Abstract
Objectives: The aim of this study was to test the association of current contextual income inequality, contextual income inequality experienced during childhood, and individual factors with oral-health-related quality of life (OHRQoL) in adolescents. Methods: Individual data of 3,854 adolescents aged 15–19 years from the Brazilian Oral Health Survey (SB Brasil Project) nested within 27 cities and contextual income inequality were analysed. OHRQoL was assessed using the Oral Impacts on Daily Performance (OIDP) questionnaire. The individual variables were demographic characteristics, socio-economic factors, and oral clinical measures. The Gini Index was used to evaluate city-level income inequality in 2003 (during childhood) and in 2010 (current) according to the tertiles of distribution. Multi-level Poisson regression was used to test the association of contextual income inequality and individual characteristics with OIDP extent according to the WHO framework on social determinants of health. Results: The prevalence of OIDP was 34.5%. In the adjusted analysis, adolescents living in the cities with high-level income inequality during childhood were 1.75 times more likely (95% confidence interval 1.23–2.48) to have a higher mean of OIDP extent. Current income inequality was not associated with OIDP extent in adolescents. Conclusions: Contextual income inequality during childhood was a structural determinant of OHRQoL among Brazilian adolescents after adjustment for individual demographic characteristics, socio-economic factors and oral clinical measures. Reducing social inequalities through inter-sectoral actions should be on the political agenda to improve adolescents’ oral health.
Key words: Social determinants of health, oral health, quality of life, socio-economic status, adolescent
INTRODUCTION
Oral epidemiology has recently acknowledged the importance of patient-reported outcome measures, reflecting a more comprehensive measurement of oral health, and a shift from the normative approach to a patient-centred perspective in dental research1., 2.. Oral-health-related quality of life (OHRQoL) indicators have been developed to assess people’s satisfaction and comfort in relation to their oral health whilst performing their daily life activities3. The association between demographic and socio-economic factors, dental clinical status, psychosocial factors and OHRQoL indicators has been investigated1., 3., 4., 5., 6., 7.. Adults from low socio-economic background and with poor dental status showed worse OHRQoL than those from high socio-economic status and with good oral health2., 7.. Similarly, poor OHRQoL in children and adolescents was associated with unfavourable socio-economic conditions, poor oral health and psychosocial factors4., 5., 6..
Evidence on the role of contextual social determinants in adolescents’ OHRQoL is scarce. Nonetheless, previous findings suggest the possible influence of contextual factors on adolescents’ clinical dental status8., 9., 10.. Adolescents from neighbourhoods with low levels of empowerment were more likely to have higher dental caries experience than those from neighbourhoods with high empowerment8. Income inequality at the municipal level and lower access to domestic sewage were associated with untreated dental caries in adolescents9., 10.. Dental pain was more common among adolescents living in deprived areas than those from better-off areas in the most populated Brazilian city11. In addition, levels of periodontal disease varied substantially between schools depending on their annual tuitions and fees12. More recently, a population-based study revealed that poor contextual school and home environmental characteristics were associated with poor OHRQoL in adolescents13.
In the 1990s, the seminal work of Wilkinson emphasised the impact of the unequal distribution of income and wealth at a population level on the population’s health, resulting in the so-called income inequality hypothesis14. The income inequality hypothesis argues that income distribution is of vital importance in the distribution of diseases and health outcomes. In addition, people’s overall health in more egalitarian societies, where the social or economic divide between the rich and the poor is smaller, is much better than in the societies where the wealth is concentrated in a small community of individuals14., 15..
The influence of income inequality on health might be mediated by different pathways. They may include underinvestment in social goods, such as in health services and public education; disruption of social cohesion and collective trust, and the erosion of social capital; and the direct psychosocial effects of social comparisons16., 17..
The income inequality hypothesis has been explored in relation to clinical oral-health outcomes18., 19. and OHRQoL2. The decrease in income inequality over a 10-year period was inversely associated with traumatic dental injuries among 12 year old children18. It was also reported that income inequality was a relevant determinant of tooth loss in adults19. Furthermore, contextual income inequality was associated with poor OHRQoL in middle-aged adults2. These results indicate that income inequality-health hypothesis seems relevant to oral-health research. However, little attention has been given to the possible role of income inequality on adolescent’s OHRQoL.
Recently, the World Health Organisation (WHO) proposed a Conceptual Framework for Action on the Social Determinants of Health20, which serves as a theoretical model to identify potential independent variables to evaluate the role of structural and intermediate social determinants of oral health2., 13.. According to this model, the structural and intermediary social determinants of health follow a hierarchical structure involving contextual factors and individual characteristics. The former is related to the socio-economic and political context, while the latter refers to measures of socio-economic stratification and demographic characteristics at the individual level20. The WHO framework on social determinants of health was used in this study (Figure 1). Current city-level income inequality and city-level income during childhood were assessed as second-level variables reflecting the contextual structural social determinants of quality of life. Individual socio-economic position and demographic factors were included as individual-level structural determinants. Clinical oral-health measures were the biological factors considered as intermediary determinants of OHRQoL.
Figure 1.
Theoretical model of the association of contextual income inequality and individual determinants with oral-health-related quality of life (OHRQoL) in adolescents adapted from WHO29.
The aim of this study was to evaluate the relationship of current city-level income inequality, city-level income inequality during childhood, and individual socio-demographic factors with OHRQoL in Brazilian adolescents. The first hypothesis of this study was that adolescents living in cities with higher levels of income inequality during their childhood have worse OHRQoL than those living in cities with lower levels of income inequality during their childhood. The second hypothesis was that adolescents currently living in cities with higher levels of income inequality have worse OHRQoL than those currently living in cities with lower levels of income inequality.
