Geographic information system and multilevel analysis: gingival status among 12-year-old schoolchildren in Sao Paulo, Brazil/Sistema de informacion geografica y analisis de niveles multioles: estado gingiva en escolares de 12 anos de edad en Sao Paulo, Brasil.
Geographic information systems (GIS) are used to aggregate and analyze variables necessary for ecological studies. This type of system recognizes the relationship between the distribution and impact of diseases in the spaces in which they occur. It also helps clarify phenomena that cannot be adequately visualized using data displayed in tables, thus reaffirming the relationship between epidemiology and cartography (4).
Isolated efforts have characterized the recognized centuries-long relationship between geography and human health. In the last 10 years, international interest in this relationship has been renewed in academic and health care management fields. Such interest stresses the positive, progressive impact of the use of spatial analysis in public health care, as well as the pertinence of a prudent, careful way of evaluating results (5).
The aim of the present study was to evaluate gingival status in 12-year-old schoolchildren in the city of Piracicaba using a GIS and multilevel analysis.
MATERIALS AND METHODS
This study received approval from the Research Ethics Committee of the Piracicaba Dental School, State University of Campinas (UNICAMP), Piracicaba, Sao Paulo, Brazil) (protocols #148/2003; #098/2006). Children were given an informed consent form to be taken home for their parent's signature; only those with signed forms were permitted to participate in the study.
The city of Piracicaba (Sao Paulo state, Brazil) has a population of 385 287 inhabitants, and a city area of 1 3 78 501 [km.sup.2]. The Human Development Index (HDI) is 0.785 and monthly income per capita for the urban area population is around R$ 755 (approximately US$ 324) (6).
Prior to the exams, a questionnaire addressing socioeconomic status and the child's behavioral habits was sent to the parents. The questionnaire consisted of items on monthly household income, number of people living in the household, home ownership, mother's education, father's education, car ownership, toothbrushing habits, and access to oral health care (7).
The sample size was calculated based on a prior study (8), considering a statistical power of 0.80, significance level of 0.05, and odds ratio of 1.5. Cluster sampling was employed to select 1 100 student 12 years of age from public and private schools in 18 districts of the municipality of Piracicaba, Sao Paulo, Brazil, based on geographic location. The proportion between public and private schools in the municipality was taken into account. Twenty-five public and private schools took part in the present study and the sample was made up of 1 002 students (loss < 10%).
The city has an Oral Health Program called "Sorria Piracicaba" ("Smile Piracicaba") for children of 0-10 years of age registered in public schools. Four times a year the children brush their teeth under supervision of dental hygienists or dentists from Brazil's Family Health Program, and receive oral hygiene kits with toothpaste, toothbrush, and floss. Furthermore, an educational program in oral health, recreational, theatrical activities, and presentations are provided. During the annual meetings, school teachers are given oral health training and education. The schoolchildren from the present study had received preventive care and oral health education in the years leading up to the study.
Piracicaba has two water treatment stations. Since 1971, the community water supply is fluoridated and the control of the optimum levels (0.6-0.8 ppm F, considering the highest temperature of 37.5[degrees]C and the lowest 18[degrees]C) are done by continuously monitoring the levels and making necessary adjustments when needed.
The exams were carried out by a single examiner who followed the criteria recommended by the World Health Organization (WHO) (9). The exams were conducted in an outdoor setting under natural light immediately after tooth brushing and flossing to remove dental biofilm, using a Community Periodontal Index (CPI) probe with a diameter of 0.5 mm (ballpoint). The students received a tooth brushing kit (a regular toothbrush, toothpaste, and floss) and performed the tooth brushing and flossing under the supervision of a dental hygienist.
The examiner was previously trained (calibrated) by a "gold standard" expert epidemiologist. The examiner was calibrated at the beginning of the 12-month study regarding gingival evaluation and dental caries diagnosis (10). The five calibration exercises (baseline 3, 6, 9, and 12 months) were performed during the exam of the schoolchildren, as well as the calculation of the agreement rate and Kappa statistics.
