Noncommunicable diseases and changes in lifestyles during the COVID-19 pandemic in Brazil

REV BRAS EPIDEMIOL 2021; 24: E210009 ABSTRACT: Objective: To compare lifestyle changes during the COVID-19 pandemic according to the presence or absence of noncommunicable diseases (NCDs) in Brazilian adults. Methods: Cross-sectional study, using data from the ConVid survey, between April and May 2020. The following variables were evaluated: lifestyle and presence of one or more NCDs (diabetes, hypertension, respiratory disease, heart disease, and cancer). Sociodemographic characteristics were used as adjustment. Relative frequencies and confidence intervals (CI) of 95% of the explanatory variables were calculated before and during the pandemic. For the comparison of groups, with or without NCDs, crude and adjusted (PRadj) prevalence ratios were estimated by Poisson regression. Results: There was a reduction in physical activity (60% in those without NCDs and 58% in those with NCDs) and in vegetable consumption (10.8% in those without NCDs and 12.7% in those with NCDs). On the other hand, there was an increase in the time spent watching television and on screens of computer/tablet (302% and 43.5% in those without NCDs and 196.5% and 30.6% with NCDs, respectively); consumption of frozen meals (43.6% in those without NCDs and 53.7% with NCDs), snacks (42.3% without NCDs and 31.2% with NCDs), and chocolate (14.8% without NCDs). During the pandemic, patients with NCDs were less active (PRadj = 0.77; 95%CI 0.65 – 0.92), had greater habit of watching TV (PRadj = 1.16; 95%CI 1.08 – 1.26), and consumed less vegetables (PRadj = 0.88; 95%CI 0.81 – 0.96). Conclusion: It was evident that adults with NCDs had their lifestyles more altered during the COVID-19 pandemic.


INTRODUCTION
Globally, it is estimated that noncommunicable diseases (NCDs) are responsible for 41 million deaths (70% of all deaths) per year 1 . In Brazil, the high burden of NCDs is similar, corresponding to 76% of the causes of death 2 . NCDs result in devastating consequences for individuals, families, and communities in addition to overloading healthcare systems 1,3,4 .
Several national and global initiatives have been proposed for the prevention and control of NCDs. The United Nations has included goals for the reduction of NCDs and risk factors by 2030 in the Sustainable Development Goals (SDGs) 5 . It should be noted that this commitment was already challenging and, considering the pandemic of the new coronavirus (Sars-CoV-2), such goal may not be achieved 6 , as the pandemic can affect lifestyles and compromise the health care of people with NCDs 7 .
NCDs are the main comorbidities of patients with COVID-19 7-9 , being responsible for the worsening of the clinical condition and for the increase in hospital stay and mortality rates 10,11 . Furthermore, social distancing measures have a potential impact on the health and quality of life of people with NCDs 6 , although they are essential to reduce the spread of the virus 12,13 .

Palavras
Within the context of NCDs, in view of the high burden and transcendence of chronic conditions, the scenario imposed by the COVID-19 pandemic reinforces the importance in prioritizing a comprehensive approach to the population's health, of disease surveillance and care actions during and after the pandemic 18 . The health crisis caused changes in the forms of care in the health system 19 , which culminated in lack of medication, difficulties in accessing services [19][20][21] and increase in risk factors for NCDs 14 .
Therefore, considering the need to monitor the health status and behaviors of patients with NCDs, especially in the context of the COVID-19 pandemic, it is relevant to investigate whether changes in lifestyle have occurred among this population, as there are scarce studies addressing this topic in the Brazilian context. Thus, this investigation aimed to compare lifestyle changes in Brazilian adults during the COVID-19 pandemic, according to the presence or absence of NCDs.

METHODS
This is a cross-sectional study, which used data from ConVid -Pesquisa de Comportamentos (Behavior Survey), a virtual health survey carried out during the COVID-19 pandemic to assess changes that have occurred in the lives of Brazilian adults 22 . The questionnaire 22 addressed sociodemographic characteristics, adherence to measures of social restriction, economic losses, chronic diseases, changes in lifestyles and mood, among other topics.
