Factores sociodemográficos y de salud asociados con la fragilidad en ancianos: una revisión sistemática de la literatura

Mello, Amanda de Carvalho; Engstrom, Elyne Montenegro; Alves, Luciana Correia

doi:10.1590/0102-311X00148213

Acessibilidade / Reportar erro

Brasil

Cadernos de Saúde Pública

Español English

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

REVIEW • Cad. Saúde Pública 30 (6) • Jun 2014 • https://doi.org/10.1590/0102-311X00148213 copiar

Factores sociodemográficos y de salud asociados con la fragilidad en ancianos: una revisión sistemática de la literatura

Autoría SCIMAGO INSTITUTIONS RANKINGS

La fragilidad es un síndrome que causa daño en la vida práctica de ancianos, ya que está relacionada con un mayor riesgo de dependencia, caídas, hospitalización, institucionalización y muerte. El objetivo de esta revisión sistemática fue identificar factores sociodemográficos, psicoconductuales, de condiciones de salud, nutrición y estilo de vida asociados a fragilidad en ancianos. Se detectaron 4.183 estudios publicados entre 2001 y 2013 en las bases bibliográficas y se seleccionaron 182 artículos completos. Después de la lectura y aplicación de los criterios de selección, quedaron 35 artículos elegibles para análisis. Los principales factores asociados fueron edad, sexo femenino, raza negra, educación, ingresos, enfermedad cardiovascular, número de comorbilidades/enfermedad, incapacidad funcional, autopercepción de mala salud, síntomas depresivos, función cognitiva, índice de masa corporal, tabaquismo y consumo de alcohol. El conocimiento de la complejidad de los determinantes de fragilidad ayuda en la formulación de medidas preventivas e intervención temprana, asegurando mejor calidad de vida.

Anciano Frágil; Calidad de Vida; Factores de Riesgo

Exposure
Independent variables	a) Secure recording + primary measurements * (low risk of bias)
	b) Structured interview + primary measurements, without knowledge of outcome * (low risk of bias)
	c) Interview with knowledge of outcome (high risk of bias)
	d) Non-secures sources and self-assessment (high risk of bias)
	e) Does not describe clearly (uncertain risk of bias)
Outcome
Is the assessment of frailty adequate?	a) Yes, according to Fried et al.² * (low risk of bias)
	b) Yes, with some changes (2 or 1 components) (uncertain risk of bias)
	c) Yes, with many changes (3 or more components) (high risk of bias)
	d) Does not describe clearly (uncertain risk of bias)
Representativeness of sample	a) Representative of local population * (low risk of bias)
	b) Possibility of selection bias (high risk of bias)
	c) Does not describe clearly (uncertain risk of bias)
Selection of participants	a) Community * (low risk of bias)
	b) Hospital or part of hospital sample (high risk of bias)
	c) Does not describe clearly (uncertain risk of bias)
Definition of the control group or cohort (only for longitudinal studies)	a) Without previous history of the syndrome * (low risk of bias)
	b) Does not describe clearly (uncertain risk of bias)

