Autopercepção positiva de saúde em idosos : estudo populacional no Sul do Brasil Positive self-rated health in the elderly : a population-based study in the South of Brazil

The objective was to identify factors associated with positive self-rated health in the elderly in Florianópolis, Santa Catarina State, in the South of Brazil. This population-based cross-sectional study evaluated 1,705 elderly. Self-rated health was classified as positive (very good or good) or negative (fair, poor, and very poor). Crude and adjusted Poisson regression was used to identify associated factors. Prevalence of positive selfrated health was 51.2%, associated with male gender (PR = 1.13), more than 5 years of schooling, moderate (PR = 1.33) or high alcohol intake (PR = 1.37), leisure-time activity (PR = 1.20), Internet use (PR = 1.21), fewer diseases, mild/moderate dependence (PR = 2.20) or no dependence (PR = 2.67), no falls (PR = 1.19), and non-use of polypharmacy (PR = 1.27). Several modifiable factors were identified that can affect positive self-rated health in the elderly and contribute to the development of strategies to improve their quality of life. Health of the Elderly; Self-Assessment; Cross-Sectional Studies Resumo O objetivo foi identificar fatores associados à autopercepção positiva de saúde em idosos de Florianópolis, Santa Catarina, Sul do Brasil. Tratase de estudo transversal, de base populacional, com 1.705 idosos. A autopercepção de saúde foi classificada como positiva (muito boa e boa) e negativa (regular, ruim e muito ruim). A regressão bruta e ajustada de Poisson foi utilizada para identificar os fatores associados. A prevalência do desfecho foi de 51,2%, associada positivamente ao sexo masculino (RP = 1,13), ter mais de 5 anos de estudo, consumo moderado (RP = 1,33) ou alto de álcool (RP = 1,37), ser ativo no lazer (RP = 1,20), utilizar Internet (RP = 1,21), menor número de morbidades, dependência leve/moderada (RP = 2,20) ou nenhuma (RP = 2,67), não sofrer quedas (RP = 1,19) e não fazer uso de polifarmácia (RP = 1,27). Foram identificados diversos fatores modificáveis que podem interferir na autopercepção de saúde positiva de idosos e contribuir para o desenvolvimento de estratégias para melhorar a qualidade de vida desses. Saúde do Idoso; Autoavaliação; Estudos Transversais 1049 ARTIGO ARTICLE http://dx.doi.org/10.1590/0102-311X00132014 Confortin SC et al. 1050 Cad. Saúde Pública, Rio de Janeiro, 31(5):1049-1060, mai, 2015 Introdução A autopercepção de saúde é considerada bom indicador do estado de saúde em idosos, pois incorpora tanto componentes físicos, cognitivos e emocionais quanto aspectos relacionados ao bem-estar e à satisfação com a própria vida 1,2,3. Essa medida tem sido amplamente utilizada em estudos populacionais com idosos por estar associada, de forma consistente, à mortalidade e ao declínio funcional nesse grupo etário 4,5,6 e ainda por servir como ferramenta para o desenvolvimento de políticas de saúde que visem melhorar a situação de saúde da população idosa 6. A prevalência de autopercepção positiva de saúde difere de forma considerável entre os estudos. Embora as perguntas sobre autopercepção de saúde, assim como as opções de resposta, sejam similares entre as investigações, os achados não são unânimes. Essa discrepância nas prevalências pode ser devido a flutuações de curto prazo na saúde ou doença, ocasionadas por variações cíclicas relacionadas ao bem-estar. No âmbito nacional, alguns estudos 1,2,3,4 têm investigado a autopercepção de saúde em idosos, abordando a variável, principalmente, na forma de autopercepção negativa 2. No entanto, avaliar a autopercepção na sua forma positiva facilita o entendimento dos fatores que podem ser modificados na busca dos motivos que determinem positivamente a saúde. As prevalências de autopercepção de saúde positiva entre as investigações de base populacional brasileiras variam de 24,7 a 89,1% 1,4,7. Enquanto, no âmbito internacional, variam de 35,7% a 63% 8,9,10. Haja vista que a autopercepção de saúde tem relação com vários aspectos fundamentais da saúde do idoso, como, por exemplo, sexo 11, idade 11, estado civil 11, escolaridade 3,8,11,12, renda domiciliar mensal 1, atividade física 1,3,7,13 e o consumo de álcool 14, fica fácil entender as discrepâncias nas prevalências entre os estudos. Além disso, condições de saúde também apresentam relação com a autopercepção de saúde, dentre elas, as quedas 15, a capacidade física 4,7,16, os sintomas depressivos 1,16, e as morbidades 3,4,17,18, além de apresentarem impacto na mortalidade 4,6,19. O uso de medicamentos 1,11 e o uso de serviços de saúde também estão associados, com destaque para o número de consultas médicas 1, número de internações hospitalares e o tipo de seguro de saúde 11. Compreender os aspectos envolvidos na autopercepção da saúde pode evidenciar o perfil dos idosos que a referem de forma positiva, o que ainda é pouco abordado na literatura brasileira, conforme visto em revisão de literatura sobre o tema, na qual todos os estudos investigados abordavam a variável na forma negativa 2. Ao considerar a relevância do tema como importante indicador para a vigilância da saúde geral do idoso, o objetivo desse estudo é identificar a prevalência de autopercepção positiva de saúde e sua associação com fatores demográficos, socioeconômicos, de estilo de vida e condições de saúde em idosos de Florianópolis, capital de Santa Catarina, Brasil.


