Cluster analysis of risk factors for chronic non-communicable diseases in elderly Brazilians: population-based cross-sectional studies in a rural town

Negative health behaviors incorporated into lifestyle are considered the main risk factors for chronic non-communicable diseases (NCDs) in adults and the elderly. However, the relationship between the aggregation of these factors and the sociodemographic conditions of the elderly needs to be better elucidated. The aim of this study was to analyze the simultaneity of the five risk factors for NCDs in the elderly with low economic status living in a rural city in Brazil, and their association with sociodemographic variables. Cross-sectional study was conducted with elderly people from Family Health Units of the city of Ibicui-Bahia, Brazil, where 310 elderly were enrolled. Rates of physical inactivity in leisure (PIL), alcohol consumption, sedentary behavior, overweight/obesity and tobacco consumption were collected through a questionnaire in an individual interview. The average age among participants was 71.62 (± 8.16) years. The group presenting the five behaviors had high scores in both sexes (men O/E = 242.5; women O/E = 161.7). Among men and women, the highest scores found through clustering of simultaneous NCD risk factors were for the consumption of alcohol with smoking, and physical inactivity with smoking. When analyzing the association between groups and sociodemographic characteristics, men were more physically inactive than women (OR = 0.96, CI = 0.92-0.98) and concomitantly had unhealthy habits (smoking). In conclusion, the elderly are exposed to health risk factors for concurrent CNCDs and the accumulation of these risk factors was not associated with sociodemographic variables, it is suggested that further studies be carried out with the prevalent variables as well as to analyze why the elderly population presents these levels. tasas de inactividad física en el tiempo libre (PIL), consumo de alcohol, comportamiento sedentario, sobrepeso / obesidad y consumo de tabaco mediante un cuestionario en entrevista individual. La edad promedio de los participantes fue de 71,62 (± 8,16) años. El grupo que presentó las cinco conductas tuvo puntuaciones altas en ambos sexos (hombres O / E = 242.5; mujeres O / E = 161.7). Entre hombres y mujeres, las puntuaciones más altas encontradas mediante la agrupación de factores de riesgo de ENT simultáneos fueron para el Al y características inactivos físicamente que las (OR y concomitantemente poco (tabaquismo). factores variables variables qué la


Introduction
The human aging has been a historically pronounced worldwide trend. Approximately 22% of the elderly population has increased in recent decades worldwide (Bloom et al., 2015). Although, in Brazil, the trend is for a faster demographic aging when compared to world projections, assuming an aging growth of 252% in 2020 compared to 2010 data (United Nations, 2015;IBGE, 2018). Although aging is common in the population, it is still challenging to distinguish cases of chronic noncommunicable diseases (NCDs) in this population.
It is known that NCDs are varied (cancer, diabetes, cardiovascular diseases, chronic respiratory diseases, etc.) they are the main cause of death worldwide (Williams et al., 2018). Numerous characteristics, such as being 70 years of age or older and residing in low-and middle-income countries, lead to the risk of death (Bennett, et al., 2018). Moreover, some strategies to prevent and control NCDs are under development to understand the etiology of these diseases, such as investing some countries in preventing NCDs and obtaining economic and social returns (Bertram et al., 2018), on the other hand, actions are used to modify risk factors (tobacco use, harmful use of alcohol, physical inactivity, unhealthy eating behaviors, overweight/obesity, among others) (World Health Organization, 2018). However, some investigations are focused on analyzing the clustering of multiple behavioral risk factors for NCDs, providing supportive interventions that target multiple risk behaviors, helping to plan future preventive strategies (Meade et al., 2016;Noble et al., 2015).
In addition, sociodemographic conditions, and their relationship with NCDs and how lifestyle can influence the quality of life of the elderly. The problems related to overweight are common and are directly associated with socioeconomic conditions and high consumption of carbohydrate-rich foods, in addition to the automation of daily activities and sedentary lifestyle (Silva et al, 2018). Therefore, it is necessary to investigate the factors that trigger possible NCDs in the elderly to understand the relationship of modifiable risk factors with sociodemographic factors for epidemiology (Bertram et al., 2018).
The development of NCDs is quite common in cases of excess of modifiable risk factors and even more frequent due to sociodemographic factors. Through combinations of two or more risk factors, the elderly is more exposed to NCDs, therefore, it was defined that those individuals who have more than one risk factor are in a worse health situation and whose sociodemographic factors can interfere with the conditions of life (Cruz et al. 2017). However, the combination of five modifiable factors and their impact on sociodemographic conditions and on the prognosis of elderly individuals is still unknown. Therefore, our objective is to analyze the simultaneity of the five risk factors for NCDs in low-income elderly people living in a rural city in Brazil and to show their impact on the association with sociodemographic variables. Our hypothesis is that there are different risk factors and that there is a correlation between them, as observed in Cruz et al., 2017. It is expected that there are conglomerations in relation to modifiable risk factors associated with sociodemographic factors, as significant clusters NCDs can lead to this condition and the prognosis of these diseases in the elderly population.