Data collection was carried out between April and May 2020. Participants were invited to participate in the research through a chain sampling process 23 . In the first stage, the involved researchers chose a total of 200 other researchers from different states in Brazil and, additionally, each one of them selected 20 people from their respective social networks. People chosen in the first stage invited at least 12 people from their own social networks, following the stratification by sex, age group (18-39; 40-59; ≥ 60 years) and education level (some high school or less; high school or more). Each guest was asked to invite at least three other people from their social networks. With this process, the final sample consisted of 45,160 individuals. Post-stratification weights were used based on data from the National Household Sample Survey (PNAD) -2019 24,25 to obtain the same distribution of the Brazilian population per Federation Unit, sex, age group, ethnicity/skin color, and education level.
The study is a partnership between researchers from Fundação Oswaldo Cruz, Universidade Federal de Minas Gerais, and Universidade Estadual de Campinas. Participants filled out the informed consent form and all responses were anonymous. The research was approved by the National Research Ethics Committee (opinion number: 3.980.277).
In this study, the variable "having one or more NCDs" was assessed by the question: "Has any doctor ever diagnosed you with any of these diseases (diabetes, hypertension, respiratory disease, heart disease, cancer)?" REV BRAS EPIDEMIOL 2021; 24: E210009 The following variables, related to lifestyle before and during the pandemic, were analyzed. • Physical activity (PA), assessed by the following questions: "Before the pandemic, how many days a week did you do any type of physical exercise or sport?"; "During the pandemic, how many days a week have you been doing/did you do physical exercise or sport?" The answers to these questions were: a) less than one day a week; b) one to two days; c) three to four days; and d) five or more days. "Before the pandemic, how long did this activity last?"; "During the pandemic, how long has this activity been lasting/did this activity last?" The response options were: a) less than 30 minutes; b) 30 to 45 minutes; c) 46 to 60 minutes; and d) an hour or more. Individuals who reported at least 150 minutes of PA per week were considered sufficiently active 26 . • Sedentary behavior was assessed by time spent watching television and on screens of computer/tablet. As for the time spent watching television, the questions were: "Before the pandemic, how many hours a day did you watch television?"; "During the pandemic, how many hours a day have you been watching/did you watch television?" The answer options were: a) did not watch television; b) less than 1 hour; c) between 1 hour and less than 2 hours; d) between 2 hours and less than 3 hours; e) between 3 hours and less than 4 hours; f ) between 4 hours and less than 5 hours; g) between 5 hours and less than 6 hours; and h) 6 hours or more. For computer/tablet use, the questions were: "Before the pandemic, how many hours a day did you use to use a computer or tablet?"; "During the pandemic, how many hours a day have you been using your computer or tablet?" This was an open-ended question, and the respondent indicated the number in hours. Both indicators were classified using the cut-off point of 4 hours/day 27 . • Consumption of healthy (greens or vegetables, fruits and beans) and unhealthy foods (frozen pizza or frozen lasagna or other frozen meals; packaged snacks; processed meats chocolates and sweets) were assessed by indicating the consumption of these foods according to the following question: "Usually, before the pandemic, how many days a week did you use to eat these foods?" and "During the pandemic, how many days a week have you been eating/did you use to eat these foods?" The response options were: a) five days or more; b) two to four days; c) one day or less. For healthy foods, consumption on five or more days a week was considered adequate 28 . For unhealthy foods, the frequency of two or more days a week was considered inadequate, as the consumption of these foods is not indiscriminately recommended 29 ; • Increased consumption of alcoholic beverages was analyzed by the question: "During the pandemic, how have you been consuming/did you consume alcoholic beverages?" The response options were: a) I continued drinking with the same frequency as before; b) I am drinking more; c) I am drinking less; and d) I had stopped drinking, but I started again. The alternatives "b" and "d" were considered as an increase in the consumption of alcoholic beverages, and "a" and "c" were categorized as "no increase." • Increase in cigarette consumption was assessed by the questions: "Are you a smoker?" and, if so: "Before the pandemic, on average, how many cigarettes did you use to REV BRAS EPIDEMIOL 2021; 24: E210009 smoke per day?" and "During the pandemic, on average, how many cigarettes have you been smoking/did you smoke per day?" The answer options were: a) I did not smoke cigarettes; b) less than one per day; c) one to nine cigarettes; d) 10 to 19 cigarettes; e) 20 to 29 cigarettes; f ) 30 to 39 cigarettes; g) 40 or more cigarettes.
To assess the number of cigarettes, the midpoint of the response category was used and the difference between the two moments was calculated. The positive difference was considered an increase in cigarette consumption.