Article (year)	Study design	Independent variables	Statistical technique	Variables used in adjustment of final model	Principal results		Methodological limitations according to the authors
Article (year)	Study design	Independent variables	Statistical technique	Variables used in adjustment of final model	Positive association with frailty	Inverse association with frailty	Methodological limitations according to the authors
Fried et al. ² (2001)	Longitudinal	Demographic: age, gender, race Socioeconomic: schooling, income, living alone Diseases and health conditions: number of diseases, comorbidities (CVD, COPD, SAH, diabetes mellitus, arthritis, cancer) Psycho-behavioral cognitive function, depressive symptoms Other: ADL, IADL, self-rated health, functional incapacity	Cochran-Mantel-Haenszel test	Age	Female gender, black race, poor self-rated health, living alone, number of diseases (2 or more), CVD, lung diseases, arthritis and diabetes, functional incapacity and depressive symptoms	Income, schooling and cognitive function	Measures used in operationalization of the frailty criterion were limited to those used during the data collection for other study purposes; weight at the beginning of study was self-reported
Newman et al. ²⁷ (2001)	Longitudinal	Demographic: age, gender, race Socioeconomic: schooling, income Diseases and health conditions: CVD Psycho-behavioral: cognitive function, depressive symptoms Other: ADL, IADL, self-rated health, functional incapacity	Bivariate analysis and multinomial logistic regression	Age, gender and race	Female gender, black race, CVD	Income, schooling and cognitive function	Use of self-reported measures; few details on fatigue and energy expenditure (Kcal); cross-sectional study, does not allow causal inferences
Blaum et al. ³³ (2005)	Cross-sectional	Demographic: age, gender, race Socioeconomic: schooling Diseases and health conditions: CVD, COPD, osteoarthritis Nutritional status: BMI Other: functional status (ADL and IADL)	Multinomial logistic regression	All variables listed	Pre-frailty and frailty: overweight/ obesity	-	-
Woods et al. ²⁹ (2005)	Longitudinal	Demographic: age, race Socioeconomic: schooling, income, living alone Diseases and health conditions: comorbidity (CVD, SAH, diabetes mellitus, fractures, COPD, arthritis, stroke) Lifestyle: smoking, alcohol use Nutritional status: BMI Psycho-behavioral: depressive symptoms Other: functional status (ADL)	Bivariate analysis and multivariate logistic regression	All	Age, black and Hispanic race, underweight and overweight/obesity (BMI), comorbidities, depressive symptoms, history of fall, functional incapacity, poor self-rated health, smoking	Income, schooling, living alone, alcohol use	Lack of information on physical activity and unintentional weight loss; study limited to sample of women, not possible to make inferences for men; low cognitive function may be confounder
Hirsch et al. ²⁵ (2006)	Cross-sectional	Demographic: age, gender, race Socioeconomic: schooling, income Diseases and health conditions: CVD, SAH, COPD, diabetes mellitus and arthritis Psycho-behavioral: cognitive function Nutritional status: BMI	Multinomial logistic regression	Gender and obesity	Non-obese blacks	-	Cross-sectional study, does not allow causal inferences; possible selection bias in subject recruitment; introduction of bias by excluding individuals with missing data
Michelon et al. ¹⁷ (2006)	Cross-sectional	Demographic: age, race, marital status Socioeconomic: schooling, income Lifestyle: smoking, alcohol use Nutritional status: BMI, self-reported weight loss and low appetite Other: functional incapacity, self-rated health	Bivariate analysis	-	Age, black race, widowhood, smoking, self-reported weight loss, low appetite, use of dentures, problems with swallowing, poor self-rated health, BMI, functional incapacity	Income, schooling, alcohol use	Cross-sectional study, does not allow causal inferences; non- inclusion of dietary analyses
Semba et al. ³⁴ (2006)	Longitudinal	Demographic: age, race, marital status Socioeconomic: schooling, income Lifestyle: smoking, alcohol use Nutritional status: BMI, self-reported weight loss and low appetite Other: functional incapacity, self-rated health	Bivariate analysis	-	Age, CVD, low appetite	BMI, schooling	-
Ávila-Funes et al. ¹⁶ (2008)	Longitudinal	Demographic: age, gender, race, marital status Socioeconomic: schooling, income, living alone Diseases and health conditions: CVD, COPD, diabetes mellitus, SAH, cancer and arthrosis Lifestyle: smoking, alcohol use Psycho-behavioral: cognitive function and depressive symptoms Nutritional status: food intake Other: functional incapacity, self-rated health	Bivariate analysis	-	Age, female gender, chronic diseases, functional incapacity, depressive symptoms, poor self-rated health	Schooling, income, cognitive function, alcohol use and smoking	Adaptation of scale from Fried et al. ² for weight loss and grip strength