MATERIALS AND METHODS
Study design and sample
The Brazilian Oral Health Survey (SB Brasil Project) was a cross-sectional study conducted in 2010 to estimate the occurrence of oral-health problems in a representative sample of children, adolescents, adults and elderly people21. Cross-sectional information on demographics, socio-economic factors, oral clinical measures and OHRQoL were collected in the urban areas of Brazil.
Brazil consists of 27 federative units, including 26 Brazilian states and the Federal District. Each state has a capital city where the state administrative and political headquarters is located. The Federal District is located at the Central-West region of the country and contains the Brazilian capital city, Brasilia, which is the seat of the federal government. The participants of the SB Brasil Project were selected from 32 geographical domains of the country. They included 26 state capital cities, the Brazilian capital city located at the Federal District, and five domains representing the inner cities of the five geographical regions of the country (Central-West, North, Northeast, Southeast and South)22. A stratified multi-stage cluster sampling approach was employed to select participants with a proportional probability to the number of households in each geographical domain. The primary stage of the sampling process was 30 census tracts for each state capital and the Brazilian capital city in the Federal District, and 30 cities in the interior of each state. The second sampling units were households in the census tracts of each capital and the Brazilian capital city, and two census tracts in the inner cities. Each geographical region was composed of 30 tracts for each state capital city and the Brazilian capital city, and 60 for the sample of the inner cities. In the inner cities, the third stage of selection was applied to randomly select households within each of the sectors selected in the previous stage. Additional information about the sampling process is described elsewhere22. The inverse probability equation was used to estimate the sampling weights that were incorporated into the dataset afterwards for analytical purposes.
The sample of the oral-health survey was representative of the state capitals and the Brazilian capital city in the Federal District. Thus, in the present study, data from participants from the inner cities of the country were not included in the analysis. The studied sample included adolescents aged between 15 and 19 years from the state capitals and the Brazilian capital city at the Federal District in Brazil in which individual data were linked to city-level income inequality measures23. Participants from the interior municipalities, those younger than 15 and older than 19 years old, and those with missing data for any investigated variable were excluded.
Data collection and clinical calibration
Individual data were collected through structured interviews and clinical oral examinations in the participants’ households. A total of 270 fieldwork teams were involved where each team was composed of one dentist (examiner) and one healthcare worker (interviewer) from the Health Care System21. The fieldwork data collection was preceded by 40 hours of training for interviewers to conduct standardised interviews. Oral clinical examinations were conducted by calibrated dentists who achieved a minimum Kappa coefficient of 0.65 during the calibration study according to the consensus technique22., 24..
Oral-health-related quality of life
Oral-health-related quality of life was measured using the Oral Impacts on Daily Performance (OIDP) questionnaire validated for the Brazilian population25., 26.. The OIDP consists of nine items that evaluate the impact of oral status on physical, psychological, emotional and social performances in the previous 6 months25. The impact of oral status on quality of life considers the following daily life activities: ‘eating’, ‘speaking’, ‘cleaning’, ‘sleeping’, ‘smiling’, ‘emotional state’, ‘work or social role’, ‘contact with people’ and ‘sports’25., 26.. The participants reported whether each daily performance was affected by their oral conditions (yes = 1) or not (no = 0). The OIDP was analysed as a discrete outcome measure (OIDP extent) according to the number of performances affected by oral conditions, ranging from zero to nine. The Cronbach’s α for the OIDP questionnaire was 0.78 [95% confidence interval (CI) 0.77–0.79].
Contextual income inequality
City-level income inequality in the state capitals and the Brazilian capital city in the Federal District was assessed using the Gini Index in 2003 and 2010, representing contextual income inequality during childhood and current income inequality, respectively. Overall, the Gini Index reduced in Brazil from 0.5815 to 0.5401 between 2003 and 2010. However, the Gini Index varied considerably between the studied periods across the 27 cities included in this study. Gini Index measures were gathered from the United Nations Development Program23. The Gini Index varies between 0 and 1, and the higher the index the greater the income discrepancies between people in a society. The Gini Index was analysed as an ordinal variable using the tertiles of the distribution to analyse a potential gradient in the relationship between income inequality and OHRQoL. This approach has been applied in previous studies in oral-health research2., 8..
Individual demographic and socio-economic characteristics
Age, sex and ethnicity were the demographic characteristics. Participant’s self-perceived skin colour was used to asses ethnicity as proposed by the Brazilian Institute of Geography and Statistics using the following options: white, yellow, indigenous, brown and black27. Family income, schooling and number of goods in the house were the individual socio-economic characteristics. Family monthly income was provided by the head of the family. Income was originally registered in Brazilian Reais (R$) and converted into American dollars (US$) according to the following categories: <US$ 294, US$ 294–1,465 and >US$ 1,465. The Brazilian minimum wage was US$ 293 when the data were collected. Adolescent’s schooling was measured according to the number of years at school concluded without failure.
Oral clinical measures
Dental examinations were carried out according to WHO guidelines for oral-health surveys to collect measures of untreated decayed teeth, missing teeth and periodontal conditions24. The ‘decayed’ and ‘missing’ components of the DMFT index were used to assess the number of untreated decayed teeth and number of missing teeth. The Community Periodontal Index (CPI) was used to register the number of sextants with bleeding on probing (gingivitis), dental calculus and periodontal pockets using a ball-end CPI probe24.
Statistical analysis
Demographic and socio-economic characteristics were presented using proportions with their respective 95% CIs for the studied sample and according to OIDP prevalence. The mean of OIDP and 95% CIs were also reported according to the individual independent variables.