In the present study, approximately 10% of the sample was reexamined in order to determine intra-examiner reproducibility; the mean Kappa value for the periodontal exam was >0.86.
The gingival status of the students was evaluated using the Community Periodontal Index (CPI) to detect gingival bleeding and dental calculus. Based on WHO criteria for individuals less than 20 years of age, only six index teeth (#16, #11, #26, #36, #31, and #46) were examined. Six sites were evaluated on each tooth. The periodontal pocket was not included in the evaluations due to the age of the individuals. The DMFS index (Decayed, Missing, and Filled surfaces of the tooth index) was used to evaluate dental cavities and the WHO criteria (9) was applied.
The presence or absence of bleeding sites and gingival calculus were used as dependent variables. The chi-square test was used to compare these variables to the dental caries index, socioeconomic, and behavioral variables. Multilevel statistical modeling was performed to identify variables associated with areas of social privation on two levels--individual and contextual (districts) at a 5% level of significance. The socioeconomic and behavior variables were the independent variables for the individual level; for the contextual level, the following demographic variables were the independent variables: percentage of heads of families without income and percentage of illiterate heads of families. In the present study, areas with heads of family with no income (No. > 4) and percentage of illiterate heads of families (No. > 3) were considered social deprivation areas. The cluster variables were obtained from the Research and Planning Institute of Piracicaba (IPPLAP, Piracicaba, Sao Paulo, Brazil) (11). Odds ratios and respective 95% confidence intervals (95%CI) were estimated. Associations between the independent variables were evaluated in order to avoid the multicollinearity during the chi-square test. All the statistics were performed using SAS 9.1.2 (SAS Institute Inc., Cary, North Carolina, United States).
A GIS was constructed using the entire urban area of the municipality of Piracicaba, including a district 21.8 km from Piracicaba. Eighteen neighborhoods were sampled. The data were georeferenced using thematic maps. The system was constructed to map the data using ArcView 3.1 software (Environmental Systems Research Institute Inc., Redlands, California, United States). First, the districts were drawn with the aid of satellite photographs and maps of the municipality of Piracicaba. A georeferenced database was then created to construct the GIS (system composed of database, variables, themes, and maps). Areas at risk for gingival bleeding were identified and mapped. Due to low frequency of calculus in the present sample, the results were presented only in tables.
Associations were found in the present study between dental caries (DMFS Index) (P < 0.0001), monthly household income (P < 0.0001), father's education (P < 0.0001), mother's education (P < 0.0001), car ownership (P = 0.027), tooth brushing frequency (P = 0.0305), age at first tooth brushing (P = 0.0046), and absence or presence of gingival bleeding sites (Table 1).
The DMFS index was 1.88 for the overall sample, and 0.63 and 2.05 for private and public schools, respectively. Independent variables with a P-value < 0.15 were included in the multilevel model.
On the individual level, the schoolchildren with caries experience (presence of dental caries in the past, i.e., filled or missing teeth, and/or current presence of dental caries DMFS > 0 (Odds Ratio [OR] = 4.68; 95%CI = 3.53-6.20) and those who began tooth brushing at a later age (OR = 1.49; 95%CI = 1.04-2.14) had a greater chance of exhibiting gingival bleeding sites. On the contextual level, only the variable "percentage of illiterate heads of families" was associated to gingival bleeding (Table 2).
There was a significant association between the presence or absence of sites with calculus and the following variables: gender (P = 0.030), visits to the dentist (P = 0.061), car ownership (P = 0.072), and home ownership (P = 0.050) (Table 3). To avoid multicollinearity in this data, an association analysis was performed and a highly significant association was found between car ownership and home ownership. Thus, only the latter was tested in the multivariate model (Table 3).
Calculus was associated to the variables "gender" (OR = 1.64; 95%CI = 1.04-2.58) and "home ownership" (OR = 1.76; 95%CI = 1.11-2.78) on the individual level (P < 0.05), whereas calculus was only associated to the variable "percentage of illiterate heads of families" on the contextual level (Table 4).