Abstract
The objective was to identify factors associated with positive self-rated health in the elderly in Florianópolis, Santa Catarina State, in the South of Brazil.This population-based cross-sectional study evaluated 1,705 elderly.Self-rated health was classified as positive (very good or good) or negative (fair, poor, and very poor).Crude and adjusted Poisson regression was used to identify associated factors.Prevalence of positive selfrated health was 51.2%, associated with male gender (PR = 1.13), more than 5 years of schooling, moderate (PR = 1.33) or high alcohol intake (PR = 1.37), leisure-time activity (PR = 1.20),Internet use (PR = 1.21), fewer diseases, mild/moderate dependence (PR = 2.20) or no dependence (PR = 2.67), no falls (PR = 1.19), and non-use of polypharmacy (PR = 1.27).Several modifiable factors were identified that can affect positive self-rated health in the elderly and contribute to the development of strategies to improve their quality of life.

Introduction
Self-rated health is considered good indicator of health status in the elderly, since it incorporates physical, cognitive, and emotional components as well as aspects related to well-being and satisfaction with one's own life 1,2,3 .This measure has been widely used in population-based studies of the elderly, since it is consistently associated with mortality and functional decline in this age group 4,5,6 , besides serving as a tool for developing health policies aimed at improving the elderly population's health status 4 .
Prevalence of positive self-rated health differs considerably between studies.Although the question on self-rated health and the options for answers are similar between studies, the findings are not unanimous.This discrepancy in prevalence rates may be due to short-term fluctuations in health or disease caused by cyclical variations related to well-being.Some Brazilian studies 1,2,3,4 have analyzed self-rated health in the elderly, mainly approaching the variable as negative self-rated health.However, self-rated health in its positive form facilitates understanding the factors that be modified in the search for positive determinants of health.
Since self-rated health bears a relationship to fundamental health issues in the elderly, such as gender 11 , age 11 , marital status 11 , schooling 3,8,11,12 , monthly household income 1 , physical activity 1,3,7,13 , and alcohol consumption 14 , it is easy to understand the discrepancies in prevalence rates between studies.
In addition, health conditions also bear a relationship to self-rated health, for example falls 15 , physical capacity 4,7,16 , depressive symptoms 1,16 , and diseases 3,4,17,18 , in addition to having an impact on mortality 4,6,19 .The use of medicines 1,11 and healthcare services are also associated, especially the number of medical consultations 1 , number of hospitalizations, and type of health insurance 11 .
Understanding the aspects involved in selfrated health can reveal the profile of elderly that refer to it positively, which has still received little attention in Brazilian research, as shown in a literature review on the theme, in which all the studies that were examined approached the variable in its negative form 2 .By considering the theme's relevance as an important indicator for overall health surveillance in the elderly, the current study aims to identify the prevalence of positive self-rated health and its association with demographic, socioeconomic, and lifestyle factors and health conditions in the elderly in Florianópolis, the capital of Santa Catarina State, Brazil.