Methodology
This cross-sectional and population-based study is part of a macro project entitled Monitoring the health conditions of the elderly in a small municipality (MONIDI) and was approved by the Ethics Committee of Human Science (CAAE: 22969013.0.0000.0055) and it is in accordance with The Declaration of Helsinki. Informed consent was obtained from each participant. The study area was developed in 2014 in the city of Ibicuí. and it is in accordance with The Declaration of Helsinki.
Informed consent was obtained from each participant. The study area was developed in 2014 in the city of Ibicuí. This municipality is located in a rural area of Brazil, in the southwest of the state of Bahia, it has a territorial area of 1,139,378 km 2 , and an estimated population of 16,640 inhabitants, in the sampling period (IBGE, 2019).
To compose the sample, 335 individuals were selected, and the sample size was calculated considering a significance level of 5%, confidence interval of 95% and tolerable error of 3%. In addition, there was a 7.4% rate (n = 25) of sample loss and the response rate was 92.6%, with a final sample comprising 310 individuals, who were registered in the Family Health Strategy Data were collected through a questionnaire called the Health Assessment Instrument of the Elderly (IASI), previously validated instrument (Pedreira et al., 2016), applied in the form of an individual interview, followed by an anthropometric assessment.
Anthropometric measurements were taken (body mass and height) and evaluated by a previously trained professional.
The Body Mass Index (BMI) was evaluated with a portable digital scale (OMRON®) for weight and with a stadiometer (Sanny®) for height and calculated by dividing weight (Kg) by height squared (m 2 ). Overweight values were those equal to or greater than 25.0 kg / m (Gibson, 1990).

Statistical analysis
Statistical analysis of the data was performed using descriptive statistics procedures, association measures for categorical variables (Chi-square Test) and cluster analysis using Statistical Package for Social Sciences -SPSS for Windows® version 22.0. A 95% confidence interval and significance level of 5% were adopted for data analysis and interpretation. The clustering of behavioral risk factors was analyzed, stratifying by sex. When the prevalence of the combination of observed behaviors (O) was greater than the expected prevalence (E), it was considered a new grouping. To identify the expected prevalence values of each combination, the individual probability of occurrence of each behavior was multiplied by the occurrence observed in the study; when the ratio between observed/expected (O/E) was > 1, then a grouping was considered (Tassitano et al., 2014;Schuit et al., 2002;Galán et al., 2005). Multivariate analysis (logistic regression) was performed to evaluate the association between the accumulation of risk factors and the independent variables.

