It is important to explore whether social disparities in adults' oral health present also in adolescents and whether the effects of the material distribution are mediated differently in adolescence and adulthood. This study revealed that a social gradient was present with respect to dental caries, treatment need, and reported oral health status as well as sugar intake, tooth brushing, use of fluoridated toothpaste and dental attendance patterns among adolescents in Kilwa, a generally deprived district with limited access to oral health care services. Differences across educational level, household wealth and place of residence groups were statistically significant for most oral health outcomes and oral health behaviors investigated, both in unadjusted and adjusted analyses. The results confirm previous findings in adolescent-and adult populations, globally [1
]. Adolescents belonging to the less poor households presented with treatment need and dental attendance more frequently than their counterparts in the poorest households. According to Grytten and Holst [27
], several US studies reported a positive association between income and demand for dental care, particularly when treatment need was high. Although dental attendance, oral hygiene behavior and sugar intake varied systematically with oral health outcomes (Table ), social disparities in caries experience and self reported oral health were not attenuated whilst adjusting for those lifestyle patterns (Table ). In spite of some attenuation of the relationship between household wealth and treatment need after controlling for dental attendance (OR reduced from 2.6 to 2.2), a direct relationship persisted that was unexplained by adolescent's dental attendance profile. Thus, the present results accord at best partly with findings from industrialized countries, suggesting that unequal access to dental care explains socio-economic disparities in adults' oral health [7
]. In contrast to the experience from US, suggesting that children with untreated caries are less likely to obtain regular dental care, the present study indicate that Tanzanian adolescents who had attended a dentists were those with the most severe caries in terms of moderate to high treatment need [29
]. This is a common finding in developing countries [30
] and might reflect that dental attendance follows from a high treatment need rather than being an unexpected outcome of dental care. It might also reflect delayed treatment demand and limited access to dental care (only one assistant dental officer in Kilwa).
A low caries prevalence of about 20% presented among a minority of the participants is consistent with the caries trends of younger age groups in Tanzania [18
]. In accordance with a suggested positive relationship between dental caries and the level of social development, previous studies have provided evidence of a higher caries prevalence in urban than in rural societies [10
]. In contrast, the present findings accord to some extent with those observed among Brazilian schoolchildren, where living in a rural area almost doubled the odds of having dental caries [35
]. Rural participants showed the highest prevalence of dental caries, but they were less dissatisfied with their oral health and visited a dentist less frequently than their urban counterparts. This is consistent with previous studies of schoolchildren in sub Saharan Africa [12
]. Affordability, accessibility and structural barriers reflected by place of residence as a more area based measure of deprivation, have previously been reported to be important reasons for rural dwellers' non-use of dental services in Tanzania [36
]. Nevertheless, independent of socio-economic position, students with caries experience and moderate to high need for dental treatment were the most frequent dental attendees suggesting a demand for dental care among young people in Kilwa.
The Lorenz curves presented in this study were based on the participating rural and urban population groups. Each group showed almost similar skewness dominance. Previously polarization mostly has been reported from western industrial populations. Macek et al
] described highly skewed caries data from young children, while the caries data from adolescents were more dispersed. However, more recent data of 15-yr-old Danish children collected from the Danish National Board of Health in 2006 [38
], were more comparable to the present findings. Seventy-five percent of the total number of DMF-surfaces was found in 13% of Danish 15-yr-old children with the highest DMFS. It is interesting that the caries distribution among adolescents in a poor Tanzanian district is comparable to western industrial populations.
Socio-economic status of parents as assessed in terms of income, education and occupation, might be difficult to assess because of unawareness and unwillingness to reveal such information on the part of the adolescents, resulting in high rates of non responses particularly among lower socio-economic status groups [39
]. The present study assessed socioeconomic status using a wealth index based on a weighted sum of self-reported household assets and the more conventional measure of parental education [1
]. The wealth index showed good discriminate power against moderate to high need for treatment and oral health related behaviors even in the small, social homogeneous and generally deprived district of Kilwa. It bears similarity to the family affluence scale, FAS, developed as a supplementary measure of socioeconomic status for adolescents by the WHO Collaborative cross national study of Health Behavior in School aged Children [26
]. Both the weighted wealth index and the FAS contain items reflecting family expenditure and consumption that are relevant to the family circumstances. In addition, this study used a surrogate area based social indicator of place of residence, as suggested by Locker [1
The present study contributes to the knowledge of adolescents' oral health situation in deprived areas of Tanzania. However, the present results should be interpreted in the light of limitations that include a cross- sectional design and use of self-reported measures. Some schools in the selected wards were not accessible due to natural calamities in the area at the time of data collection and the number of enrolled standard six adolescents and attendance rates in rural schools were particularly low. Nevertheless, use of an unequal sampling fraction in urban and rural parts of Kilwa was compensated by providing weighted estimates for all oral health outcomes and oral health behaviors investigated. Since the present data rely on self reporting, they might have been biased by under-and over reporting due to socially desirable responses and poor recall effect. However, the core questions utilized in this study have shown a good validity and reliability in previous studies focusing children, adolescents and adults in sub Saharan Africa. The self-reported use of fluoridated toothpaste is questionable as we did not validate whether the toothpaste used contained the recommended free fluoride ions concentration. As it has been reported that toothpaste manufactured and sold in Tanzania has free fluoride concentrations below the recommended levels for prevention of caries [40
]. Given that the data collection was conducted in a relatively short period of time, the presence of any temporal changes could hardly have confused the data.