Spearman’s coefficients were used to test the correlations between OIDP extent, number of decayed teeth, number of missing teeth, number of sextants with bleeding on probing, number of sextants with dental calculus, and number of sextants with dental pockets.
Multi-level Poisson regression using fixed-effect models with a random intercept was used to estimate rate ratios (RRs) and 95% CIs between contextual and individual variables and OHRQoL. All independent variables were included in the multi-variable analysis. The multi-level structure for the analysis considered two levels, namely 3,584 adolescents (first-level) nested within 27 cities (second-level), according to the WHO framework (Figure 1). The first model was composed of Gini Index in 2003 and Gini Index in 2010 as contextual variables. The association of Gini Index in 2003 and Gini Index in 2010 with OHRQoL was adjusted for individual demographic and socio-economic confounding variables in the second model. In the final model, oral clinical measures (mediators) were included for further adjustment. The OIDP extent variance and standard error across cities (random effects) were used to assess the variation of the OIDP at the contextual level. The statistical significance of the variance of OIDP at the city-level was tested using Wald statistic through the ratio of the variance and the standard error following a Chi-square distribution.
IBM SPSS Statistics version 24.0 (SPSS, Chicago, IL, USA) was used for the descriptive analysis taking into account the complex samples and sampling weights. Multi-level Poisson regression was conducted using Stata version 14.0 (Stata, College Station, TX, USA).
Ethical considerations
The SB Brasil Project study was approved by the Brazilian National Council of Ethics in Research, protocol no. 15498, 7 January 2010. The present research was conducted in accordance with the World Medical Association Declaration of Helsinki ethical standards. All participants aged 18 years and over provided written informed consent. Participants’ parents or guardians have given written informed consent for those under 18 years of age.
RESULTS
Initially, 5,887 adolescents were invited to participate in the SB Brasil Project. Of these, 5,445 participants were interviewed and examined (response rate = 92.5%). Adolescents from the interior municipalities were excluded (N = 1,238), resulting in 4,207 participants from the state capital cities and the Brazilian capital city in the Federal District. A further 353 participants were also excluded due to missing data on socio-economic variables (N = 296) and oral clinical measures (N = 57). The final sample was composed of 3,854 adolescents.
The mean of oral impacts affecting daily life activities (OIDP) in the last 6 months was 0.92 (95% CI 0.74–1.09), and the prevalence of oral impacts (OIDP ≥ 1) was 34.5% (95% CI 29.6–39.6). ‘Eating’, ‘cleaning teeth’ and ‘smiling’ were the most common performances affected by oral-health status.
Table 1 presents the demographic and socio-economic characteristics of the sample according to OIDP measures. The majority of the participants were 15 years old, females, had brown skin colour, monthly family income between US$ 294 and US$ 1465, and at least 9 years of schooling. The direction and strength of the correlations between OIDP extent and oral clinical measures are reported in Table 2. Significant correlations were identified between OIDP extent and worse oral clinical measures.
Table 1.
Demographic and socio-economic characteristics of the study sample (N = 3,854)
| Variables | Study sample % (95% CI) | OIDP |
||
|---|---|---|---|---|
| OIDP = 0 % (95% CI) | OIDP ≥ 1% (95% CI) | Mean (95% CI) | ||
| Age (years) | ||||
| 15 | 24.6 (21.5–28.0) | 24.7 (20.3–29.7) | 24.3 (21.0–28.0) | 0.6 (0.7–1.0) |
| 16 | 18.2 (16.1–20.6) | 19.9 (17.2–23.0) | 15.1 (12.0–18.8) | 0.6 (0.4–0.8) |
| 17 | 18.2 (16.1–20.6) | 17.8 (15.2–20.7) | 19.0 (15.8–22.6) | 0.9 (0.7–1.2) |
| 18 | 18.1 (14.9–21.8) | 18.2 (13.9–23.5) | 18.0 (14.8–21.6) | 1.0 (0.6–1.3) |
| 19 | 20.9 (18.7–23.2) | 19.4 (17.0–22.1) | 23.6 (19.7–28.0) | 1.2 (0.9–1.6) |
| Sex | ||||
| Male | 45.8 (43.3–48.3) | 49.1 (46.4–51.9) | 39.6 (35.2–44.2) | 0.7 (0.5–0.8) |
| Female | 54.2 (51.7–56.7) | 50.9 (48.1–536.) | 60.4 (55.8–64.8) | 1.1 (0.9–1.4) |
| Skin colour | ||||
| White | 39.4 (34.9–44.0) | 39.4 (33.4–45.8) | 39.3 (33.9–45.0) | 0.9 (0.7–1.1) |
| Yellow | 0.6 (0.4–1.0) | 0.5 (0.3–0.9) | 0.9 (0.5–1.6) | 1.1 (0.6–1.7) |
| Indigenous | 0.6 (0.2–1.5) | 0.4 (0.2–0.7) | 1.1 (0.3–4.3) | 3.2 (0.2–6.2) |
| Brown | 43.7 (40.1–47.4) | 42.8 (38.7–46.9) | 45.5 (40.0–51.1) | 1.0 (0.8–1.2) |
| Black | 15.7 (11.6–20.8) | 16.9 (11.4–24.4) | 13.3 (10.1–17.2) | 0.8 (0.4–1.2) |
| Monthly family income (US$) | ||||
| < 294 | 14.9 (12.2–18.1) | 13.8 (10.5–17.9) | 17.1 (14.0–20.7) | 1.1 (0.8–1.4) |
| 294– 1,465 | 72.0 (67.4–76.1) | 71.8 (65.5–77.4) | 72.3 (67.9–76.2) | 0.9 (0.7–1.1) |
| > 1,465 | 13.1 (10.5–16.3) | 14.4 (10.9–18.8) | 10.7 (8.1–13.9) | 0.6 (0.4–0.8) |
| Years of schooling | ||||
| < 9 | 41.3 (36.4–46.4) | 36.5 (30.7–42.9) | 50.1 (45.2–55.0) | 1.3 (1.0–1.5) |
| ≥ 9 | 58.7 (53.6–63.6) | 63.5 (57.1–69.3) | 49.9 (45.0–54.8) | 0.7 (0.5–0.8) |
| Number of durable goods | ||||
| 0–7 | 56.5 (48.7–63.9) | 52.9 (43.0–62.6) | 63.1 (56.2–69.4) | 1.1 (0.9–1.3) |
| 8–11 | 43.5 (36.1–51.3) | 47.1 (37.4–57.0) | 36.9 (30.6–43.8) | 0.7 (0.5–1.0) |
CI, confidence interval; OIDP, Oral Impacts on Daily Performance.