Study area and population
The data in this study are part of the research project on health conditions in the elderly population in Florianópolis called EpiFloripa Idoso (http://www.epifloripa.ufsc.br/category/inqueritos/epi_idoso/epi_idoso_10/fin_idoso_09, accessed on 15/Aug/2014).This is a cross-sectional, population-based, household survey with elderly (≥ 60 years) living in the urban area of the city of Florianópolis, in 2009-2010.
Details on the location, study population, and sampling have been published previously 20 and will be presented briefly here.A two-stage sampling strategy was used to reach the initial sample size calculated at 1,599 elderly individuals.In the first stage, 420 urban census tracts were stratified upwardly according to the headof-family's mean monthly income (BRL 314.76 to BRL 5,057.77),aggregated in deciles with 42 tracts each, systematically picking eight tracts in each decile in order to cover all the groups.The units in the second stage were the households, picked systematically.All the elderly living in the selected households were invited to participate in the study.Losses were defined as interviews that had not been held after four attempts (including in evenings and on weekends), and refusals were defined as persons that chose not to answer the questionnaire.

Data collection
Data collection employed a structured face-toface interview, using personal digital assistant (PDA) to record the data.
Data consistency was verified weekly by simple frequency and comparison with expected values.Incongruent responses were identified, corrected by the supervisor and the interviewer responsible for the inconsistencies, followed by reporting back to the person responsible for the final databank.Quality control was conducted by telephone with a short version of the questionnaire in 10% of the interviews, which were randomly selected.
Self-rated health status was obtained with the question "In general, would you say that your health is: very good, good, fair, poor, or very poor?" 21 .
The five categories for the answer were dichotomized as positive ("very good" and "good") and negative ("regular", "poor", and "very poor").
• Lifestyle: smoking (current smoker, former smoker, or never smoked), alcohol consumption (never, moderate, frequent, or very frequent) verified by the first three questions of the AUDIT questionnaire (The Alcohol Use Disorders Identification Test) 22 , leisure-time physical activity verified by the long version of the International Physical Activity Questionnaire (IPAQ) (insufficiently active: 0 to 149 minutes of physical activity/week and physically active: ≥ 150 minutes of physical activity/week) 23 .
• Internet use: self-reported capacity to send and receive e-mail messages 24 .
• Health conditions: number of diseases (none; 1 to 2; or ≥ 3 self-reported diseases) verified by the question "Has some doctor or other health professional ever told you that you have...?", followed by a list of 16 health problems (the questionnaire from the Brazilian National Household Sample Survey -PNAD 25 ), functional dependence 26 was assessed by the Brazilian Questionnaire for Multidimensional Functional Assessment, adapted from the questionnaire Older Americans Resources and Services (BOMFAQ/OARS), categorized as no functional dependence 26 (absence of dependence), mild dependence (dependence in one to three activities), and moderate/severe (dependence in four or more activities).Cognitive status (normal versus presence of probable cognitive deficit) was investigated by the Mini-Mental State Examination (MMSE), with cutoff points that took schooling into account, according to Brucki et al. 27 .A fall in the previous year was verified by the question: "Did you suffer any fall in the last year?",categorized as yes or no.
• Use of medicines and healthcare services: health plan coverage was verified by the question "Do you have a private, company, or government health plan?" (yes; no); hospitalization in the previous 6 months (yes; no); and use of polypharmacy 28 (yes; no), obtained from packages and/or physician's prescriptions for medicines, recorded by the brand name and later classified according to the Anatomical Therapeutic Chemical Code (ATC) 29 , defined as concurrent use of four or more medicines.
Interviews answered by informants (n = 49) were excluded, since self-rated health is subjective and can only be provided by the elderly themselves.

Data analysis
Descriptive analyses were performed for all the variables.Prevalence rates and respective 95% confidence intervals (95%CI) were calculated for positive self-rated health according to the nature of the exposures.Bivariate and adjusted analysis used Poisson regression 30 estimating crude and adjusted prevalence ratios (PR) with the respective 95%CI.Adjusted analysis used a hierarchical model that included the demographic and socioeconomic variables (first level), lifestyle (second level), Internet use (third level), health conditions (fourth level), and use of medicines and healthcare services (the fifth and last level, more proximal in relation to the dependent variable).The statistical significance of each variable in the model was verified by the Wald test for heterogeneity or trend.Variables were entered into the model by level, and the adjusted analysis included all those with p < 0.20.Statistical significance was set at 5%.Data analysis used Stata SE 12.0 (Stata Corp., College Station, USA).All the analyses considered the cluster sample design effect, incorporating sample weights with the svy command.

Ethical issues
The project was approved by the Ethics Research Committee at the Federal University of Santa Catarina, protocol n. 352/2008.Free and informed consent was obtained from all the subjects.The authors reported no conflicts of interest.

