Oral health, like general health, is strongly influenced by socioeconomic status, or SES.1–5 Disparities in oral health status between low-SES and high-SES children have been documented extensively in the dental literature.6 However, we have limited information on which policies and interventions are effective in reducing or eliminating disparities in general or oral health.
Promoting access to dental care by itself may not lead to the elimination of disparities in dental caries severity among children at different socioeconomic levels.
Attempts to reduce disparities through increasing access to medical treatment have not reduced the inequalities in health status between people with low and high SES.1 Recent legislative initiatives in the United States (such as the Children’s Health Insurance Program) have focused on increasing access of low-income children to health and dental care. These initiatives are based on the assumption that if access to dental care of low-income children increases, then inequalities in oral health status may be reduced. This contention is supported by the findings from the Rand Health Insurance Experiment, or RHIE, in which researchers found that preschool children enrolled in a "free" dental insurance program had a significantly lower mean number of decayed teeth and a significantly higher prevalence of caries-free status than those enrolled in plans with copayments.7 The RHIE researchers also found that children from "low-income families benefited most from having access to free dental care." While the RHIE study showed the efficacy of having free dental care, the concept’s effectiveness has not been thoroughly evaluated in a real-life situation in North America. Medicaid data are not useful for evaluating the impact of full coverage on reducing disparities in caries experiences between recipients and nonrecipients of Medicaid because the Medicaid dental program does not cover all the children with the same benefits in the United States (in other words, it is not universal). The Medicaid dental program also had a low rate of utilization of dental services among its participants, making the data even less useful in an evaluation of the program’s impact on oral health disparities.8
Data from other countries that have universal dental insurance programs for children (such as the United Kingdom) do not support the assumption that access to dental care can reduce the relative inequalities in dental caries prevalence between low and high socioeconomic groups.9,10 Even in an environment wherein all children had "free-of-charge" coverage for diagnostic, preventive and basic restorative services, dental caries experience remained clustered in children from low-SES backgrounds.
The province of Nova Scotia has had a publicly financed universal dental insurance program for children since 1975, providing basic preventive, restorative and surgical services.
To determine the impact of access to dental care, we undertook a study using data from one North American site: Nova Scotia, Canada, where a universal dental insurance program for children has been in operation since 1975. All children, regardless of income, are covered by one insurance program that provides basic preventive, restorative and surgical services. The program is financed from the operating budget of the government of Nova Scotia. In this program, dentists are paid on a fee-for-service basis using one provincial fee schedule. The fee schedule is revised periodically via negotiation between the government and the Nova Scotia Dental Association. In essence, the program is equivalent to a preferred provider organization, or PPO, in which all dentists charge the same negotiated (and usually reduced) fees. This article presents, for the first time, data on dental caries scores of a representative sample of 6- and 7-year-old children in Nova Scotia, who have had full access to and relatively high utilization of dental services throughout their lives.
 |
METHODS
|
|---|
The goal of our study was to answer the following question: in an environment of universal access to dental care with a high rate of utilization among children since birth, does low SES remain a significant risk factor for development of dental caries in primary teeth?
The province of Nova Scotia is located on the Atlantic coast of Canada. It has a population of about 950,000 people. The province has had a publicly financed universal dental insurance program for children since 1975. The dental program, like all other health programs in Canada, has undergone significant changes since its inception; however, its basics have remained the same. Dentists are the sole providers of dental care to children, and they were paid for rendered services following a fee schedule. (At the time of the study, the program covered all children until their 12th birthdays). Dentists are not allowed to extra-bill—or charge the parents for the difference between the fees that were negotiated with the government of Nova Scotia and their usual and customary fees—for the dental services covered by the program. However, there were no restrictions on billing for noncovered services.
Study design.
In 1995 and 1996, an epidemiologist (A.I.I.) and a statistician selected a representative sample of all first-grade children in Nova Scotia, Canada, for each of the four health regions in the province. We chose first-graders for this analysis because we had information on their residence in the province since birth. We had data on sixth- and ninth-graders, but we could not determine with confidence whether they had resided in the province all of their lives.
Of a total population of about 12,000 first-grade children, we sampled 1,614 children; 1,342 of those consented to complete a questionnaire and did so. Of those who completed the questionnaire, 1,271 were examined by two trained full-time dentists hired specifically to conduct the examinations. Of those children, 955 were lifelong residents in the province. All included children were 6 or 7 years old.
For each health region, we classified schools into quartiles according to the size of their student population. A sample of equal size was systematically selected from each quartile. We developed this stratification to select a large enough sample from small rural areas in each health region. The dentists examined all children in the selected schools. We used weights in the analysis to adjust for disproportionate selection probabilities of students from the different quartiles and regions.
Questionnaire.
