| Sign In to gain access to subscriptions and/or personal tools. |
Social Gradients in Oral and General HealthDepartment of Epidemiology and Public Health, University College London, 1-19 Torrington Place, London, WC1E 6BT, UK Correspondence: * corresponding author, w.sabbah{at}ucl.ac.uk
There are social gradients in general health and oral health. However, there have been few studies addressing whether similarities exist in the gradients in oral and general health in the same individuals. We set out to test, using data from NHANES III, whether there are social gradients in oral health, and whether they resemble the gradients in general health. Income, indicated by poverty-income ratio, and education gradients were examined in periodontal diseases, ischemic heart disease, and perceived oral/general health. Our analysis demonstrated consistent income and education gradients in all outcomes assessed. In the adjusted regression models, the probabilities of having poorer clinical and perceived health were attenuated, but remained significantly higher at each lower level of income and education for most outcomes. The results showed similar income and education gradients in oral and general health, implying commonalities of the social determinants of both oral and general health.
Key Words: social gradients • periodontal disease • ischemic heart disease • perceived oral/general health
Social gradients exist in mortality and in most common chronic diseases (Fuhrer et al., 2002; Ferrie et al., 2003; Marmot, 2003). Higher morbidity and mortality rates from all causes have been reported at each lower level of the social hierarchy (Adler and Ostrove, 1999). Different plausible mechanisms have been postulated to explain the gradients. They include material, behavioral, environmental, and psychosocial characteristics (Marmot and Wilkinson, 2006). Social class and deprivation gradients exist in dental caries, periodontal diseases, tooth loss, and edentulousness in adults and children in the UK, Australia, New Zealand, and Chile (Watt and Sheiham, 1999; Locker, 2000; Nuttall, 2003; Thomson and Mackay, 2004; Lopez et al., 2006; Sanders et al., 2006a,b). Some have argued that the pathways to the gradients in oral health, especially periodontal disease, are similar to those in general health (Sheiham and Nicolau, 2005). A few studies have addressed the similarities between social gradients in general and oral health (Poulton et al., 2002; Borrell et al., 2004). However, the subject has not been approached in a comprehensive manner, based on both clinical and subjective measures of health, reliable measures of specific general conditions, and precise and validated measures of periodontal health, in the same individuals and for a nationally representative sample. The emphasis on similarities between determinants of oral health and general health is consistent with the US Surgeon General’s report on oral health, namely, that oral health is an integral part of general health (US Department of Health and Human Services, 2000). The objective of this study was to examine and compare the social gradients in oral health, here indicated by periodontal disease and perceived oral health, on the one hand, and in general health, indicated by ischemic heart disease and perceived general health, on the other.
Data for this research are from the Third National Health and Nutrition Examination Survey (NHANES III) (US Department of Health and Human Services, 1997), conducted in 1988–1994. NHANES used a stratified multistage probability sampling design with a sample representative of the non-institutionalized civilian American population. We used data pertaining to adults aged 17 years and older. The survey included a comprehensive dental and medical examination. Periodontal measures were assessed on randomly assigned half-mouths, one upper quadrant and one lower quadrant, selected at the beginning of the examination, according to the NIDCR protocol (Miller et al., 1987). NHANES included questions about perceived oral and general health, and a questionnaire for angina pectoris based on the WHO Rose questionnaire (Rose, 1982). In addition, the survey included comprehensive demographic, socio-economic, and health data. This included variables on years of education, poverty-income ratio, ethnicity, diagnosis of heart attack, diabetes, medical/dental insurance, smoking, blood pressure, body mass index (BMI), questions pertaining to frequency of exercise per month, and the consumption of fattening foods per day.
Outcome Variables Individuals were defined as having ischemic heart disease if they reported a diagnosis of heart attack or had angina pectoris according to the WHO Rose questionnaire (Rose, 1982). Participants had angina pectoris when they had all of the following symptoms: any chest pain or discomfort, had the pain or discomfort while walking uphill or in a hurry, the pain caused them to stop or slow down, the pain was relieved by standing still, the pain was relieved within 10 minutes, and the pain was around the sternum, left anterior chest, or left arm. Participants who responded that they never walked uphill or in a hurry were considered as having angina if they met the other criteria.
Explanatory Variables Other covariates included ethnicity (White, African, Hispanic Americans, and Other Ethnicities), age, sex, having any medical/dental insurance, diagnosis of diabetes, smoking (current smoker, non-smoker, non-respondent), high blood pressure (more than or equal to 130 systolic or 85 diastolic), and BMI. Frequency of exercise and the consumption of fattening foods were calculated by summing responses from various questions on the frequencies of exercise per month and the consumption of fattening foods per day.
Data Analysis
Consistent gradients in clinical and subjective measures of both oral and general health were found. The prevalence of poorer perceived oral and general health, periodontal disease, and ischemic heart disease was greater at each lower level of poverty-income ratio and education. Similarly, the extent of gingival bleeding, loss of attachment, and pocket depth were greater at each lower poverty-income ratio and education level (Table 1  ).