Alvarado et al. ³⁹ (2008)	Cross-sectional	Demographic: age, gender, marital status Socioeconomic: schooling, socioeconomic status Diseases and health conditions: comorbidities, childhood and adulthood diseases Lifestyle: physical activity Other: social history	Logistic regression	Social history, comorbidity, gender, childhood diseases, schooling, occupation, marital status, socioeconomic status	Female gender	Living conditions in childhood	Adaptation of scale from Fried et al. ² for weight and gait speed; absence of mortality study; recall bias for information on poverty in childhood
Chaves et al. ³⁵ (2008)	Cross-sectional	Demographic: age, race Diseases and health conditions: CVD, diabetes mellitus Psycho-behavioral: cognitive function, depressive symptoms Other: self-rated health	Multivariate logistic regression	-	Age, depressive symptoms, CHF	-	Cross-sectional study, does not allow causal inferences
Endeshaw et al. ⁴⁰ (2009)	Cross-sectional	Demographic: age, race Diseases and health conditions: SAH, diabetes mellitus, CVD, stroke Psycho-behavioral: cognitive function Nutritional status: BMI Other: functional status (ADL)	Multivariate logistic regression	-	In men: age, black race, CVD, stroke In women: age, obesity	-	Cross-sectional study, does not allow causal inferences
Masel et al. ²⁶ (2009)	Cross-sectional	Demographic: age, gender, marital status Socioeconomic: schooling, difficulty paying expenses Diseases and health conditions: CVD, stroke, arthrosis, SAH, diabetes mellitus, fractures, number of comorbidities Nutritional status: BMI Other: quality of life	Bivariate analysis	-	Age, chronic diseases, arthritis, underweight (BMI), low quality of life	Married, schooling, overweight (BMI)	Cross-sectional study, does not allow causal inferences; ethnic homogeneity; some variables self-reported
Ottenbacher et al. ⁴¹ (2009)	Longitudinal	Demographic: age, gender, marital status Socioeconomic: financial status, schooling Diseases and health conditions: CVD, stroke, arthritis, cancer, diabetes mellitus Lifestyle: smoking Nutritional status: weight, height Psycho-behavioral: cognitive function and depressive symptoms Other: functional status (ADL, IADL)	Multiple linear regression	All, in 3 models	Age, history of smoking, diabetes mellitus, arthritis, BMI, depression, number of comorbidities	Cognitive function	Conditions and comorbidities self-reported; original sample excludes individuals that did not complete the performance tests required to calculate frailty construct. Persons that remained in the study represented the healthier members of the original sample
Szanton et al. ³⁶ (2009)	Cross-sectional	Demographic: age, race Socioeconomic: schooling Diseases and health conditions: number of chronic diseases Nutritional status: BMI Lifestyle: smoking	Bivariate analysis	-	Age, BMI	Schooling	Cross-sectional study, does not allow causal inferences
Wu et al. ⁴² (2009)	Cross-sectional	Demographic: age Diseases and health conditions: SAH, diabetes mellitus, CHF, osteoarthritis Psycho-behavioral: depression Nutritional status: BMI, waist/hips ratio Lifestyle: smoking	Bivariate analysis and logistic regression	-	Age, depression, waist/hips ratio (> 95cm)	-	Cross-sectional study, does not allow causal inferences; sample considered small
Alcalá et al. ¹⁸ (2010)	Cross-sectional	Demographic: age, gender, marital status Socioeconomic: schooling Diseases and health conditions: chronic diseases Other: functional status (ADL, IADL)	Logistic regression	Age, schooling, comorbidities, functional incapacity	Age (> 85 years), comorbidities, functional incapacity	Schooling	Advanced age of participants (> 71 years), different socioeconomic characteristics and lack of homogeneity in measurement of frailty criteria may have influenced the observed differences
Chang et al. ³⁷ (2010)	Cross-sectional	Demographic: age, race Socioeconomic: schooling Diseases and health conditions: CKD, lung disease, CVD, diabetes mellitus, anemia, arthritis, peripheral artery disease, total number of inflammatory diseases Psycho-behavioral: depressive symptoms	Multivariate logistic regression	Age, race, schooling	All diseases, depressive symptoms	-	Cross-sectional study, does not allow causal inferences; study limited to sample of women, cannot make inferences for men; sample with higher percentage of black women than the general population; imprecise estimate of CKD due to the diagnostic method used
Chen et al. ²⁸ (2010)	Cross-sectional	Demographic: age, gender, marital status Socioeconomic: schooling Diseases and health conditions: CVD, gout, diabetes mellitus, kidney disease, COPD, osteoarthritis, osteoporosis, arthritis, osteoporosis, ulcer, cancer, SAH, cataract, liver and gall bladder disease, number of comorbidities, geriatric syndromes Psycho-behavioral: depressive symptoms Other: functional incapacity	Bivariate analysis	-	Age, female gender, absence of spouse, functional incapacity, comorbidity for chronic diseases, depression, geriatric syndromes	Schooling	-