The maps allow the visualization of a tendency toward better oral health (lower scores of gingival bleeding) in the downtown area of the municipality, which is considered to be a privileged area with the highest percentage of individuals earning a higher income and more years of education (Figure 1). However, on the contextual level of the multilevel analysis, only "percentage of illiterate heads of families" was significantly associated to both outcomes, whereas "income" was not statistically associated to these oral heath problems.
Analysis of the social and spatial phenomena involved in the health and disease processes can be done in the more important ecological studies conducted in recent years. These phenomena are complex and methodological refining is needed not to simplify them, but reproduce the static, isolated notion of space (2).
The use of geoprocessing can identify relationships between the distribution of health problems and environmental conditioning factors by aggregating and mapping variables from different sources (4). Determining associations between space and health problems has been investigated in the non-dental literature (12, 13). In dentistry, however, studies on this subject are scarce (14).
A geographic vision is a key aspect in public health. The population and communities tend to have their own characteristics according to their geographical distribution. Factors that influence the health of a community are classified into four groups: inherited conditions; environment (air and water quality, soil characteristics, radiation, and socioeconomic conditions); lifestyle; and health care (15).
GIS provides a broad source of tools for exploring data. These systems can be defined as a set of tools for the acquisition, storage, retrieval, analysis, and interpretation of spatial data (16, 17). The incorporation of mapping revealed inequalities in relation to the oral health status in the municipality studied. This tool is useful for evaluating and easily visualizing the areas that should receive more attention from health care managers. GIS applied to track the oral health of a population has been extremely effective at revealing the distribution of oral health problems and assisting in the planning and evaluation of the effectiveness of tailored health promotion and preventive care in underprivileged areas (14, 18-21).
It should be stressed that oral health actions should be considered, discussed, and carried out by different sectors. Moreover, actions directed at oral health alone are unable to solve problems related to the distribution of disease. The territory is a reflection of the socioeconomic condition of its inhabitants and is influenced by social and political inequalities. Thus, a territory can adversely influence the health status of its inhabitants and perpetuate poverty (2).
In Brazil, as well as the municipality studied here, most of the individuals enrolled in public schools are from underprivileged families that live in areas of social deprivation and do not have financial resources to afford private school (7). These individuals are distinguished by having a greater number of caries and lower access to the health services (7, 21). Although social inequality in Brazil is evident (22, 23), especially the contrast between very wealthy families extremely poor families, and this inequality is visible in the data from the map constructed of the area studied (Figure 1)--this social difference is not always statistically significant from a geographical standpoint, as people with different socioeconomic levels often occupy the same area. It is possible to find rich and poor individuals living in the same geographical space due to fast urban growth, as well as the growth of inequality itself.
This complexity of social factors is related to oral health problems, which means that analysis and actions that respect this complexity are required. Careful analysis is, therefore, important and should unify all the tools that can clearly explain health and disease processes in different populations. The application of geographical and multilevel analysis can offer more consistency and reliability to epidemiological data. The latest periodontal studies have used Multilevel Modeling (24), which results in a better comprehension of the structure of the complex data found in health, such as the complexity of the progression of periodontal disease and the factors predicting (25, 26) its impact on general health and the importance of surveillance regarding periodontal disease (26, 27).
Further studies that take geographical areas into account are needed in order to enhance the understanding of social inequalities in oral health and lead to the development of interventions for diminishing these inequalities. Progress in understanding the causes of social inequalities will remain limited unless there is a general change in the manner in which health is measured (28). This complexity was first recognized in the 1990s (29) by a study that considered caries to be a complex disease, with a number of social characteristics that may influence its distribution. This suggests that a detailed analysis of the data found in literature is necessary in order to identify groups that are more exposed to disease.
The present study used spatial analysis and statistical model analysis on multiple levels with the purpose of obtaining a better understanding of the data. Bivariate analysis (chi-square test) was used to determine associations between the variables studied (social, economic, and behavioral) and gingival and calculus status. From the data of this first analysis, multilevel models were constructed, in which only the variables that were associated to the outcomes were included. The mapping of the areas at risk for gingival problems was then carried out by means of a GIS.
