not an objective of this study for us to make inferences to the entire state of Florida. We focused on specific census tracts that maximized the probability of contacting black households. We found that this sample design ensured that we would interview whites with similar socioeconomic levels as the blacks in those areas. That is, those whites who were living in high-density black areas typically also were disadvantaged economically, with a few exceptions. Thus, we could compare whites and blacks for differences in their health care behaviors without being as concerned about intervening variables such as education level and income.

For this survey, we initially defined two strata, one in Duval County and one in Miami-Dade County, in which at least 50 percent of the households were black. Then we generated and released 2,500 telephone numbers for each stratum. When it became apparent that there were not enough white respondents, we created two additional strata, one in Duval County and one in Miami-Dade County, in which at least 30 percent of the households were black, and we released an additional 2,000 telephone numbers for each stratum. Given that this was a telephone survey and that we used RDD to contact both listed and unlisted households, we had to use a database of valid telephone banks to generate the actual numbers called.

We recognized that the black culture might not have been homogenous across communities. Thus, we selected samples in Duval County and in Miami-Dade County. We expected respondents in Duval County, which includes the city of Jacksonville, to be more representative of North Florida communities culturally, with possible influences from the large naval base. And we expected respondents in Miami-Dade County to reflect the heavy influence of Latino culture. We attempted to sample 500 respondents in Miami-Dade County and in Duval County, with approximately 250 black respondents and 250 white respondents in each county.

The University of Florida Bureau of Economic and Business Research conducted the telephone-based survey, using RDD generated with Win-CATI (Version 4.1, Sawtooth Technologies, Northbrook, Ill.). Each telephone interview lasted approximately 20 minutes and the mean yield was .7 completed surveys per hour. Approximately 20 percent of the telephone interviewers were black and 60 percent were female. The survey software used call rules to rotate sampling across days and times. Numbers were called a maximum of 10 times before being finalized as unproductive. Interviewers called people who refused to participate twice.

Response rates varied, depending on the definition. Given the ambiguity in the field surrounding response rates, we present the following to show the distribution of dispositions. To complete 1,005 surveys, 9,154 telephone numbers were selected of which 1,106 were for businesses, institutions or group quarters, which were ineligible to participate. There were technical problems, no answer or consistent busy signals for an additional 5,494 telephone numbers. A total of 349 telephone numbers reached households in which physical, mental or language barriers made the potential respondents ineligible to participate. No eligible respondent was available at 353 telephone numbers, and 847 people refused to participate or did not complete the telephone survey.

Analytic variables. The dependent variable in our study was whether a respondent reported having a regular dentist, which we used as a proxy for having a dental home. We selected variables that might predict who has a regular dentist based on the findings of a national survey.¹⁵ We included the variable for trust of medical and dental care providers to test our hypothesis that people who trust medical and dental care providers are more likely to have a dental home. The interviewers asked respondents, "How much do you trust the people who provide medical and dental care?" They scored the responses on a five-point Likert-type scale, ranging from "completely" to "not at all," and then recoded them as high level of trust (completely and quite a bit), moderate level of trust (some) and low level of trust (a little and not at all). Interviewers also asked respondents to rate their general health as excellent, very good, good, fair or poor. The variables then were recoded as "excellent," "good," "average," "fair" and "poor."

Other predictor variables we were interested in were current employment status, education level (less than high school, high school or more than high school), race/ethnicity, sex, age (18–24, 25–40, 41–64 or 65 years and older) and median household income. The director of the survey matched the telephone number for each respondent to the census tract as defined by the U.S. Bureau of the Census. We used the median household income for the respondent’s census tract as the proxy for each person’s income. For analysis purposes, we split the census tract median household income variable at the median, which resulted in median household incomes of $31,350 or less and greater than $31,350.

Data analysis. We confined our data analyses to respondents 18 years and older for whom we had complete data. We conducted a bivariate analysis using ² tests. We examined the association between the selected predictor variables and having a dental home simultaneously by including them in a multiple logistic regression model. We considered the resulting odds ratios, or ORs, statistically significant if the confidence intervals did not include 1.

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Characteristics of respondents. Table 1 shows the description of the study group. Respondents with complete data (n = 962) included 515 blacks, 255 whites, 128 Hispanics and 64 people of other or unspecified race/ethnicity. A total of 575 participants (59.7 percent) were women, and 387 (40.2 percent) were men.