Initially, a descriptive analysis of the lifestyle variables was performed before and during the pandemic by calculating relative frequencies and 95% confidence intervals (95%CI), followed by the calculation of the percentage of behavioral change during the pandemic. Statistically significant differences were identified by the non-overlapping of the 95%CI of the prevalence values in question.
The two population groups (those with one or more NCDs and those without NCDs) were also compared in relation to lifestyles during the pandemic by calculating the prevalence and the crude and adjusted prevalence ratio (PR) by sex (men and women), age group (18 to 39; 40 to 59; ≥ 60 years), and education level (elementary school or less; high school; higher education or more), and respective 95%CI.
The Poisson regression with robust variance was applied. The Software for Statistics and Data Science (Stata) version 14 was used to process the information, using the survey module, which considers the post-stratification weights.
The prevalence of hypertension, diabetes, respiratory disease, heart disease, and cancer was 19.3% (95%CI 18.2 -20.4); 7.2% (95%CI 6.5 -8); 11.3% (95%CI 10.4 -12.2); 4.3% (95%CI 3.7 -5); and 2.4 (95%CI 2.1 -2.9), respectively ( Figure 1A). The prevalence of one or more NCDs was 33.9% (95%CI 32.5 -35.3), being higher with increasing age (38.3%; 95%CI 36 -40.6 between 40 and 59 years; and 58.6%; 95%CI 55.3 -61.9 for those aged 60 years and older) and among those with lower education level (39.3%; 95%CI 34.4 -44.5), with no differences according to sex ( Figure 1B). Table 1 shows the prevalence and the percentage (%) of changes in the PA and food consumption indicators before and during the pandemic according to the presence of NCDs. There was a 60% reduction in PA in groups without NCDs, and 58% in those with NCDs. The increase in time spent watching television and using computer/tablet was 302 and 43.5%, respectively, in the group without NCDs; and 196.5 and 30.6% in those with NCDs.  Regarding the consumption of healthy foods, in the population without NCDs, the consumption of vegetables decreased by 10.8% and that of fruits by 3.5%; and, in those with NCDs, it decreased by 12.7% and increased by 37.4%, respectively. It is noteworthy that the reduction in fruit consumption was not significant in both groups. Regarding the consumption of unhealthy foods, in the population without NCDs, the consumption of frozen meals increased 43.6%; of snacks, 42.3%; of chocolate, 14.8%; and of processed meats, 0.72%. There were statistically significant differences in all foods, except processed meats. For people with NCDs, the increase in the consumption of frozen meals, snacks, and chocolates was 53.7, 31.2, and 12.5%, and the reduction was 6% for processed meats; however, the change was not significant for consumption of processed meats and chocolates (Table 1). Table 2 shows the prevalence and crude and adjusted (PRadj) prevalence ratio of PA, sedentary behaviors, and eating habits during the pandemic, among the population with and without one or more NCDs. It was verified that, during the pandemic, people with one or more NCDs were less active (PRadj = 0.77; 95%CI 0.65 -0.92) and had greater habit of watching television for 4 hours or more/day (PRadj = 1.16; 95%CI 1.08 -1.26). The time spent using the computer/tablet did not differ between groups (PRadj = 0.96; 95%CI 0.90 -1).
As for the consumption of healthy and unhealthy foods, it was observed that, during the pandemic, people with NCDs reported inadequate intake of vegetables (PRadj = 0.88; 95%CI 0.81 -0.96). In addition, people with NCDs had a lower prevalence for the consumption of unhealthy foods, but there were no significant differences after adjusting for age, education level, and sex (Table 2).
In Table 2, when analyzing the increase in consumption of alcoholic beverages and cigarettes during the pandemic according to the diagnosis of NCDs, it is evident that people with at least one NCD had a lower increase in the consumption of alcoholic beverages (15.1%; 95%CI 13.3 -17.2) compared with those without NCDs (18.9%; 95%CI 17.3 -20.6). Nevertheless, when the analyses were adjusted, there was no difference between groups. The increase in cigarette consumption was 34.8% (95%CI 28.9 -41.2) among people with NCDs and 33.7% (95%CI 29.1 -38.6) among those who did not report NCDs, but without statistically significant differences between groups (Table 2).

DISCUSSION
The study indicated there were changes in the lifestyles of the population with and without NCDs after the pandemic such as reduced PA and increased sedentary time, reduced consumption of vegetables and increased consumption of ultra-processed foods -such as frozen meals and snacks. When analyzing the prevalence of lifestyles during the pandemic between the two groups, it was observed that people with NCDs were less active, had inadequate consumption of vegetables, and spent more sedentary time watching television. Regarding the consumption of tobacco, alcohol, fruits, and unhealthy foods, there was no difference between groups.