From the data evaluated in the different analyses, social, economic, and behavioral variables are directly associated to levels of disease (chi-square test, Tables 1 and 3). This can be seen on the map (Figure 1), which shows a tendency toward better oral health in the privileged downtown areas. Regarding multilevel analysis, on the individual level, the experience with caries and age at first tooth brushing were associated to gingival bleeding, whereas gender and home ownership were associated to calculus. However, on the contextual level (districts), only the percentage of illiterate heads of families was statistically significant in both outcomes, and income was not associated with these oral health problems. These results suggest that individuals who live in places in which the percentage of illiterate heads of families is higher will have a greater proneness toward gingival problems (Tables 2 and 4). However, areas in which the income of the head of family is low were not associated to worse oral health status. This result disagrees with a large number of previous studies, which have found income to be a major factor of social deprivation and strongly associated to oral health problems, especially in Brazil (30-33). However, income alone is insufficient to test materialist theories regarding health inequalities (28).
This disagreement can be related to the presence of an oral health program directed at schoolchildren that involves a partnership between the dental school, a local company, city hall, and the state dentistry association. Treatment, along with preventive and educative measures, is offered to 5 000 students from the public schools administered by the local Board of Education of Piracicaba, Sao Paulo. The schoolchildren examined in the present study were enrolled in preventive and educational oral health programs during the years preceding this study. In addition, the community water supply has been treated with fluoride since 1971. These procedures reach the less privileged and minimize the impact of social and environmental factors on individuals. They help to reduce the existing differences regarding oral health between schoolchildren from low-income and high-income families. The study suggests that these individuals are reasonably protected from the impact of social deprivation by the municipality's public health care services and actions regarding the oral health prevention and treatment. The approach taken in the study reveals an understanding that certain socioeconomic inequalities may be sidestepped by means of oral health care programs that reach the socially underprivileged. It is important to report the efforts of the public health care services in the municipality studied.
[FIGURE 1 OMITTED]
The results of the present study could be better explained by long-term prospective ecological studies in which there are a greater number of groups and individuals (3). An authentic causal relationship could only be confirmed by randomized, controlled, prospective studies (24) that test the predictive value of variables associated to gum disease and onset of lesions with the aim of finding viable mechanisms of disease surveillance (27). The distribution of disease is commonly determined by three pillars: person, time, and space (15, 34).
Another study limitation is that in some areas, the physical environment of the municipality does not present a clear division between the different social strata. Disorderly expansion of urban areas gives rise to areas where luxury condominiums coexist with areas of social deprivation. This panorama is still evolving; however, it can already be seen as a limitation of this study.
The present study confirms better oral health status among schoolchildren from privileged families, but it does not confirm data regarding the socioeconomic cluster variable "income." Individuals whose head of family had no income were not associated to a higher prevalence of gingival problems. From the findings of the present study, it is possible to conclude that, on the individual level, socioeconomic variables were associated with a higher prevalence of the gingival bleeding and calculus, but this relationship was not observed in the multilevel model. By means of multilevel analysis and the use of a geographic information system, it was possible to better evaluate oral health status and construct a panorama of this condition in a small area.
Conflict of interests. None.
(1.) Newton JT, Bower EJ. The social determinants of oral health: new approaches to conceptualizing and researching complex causal networks. Community Dent Oral Epidemiol. 2005;33(1):25-34.
(2.) Moreira RF, Nico LS, Tomita NE. A relacao entre o espaco e a saude bucal coletiva: por uma epidemiologia georreferenciada. Cien Saude Colet. 2007;12(1):275-84.
(3.) Bower E, Gulliford M, Steele J, Newton T. Area deprivation and oral health in Scottish adults: a multilevel study. Community Dent Oral Epidemiol. 2007;35(2):118-29.
(4.) Lopes FS, Ribeiro H. Mapping of hospitalizations due to respiratory problems and possible associations to human exposure to burnt sugar-cane straw products in the state of Sao Paulo. Rev Bras Epidemiol. 2006;9(2): 215-25.