Results
Of all the elderly eligible for the study (n = 1,911), 1,705 were interviewed, or 89.1% response rate, including the interviews with informants.Still, for the current analysis, the latter interviews were excluded (n = 49), totaling an analytical sample of 1,656 elderly.Median age was 69 years (range: 60 to 102).More than half were females (62.5%), belonged to the 60 to 69 year age group (51.7%), reported being married or having a partner (58.9%), and were not working at the time of the interview (86.5%), and 32.5% had 1 to 4 years of schooling.
The proportion of elderly with positive selfrated health was 51.2% (95%CI: 48.8-53.6).Table 2 shows the associations in the crude and adjusted analysis between positive self-rated health and independent variables.All independent variables presented significant associations with the outcome.Adjusted analysis showed that prevalence of positive self-rated health was 1.13 times higher in men than in women.Prevalence of positive self-rated health was also directly associated with schooling.
Prevalence rates for the outcome were 33% and 37% higher in the elderly with moderate and high alcohol consumption, respectively, compared to non-consumers.Elders that were active in their leisure time showed 20% higher prevalence of positive self-rated health when compared to their inactive peers.
Internet users presented 21% higher prevalence of positive self-rated health when compared to non-users.Elderly that reported no diseases or only 1 to 2 diseases showed 52% and 45% higher prevalence of the outcome when compared to those with 3 or more diseases.Those with mild dependence or absence of dependence presented 2.20 and 2.67 times the prevalence of positive self-rated health when compared to elderly with moderate to severe dependence.
Prevalence of the outcome was 19% and 27% higher in the elderly that had not suffered falls in the previous year and those not using polypharmacy, respectively, when compared to their peers.