Hubbard et al. ⁴³ (2010)	Cross-sectional	Nutritional status: BMI, WC	Logistic regression	Age, gender, schooling, wealth/assets, smoking	WC ≥ 88cm in women and ≥ 102cm in men	-	Adaptation of scale from Fried et al. ² on criterion for weight loss
Syddaal et al. ⁴⁴ (2010)	Longitudinal	Demographic: age, gender Socioeconomic: socioeconomic status, schooling, number of cars, own home Lifestyle: physical activity, smoking, alcohol use	Multivariate logistic regression	Age, comorbidity, lifestyle, socioeconomic status, car	In men: number of cars, age	In men: schooling, own home In women: without own home	Cross-sectional study, does not allow causal inferences; did not assess other socioeconomic variables like income; analysis of information bias: age and behaviors such as smoking and social factors such home ownership suffer neighborhood influence
Szanton et al. ³⁸ (2010)	Cross-sectional	Demographic: age, race Socioeconomics: schooling, income, number of home, poverty Diseases and health conditions: number of chronic diseases Lifestyle: smoking	Multinomial logistic regression	Race, schooling, smoking, state of insecurity, number of chronic diseases	Black race	Schooling, income	Cross-sectional study, does not allow causal inferences; only considered white and black races
Drey et al. ⁴⁵ (2011)	Cross-sectional	Demographic: age, gender Socioeconomic: schooling Psycho-behavioral: cognitive function, depressive symptoms Other: quality of life	Bivariate analysis	-	Number of diseases, depressive symptoms	Cognitive function, quality of life	-
Giménez et al. ⁴⁶ (2011)	Cross-sectional	Demographic: age, gender Socioeconomic: marital status, schooling, income, living alone Nutritional status: not malnourished /malnourished (MNA) Diseases and health conditions: comorbidity Psycho-behavioral: depressive symptoms Other: functional status (ADL, IADL)	Bivariate analysis	-	Female gender, comorbidity, depressive symptoms, functional dependency, malnutrition	-	Cross-sectional study, does not allow causal inferences
Garcia-Garcia et al. ⁴⁷ (2011)	Cross-sectional	Demographic: age, gender Socioeconomic: conjugal status, mean years of schooling, educational level Diseases and health conditions: comorbidities (SAH, diabetes mellitus, CVD, COPD, peptic ulcer, fractures, osteoporosis, arthritis, dementia, Parkinson, cancer, osteoporosis, osteoarthritis, kidney, thyroid, and liver diseases, high cholesterol, stroke) Psycho-behavioral: cognitive function, depressive symptoms Other: functional status (ADL, IADL)	Bivariate analysis	-	Age, CVD, Parkinson, stroke, dementia, COPD, hip fracture, depressive symptoms, functional dependency, cognitive impairment	-	-
Hoeck et al. ⁴⁸ (2011)	Cross-sectional	Demographic: age, gender Socioeconomic: socioeconomic status, schooling, income familiar, housing situation Diseases and health conditions: comorbidities Other: year of interview	Bivariate analysis and multivariate logistic regression	Age, gender, comorbidity	Comorbidities, Wallonia (Belgium), rented housing	More recent interview year, schooling, family income	-
Danon-Hersch et al. ⁴⁹ (2012)	Cross-sectional	Demographic: year of birth, gender Diseases and health conditions: number of chronic diseases (CVD, stroke, diabetes mellitus, SAH, cancer, chronic respiratory disease, arthritis), self-reported disease Other: functional status (ADL, IADL)	Bivariate analysis and multivariate logistic regression	Gender, components of frailty	Number of chronic diseases, diseases self-reported (except hypercholestero-lemia)	Functional capacity, year of birth	Since the analysis was from a cohort, losses of information may have occurred; differences in operationalization of the components in criterion from Fried et al. ²
Lakey et al. ⁵⁰ (2012)	Longitudinal	Psycho-behavioral: depressive symptoms	Multinomial logistic regression	Age, income, schooling, race, living alone, BMI, self-rated health, ADL, smoking, alcohol consumption, hormone replacement therapy, SAH, diabetes mellitus, CVD, COPD, hip fracture, falls, arthritis, cancer, stroke, number of comorbidities	Depressive symptoms	-	Lack of information on indication for antidepressants, dosage, and treatment adherence
Bastos-Barbosa et al. ⁵¹ (2012)	Cross-sectional	Demographic: age, gender, race Socioeconomic: schooling, income, conjugal status, work Diseases and health conditions: blood pressure, number of diseases, comorbidities (SAH, dyslipidemia, osteoporosis, osteoarthritis, diabetes mellitus, hypothyroidism) Psycho-behavioral: cognitive function Nutritional status: BMI, waist circumference	Fixed-effects linear models and ANOVA	Age, gender	Systolic and diastolic pressure, waist circumference	-	Selection of frail and non-frail participants according to specific criteria, which does not allow extrapolating to other elderly; small sample: results cannot be extrapolated to populations with very different living conditions