(5.) Rojas LI, Barcellos C. Geografia y salud en america latina: evolucion y tendencias. Rev Cubana Salud Publica. 2003;29(4):330-43.
(6.) Brazilian Institute of Geography and Statistics. Available at: http://www.ibge.gov. br/english/disseminacao/eventos/missao/ default.shtm Accessed 13 December 2013.
(7.) Pereira SM, Tagliaferro EPS, Ambrosano GMB, Cortellazzi KL, Meneghim MC, Pereira AC. Dental caries in 12-year-old schoolchildren and its relationship with socioeconomic and behavioral. Oral Health Prev Dent. 2007;5(4):299-306.
(8.) Demidenko E. Sample size and optimal design for logistic regression with binary interaction. Stat Med. 2008; 27(1):36-46.
(9.) World Health Organization. Oral health surveys: basic methods. 4th ed. Geneva: WHO; 1997.
(10.) Assaf AV, de Castro Meneghim M, Zanin L, Tengan C, Pereira AC. Effect of different diagnostic thresholds on dental caries calibration--a 12 month evaluation. Community Dent Oral Epidemiol. 2006;34(3):213-9.
(11.) Instituto de Pesquisas e Planejamento de Piracicaba. Available at: http://www.ipplap. com.br. Accessed 16 December 2013.
(12.) Tiwari N, Adhikari CMS, Tewari A, Kandpal V. Investigation of geospatial hotspots for the occurrence of tuberculosis in Almora district, India, using GIS and spatial scan statistic. Int J Health Geogr. 2006;10;5:33.
(13.) Harris DE, Aboueissa A, Hartley D. Myocardial infarction and heart failure hospitalization rates in Maine, USA--variability along the urban-rural continuum. Rural Remote Health. 2008;8(2):980.
(14.) Antunes JLF, Frazao P, Narvai PC, Bispo CM, Pegoretti T. Spatial analysis to identify differentials in dental needs by area-based measures. Community Dent Oral Epidemiol. 2002;30(2):133-42.
(15.) Maheswaran R, Craglia M. GIS in public health practice. 1st ed. Boca Raton: CRC Press; 2004.
(16.) Graham AJ, Atkinson PM, Danson FM. Spatial analysis for epidemiology. Acta Trop. 2004;91(3):219-25.
(17.) Ruankaew N. GIS and epidemiology. J Med Assoc Thai. 2005; 88(11):1735-8.
(18.) Carvalho ML, Moyses SJ, Bueno RE, Shimakura S, Moyses ST. A geographical population analysis of dental trauma in school-children aged 12 and 15 in the city of Curitiba-Brazil. BMC Health Serv Res. 2010;13(10):203.
(19.) Stromberg U, Holmn A, Magnusson K, Twetman S. Geo-mapping of time trends in childhood caries risk--a method for assessment of preventive care. BMC Oral Health. 2012; 12:9.
(20.) Chiang CT, Lian IeB, Su CC, Tsai KY, Lin YP, Chang TK. Spatiotemporal trends in oral cancer mortality and potential risks associated with heavy metal content in Taiwan soil. Int J Environ Res Public Health. 2010;7(11): 3916-28.
(21.) Lopes RM, Domingues GG, Junqueira SR, Araujo ME, Frias AC. Conditional factors for untreated caries in 12-year-old children in the city of Sao Paulo. Braz Oral Res. 2013;27(4):376-81.
(22.) Antunes JL, Peres MA, de Campos Mello TR, Waldman EA. Multilevel assessment of determinants of dental caries experience in Brazil. Community Dent Oral Epidemiol. 2006;34(2):146-52.
(23.) Pabayo R, Chiavegatto Filho AD, Lebrao ML, Kawachi I. Income inequality and mortality: results from a longitudinal study of older residents of Sao Paulo, Brazil. Am J Public Health. 2013;103(9):43-9.
