Some of these findings are consistent with those from a survey of the frequency of dental care service utilization we conducted among women with HIV participating in the northern California site of the Women’s Interagency HIV Study (WIHS) from May 1, 1995, through August 1996.¹³ This prior study of 213 HIV-infected women in the San Francisco Bay area found that women with HIV were underutilizing dental care services, and that fear of dentists and lack of information regarding available resources were important barriers to getting this care. One limitation of this study was that it did not include as a comparison group women who were not infected with HIV. The absence of a comparison group did not allow us to infer any conclusion on how HIV infection may affect utilization of dental care.

Therefore, to explore the frequency of dental care utilization and identify the main barriers to access to dental care among WIHS participants over time, as well as how these may vary with geographic location and by HIV infection status, we conducted an expanded follow-up study including HIV-infected participants from another WIHS site (Chicago) and HIV-uninfected participants from both the northern California and Chicago WIHS sites. One hypothesis was that the participants in this kind of prospective study of HIV-related outcomes may become more aware of their health, including oral health, and increase their utilization of preventive dental care services.

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Study population. Each survey participant was enrolled in the northern California site or the Chicago site of the WIHS, a prospective multi-center study of the natural history of HIV infection among adult women in five large U.S. cities.^14,15 As previously described, WIHS participants were recruited from AIDS clinics and hospital-based clinics that provide primary care services to a large proportion of women living with HIV infection in both northern California and Chicago.^13–15 Participants also were recruited from general medical clinics, from other cohort studies, via word-of-mouth referral, through distribution of flyers and notices in local newsletters, and by personal contact made by outreach workers employed by local advocacy organizations. Women infected and not infected with HIV were recruited from similar sources and were frequency-matched on demographics and key risk factors, including age, race or ethnicity, level of education, injected-drug use (IDU) since 1978 and number of sexual partners since 1980.¹⁵ WIHS participants were followed at six-month intervals, and project assistants obtained informed consent to participate in our survey at the time of the WIHS baseline visit or at one of the follow-up visits. All enrollees had participated in the WIHS for at least three years. We enrolled HIV-infected and -uninfected women who had attended a WIHS visit between Aug. 1, 1998, and April 15, 1999, in the northern California site, and between June 1, 1999, and Aug. 30, 2000, in the Chicago site. Some of the respondents had participated in our earlier survey in northern California in 1995, while others had not.

Variables and measurements. A trained interviewer administered a standardized questionnaire via a 25-minute telephone interview. Each participant received a small monetary compensation ($10) or a $10 food voucher and a brochure about clinics that deliver free or low-fee dental care by county of residence. We designed the questionnaire to assess the frequency of and factors associated with the utilization of dental care services during the one-year period that preceded the interview. We adapted measures of dental care utilization from the 1989 National Health Interview Survey¹⁶ and the 1975–1976 U.S. population household survey by the Center for Health Administrative Studies.¹⁷

We assessed dental care utilization using two outcomes:

– whether participants visited a dentist in the preceding year (dental use);

– whether participants had a dental cleaning in the preceding year (as a measure of preventive dental care).

We also assessed unmet dental care needs by asking the respondents whether they felt they had needed dental care in the past year but had failed to get it. To assess factors that predicted dental service utilization, we employed the behavioral model, developed by Andersen and colleagues,^18–21 in which predisposing, enabling and need (PEN) variables are explored in relation to utilization of health care services over a given period.^18–23

The predisposing factors (which reflect the propensity of a person to seek care) measured in this survey were race, age, education, employment, IDU, dental fear, health status (HIV status and, for HIV-positive women, history of an AIDS-defining condition, CD4 lymphocyte count and current antiretroviral medication) and perception of general health. "AIDS-defining conditions" referred to the specific illnesses that are comprised in the case definition of AIDS,²² but not to a CD4 count of less than 200 cells per cubic millimeter, because we explored that variable independently of AIDS-defining conditions. The enabling variables we considered were annual income and dental insurance status. The need variables included oral disease symptoms experienced in the past year (including toothache, broken or loose teeth, bleeding gingiva, severe pain in gingiva, a swelling in the mouth or a sore on the oral soft tissue) as well as perceived oral health and dentate status. We obtained data on general health status, HIV disease stage, laboratory parameters and sociodemographic variables such as race/ethnicity, age, education, employment history and IDU history from the WIHS database.