The prevalence of reporting poorer perceived oral and general health was significantly higher at each lower level of education and poverty-income ratio. After adjustment for age, ethnicity, sex, smoking, and medical/dental insurance, the odds ratios were attenuated, but remained significant (Table 2 ). Other factors that were significantly associated with reporting poorer perceived oral health included ethnicity (African and Hispanic Americans), older age, lack of dental insurance, and smoking. Other factors associated with poorer perceived general health included ethnicity (African and Hispanic Americans), sex (female), and older age.
For both the prevalence of ischemic heart disease and periodontal disease, there were also consistent income and education gradients. In the adjusted models, the odds ratio retained the same overall pattern and remained significant, except for the second highest level of education in the ischemic heart disease model (Table 2  ). Other factors significantly associated with periodontal disease included older age, sex (male), ethnicity (African Americans and other ethnicities), diabetes, and lack of dental insurance. Additional factors significantly associated with ischemic heart disease included older age, diabetes, and high blood pressure. Overall, the gradients in perceived oral and general health and the prevalence of ischemic heart disease and periodontal disease were very similar after adjustment for relevant covariates (Fig.).
There were also consistent gradients in the measures of the extent of periodontal disease, namely, extent of gingival bleeding, loss of attachment and pocket depth. The differences in the disease levels were all significant, with the exception of the second and fourth levels of poverty-income ratio in the loss-of-attachment model (Table 3 ). In the adjusted models, the gradients persisted and were significant, except for the second highest level of poverty-income ratio in the pocket-depth models (Table 3). Other factors significantly associated with extent of gingival bleeding were older age, sex (male), lack of dental insurance, diabetes, and being a non-smoker. In the ’loss of attachment’ model, older age, sex (male), ethnicity (African Americans, other ethnicities), and smoking were associated with greater extent of the disease, while being Hispanic American was significantly associated with a lower extent of the condition. In the ’pocket depth’ model, older age, sex (male), ethnicity (African Americans), and smoking were significantly associated with greater extent of the disease level.
This study showed that income and education gradients exist in perceived oral and general health, periodontal disease, and ischemic heart disease in the same individuals in a nationally representative sample of adult Americans. At each lower level of poverty-income ratio and education, individuals had poorer oral and general health. The health differences across education and poverty-income groups were significant for most outcomes in the binary and adjusted analyses. Social gradients have been reported in perceived oral health (Stahlnacke et al., 2003), in periodontal diseases (Watt and Sheiham, 1999; Unell et al., 2000; Zurriaga et al., 2004; Dye and Selwitz, 2005; Lopez et al., 2006; Sanders et al., 2006a), and in perceived general health and ischemic heart disease (Marmot et al., 1991; Brunner et al., 1997; Ferrie et al., 2002; Banks et al., 2006; Marmot and Wilkinson, 2006; Singh-Manoux et al., 2006). Borrell et al.(2004) found similarities in the social gradients in perceived general and oral health. The findings reported here were consistent with those found by the above-mentioned authors. The present study examined and found similarities between the social gradients in oral health and general health in the same people. In addition, using both clinical and subjective health outcomes, we reported similarities in the gradients. Income and education gradients were attenuated after adjustment for ethnicity, sex, age, medical/dental insurance, smoking, and diabetes, but remained significant for most outcomes. In the adjusted ischemic heart disease model, the probability of having the condition was still higher but lost significance in the second highest level of education. Similarly, the adjusted models for extent of periodontal disease demonstrated the persistence of education and income gradients. Use of the percentage of sites with periodontal disease is important, because it accounts for all those examined for severity of periodontal disease. The fact that the gradients were attenuated, but did not disappear, in the adjusted models implies that confounders used in these models explained part, but not all, of the factors affecting the gradients. Further studies are needed to explore other mechanisms causing the gradients. This study cannot support conclusions about causal effects of socio-economic position, since it is based on a cross-sectional survey. Additionally, the dichotomous periodontal variable used here is consistent with a definition of mild periodontitis (Offenbacher et al., 2001). However, the dichotomous variable was used with other variables indicating extent of periodontal diseases. Lack of data on oral health-related behaviors (such as tooth cleaning) in the NHANES III is another weakness in this study, since that could influence the gradients in periodontal health. However, another study showed that adjustment for behavioral variables did not eliminate oral health gradients in an Australian population (Sanders et al., 2006a). This study demonstrated that there are social gradients in particular measures of oral and general health. The study also highlighted the similarities of the social gradients in oral and general health in the same adult population, by both clinical and subjective measures of health.
The authors thank the US Department of Health and Human Services for making publicly available the data from the NHANES III. No external funding was received to undertake the data analysis. Received for publication August 7, 2006. Revision received May 1, 2007. Accepted for publication May 31, 2007.
Journal of Dental Research, Vol. 86, No. 10,
992-996 (2007) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
4 mm, extent of loss of periodontal attachment 