Table 2.
Correlation matrix (Spearman coefficient) between OIDP extent and oral clinical measures
| Variables | OIDP | Decayed teeth | Missing teeth | Gingivitis | Dental calculus | Pocket depths |
|---|---|---|---|---|---|---|
| OIDP | 1 | |||||
| Decayed teeth | 0.239** | 1 | ||||
| Missing teeth | 0.117** | 0.255** | 1 | |||
| Gingivitis | 0.129** | 0.229** | 0.070** | 1 | ||
| Dental calculus | 0.181** | 0.240** | 0.091** | 0.544** | 1 | |
| Pocket depths | 0.115** | 0.139** | 0.107** | 0.337** | 0.302** | 1 |
P < 0.01.
OIDP, Oral Impacts on Daily Performance.
The distribution of demographic and socio-economic characteristics according to OIDP is summarised in Table 3. Overall, the frequency of all oral impacts was higher among those who were 19 years old, with indigenous skin colour, and among adolescents of poor socio-economic status.
Table 3.
Distribution of OIDP according to demographic and socio-economic characteristics (N = 3,854)
| Eating % (95% CI) | Speaking % (95% CI) | Cleaning % (95% CI) | Sleeping % (95% CI) | Smiling % (95% CI) | |
|---|---|---|---|---|---|
| Age (years) | |||||
| 15 | 22.9 (18.8–27.7) | 22.4 (15.4–31.5) | 23.0 (18.2–28.7) | 22.4 (16.9–29.1) | 24.9 (19.0–31.9) |
| 16 | 14.3 (10.9–18.7) | 9.9 (4.8–19.4) | 13.1 (9.1–18.6) | 9.8 (6.0–15.6) | 10.9 (7.1–16.4) |
| 17 | 21.3 (17.3–26.0) | 14.9 (8.3–25.3) | 17.8 (13.2–23.5) | 17.3 (12.3–23.8) | 18.2 (12.4–25.9) |
| 18 | 16.4 (12.4–21.4) | 17.2 (10.5–26.9) | 17.8 (13.3–23.5) | 22.4 (16.3–30.0) | 20.2 (15.0–26.8) |
| 19 | 25.0 (19.5–31.5) | 35.5 (25.8–46.6) | 28.2 (21.9–35.5) | 28.0 (21.9–35.1) | 25.7 (19.7–32.9) |
| Skin colour | |||||
| White | 18.3 (14.1–23.3) | 7.5 (4.9–11.4) | 14.5 (10.9–18.9) | 10.4 (7.9–13.5) | 13.4 (10.2–17.2) |
| Yellow | 23.7 (13.4–38.5) | 4.3 (1.5–12.0) | 20.3 (8.1–42.2) | 13.5 (4.9–32.1) | 18.4 (8.5–35.4) |
| Indigenous | 49.0 (13.2–85.8) | 26.5 (10.6–52.3) | 24.9 (8.9–52.9) | 46.4 (11.2–85.6) | 30.6 (15.1–52.3) |
| Brown | 21.3 (16.8–26.6) | 6.4 (4.4–9.2) | 14.8 (11.6–18.6) | 13.0 (10.0–16.6) | 12.1 (9.4–15.5) |
| Black | 15.6 (9.6–24.3) | 5.0 (2.4–9.8) | 10.7 (6.2–17.8) | 8.8 (5.0–14.9) | 12.5 (7.1–21.1) |
| Monthly family income (US$) | |||||
| < 294 | 20.3 (15.1–26.8) | 9.4 (5.5–15.5) | 16.3 (11.5–22.6) | 12.0 (9.3–14.1) | 16.2 (11.6–22.1) |
| 294– 1,465 | 20.1 (16.1–24.8) | 6.6 (4.8–9.0) | 14.2 (11.2–17.9) | 12.1 (9.4–15.3) | 13.0 (10.2–16.5) |
| > 1,465 | 14.6 (10.2–20.4) | 4.0 (1.8–8.9) | 11.0 (6.9–17.1) | 8.0 (4.7–13.3) | 7.9 (4.6–13.1) |
| Years of schooling | |||||
| < 9 | 25.4 (21.3–29.9) | 10.0 (7.1–13.8) | 17.4 (13.7–21.7) | 17.0 (13.3–21.6) | 18.0 (14.4–22.2) |
| ≥ 9 | 15.2 (11.8–19.4) | 4.4 (3.1–6.1) | 11.8 (9.0–15.5) | 7.6 (5.6–10.3) | 9.2 (7.1–11.9) |
| Number of durable goods | |||||
| 0–7 | 21.7 (18.3–25.6) | 7.8 (5.6–10.7) | 16.0 (13.2–19.3) | 11.5 (9.3–14.2) | 15.1 (12.3–18.5) |
| 8–11 | 16.4 (11.4–22.9) | 5.3 (3.3–8.5) | 11.6 (7.8–17.0) | 11.5 (7.9–16.3) | 9.8 (6.9–13.9) |
| Emotional status % (95% CI) | Study % (95% CI) | Social contact % (95% CI) | Sports % (95% CI) | |
|---|---|---|---|---|
| Age (years) | ||||
| 15 | 24.3 (17.8–32.3) | 23.6 (17.9–37.0) | 17.3 (11.0–26.2) | 18.6 (10.3–31.1) |
| 16 | 12.6 (8.6–18.1) | 8.0 (4.9–12.9) | 10.1 (5.4–18.3) | 6.9 (3.7–12.6) |
| 17 | 17.2 (12.2–23.6) | 14.5 (8.1–24.7) | 19.1 (11.8–29.5) | 29.7 (16.3–47.8) |
| 18 | 20.9 (15.1–28.2) | 20.3 (11.6–33.1) | 21.9 (14.7–31.3) | 21.3 (12.1–34.6) |
| 19 | 25.0 (18.8–32.3) | 30.8 (19.5–44.9) | 31.6 (21.4–43.9) | 23.6 (12.9–39.3) |
| Skin colour | ||||
| White | 11.1 (7.7–15.6) | 3.8 (2.3–6.3) | 5.3 (3.5–8.1) | 2.0 (1.1–3.5) |