Discussion
In the current study, prevalence of positive selfrated health was 51.2%.This result is consistent with a previous study by Silva et al. 7 , who found 50.4% prevalence in three municipalities (counties) in the Northeast, Southeast, and South of Brazil.The study in Bambuí, Minas Gerais State 1 , showed a prevalence rate of 24.7% for this outcome, while a study by Borim et al. 4 in the elderly in Campinas, São Paulo State, found a higher estimated prevalence of positive self-rated health (80.9%) than in the current study.Two possibilities may explain the differences between prevalence rates in the various studies: first, different options for answers to the question on self-rated health or even the kind of categorization, which is not homogeneous, and second, potential differences between the target regions in socioeconomic and demographic terms, both of which can affect the prevalence of self-rated health.
Positive self-rated health was associated with male gender, corroborating previous studies 31,32 .The aging process is marked by differences in health status between men and women, which may result from a combination of biological, social, and behavioral factors, leading to different perceptions of health status 33,34 .Women also tend to use healthcare services more, have a higher life expectancy, are susceptible to more diseases and functional decline, which can influence and favor the discovery of health problems 35 , and are more prone to less lethal conditions 35,36 , so that they perceive their own health worse than men.
According to the current study, positive selfrated health was associated with more schooling, corroborating a previous study 12 .Individuals with more schooling and higher income are known to have more access to information, leading to the adoption of healthy habits like adequate, balanced diet and regular physical activity 37 .Elderly with more schooling thus have better health-related quality of life 31 and therefore better self-rated health.
As for health behaviors, the current study found that positive self-rated health was associated with alcohol consumption.According to Poikolainen et al. 14 , after adjusting for sociodemographic and lifestyle variables, moderate alcohol intake was associated with positive selfrated health.Lang et al. 36 found that individuals 50 years or older that drank moderately showed better cognition and quality of life and fewer depressive symptoms.The latter authors reported that moderate alcohol intake is frequently associated with more active social life, which may explain such associations.Still, the results should be interpreted with caution, given the risk of excessive alcohol intake, especially by the elderly.Further research is thus needed to elucidate this relationship.
Two studies 3,13 found an association between leisure-time activity and positive selfrated health, corroborating the current study's findings.Physical activity provides benefits such as the maintenance and/or improvement of functional capacity, maintenance of indepen-  dence and autonomy, reduction of diseases 38 , and improvement in quality of life 13 , which fosters more social activity and thus biopsychosocial well-being and preservation of physical and mental health 38 .
As for Internet use, no studies were found that demonstrate this association.This is a new and promising field of study, since digital inclusion of the elderly is a recent phenomenon with various positive effects on health and quality of life.Studies show that elderly Internet users have healthier habits 24,39 and better functional capacity 24,40 and cognitive status 41 , which can foster better self-rated health.Elderly that use new technologies are also motivated to improve their connection to the world, seek communication and interaction, especially with family and friends, pursue possibilities for leisure, and want to feel less excluded from society in the digital age 42 .This may be linked to better self-esteem, self-confidence, and mental health 43 , which could explain the relationship between Internet use and more optimistic self-rated health.
The current study found that the absence or lower number of diseases increased the prevalence of positive self-rated health.This finding may be explained by the fact that elderly that reported 1 or 2 chronic diseases might have more ease in controlling them, and thus see themselves as healthy and enjoy better self-rated health.Previous studies found an association between diseases 44 and negative self-rated health 3,17 (the more the diseases, the higher the prevalence of negative self-rated health).Elderly with three or more diseases present worse quality of life 45,46 , worse health conditions 45 , and more difficulty in performing ADL and consequently worse functional capacity 45 .These elements may lead older individuals with 3 or more diseases to rate their health less optimistically.
The association between self-rated health and less dependence in ADL was found in other studies 7,44 .Greater difficulty in performing ADL is known to lead to worse functional capacity 45 , negatively impacting quality of life in the elderly 47 .In addition, dependence in ADL interferes in the social activities in which the elderly participate, such as attending church, visiting friends 46 , and attending social groups, among others.This may explain the lower prevalence of positive self-rated health in more dependent elderly.
The association between better self-rated health and not suffering falls may be understood by the loss of autonomy and independence in ADL generated by falls, which involve feelings of frailty and insecurity, decrease in social activities, and low self-esteem 47 .Falls can negatively affect quality of life in the elderly and lead to lower prevalence of positive self-rated health.Another study found similar results 15 .
Non-use of polypharmacy is associated with positive self-rated health.Among the elderly, medication is one of the most widely used interventions to treat diseases.Thus, the use of medicines can indicate that something is wrong and interfere in their self-rated health 48 .Polypharmacy can also be attributed to one or more self-reported diseases, which can contribute to clinical alterations, in addition to higher frequency of drug-drug interactions with potential adverse effects, worse health status 3,45 , and less optimistic self-rated health.
This study has some limitations to be considered when interpreting the results.The crosssectional design does not allow establishing causal relations between self-rated health and the exposure variables, but indicates the magnitude of the associations and can raise new hypotheses for developing this area of study.Second, the use of self-reported measures (socioeconomic and lifestyle variables, health conditions, and use of medication and healthcare services) may involve an information bias.However, various precautions were taken to avoid biases, using standardized instruments and extensive training of the field team.
Among the study's strengths, the results are population-based data with a representative sample of the elderly population in Florianópolis.These findings may differ from those of other studies, due to the reasons presented in the introduction, since the city has one of the highest Municipal Human Development Indices (IDH-M) in Brazil, in addition to high income and educational indicators (Programa das Nações Unidas para o Desenvolvimento.Atlas do Desenvolvimento Humano no Brasil 2013.Ranking -todo o Brasil (2010).http://atlasbrasil.org.br/2013/ranking,accessed on 08/Nov/2014).This may influence residents' quality of life and self-rated health.
Self-rated health is a subjective measure that is considered a good indicator of health status in the elderly.This study can reveal the profile of older persons with positive self-rated health, essential for monitoring their overall health.Based on this, the development of health policies can support measures on issues associated with positive self-rated health and indirectly promote strategies to improve quality of life in this population group.
In conclusion, more schooling, more alcohol consumption, leisure-time physical activity, Internet use, fewer chronic illnesses, less dependence in ADL, absence of falls, and non-use of polypharmacy were associated with positive self-rated health in the elderly in the city of Florianópolis.

Salud del Anciano; Autoevaluación; Estudios Transversales
Contributors S. C. Confortin, M. W. C. Giehl, D. L. Antes, I. J. C. Schneider, and E. d'Orsi participated in the study's conceptualization, data analysis and interpretation, writing of the article, and approval of the final version for publication and are responsible for all aspects pertaining to the entire article's accuracy and integrity.

Table 1
Description of sample and prevalence rates for positive self-rated health according to demographic, socioeconomic, and lifestyle variables, Internet use, health conditions, and use of medicines and healthcare services among the elderly in Florianópolis, Santa Catarina State, Brazil, 2009-2010.

Table 2
Crude and adjusted analysis of factors associated with positive self-rated health in the elderly.Florianópolis, Santa Catarina State, Brazil, 2009-2010.