Jürschik et al. ⁵² (2012)	Cross-sectional	Demographic: age, gender, race Socioeconomic: schooling, income, conjugal status, work, living alone Diseases and health conditions: number of comorbidities Psycho-behavioral: cognitive function, depressive symptoms, quality of life, social relations Lifestyle: smoking, alcohol consumption Nutritional status: nutritional risk (MNA), BMI, waist circumference Other: functional status (ADL, IADL), self-rated health	Bivariate analysis and logistic regression	-	Age, female gender, conjugal status widowed or single, smoking, alcohol consumption, number of comorbidities, functional incapacity, depressive symptoms, cognitive decline, risk of malnutrition, visual impairment, poor self-rated health	Social interaction, quality of life	Modification of criterion for unintentional weight loss; cross-sectional study, does not allow causal inferences
Chang et al. ⁵³ (2012)	Cross-sectional	Demographic: age, gender Socioeconomic: Living alone Diseases and health conditions: blood pressure, number of comorbidities, diseases (SAH, diabetes mellitus, CVD, arthritis, peptic ulcer) Psycho-behavioral: cognitive function, depressive symptoms, health-related quality of life, social relations	Bivariate analysis and multivariate regression analysis	Age, number of comorbidities, living alone, falls in previous year, arthritis, peptic ulcer, depression	Age, number of comorbidities, living alone, arthritis, peptic ulcer, depression	Components of health-related quality of life scale, like physical and mental	Cross-sectional study, does not allow causal inferences; study population predominantly urban, which does not allow extrapolating the findings to the rural population; small sample of frail elderly; lack of specific cutoff points for components of the diagnostic criterion for frailty in the study population; low response rate due to exclusion of many comorbidities
Sousa et al. ⁵⁴ (2012)	Cross-sectional	Demographic: age, gender, race Socioeconomic: conjugal status, family arrangement, household situation, schooling, monthly income, occupation Diseases and health conditions: comorbidities, chronic diseases (SAH, diabetes mellitus, CVD, malignant tumors, arthritis or rheumatism, lung diseases, stroke, osteoporosis) Psycho-behavioral: cognitive function, depressive symptoms Other: falls, functional status (ADL, IADL), self-rated health	Bivariate analysis and binary logistic regression	Socio-demographic factors, functional status, chronic diseases, self-rated health	Advanced age osteoporosis, stroke, depression, falls, presence of comorbidities, functional dependency, poor self-rated health	-	Cross-sectional study, does not allow causal inferences
Neri et al. ⁵⁵ (2012)	Cross-sectional	Demographic: age, gender Socioeconomic: family income, family arrangement Diseases and health conditions: number of chronic diseases (SAH, diabetes mellitus, CVD, cancer, arthritis or rheumatism, ischemia, stroke, depression, osteoporosis) Psycho-behavioral: cognitive function, depressive symptoms Other: perception of social support, social isolation	Bivariate analysis and univariate and multivariate logistic regression	-	No factors significantly associated with frailty were found		Cross-sectional study, does not allow causal inferences; difficulty in controlling each variable’s effect; limitations in design and sample size; unequal number of men and women in sample; lack of information on care
Schnittger et al. ⁵⁶ (2012)	Cross-sectional	Demographic: age, gender Socioeconomic: educational level, living alone Diseases and health conditions: age-adjusted comorbidity index Psycho-behavioral: cognitive function, psychological status Nutritional status: nutritional risk (MNA), BMI Other: functional status (ADL, IADL)	Bivariate analysis and Kaiser-Meyer-Olkin test	-	Psychological distress (mood, stress, neuroticism, and emotional loneliness)	-	Cross-sectional study, does not allow causal inferences; dimensions of psychological distress measured are specific to the study population, and cannot be generalized to other population
Casale-Martínez et al. ⁵⁷(2012)	Cross-sectional	Socioeconomic: educational level, parents’ educational level, conjugal status, household assets and situation, socioeconomic status, employment benefits, occupational history, friends and family members living in the same neighborhood, financial support Diseases and health conditions: childhood history Psycho-behavioral: ability to make important decisions Other: religion, volunteer work, abuse	Multinomial logistic regression	-	Not having companion, not making important decisions, poor economic status, abuse	Employment benefits	-
Macuco et al. 58 (2012)	Cross-sectional	Demographic: age, gender Socioeconomic: years of schooling, monthly family income Psycho-behavioral: cognitive function, loneliness, adverse life events Other: functional status (IADL)	Bivariate analysis and univariate and multivariate linear regression	-	Age	Cognitive function, years of schooling, monthly family income	Rigid exclusion criteria; high number of younger elderly in the ample; cross-sectional study, no follow-up of cohort