(24.) Tu YK, Gilthorpe MS, Griffiths GS, Maddick IH, Eaton KA, Johnson NW. The application of multilevel modeling in the analysis of longitudinal periodontal data--part II: Changes in disease levels over time. J Periodontol. 2004;75(1):137-45.
(25.) Gilthorpe MS, Griffiths GS, Maddick IH, Zamzuri AT. An application of multilevel modelling to longitudinal periodontal research data. Community Dental Health. 2001;18(2):79-86.
(26.) Wan CP, Leung WK, Wong MC, Wong RM, Wan P, Lo EC, Corbet EF. Effects of smoking on healing response to non-surgical periodontal therapy: a multilevel modelling analysis. J Clin Periodontol. 2009;36(3):229-39.
(27.) Tomar SL. Public health perspectives on surveillance for periodontal disease. J Periodontol. 2007;78(7):1380-6.
(28.) Sisson KL. Theoretical explanations for social inequalities in oral health. Community Dent Oral Epidemiol. 2007;35(2):81-8.
(29.) Ellwood RP, O'Mullane DM. Identification of areas with high levels of untreated dental caries. Community Dent Oral Epidemiol. 1996; 24(1):1-6.
(30.) Peres KGA, Bastos JRM, Latorre MRDO. Relationship between severity of dental caries and social and behavioral factors in children. Rev Saude Publica. 2000;34(4):402-8.
(31.) Pattussi MP, Marcenes W, Croucher R, Sheiham A. Social deprivation, income inequality, social cohesion and dental caries in Brazilian school children. Soc Sci Med. 2001;53(7):915-25.
(32.) Cortellazzi KL, Pereira SM, Tagliaferro EP, Ambrosano GM, Zanin L, Meneghim MC, et al. Risk indicators of gingivitis in 5-year-old Brazilian children. Oral Health Prev Dent. 2008;6(2):131-7.
(33.) Antunes JL, Peres MA, Frias AC, Crosato EM, Biazevic MG. Gingival health of adolescents and the utilization of dental services, state of Sao Paulo, Brazil. Rev Saude Publica. 2008;42(2):191-9.
(34.) Werneck GL, Struchiner CJ. Studies on space-time disease clusters: concepts, techniques, and challenges. Cad. Saude Publ. 1997;13: 611-24.
Manuscript received on 30 November 2012. Revised version accepted for publication on 29 January 2014.
Stela Marcia Pereira,  Vanessa Pardi,  Karine L. Cortellazzi,  Glaucia Maria Bovi Ambrosano,  Carlos Alberto Vettorazzi,  Silvio F.B. Ferraz,  Marcelo de Castro Meneghim,  and Antonio Carlos Pereira 
 Universidade Federal de Lavras, Area de Epidemiologia e Saude Publica/Setor de Preventiva, Lavras, Minas Gerais, Brazil. Send correspondence to Stela Marcia Pereira, email: email@example.com
 Herman Ostrow School of Dentistry, University of Southern California, Division of Periodontology, Diagnostic Sciences and Dental Hygiene, Los Angeles, California, United States of America.
 Faculdade de Odontologia de Piracicaba, Universidade Estadual de Campinas, Piracicaba, Sao Paulo, Brazil.
 Universidade de Sao Paulo, Departamento de Engenharia Rural, Piracicaba, Sao Paulo, Brazil.