The prevalence of having a regular dentist differed significantly by education level and employment status. A total of 71.4 percent of respondents with more than a high-school education, 53.3 percent of respondents with a high-school education and 33.1 percent of respondents with less than a high-school education reported having a regular dentist (P < .01). Among employed respondents, 66.1 percent reported having a regular dentist compared with 47.1 percent of unemployed respondents. Respondents living in census tracts with median annual household incomes more than $31,350 were more likely than those residing in census tracts with median incomes of $31,350 or less (62.8 percent and 55.9 percent, respectively; P < .05) to have a regular dentist.

Multivariate analysis. Table 3 (page 1555) shows our findings from the final multiple logistic regression model of having a regular dentist. After adjusting for other variables in the model, we found that respondents with moderate levels of trust in physicians and dentists were 52 percent less likely (OR = 0.48; 95 percent confidence interval, or CI, 0.26 to 0.89) than those with high trust to have a regular dentist. Similarly, those with low trust were 54 percent less likely (OR = 0.46; 95 percent CI, 0.28 to 0.75) than those with high trust to have a regular dentist. Race/ethnicity, sex, age, education level and employment status remained significant correlates of having a regular dentist in the multivariate model. After controlling for other factors in the model, we found that median household income was not a significant predictor of having a dental home. Self-reported health status showed a trend consistent with that seen in bivariate analysis, that the likelihood of having a regular dentist tended to decline with decreasing levels of self-reported general health, though the confidence intervals for the individual OR estimates were wide and generally included 1.0.

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

The importance of having a dental home and the frequency of dental visits has been reported previously.^16–18 Data from the International Collaborative Study of Oral Health Outcomes provided evidence that having a usual source of dental care was the strongest and most consistent predictor of a dental visit in the past 12 months, regardless of geographic location, dental care delivery system or cultural diversity of the population.¹⁹ Further, Nowak and Casamassimo³ reported that having such an established relationship between the practitioner and the patient would foster care that is accessible, coordinated and compassionate, and would encourage mutual responsibility and trust. Only a few other studies,^17,19,20 however, have looked at factors related to reporting a usual source of dental care or a dental home, and to our knowledge, ours is the first to examine the role that trust in the health care system might play in having a dental home.

The growing concerns about racial/ethnic disparities in dental health^1,21 and strong evidence that minorities lack dental homes make it reasonable to assume that increasing the number of minorities who have dental homes or usual sources of general health services might improve their oral health.^17,22,23 Although establishing causality with cross-sectional data is problematic, our data suggest that an important step toward facilitating dental homes among minorities would be to increase the level of trust that many minorities have in the health care system.

Mechanic¹⁰ defined trust as the expectation that people and institutions will meet their responsibilities to the person. Terrell and Terrell⁸ described cultural mistrust as blacks’ tendency to distrust white Americans and traditional American systems. It follows that a step toward increasing trust, while simultaneously increasing the likelihood of having a dental home, will come when the minority community perceives that the dental profession is meeting its responsibility to them. This most likely will occur through an ongoing dialogue between policy-makers, practitioners and the diverse groups of patients within the communities.

An explanation for the association between trust and having a dental home is that people who have a usual source of dental care develop trust in the overall health care system.

An alternative explanation for the association between trust and having a dental home is that people who have a usual source of dental care develop trust in the overall health care system. This explanation has important policy implications. Whether services are provided in the public or private setting, having a dental home that includes a regular provider may increase patients’ trust in the health care system and increase patients’ overall willingness to participate in their own care.¹⁹

Although the goal of our study was to identify predictors of having a dental home, the results are similar to those that predict dental care utilization. For example, analysis of a comprehensive, nationally representative data set from the 1989 National Health Interview Survey, or NHIS, showed that dental care utilization was related strongly to race, sex, income, education level and dental insurance coverage.¹⁵ Large differences in the use of dental services among blacks, Hispanics and whites were reported, even after controlling for education, income, age and other variables.

An Alabama study found that a statistically significantly higher (P < .05) proportion of Medicaid-eligible whites (31.2 percent) used dental services compared with Medicaid-eligible blacks (24.0 percent) or Medicaid-eligible people in other racial groups (21.8 percent).²⁰ This finding suggests that there are other variables besides socioeconomic status that may account for disparities in use of dental care services. Similarly, Watson and colleagues²⁴ studied the impact of income on children’s and adolescents’ preventive dental visits and reported that increasing financial eligibility alone has been found to be unlikely to sufficiently improve low-income children’s access to care. Another study found that blacks and Hispanics with dental insurance coverage and high incomes also had low utilization rates.¹⁵ The authors of the study speculated that barriers to care (such as lack of transportation or the inability to take time off from work to seek care), language barriers and cultural mistrust were possible explanations.