| Yellow | 12.3 (4.1–21.3) | 4.1 (0.9–16.5) | 7.6 (1.7–28.6) | 6.7 (1.2–30.0) |
| Indigenous | 27.2 (11.3–52.3) | 23.4 (7.5–53.7) | 45.4 (10.1–86.0) | 44.7 (9.9–85.5) |
| Brown | 12.8 (9.6–16.9) | 5.7 (3.8–8.4) | 8.1 (5.4–11.8) | 3.7 (2.3–6.0) |
| Black | 9.9 (5.8–16.3) | 4.8 (2.3–9.4) | 7.1 (3.5–14.0) | 4.3 (1.6–11.3) |
| Monthly family income (US$) | ||||
| <294 | 16.9 (11.7–23.7) | 8.0 (4.9–13.0) | 8.5 (5.6–12.6) | 4.5 (2.7–7.4) |
| 294–1,465 | 11.8 (8.9–15.6) | 4.8 (3.5–6.6) | 7.2 (5.1–10.2) | 3.2 (2.0–5.1) |
| >1,465 | 5.3 (2.8–10.1) | 1.6 (0.8–3.6) | 4.5 (1.9–10.0) | 3.1 (1.1–8.5) |
| Years of schooling | ||||
| <9 | 15.6 (11.8–20.3) | 7.9 (5.8–10.8) | 10.6 (7.7–14.6) | 6.0 (3.9–9.1) |
| ≥9 | 9.0 (6.7–12.0) | 2.8 (1.9–4.1) | 4.5 (3.0–6.8) | 1.6 (1.0–2.5) |
| Number of durable goods | ||||
| 0–7 | 14.4 (10.9–18.6) | 5.9 (4.3–8.1) | 8.9 (6.6–12.0) | 4.3 (2.9–6.4) |
| 8–11 | 8.4 (5.6–12.2) | 3.6 (2.2–5.7) | 4.6 (2.7–7.7) | 2.2 (1.1–4.6) |
CI, confidence interval.
Table 4 presents the unadjusted associations of contextual and individual variables with OIDP extent. A high Gini Index during childhood (Gini Index in 2003) was associated with greater OIDP extent (RR = 1.66, 95% CI 1.13–2.45). The relationship between current Gini Index (2010) and OIDP extent was not statistically significant. Adolescents aged between 18 and 19 years, females and those with non-white skin colour were more likely to have higher OIDP extent. The mean of OIDP extent was significantly greater in adolescents from low-income families, with low schooling and low number of durable goods. Poor oral clinical measures were associated with OIDP extent.
Table 4.
Unadjusted association of contextual income inequality and individual variables with OIDP extent, determined using multi-level Poisson regression (N = 3,854)
| Variables | Variance† | β | SE | RR | 95% CI | P |
|---|---|---|---|---|---|---|
| Contextual variables | ||||||
| Gini Index (2003) | 0.378 (0.056)* | |||||
| Moderate | 0.197 | 0.175 | 1.22 | 0.86–1.72 | 0.262 | |
| High | 0.507 | 0.198 | 1.66 | 1.13–2.45 | 0.011 | |
| Gini Index (2010) | 0.423 (0.061)* | |||||
| Moderate | 0.077 | 0.205 | 1.08 | 0.72–1.62 | 0.706 | |
| High | 0.099 | 0.205 | 1.11 | 0.74–1.65 | 0.628 | |
| Individual variables | ||||||
| Age | ||||||
| 15 | 1 | |||||
| 16 | 0.081 | 0.054 | 1.08 | 0.98–1.21 | 0.136 | |
| 17 | 0.089 | 0.054 | 1.09 | 0.98–1.22 | 0.096 | |
| 18 | 0.276 | 0.052 | 1.32 | 1.19–1.46 | <0.001 | |
| 19 | 0.312 | 0.051 | 1.37 | 1.24–1.51 | <0.001 | |
| Sex | ||||||
| Male | 1 | |||||
| Female | 0.387 | 0.036 | 1.47 | 1.37–1.59 | <0.001 | |
| Skin colour | ||||||
| White | ||||||
| Yellow | 0.401 | 0.122 | 1.49 | 1.18–1.90 | 0.001 | |
| Indigenous | 0.433 | 0.150 | 1.54 | 1.15–2.07 | 0.004 | |
| Brown | 0.315 | 0.041 | 1.37 | 1.26–1.49 | <0.001 | |
| Black | 0.364 | 0.059 | 1.44 | 1.28–1.61 | <0.001 | |
| Monthly family income (US$) | ||||||
| >1,465 | ||||||
| 294–1,465 | 0.538 | 0.064 | 1.71 | 1.51–1.94 | <0.001 | |
| <294 | 0.793 | 0.075 | 2.21 | 1.91–2.56 | <0.001 | |
| Education | −0.104 | 0.007 | 0.90 | 0.89–0.91 | <0.001 | |
| Number of durable goods | −0.086 | 0.008 | 0.92 | 0.90–0.93 | <0.001 | |
| Untreated decayed teeth | 0.126 | 0.005 | 1.13 | 1.12–1.14 | <0.001 | |
| Missing teeth | 0.116 | 0.007 | 1.12 | 1.11–1.14 | <0.001 | |
| Gingivitis | 0.126 | 0.009 | 1.13 | 1.12–1.15 | <0.001 | |
| Dental calculus | 0.115 | 0.009 | 1.12 | 1.10–1.14 | <0.001 | |
| Pocket depths | 0.145 | 0.015 | 1.16 | 1.12–1.19 | <0.001 | |
CI, confidence interval; RR, rate ratio.