TABLE 1. Bivariate analysis of the association between gingival bleeding sites and gender, socioeconomic characteristics, and behavioral variables related to the individual level (subjects), among students 12 years of age, from Piracicaba, Sao Paulo, Brazil, in 2005 Variable IPC (a) no IPC (a) bleeding bleeding No. % No. % Individual level: Gender Female 259 44.9 318 55.1 Male 206 48.5 219 51.5 DMFS indexb = 0 338 63.6 194 36.4 > 0 126 26.9 343 73.1 People living in the household < 4 people 251 48.2 270 51.8 > 4 people 209 44.4 262 55.6 Monthly family income [less than or 400 44.2 505 55.8 equal to] 3 times minimum wage (c) > 3 times 56 67.5 27 32.5 minumum wage Father's education Completed 174 39.2 270 60.8 middle school Completed 129 53.8 111 46.2 high school Completed college 76 58.9 53 41.1 Mother's education Completed 238 40.0 357 60.0 middle school Completed 154 54.0 131 46.0 high school Completed 67 59.9 45 40.2 college Visits to the dentist Never/ 187 43.8 240 56.2 irregularly Regularly 267 47.7 293 52.3 Home ownership Yes 298 47.2 334 52.8 No 164 44.9 201 55.1 Car ownership No car 173 40.8 251 47.4 [greater than 283 50.4 278 49.6 or equal to] 1 Toothbrushing frequency [less than or 47 37.3 414 47.6 equal to] once/day [greater than 79 62.7 456 52.4 or equal to] twice/day Onset of toothbrushing [less than or 389 48.6 412 51.4 equal to] 1 year old > 1 year old 71 37.2 120 62.8 Variable Non-adjusted P-value Odds Ratio (95CI%) Individual level: Gender Female 1.00 0.2610 Male 0.86 (0.67-1.11) DMFS indexb = 0 1.00 < 0.0001 > 0 4.76 (3.63-6.23) People living in the household < 4 people 1.00 0.2304 > 4 people 1.16 (0.91-1.50) Monthly family income [less than or 1.00 < 0.0001 equal to] 3 times minimum wage (c) > 3 times 0.38 minumum wage (0.24-0.62) Father's education Completed 1.00 middle school Completed 0.55 < 0.0001 high school (0.40-0.76) Completed college 0.45 (0.30-0.67) Mother's education Completed 1.00 middle school Completed 0.57 < 0.0001 high school (0.43-0.75) Completed 0.45 college (0.30-0.68) Visits to the dentist Never/ 1.00 0.2251 irregularly Regularly 0.85 (0.66-1.10) Home ownership Yes 1.00 0.4982 No 1.09 (0.84-1.42) Car ownership No car 1.00 0.027 [greater than 0.68 or equal to] 1 (0.52-0.87) Toothbrushing frequency [less than or 0.65 0.0305 equal to] (0.45-0.96) once/day [greater than 1.00 or equal to] twice/day Onset of toothbrushing [less than or 1.00 0.0046 equal to] 1 year old > 1 year old 1.60 (1.15-2.21) (a) IPC = Community Periodontal Index. (b) Decayed, Missing, and Filled surfaces of the tooth index. (c) Minimum wage at the time of the data collection (2005), approximately US$ 164. TABLE 2. Multilevel model of logistic regression analysis of gingival bleeding sites among schoolchildren 12 years of age, Brazil, 2005 Variable Estimate Standard Adjusted error (SE) odds ratio Individual level: Bleeding Bleeding 0.77 0.07 4.68 No bleeding Reference Frequency of tooth brushing: [less than or 0.20 0.09 1.49 equal to] once/day [greater than Reference or equal to] twice/day -2 loglikelihood 1 203.48 (individual level) Contextual level: % Illiterate/cluster 0.10 0.03 0.11 (b) -2 loglikelihood 1 157.98 full model) Variable 95%CI (a) P Individual level: Bleeding Bleeding 3.53-6.20 < 0.0001 No bleeding Frequency of tooth brushing: [less than or 1.04-2.14 0.0303 equal to] once/day [greater than or equal to] twice/day -2 loglikelihood (individual level) Contextual level: % Illiterate/cluster 0.04 (SE) 0.0081 -2 loglikelihood full model) (a) Confidence interval. TABLE 3. Bivariate analysis of the association between sites with calculus and gender, socioeconomic characteristics, and behavioral