Some women who enrolled in our survey also participated in the WIHS Oral Study, as part of which they received a comprehensive oral examination including oral mucosa, periodontal and dental assessments. However, when we asked them if and how many times they had visited a dentist in the preceding year as part of our access-to-care interview, we asked them to exclude any Oral Study visit(s) in their answers. Oral health outcomes measured as part of the WIHS Oral Study will be reported elsewhere. We did not include specific oral health outcomes measured in the Oral Study in this analysis because only a subsample (62 percent) participated in it. However, in our analyses, we did account for whether women participated in the Oral Study (as a predisposing variable) to assess any effect such participation might have on their utilization of dental care services.

Statistical analysis. We applied standard statistics (contingency table analysis) to compare, within the WIHS cohort, sociodemographic variables of women who participated in our survey versus the variables of those who did not, with the aim of exploring whether our subsample was representative of the larger WIHS cohort in the northern California and Chicago sites. We then summarized sample characteristics and the extent to which they did not use dental care services in both sites using proportions. We explored the association between nonuse of dental care services in the preceding year and PEN variables in both sites with contingency tables and ² statistics (or, when relevant, the Fisher exact test). We examined utilization of dental care services among HIV-infected and -uninfected women in both geographic sites by calculating proportions of reported number of dental visits in the past year and the proportion of dentate women who reported having had a dental cleaning in the past year (which reflects use of preventive dental care). We also used contingency table methods and ² statistics (or, when relevant, the Fisher exact test) to investigate any association between use of preventive dental care in relation to PEN variables in both northern California and Chicago sites.

To further explore these associations while controlling for potential confounders, we fit logistic regression models to the following outcomes: no use of dental care services versus one or more dental visits in the preceding year; and no dental cleaning versus one or more dental cleanings in the preceding year among dentate women. We also fit a logistic regression model to examine predictors of unmet dental care needs.

The initial logistic model for each outcome included all independent variables that were associated with the outcome at the .1 level of significance in the contingency table analysis and suspected confounders. We based model selection on comparison of deviances (using ² tests) and on comparison of regression parameter estimates among nested models. We assessed goodness of fit with the Hosmer-Lemeshow goodness-of-fit test.²³

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Study sample. Among the 196 women from the northern California WIHS, 160 were HIV-infected and 36 were uninfected. One hundred nine (68 percent) of the HIV-infected women in northern California had participated in the earlier survey. In the northern California site, when comparing sociodemographic variables among the women who participated in our survey with the WIHS participants who did not enroll in our survey (n = 233), we found no statistically significant difference with respect to race or ethnicity (P = .6), HIV serostatus (P = .3), level of education (P = .07) or history of IDU (P = .4). The median age of the participants in our survey was slightly higher than that of the nonparticipants (40 years versus 37.8 years, P = .002).

Among 167 women recruited from the Chicago WIHS site, 138 were HIV-positive and 29 were HIV-negative. In the Chicago site, when comparing sociodemographic variables among the women who participated in our survey with those of the WIHS participants who did not enroll in our survey (n = 165), we found no statistically significant difference with respect to HIV-serostatus (P = .6), level of education (P = .6), history of IDU (P = .1) or median age (36.5 years versus 35.9 years, P = .9). However, there was a difference between the two groups with respect to race, with a higher proportion of African-Americans (70 percent) and a lower proportion of whites (20 percent) among the survey participants than among the nonparticipants (54 percent blacks and 33 percent whites, P = 0.02).

Predisposing and enabling variables. The majority of survey participants from both northern California and Chicago WIHS groups were black (60 percent and 69 percent, respectively), were currently unemployed (71 percent and 67 percent, respectively) and had used injected drugs at some point in their lives (60 percent and 53 percent, respectively) (Table 1). One-quarter of the women in the northern California site and 37 percent in the Chicago site did not complete high school, while more than 30 percent reported having had some college education. Nearly one-third of women in the Chicago site reported having a fear of dentists, and more than 40 percent perceived their general health as being fair or poor. Only 54 percent of the women in the northern California site and 72 percent in the Chicago site also were participants in the WIHS Oral Study. More than 80 percent had no history of an AIDS-defining clinical condition at the time of the survey. More than one-half of the HIV-infected participants (59 percent in the northern California site and 70 percent in the Chicago site) were taking antiretroviral medications at the time of the interview (the vast majority were receiving HAART, with only four participants in northern California and one participant in Chicago receiving monotherapy).

View this table: [in this window] [in a new window]   TABLE 1 CHARACTERISTICS OF WOMEN RECRUITED FROM THE NORTHERN CALIFORNIA WIHS* AND THE CHICAGO WIHS AND PREVALENCE OF NONUSE OF DENTAL CARE SERVICES IN THE PRECEDING YEAR.

Dental care needs, utilization of dental care services and barriers among nonusers. More than one-half of the women at both sites perceived their oral health as being fair or poor, and more than 70 percent reported having experienced some oral symptoms in the past year (Table 1). A higher proportion (17 percent) of women at the northern California site were completely edentulous, compared with only 7 percent in the Chicago site (P = .004).