P ≤ 0.05.
Variance at the city-level (standard error) was obtained through random effects.
The multi-level Poisson regression models for OIDP are reported in Table 5. A high Gini Index during childhood (Gini Index in 2003) was associated with OIDP extent in Model 1 and remained associated afterwards. The association of Gini Index in 2003 and Gini Index in 2010 with OIDP extent was adjusted for demographic and socio-economic characteristics, and oral clinical measures in Models 2 and 3, respectively. In Model 3 (final model), adolescents living in the cities with greater levels of income inequality during childhood (RR = 1.75, 95% CI 1.23–2.48) had a higher mean of OIDP extent. The association between high-income inequality during childhood and OIDP extent was amplified by 3% after adjustment for socio-demographic characteristics and oral clinical measures.
Table 5.
Adjusted association between contextual income inequality OIDP extent, determined using multi-level Poisson regression (N = 3,854)
| Variables | Model 1 |
Model 2 |
Model 3 |
|---|---|---|---|
| RR (95% CI) | RR (95% CI) | RR (95% CI) | |
| Contextual variables | |||
| Gini Index (2003) | |||
| Moderate | 1.17 (0.80–1.70) | 1.13 (0.80–1.61) | 1.18 (0.85–1.64) |
| High | 1.70 (1.14–2.53)** | 1.78 (1.23–2.58)** | 1.75 (1.23–2.48)** |
| Gini Index (2010) | |||
| Moderate | 1.14 (0.76–1.71) | 1.27 (0.87–1.85) | 1.21 (0.84–1.65) |
| High | 1.16 (0.79–1.70) | 1.18 (0.83–1.69) | 1.18 (0.85–1.72) |
| Variance at city-level [Ωμ (standard error)]† | 0.374 (0.055)** | 0.349 (0.052)** | 0.325 (0.049)** |
Model 1: Gini Index 2003 and Gini Index 2010.
Model 2: mutually adjusted for Gini Index 2003, Gini Index 2010 and socio-demographic variables.
Model 3: mutually adjusted for Gini Index 2003, Gini Index 2010, socio-demographic and clinical variables.
P ≤ 0.05; **P ≤ 0.01.
Variance at the city-level (standard error) was obtained through random effects.
CI, confidence interval; RR, rate ratio.
DISCUSSION
The hypothesis that adolescents living in cities with higher levels of income inequality during their childhood experience poor OHRQoL compared with those from cities with lower levels of income inequality was confirmed. On the other hand, the hypothesis that adolescents currently living in cities with higher levels of income inequality have worse OHRQoL than those living in cities with lower levels of income inequality was not supported by our findings. The differences in the relationship between contextual income inequality measures during childhood and currently and OHRQoL indicate that the time period interval between contextual income inequality and OHRQoL appear to be meaningful regarding adolescents’ oral health. Individual socio-economic characteristics and poor oral clinical measures were also important factors associated with OHRQoL. Our results suggest that contextual and individual social determinants have independent effects on adolescents’ OHRQoL.
Oral-health-related quality of life was the outcome of interest in a few studies conducted in different age groups2., 13.. Poor OHRQoL was associated with contextual social deprivation and income inequality among Brazilian adults2. In addition, poor contextual school and home environmental characteristics were associated with poor OHRQoL in adolescents13. However, as far as the authors are aware, this study is the first study to evaluate the association of contextual income inequality during childhood and currently with OHRQoL in adolescents.
In this study, the relationship between contextual income inequality and OHRQoL was evaluated using the Gini Index during childhood (2003) and currently (2010). The association between income inequality and OHRQoL was significant only in the former period. This suggests that a time span of 7 years between income inequality and OHRQoL in adolescents is necessary to detect a relevant effect. Another possible reason for these findings might be the reduction of Gini Index between 2003 and 2010 in Brazil, and income inequality in the latter period was no longer relevant for OHRQoL.