In our study, we found that higher education was associated with a greater likelihood of having a regular dentist. Moreover, the percentage of respondents with more than a high-school education was similar to that reported in the 2000 U.S. census.²⁵ Similarly, in their study of the 1989 NHIS, Manski and Magder¹⁵ found that respondents with low levels of education were less likely (P < .001) to report visiting the dentist than those with high levels of education.

Global assessments, in which a person rates his or her health as "poor," "fair," "good," "very good" or "excellent" can be reliable indicators of his or her perceived health.¹ According to Schulz and colleagues,²⁶ racial disparities in health status often remain even after adjusting for household or individual income and education. In their study of the cumulative effects of multiple stressors on women’s health, they found that black women reported having significantly poorer health statuses than did white women, regardless of where they lived (P < .01). Our study found that those who perceived themselves as being in excellent health were more likely to report having a regular dentist. Similarly, Davidson and colleagues¹⁹ found that people who believed in the importance of oral health were significantly more likely to visit the dentist in the past 12 months.

Another potential barrier to visiting a dentist regularly is dental anxiety, since anxious patients tend to delay or avoid dental treatment, which results in a strong association with extreme deterioration of oral health.²⁷ We did not measure anxiety; however, if the fairly new idea of having a dental home, which is patterned after the medical home concept, is adopted, it could expose patients to prevention and early intervention, as well as reduce levels of anxiety.³ Moreover, an approach to care that embraces cultural understanding is more effective than one that does not, leading to better outcomes as it helps promote patient autonomy, beneficence and justice.²⁸

Limitations. The income data in our study are based on census tract data from the 2000 U.S. census. Because reliable self-reported income is difficult to obtain in telephone-based surveys, we sacrificed some degree of precision. We could not make more subtle forms of income discrimination. However, there is some evidence that census tract–level measures of socioeconomic status are reasonable proxy measures for person-level measures in health studies,^29,30 though they may perform better for the majority population than for minority populations.³¹

A second limitation pertains to the cross-sectional nature of the data, which precluded causal inferences. Nevertheless, these data are consistent with other research showing that racial/ethnic differences in the use of dental service are related to a host of socioeconomic factors.

A third limitation is that telephone survey methodology assumes telephone ownership is nearly universal and that virtually all households would be in the sampling frame. We know this is an unfounded assumption; rates of telephone ownership tend to be lower in low-income areas and some racial/ethnic communities than in higher-income areas and those that are predominantly white.³² We also do not know whether the relatively large number of telephone numbers for which there was no answer resulted in some degree of sampling bias.

A fourth limitation was that women were over-represented in the group of black respondents. However, the size and racial/ethnic composition of our sample gave us some confidence in the representative nature of our respondents.

	CONCLUSIONS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
This study expands on previous studies by identifying predictors of having a dental home. The results of our community-based study have important implications for eliminating racial/ethnic disparities in the regular use of dental services. To our knowledge, no other study has shown that beyond racial/ethnic group membership, sex, age, education, income and employment status, trust in the health care system predicts having a dental home. These results suggest that efforts to reduce disparities in access to dental care must include programs to increase low-income minority patients’ confidence that their best interests are a top priority of the health care establishment. In short, while policy-makers often consider ways to improve access to dental care, dental professionals should work at the community level to earn people’s trust and educate them about oral diseases and their impact on systemic health and general well-being.

View larger version (124K): [in this window] [in a new window]   Dr. Graham was a clinical assistant professor, Department of Operative Dentistry, Division of Public Health Services and Research, University of Florida College of Dentistry, Gainesville, when this article was written. She now is a Ph.D. candidate in oral epidemiology, Harvard School of Public Health and Harvard School of Dental Medicine, Boston.

View larger version (123K): [in this window] [in a new window]   Dr. Logan is a professor, Department of Operative Dentistry, Division of Public Health Services and Research, University of Florida College of Dentistry, Gainesville, 1600 SW Archer Road, Box 100404, Gainesville, Fla. 32610, e-mail "hlogan{at}dental.ufl.edu". Address reprint requests to Dr. Logan.

View larger version (117K): [in this window] [in a new window]   Dr. Tomar is an associate professor, Department of Operative Dentistry, Division of Public Health Services and Research, University of Florida College of Dentistry, Gainesville.