For each selected school, we hired a coordinator to work with the research team. After school boards, school principals and teachers all gave their approval, the children hand-delivered introductory letters and consent forms to their parents. The school coordinators contacted the parents by telephone to answer questions and check on whether they had received a letter, a consent form and a questionnaire. In the questionnaire, parents answered questions on the use of dental services by their children (age and reason for first dental visit and frequency of dental visits), oral hygiene habits and use of fluoride products. Additionally, they were asked to indicate the highest educational attainment in their household (completed elementary school; completed high school; completed community college, including vocational training; and completed university education). One of the authors (A.I.I.) pretested the questionnaire with 30 parent volunteers in a pilot study in 1995. The questions on frequency of brushing and flossing, age at first visit, reason for first visit, frequency of dental visits and educational status had good to excellent reliability, with 
coefficients ranging between 0.55 and 1.00.
Only 3.1 percent of the children visited a dentist for the first time after the age of 5 years.
Dental examination.
All children whose parents consented were examined for presence of dental caries, sealants, restorations, missing teeth, fluorosis and gingival bleeding. The two trained survey dentists conducted the examinations at the selected schools using portable dental chairs, fiber-optic lights, compressed air syringes, plane mouth mirrors and no. 23 explorers. The dentists were trained by the epidemiologist for two months before the start of the study. The training was conducted using slides, extracted teeth and examination of children. The epidemiologist checked the quality of data throughout the study by re-examining children and comparing the dentists’ findings with his own.
Modified World Health Organization criteria11 were used in this study. The modification included scoring the "decayed" and noncavitated pits and fissures separately from cavitated surfaces. Noncavitated carious pits and fissures were defined by the presence of light- or dark-brown discoloration at the base of a pit or fissure or white demineralization at the sides of a pit or fissure that the survey dentists detected visually after cleaning and drying the teeth. These surfaces were to have had no loss of tooth structure (in other words, no cavity). Cavitated pits and fissures had a loss of tooth structure that could be detected visually, or felt with an explorer, and a softened floor or wall or undermined enamel. The intra- and interexaminer reliability in diagnosing noncavitated and cavitated pits and fissures was excellent. coefficients ranged from 0.91 to 0.97 for noncavitated lesions and from 0.88 to 1.0 for cavitated lesions.
In this study, we analyzed the mean numbers of decayed (whether noncavitated or cavitated), missing or filled surfaces, or dmfs, of primary teeth separately or as a sum. Because of the age of the children and the expected exfoliation sequence of primary teeth, the two survey dentists scored missing primary molars as "missing because of caries." All other missing primary teeth were coded as "unerupted permanent teeth."
Fluoride concentration in schools’ water supply.
The examining team—consisting of a survey dentist, a recorder and a coordinator—collected water samples from all schools using standard collection tubes. These were analyzed for fluoride concentration using an ion-selective method.12,13 The analysis was conducted by MDS Environmental Services Ltd. in Halifax, Nova Scotia. Control samples from fluoridated areas were included to help us verify the reliability of the water analysis by MDS.
We had data on each child’s water exposure only relating to the school he or she attended at the time of the survey. Data on exposure to water in previous schools that he or she may have attended were not available.
Data analysis.
The research team obtained the data used in this analysis from examinations of clusters of children in elementary schools. Because students are selected as clusters rather than as individuals, conventional statistical packages that assume the selection of subjects independently of each other can result in false positive conclusions (increased type I error). To adjust for this clustering effect, we used the survey data analysis software package SUDAAN (Release 7.5, Research Triangle Institute) to compute corrected standard errors.14
Additionally, to reduce the potential for making type I errors when conducting multiple tests using the same data set, we used the Bonferroni inequality adjustment.15 In this method, the chance of a type I error (
= .05 in this analysis) for a single statistical test was divided by the number of tests conducted using the same data to produce a group type I error level, and, hence, protect against the chance of reaching erroneous positive conclusions.
We used Poisson regression modeling16 to investigate the association between dmfs scores and potential risk factors. Poisson regression assumes that the underlying distribution of the dependent or outcome variable (dmfs scores) has the following statistical attribute: the expected mean of the dmfs scores is equal to the variance. This distribution is more suited for counts (number of dmfs). Poisson regression modeling for clustered data sets is not available in SUDAAN. Therefore, to account for clustering effects, we used an average design effect of 2 (an average selected from the descriptive analysis conducted using SUDAAN) to adjust the standard errors of the regression coefficients.
|
RESULTS
|
|---|
Dental visits.
The parents of the children in the study reported that 91.6 percent of the children had their first dental visit at or after the age of 2 years (Table 1
). Only 3.1 percent of the children visited a dentist for the first time after the age of 5 years. The majority of the parents reported that the first visit was for a dental checkup. A total of 93.7 percent of the parents reported that their children annually visited a dentist. (The percentage of actual visits reported by dentists on claims forms confirmed that about 95 percent of the children in the province saw a dentist in 1993 and 1994.17)
TOP
ABSTRACT
METHODS
0.88). Of the children who were examined (n = 1,271), 955 were lifelong residents of Nova Scotia, Canada, and so were included in this analysis. Data were weighted and adjusted for clustering (design) effects. 
. A total of 7.8 percent of the parents had only an elementary school education, slightly more than one-third had completed high school and another one-third had completed junior or community college. About one in five of the parents reported that they had completed a university education. 