Results and Discussion
Chronic non-communicable diseases are common, but little is known about the characteristics of risk factors combined with demographic variables of elderly people with low economic status. To define the simultaneity characteristics of the five risk factors for NCDs in a rural city in Brazil, we analyzed and compared the sociodemographic variables of 310 low-income elderly.
When performing this analysis, we observed that the mean age was 71.62 (± 8.16) years and most were women (n = 175), who recognized themselves as non-white (71.8%), literate (56 .3%) and lived without a partner (51.1%) ( Table 2). When analyzing each risk factor, according to sociodemographic variables, the highest percentage for men and women was overweight / obesity (47.9% and 52.1%, respectively). In all risk factors, the prevalence was predominant in ages between 60-79 years. Nonwhite individuals had higher percentages in all behaviors compared. According to a study by Xiao Li at al. 2019 obesity was also a factor with a high percentile, showing that it is directly associated with non-communicable chronic diseases in the majority of women with a multimorbid lifestyle. Which can be associated with several genetic factors such as menopause, climacteric or as shown below with sedentary lifestyle.
When analyzing all behaviors, we found that there were thirty-two possible combinations in observed, expected frequencies and the O/E ratio, related to behavioral risk factors (Table 3)   Sedentary behavior is an important marker used in the present study because 43.8% of elderly Brazilians have sedentary behavior (Vigitel Brasil, 2019) and, in this population, this behavior is associated with poor cardiometabolic risk markers (Figueiró et al., 2019). The combinations of factors show that smoking and sedentary behavior are more prevalent in women, Research, Society andDevelopment, v. 10, n. 17, e18101724202, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i17.24202 7 contrary to the study by Cruz et al., 2017. That in men the prevalence of smoking and physical inactivity was higher, in this same study it is shown that smoking and overweight influence each other, which is slightly linked to the obesity data.
When we analyze the simultaneity of risk factors for NCDs in elderly people with low economic status, it can be seen that the prevalence is in men. if something is harmful to your health, other studies suggest that the onset of some diseases has caused greater physical inactivity, which is directly linked to obesity levels (Brandão et al., 2019;Cruz et al., 2017). Table 4 shows no prevalence (gross and adjusted) of three or more risk factors with socio-demographic characteristics. Note. OR = Odds ratio; CI = Confidence interval; * = Gross for all independent variables; ** = Adjusted by all independent variables. Source: Authors (2021).
Regarding clustering of simultaneous risk factors for chronic non-communicable diseases in the elderly, simultaneous excessive consumption of alcohol with smoking, and physical inactivity with smoking were the groups with the highest scores among men and women. When analyzing the association between these groups with sociodemographic characteristics, only the association between physical inactivity with smoking and sex was observed. Among the risk behaviors, the results showed that there most prevalent combination for men was alcohol, overweight/obesity, and smoking (O/E = 751.2), and for women it was physical inactivity and smoking (O/E = 192.6) ( Table 5). These are directly linked to lifestyles interfering with the work of muscles (Brandão et al., 2019). Smoking is a variable observed in both sexes of the study that has been presented in other clustering studies of elderly people. A Chinese study showed an association between smoking and alcohol, physical inactivity, and low consumption of vegetables and fruits for both sexes (Chou, 2008). A study conducted by Meader et al., 2016 showed an association between smoking and alcohol misuse or smoking and an unhealthy diet. A Poortinga's Study, 2007 showed clustering between heavy smoking, drinking, and lack of fruit/vegetables, these factors contribute to the development of chronic diseases such as cancer and diabetes.
In addition, a study conducted by Schuit et al., 2002 showed clustering of smoking with low vegetable and fruit consumption, excessive alcohol intake, and low physical activity. Smoking was also included in a regional study with elderly Brazilians, clustering with physical inactivity for men, and physical inactivity and alcohol for women (Cruz et al., 2017).
However, Cruz's study 2017 used just four risk factors (physical inactivity, smoking, alcohol consumption and overweight/obesity). Other data found in this study relates to the association between two behaviors of health risk with the variable sex. In this way, men are more physically inactive and smoke more than women (Chou, 2008).
Furthermore, it was found that elderly people had two or more NCDs risk factors. These results encourage public health policy to enable equity and facilitate healthier choices for all ages, especially because projections of demographic aging in Brazil are faster than world projections (United Nations, 2015). Studies that analyze the simultaneity of risk factors in the elderly are still scarce in the literature since most studies target adolescents and young adults as one of the study's strengths.
The present study was carried out in a city in a rural city in the largest state in Northeastern Brazil, but the sample was small, as not all elderly people living in the region were registered at the Family Health Unit in the city. As a limitation of this study, it can be highlighted that cross-sectional studies do not ensure direct causal relationships between the studied variables.
The use of self-reported information, despite being a strategy widely used in epidemiological studies, may be susceptible to information bias. However, this study has some strengths such as 1) developed a population-based survey; 2) showed the profile of elderly people that have a low socioeconomic status; 3) adopted screening strategies and data collection instruments recommended in the literature for studies with the elderly population; 4) analyzed clustering of five risk factors for chronic noncommunicable diseases, including sedentary behavior. The use of strategies to investigate aggregated risk factors may contribute to the definition of the burden of diseases and to the establishment of prevention programs in the field of public health.

Conclusion
The present study was based on data related to the macroproject Monitoring the Health Conditions of Elderly People (MONIDI) in a small municipality in the state of Bahia, Brazil (Nascimento, et al.;. According to this can be concluded that the elderly is exposed to health risk factors for concurrent NCDs. The most prevalent simultaneous behavior in both sexes was alcoholism, smoking, and overweight/obesity. However, the accumulation of these risk factors was not associated with sociodemographic variables. The simultaneous presence of a sedentary lifestyle and smoking was lower among women when compared to men. Another important factor was the association of tobacco and obesity, with overweight being more associated with a sedentary lifestyle and smoking with other factors. We suggest that future articles and research be carried out with more variables, as well as analyzing why the elderly population presents these levels according to gender, as some studies show that women are more prevalent in these same factors (smoking and excessive weight / obese) and others show that these are men with high percentiles.