The highest prevalence of nonuse of dental care services among HIV-negative women, particularly those in Chicago (79 percent) (Table 2, page 1248), among whom more than 40 percent had not seen a dentist in more than five years and only 15 percent had a professional dental cleaning during the past year. The pattern of utilization of dental care services (time since last visit, number of dental visits in the preceding year and having had a dental cleaning in the preceding year) was similar among HIV-positive women in the northern California site in the earlier period (1995–1996) and in the later period (1998–1999), suggesting no change in dental care utilization practices over time in this group.

Utilization of dental care services in relation to predisposing, enabling and need factors: univariate analyses. In the northern California site, there was a statistically significant association between nonuse of dental care services and race or ethnicity other than white (Table 1). Nearly two-thirds of black women and 70 percent of Latina women reported a lack of dental service use in the preceding year, compared with 28 percent of white women. However, we did not observe this association between lack of dental care and nonwhite race in the Chicago site, where almost one-half of the white women also reported that they had not used dental care services. We had similar findings when we excluded from the contingency table analysis the one subject whose race was not known. Also at the Chicago site, women who were unemployed, had used injected drugs in the preceding year, reported having fear of dentists, were HIV-negative, were not taking any antiretroviral therapy, and had either Medicaid or no health care payer coverage reported a significantly lower prevalence of dental care service use than other women (Table 1). At the northern California site, the only factors, in addition to race, that were statistically significantly associated with lack of dental care use were the absence of a history of an AIDS-defining condition and the perception of fair or poor oral health. Concomitant participation in the WIHS Oral Study did not affect the use of dental care in the preceding year in either site.

At both sites, women who were not white, reported fear of dentists, were not taking anti-retroviral therapy, had an annual income of $6,000 or less or perceived their oral health as being fair or poor were significantly less likely to report having had a dental cleaning in the preceding year than those who did not have these characteristics (Table 3, page 1250). At the northern California site, women who did not complete high school and those who had a perception of poor general health also were less likely to report a dental cleaning during the preceding year. Women with no history of an AIDS-defining illness were less likely to have had a cleaning. For women enrolled in Chicago, being employed and not having used injected drugs in the preceding year were associated with having received a dental cleaning. Concomitant participation in the WIHS Oral Study was not associated with whether women had received a dental cleaning in the past year in either site.

View this table: [in this window] [in a new window]   TABLE 4 MULTIVARIATE ANALYSIS TO EXPLAIN UTILIZATION OF DENTAL CARE SERVICES IN THE PAST YEAR AMONG WOMEN RECRUITED FROM THE NORTHERN CALIFORNIA WIHS* (n = 196) AND THE CHICAGO WIHS (n= 167).

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
The 363 women participating in the northern California and Chicago WIHS sites were predominantly black, were unemployed, had an IDU history, had no history of an AIDS-defining illness and had a CD4+ lymphocyte cell count greater than 200 cells/mm³. This sample was representative of the larger WIHS cohort at both sites, except with respect to race at the Chicago site where the proportion of black women was higher in the survey participants. Previous publications also have shown that the WIHS seropositive cohort at baseline was similar in terms of race or ethnicity, exposure status and age, to national AIDS cases among U.S. women reported in 1995.¹⁵ The lowest prevalence of use of dental care services was found among HIV-negative women, particularly at the Chicago site, which could be explained partly by the higher proportion of black women in the sample recruited from this site. The strongest predictors of the failure to report use of dental care services, lack of dental cleaning in the past year and the occurrence of unmet dental needs included nonwhite race, fear of dentists and a perception of fair or poor oral health. However, a history of an AIDS-defining condition was found to be associated with greater utilization of dental care services, including a dental cleaning, and a lesser occurrence of unmet dental needs. The pattern of utilization of dental care services was similar among HIV-positive women in the northern California site in the earlier period (1995–1996) and in the later period (1998–1999). Furthermore, we found no difference with respect to use of dental care (including preventive care) between the subsample of women who participated concomitantly in our survey and in the WIHS Oral Study as part of which they received a comprehensive oral examination (and were referred for care if any oral disease was found) and the women who enrolled in our survey but not in the oral study. This suggests that participation in a longitudinal study, as part of which women receive regular health evaluations (including oral evaluations), has little or no effect on their subsequent use of dental care.