Previous studies demonstrated the relationship between socio-economic status and OHRQoL in adolescents4., 5., 6.. Although the findings provided relevant information to improve the understanding on the role of low income on poor OHRQoL, these studies adopted the absolute income hypothesis, access to health services and better housing conditions15., 28.. Another explanation is the harmful influence of social inequality on health. Unequal societies have lower social cohesion and lower levels of trust, higher unemployment and violence. However, it has been argued that only increasing one’s material resources seems insufficient to affect health through direct effect20. Possible pathways by which poor socio-economic position predisposes health outcomes are complex and should consider the social environment where people are embedded. In contrast, the income inequality hypothesis suggests the harmful influence of social hierarchies on health resulting in the ‘invidious processes of social comparison’14., 15., 20.. The association between contextual income inequality and OHRQoL can be understood through the following explanations. Cities with low-income inequality may have more sustainable public policies in place that can promote a healthy environment and therefore can potentially result in a positive impact on the health of the population compared with cities with high-income inequalities29. People living in cities with lower levels of income inequality are possibly more likely to have more access to healthy food, lower social capital and poor health than more egalitarian ones30. The abovementioned explanations are highly dependent on the country’s environmental and socio-political characteristics, and understanding such pathways was out of the scope of the present study. Social inequalities may also influence health through behavioural and psychosocial pathways. Social participation and engagement, social influence, access to resources and material goods are related to health-related behaviours31. Social isolation and low social support may impact on oral health via the physiological harmful effects of psychosocial stress and cortisol levels leading to immunosuppression that increases the susceptibility to oral diseases32, and in some way to poor OHRQoL. It should also be emphasised that the social characteristics of the country must be cautiously considered when the behavioural and psychosocial pathways are raised as possible pathways to explain the influence of income inequality on explain OHRQoL in adolescents.
This study used a large and robust sample of adolescents. The sample was representative for 27 large cities of the country, and included adolescents from different socio-economic groups and with important variations in oral clinical measures. These aspects are paramount to explore the contextual and individual factors of oral health using multi-level analysis. The large sample size also offered sufficient power to test the pre-established hypothesis and to obtain estimates on the association between independent variables and OHRQoL with good precision. In addition, the use of the WHO theoretical model was a positive aspect in order to select appropriate variables to address the aim of the study. Finally, multi-level modelling is the appropriate statistical approach to assess the association between variables from different levels (e.g. individual- and contextual-level variables) and health outcomes.
Some limitations must be acknowledged. Although the use of secondary data is legitimate in social epidemiology, the variables available in the SB Brasil project were not collected to meet the aims of this study. Thus, other relevant characteristics for OHRQoL in adolescents in accordance with the WHO’s framework for the social determinants of health, such as psychosocial factors (e.g. social support and coping styles) and behaviours, were not investigated29. The use of cross-sectional data limits the establishment of a temporal relationship between variables and causal inferences. Children may have moved to another city between the period when income inequality was assessed (2003 and 2010). It is also important to note that the Gini Index is not very sensitive to measure inequality at the top and the bottom of the income spectrum in developing countries.
CONCLUSIONS
The current study provides evidence to support the role of contextual social determinants on OHRQoL in adolescents. This study suggests that in a representative sample of Brazilian adolescents aged 15–19 years, contextual income inequality during childhood was associated with poor OHRQoL. Actions aiming to diminish social inequalities by reducing the income gap would benefit the oral health of the population. Improving social conditions through inter-sectoral actions should be on the political agenda to enhance adolescents’ oral health.
Acknowledgement
There is no acknowledgement related to this study.
Funding statement
This study was not funded.
Competing interest
The authors declare that they have no conflict of interest.
References
- 1.Turrell G, Sanders AE, Slade G, et al. The independent contribution of neighborhood disadvantage and individual-level socioeconomic position to self-reported oral health: a multilevel analysis. Community Dent Oral Epidemiol. 2007;35:195–206. doi: 10.1111/j.1600-0528.2006.00311.x. [DOI] [PubMed] [Google Scholar]
- 2.Vettore M, Aqeeli A. The roles of contextual and individual social determinants of oral health-related quality of life in Brazilian adults. Qual Life Res. 2016;25:1029–1042. doi: 10.1007/s11136-015-1118-0. [DOI] [PubMed] [Google Scholar]
- 3.Locker D, Allen F. What do measures of ‘oral health-related quality of life’ measure? Community Dent Oral Epidemiol. 2007;35:401–411. doi: 10.1111/j.1600-0528.2007.00418.x. [DOI] [PubMed] [Google Scholar]
- 4.Baker SR, Mat A, Robinson PG. What psychosocial factors influence adolescents’ oral health? J Dent Res. 2010;89:1230–1235. doi: 10.1177/0022034510376650. [DOI] [PubMed] [Google Scholar]