variables related to the individual level (subjects), Brazil, 2005 Variable No calculus Calculus No. % No. % Individual level: Gender Female 538 93.2 39 6.8 Male 380 89.4 45 10.6 DMFS index (3) = 0 487 91.4 46 8.6 > 0 431 91.9 38 8.1 People living in the household < 4 people 481 92.3 40 7.7 > 4 people 427 90.7 44 9.3 Monthly household income [less than 825 91.2 80 8.8 or equal to] 3 times the minimum wage (b) > 3 times the 79 95.2 4 4.8 minimum wage Father's education Completed 406 91.4 38 8.6 middle school Completed 220 91.7 20 8.3 high school Completed college 119 92.2 10 7.8 Mother's education Completed 540 90.8 55 9.2 middle school Completed 261 91.6 24 8.4 high school Completed 107 95.5 5 4.5 College Visits to the dentist Never/Irregularly 383 89.7 44 10.3 Regularly 521 93.0 39 7.0 Home ownership Yes 587 92.9 45 7.1 No 326 89.3 39 10.7 Car ownership No car 381 89.9 43 10.1 [greater than 522 93.0 39 7.0 or equal to] 1 Tooth brushing frequency [less than or 117 92.9 9 7.1 equal to] once/day [greater than 795 91.4 75 8.6 or equal to] twice/day Onset of tooth brushing [less than or 736 91.6 65 8.1 equal to] 1 year old > 1 year old 173 90.6 18 9.4 Variable Non-adjusted P-value odds ratio (95%CI) Individual level: Gender Female 1.00 0.0307 Male 1.63 (1.04-2.56) DMFS index (3) = 0 1.00 0.7635 > 0 0.93 (0.60-1.46) People living in the household < 4 people 1.00 0.3471 > 4 people 1.24 (0.79-1.94) Monthly household income [less than 1.00 0.2088 or equal to] 3 times the minimum wage (b) > 3 times the 0.52 minimum wage (0.19-1.46) Father's education Completed 1.00 middle school Completed 0.97 0.8583 high school (0.55-1.71) Completed college 0.90 (0.43-1.86) Mother's education Completed 1.00 middle school Completed 0.90 0.2492 high school (0.55-1.49) Completed 0.46 College (0.18-1.17) Visits to the dentist Never/Irregularly 1.00 Regularly 0.65 0.0610 (0.42-1.02) Home ownership Yes 1.00 0.0509 No 1.56 (0.99-2.45) Car ownership No car 1.00 0.0728 [greater than 0.66 or equal to] 1 (0.42-1.04) Tooth brushing frequency [less than or 1.00 0.5769 equal to] once/day [greater than 1.23 or equal to] (0.60-2.51) twice/day Onset of tooth brushing [less than or 1.00 0.5571 equal to] 1 year old > 1 year old 1.17 (0.68-2.04) (a) Decayed, Missing, and Filled surfaces of the tooth index. (b) Minimum wage at the time of the data collection (2005), approximately US$ 164. TABLE 4. Multilevel model of logistic regression analysis of sites with dental calculus, Brazil, 2005 Variable Estimate Standard Adjusted error (SE) odds ratio Individual level: Gender Female 0.25 0.12 1.64 Male Reference Home ownership No 0.28 0.12 1.76 Yes Reference -2 loglikelihood 568.25 (full model) Contextual level: % Illiterate/cluster 0.08 0.04 0.008 ([beta]) -2 loglikelihood 559.13 (full model) Variable 95%CI (a) P Individual level: Gender Female 1.04-2.58 0.0387 Male Home ownership No 1.11-2.78 0.0264 Yes -2 loglikelihood (full model) Contextual level: % Illiterate/cluster 0.008 (SE) 0.3639 -2 loglikelihood (full model) (a) Confidence interval.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Original research/Investigacion original|
|Author:||Pereira, Stela Marcia; Pardi, Vanessa; Cortellazzi, Karine L.; Ambrosano, Glaucia Maria Bovi; Vettor|
|Publication:||Revista Panamericana de Salud Publica|
|Date:||Feb 1, 2014|
|Previous Article:||Social determinants of exclusion from health services and medicines in three central american countries/Determinantes sociales de la exclusion a los...|
|Next Article:||User embracement in the family health strategy in Brazil' an integrative review/Acolhimento na estrategia saude da familia: revisao integrativa.|