Reports from a number of studies on utilization of dental care among HIV-infected adults have been published,^{11–13,24–30} but few studies have included HIV-uninfected comparison groups. Therefore, in the follow-up study we conducted three years after the initial cross-sectional study,¹³ we expanded our sample to include not only HIV-infected women from a second geographic location, but also HIV-uninfected women. Expansion of the study population revealed that HIV-uninfected women had the lowest prevalence of dental care use, particularly the women who were enrolled at the Chicago site. Furthermore, women with a more advanced stage of HIV disease (those with a history of an AIDS-defining condition) were more likely to have used dental care in the preceding year, including preventive care.

Women with a history of an AIDS-defining condition also were less likely to report having an unmet dental care need in the past year. This may be explained in several ways:

– infirm patients’ increased access to case managers, who can facilitate patients’ receipt of dental care;

– the increased likelihood that more infirm patients have seen medical providers who could have referred them for dental care;

– increased diligence by medical providers in detecting oral conditions among infirm patients;

– greater concern regarding general or oral health among patients with HIV infection.

The higher prevalence of dental care utilization among HIV-infected women as compared with HIV-uninfected women may be explained by the availability of free care for HIV-infected patients under the Ryan White Comprehensive AIDS Resources Emergency (CARE) Act.³¹ The specific lack of dental care benefits for Medicaid patients in Chicago probably explains the lower prevalence of dental care use among HIV-uninfected women at that location. In 2000, the inaugural U.S. surgeon general’s report on oral health emphasized the importance of oral health with respect to overall health among Americans.³² Our study, which indicates that lack of access to dental care is even greater for HIV-uninfected patients than for HIV-infected patients living in cities that receive Ryan White CARE funds, is yet another example of the shortcomings of the current health care system in the United States, and demonstrates the need for more thoughtful and comprehensive considerations of health care funding.

More than 70 percent of women in our study who did not use dental care services reported that they had felt they needed dental care in the preceding year. As stated by Grembowski and colleagues³³ in their review of the dental care process, a wide variety of factors—behavioral, cultural, geographic and infrastructural—influence a person’s ability to obtain care, aside from the financial aspect. Therefore, the process of evaluating a population’s access to dental care is complex, and numerous conceptual approaches have been proposed.³⁴ We selected Andersen and colleagues’ model^18–21 to explain utilization of dental care among HIV-infected women because this model is well-recognized and has been used in many studies assessing dental care use,^35–37 including studies among HIV-infected populations.^{11–13,24,29} Overall, findings in the studies of populations with HIV disease were consistent with ours and showed that low income, lower educational attainment, lack of dental insurance, being black or Hispanic and being female were predictors of low utilization of dental care. Furthermore, Heslin and colleagues,²⁶ in a probability sample of 2,864 HIV-infected adults, found that unmet dental needs were twice as prevalent as unmet medical needs. Similarly, Kenagy and colleagues,¹⁰ in a longitudinal survey of HIV-infected adults receiving primary care at Title 1–funded clinics in Chicago, found that the highest proportion of unmet needs concerned dental care (28 percent), followed by housing and transportation.

Fear of dentists was a strong and independent predictor of the three outcomes that indicated lack of use of dental care. Thus, fear and anxiety related to dental care probably constitute a significant barrier to obtaining dental care in this group of high-need patients. Others have demonstrated that dental anxiety was an important barrier to dental care utilization in a population of adults selected by random-digit dialing in Detroit³⁸ and in an elderly population.³⁹ Another survey of 1,010 randomly selected Seattle residents revealed that 25 percent of the women and 16 percent of the men interviewed reported a level of dental care fear ranging from being "somewhat afraid" to "terrified."⁴⁰ That study showed a high correlation between dental care fear and low prevalence of use, delays in making appointments and missed appointments. These findings suggest a need to not only educate dental professionals in the management of patients with dental fear, but also to promote public health campaigns that raise patient dental awareness and help alleviate fear patients may have before entering the dental practice.

The high prevalence of edentulous patients in the study population is another indicator of suboptimal dental care, since total dental extraction is really the most extreme outcome of inadequate dental care. Edentulism was more prevalent at the northern California site; this, ironically, could be the result of the availability of limited Medicaid-funded dental care at this location, which includes extraction rather than the more expensive restorative options.

	CONCLUSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
We found that the prevalence of utilization of dental care (including preventive care) remained unchanged over time in a group of HIV-infected women participating in a longitudinal study as part of which they received regular health evaluations (including oral evaluations). This suggests that information about needed oral care has little or no effect on this population’s use of dental care. However, HIV-infected women had a higher prevalence of dental care use than did HIV-uninfected women, which likely is explained by the availability of free care for HIV-infected patients under the Ryan White CARE Act.³¹ This suggests that the availability of programs that provide free dental care to vulnerable populations do increase the prevalence of use and should continue to be funded.