- 5.Foster Page LA, Thomson WM, Ukra A, et al. Factors influencing adolescents’ oral health-related quality of life (OHRQoL) Int J Paediatr Dent. 2013;23:415–423. doi: 10.1111/ipd.12011. [DOI] [PubMed] [Google Scholar]
- 6.Peres KG, Cascaes AM, Leão ATT, et al. Sociodemographic and clinical aspects of quality of life related to oral health in adolescents. Rev Saúde Pública. 2013;47(Suppl 3):19–28. doi: 10.1590/s0034-8910.2013047004361. [DOI] [PubMed] [Google Scholar]
- 7.Huang DL, Park M. Socioeconomic and racial/ethnic oral health disparities among US older adults: oral health quality of life and dentition. J Public Health Dent. 2015;75:85–92. doi: 10.1111/jphd.12072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Pattussi MP, Hardy R, Sheiham A. The potential impact of neighborhood empowerment on dental caries among adolescents. Community Dent Oral Epidemiol. 2006;34:344–350. doi: 10.1111/j.1600-0528.2006.00283.x. [DOI] [PubMed] [Google Scholar]
- 9.Celeste RK, Nadanovsky P, Ponce de Leon A, et al. The individual and contextual pathways between oral health and income inequality in Brazilian adolescents and adults. Soc Sci Med. 2009;69:1468–1475. doi: 10.1016/j.socscimed.2009.08.005. [DOI] [PubMed] [Google Scholar]
- 10.Vazquez Fde L, Cortellazzi KL, Kaieda AK, et al. Individual and contextual factors related to dental caries in underprivileged Brazilian adolescents. BMC Oral Health. 2015;15:6. doi: 10.1186/1472-6831-15-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Peres MA, Peres KG, Frias AC, et al. Contextual and individual assessment of dental pain period prevalence in adolescents: a multilevel approach. BMC Oral Health. 2010;10:20. doi: 10.1186/1472-6831-10-20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lopez R, Frydenberg M, Baelum V. Contextual effects in the occurrence of periodontal attachment loss and necrotizing gingival lesions among adolescents. Eur J Oral Sci. 2009;117:547–554. doi: 10.1111/j.1600-0722.2009.00658.x. [DOI] [PubMed] [Google Scholar]
- 13.Alwadi MAM, Vettore MV. Are school and home environmental characteristics associated with oral health-related quality of life in Brazilian adolescents and young adults? Community Dent Oral Epidemiol. 2017;45:356–364. doi: 10.1111/cdoe.12298. [DOI] [PubMed] [Google Scholar]
- 14.Wilkinson RG. Income distribution and life expectancy. BMJ. 1992;304:165. doi: 10.1136/bmj.304.6820.165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wilkinson RG, Pickett KE. Income inequality and population health: a review and explanation of the evidence. Soc Sci Med. 2006;62:1768–1784. doi: 10.1016/j.socscimed.2005.08.036. [DOI] [PubMed] [Google Scholar]
- 16.Kaplan G, Pamuk E, Lynch J, et al. Inequality in income and mortality in the United States: analysis of mortality and potential pathways. BMJ. 1996;312:999–1003. doi: 10.1136/bmj.312.7037.999. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kawachi I, Kennedy BP. Income inequality and health: pathways and mechanisms. Health Serv Res. 1999;34:215–227. [PMC free article] [PubMed] [Google Scholar]
- 18.Vettore MV, Efhima S, Machuca C, et al. Income inequality and traumatic dental injuries in 12-year-old children: a multilevel analysis. Dent Traumatol. 2017;33:375–382. doi: 10.1111/edt.12350. [DOI] [PubMed] [Google Scholar]
- 19.Goulart MDA, Vettore MV. Is the relative increase in income inequality related to tooth loss in middle-aged adults? J Public Health Dent. 2016;76:65–75. doi: 10.1111/jphd.12113. [DOI] [PubMed] [Google Scholar]
- 20.World Health Organization . World Health Organization; Geneva: 2010. A Conceptual Framework for Action on the Social Determinants of Health. [Google Scholar]
- 21.Roncalli AG, Da Silva NN, Nascimento A, et al. Relevant methodological issues from the SBBrasil 2010 Project for national health surveys. Cad Saúde Pública. 2012;28:S40–S57. doi: 10.1590/s0102-311x2012001300006. (in Portuguese) [DOI] [PubMed] [Google Scholar]
- 22.Silva N, Roncalli AG. Sampling plan, weighting process and design effects of the Brazilian Oral Health Survey. Rev Saúde Pública. 2013;47:3–11. doi: 10.1590/s0034-8910.2013047004362. [DOI] [PubMed] [Google Scholar]
- 23.PNUD. PNUD Brasil – Programa das Nações Unidas para o Desenvolvimento 2003. Available from: http://www.pnud.org.br/
- 24.World Health Organization; Geneva: 1997. Oral Health Surveys. Basic Methods. [Google Scholar]
- 25.Adulyanon S, Vourapukjaru J, Sheiham A. Oral impacts affecting daily performance in a low dental disease Thai population. Community Dent Oral Epidemiol. 1996;24:385–389. doi: 10.1111/j.1600-0528.1996.tb00884.x. [DOI] [PubMed] [Google Scholar]
- 26.Castro RAL, Cortes MIS, Leão AT, et al. Child-OIDP index in Brazil: cross-cultural adaptation and validation. Health Qual Life Outcomes. 2008;6:68. doi: 10.1186/1477-7525-6-68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Travassos C, Williams D. The concept and measurement of race and their relationship to public health: a review focused on Brazil and the United States. Cad Saúde Pública. 2004;20:660–678. doi: 10.1590/s0102-311x2004000300003. [DOI] [PubMed] [Google Scholar]
- 28.Wilkinson RG. Routledge, Taylor & Francis; London: 2005. The Impact of Inequality: How to Make Sick Societies Healthier. [Google Scholar]
- 29.Wilkinson RG. Comment: income, inequality, and social cohesion. Am J Public Health. 1997;87:1504–1506. doi: 10.2105/ajph.87.9.1504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Berkman LS, Glass T. In: Social Determinants of Health. Marmot M, Wilkinson RG, editors. Oxford University Press; London: 2006. Social integration, social networks, social support, and health; pp. 137–173. [Google Scholar]
- 31.Lawrence HP, Thomson WM, Broadbent JM, et al. Oral health-related quality of life in a birth cohort of 32-year olds. Community Dent Oral Epidemiol. 2008;36:305. doi: 10.1111/j.1600-0528.2007.00395.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Warren KR, Postolache TT, Groer ME, et al. Role of chronic stress and depression in periodontal diseases. Periodontol 2000. 2014;64:127–138. doi: 10.1111/prd.12036. [DOI] [PMC free article] [PubMed] [Google Scholar]