Article (year)	Independent variables obtained	Is the assessment of frailty adequate?	Representativeness of sample	Participant selection	Definition of control group or cohort *
Fried et al. ² (2001)	B	B	I	B	I
Newman et al. ²⁷ (2001)	B	B	I	B	I
Blaum et al. ³³ (2005)	B	A	I	B	-
Woods et al. ²⁹ (2005)	B	A	I	B	I
Hirsch et al. ²⁵ (2006)	B	B	I	B	-
Michelon et al. ¹⁷ (2006)	B	I	I	B	-
Semba et al. ³⁴ (2006)	B	I	I	B	I
Ávila-Funes et al. ¹⁶ (2008)	B	A	I	B	I
Alvarado et al. ³⁹ (2008)	B	A	I	B	-
Chaves et al. ³⁵ (2008)	B	A	A	B	-
Endeshaw et al. ⁴⁰ (2009)	B	B	I	B	-
Masel et al. ²⁶ (2009)	B	I	I	B	-
Ottenbacher et al. ⁴¹ (2009)	B	I	I	B	I
Szanton et al. ³⁶ (2009)	B	A	I	B	-
Wu et al. ⁴² (2009)	B	A	I	B	-
Alcalá et al. ¹⁸ (2010)	B	A	B	B	-
Chang et al. ³⁷ (2010)	B	A	I	B	-
Chen et al. ²⁸ (2010)	I	A	I	B	-
Hubbard et al. ⁴³ (2010)	I	A	B	B	-
Syddaal et al. ⁴⁴ (2010)	B	A	I	B	I
Szanton et al. ³⁸ (2010)	B	A	A	B	-
Drey et al. ⁴⁵ (2011)	B	B	I	B	-
Giménez et al. ⁴⁶ (2011)	B	A	I	B	-
Garcia-Garcia et al. ⁴⁷ (2011)	B	I	B	A	-
Hoeck et al. ⁴⁸ (2011)	I	A	B	B	-
Danon-Hersch et al. ⁴⁹ (2012)	B	A	B	B	-
Lakey et al. ⁵⁰ (2012)	B	A	I	B	B
Bastos-Barbosa et al. ⁵¹ (2012)	B	B	I	B	-
Jürschik et al. ⁵² (2012)	B	A	I	B	-
Chang et al. ⁵³ (2012)	B	A	I	B	-
Sousa et al. ⁵⁴ (2012)	B	B	I	B	-
Neri et al. ⁵⁵ (2012)	B	B	I	B	-
Schnittger et al. ⁵⁶ (2012)	B	I	I	B	-
Casale-Martínez et al. ⁵⁷ (2012)	I	I	I	B	-
Macuco et al. ⁵⁸ (2012)	B	B	I	B	-

Escola Nacional de Saúde Pública Sergio Arouca, Fundação Oswaldo Cruz Rua Leopoldo Bulhões, 1480 , 21041-210 Rio de Janeiro RJ Brazil, Tel.:+55 21 2598-2511, Fax: +55 21 2598-2737 / +55 21 2598-2514 - Rio de Janeiro - RJ - Brazil
E-mail: cadernos@ensp.fiocruz.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] Correspondence A. C. Mello Departamento de Ciências Sociais, Escola Nacional de Saúde Pública Sergio Arouca, Fundação Oswaldo Cruz. Rua Leopoldo Bulhões 1480, Rio de Janeiro, RJ 21041-210, Brasil.amandita.nut@gmail.com