The National Survey of Health (PNS) 30 found similar self-reported prevalence of arterial hypertension (21.4%), diabetes (6.2%), asthma (4.4%), heart disease (4.8%), and cancer (1.8%). The prevalence of one or more NCDs was higher (45.1%); however, the list of NCDs evaluated was more extensive, containing thirteen. In addition, higher prevalence of NCDs was observed with increasing age and low education level 30 . Chronic conditions are known to affect social groups with greater deprivation and more vulnerable to social determinants 31 . Consequently, including NCDs in a scenario of social and economic inequalities individually exacerbates the adverse effects of each disease, as COVID-19, being considered a syndemic, has systemic effects and increases inequities 32 . Furthermore, when addressing COVID-19, attention should also be given to NCDs and their risk factors such as hypertension, obesity, diabetes, chronic cardiovascular and respiratory diseases, and cancer 32 .

REV BRAS EPIDEMIOL 2021; 24: E210009
The worsening of lifestyles during the pandemic observed in this study was also identified in other surveys 20,21,33,34 . Such changes may impact the quality of life and health of the population, resulting in an epidemic of preventable deaths among people affected with NCDs 19 . Hence, it is essential to implement a health promotion agenda 35 and to strengthen primary health care in facing the impacts of the pandemic and in social distancing 36 , with emphasis on maintaining comprehensive follow-up actions for NCD patients, using telemonitoring and telehealth 37 .
This study identified that people with one or more NCDs had irregular practice of PA during the pandemic. It is noteworthy that PA is essential for the maintenance and promotion of health and quality of life 38,39 , and the recommendation is to maintain an active life 40 . The World Health Organization 41 encourages health professionals and managers to seek strategies to maintain active and healthy habits during the period of social distancing, especially for population groups considered at risk and, therefore, more vulnerable such as older adults and the population affected with NCDs 9,40,42 .
The results indicate that the population with NCDs spent more time watching television during the pandemic. This fact is worrying, considering that a sedentary behavior has a detrimental effect on health, which can contribute to the onset and worsening of NCDs in addition to rapidly deteriorating cardiovascular health and causing premature deaths 43,44 .
It has been discussed that social distancing implies changes in eating habits of the population in general, with reduced consumption of fruits and vegetables and increased consumption of frozen meals, snacks, and sweets. In Italy, 34.4% of individuals had their diet worsened during the pandemic 34 . In the United States of America, there was an increase in the volume of purchases in supermarkets, with a stock of ultra-processed foods with high energy density such as chips, popcorn, chocolate, and ice cream 45,46 .
People with NCDs showed lower consumption of vegetables and similar consumption of ultra-processed foods. These findings can be explained by the difficulty in supplying and replacing fresh and healthy food as well as by the ease of purchase and storage of ultra--processed products 47 . A study carried out among diabetic patients in Brazil identified an important adherence to social distancing due to the perceived increased risk for COVID-19; however, this population reduced the use of healthcare services and there was a report of difficulty in purchasing foods 20 .
The increase in alcohol consumption during the COVID-19 pandemic, identified in this study, has also occurred in other contexts and has been attributed to psychological distress and financial difficulties 33 , in addition to greater physical availability and accessibility 33,48 . Although there were no differences between the population with and without NCDs after adjustment, the risk of this finding is emphasized, as alcohol can potentially worsen the clinical course of COVID-19 infection 49 , increasing the susceptibility to viral infection and aggravating the situation of people with NCDs 50 .
Tobacco consumption increased in more than a third of smokers, with or without NCDs. Tobacco is an important risk factor for NCDs, both for cancer and respiratory diseases, being contraindicated in any situation 51  course of COVID-19 52,53 . Among factors involved in the relation between smoking and this viral infection are the increase in angiotensin-converting enzyme 2 (ECA2) among smokers 52,53 , the increase in viral contamination by taking the cigarette to the mouth 54 , and the worsening in cardiorespiratory performance 55 .
In South Africa, in order to tackle smoking during the COVID-19 pandemic, the sale of tobacco and nicotine products has been banned, which should be understood as a double attempt to protect the population against COVID-19 and the avoidable burden of tobacco 56 . Therefore, advancing new regulatory measures 57 and monitoring the actions of tobacco industries 58 should be a priority in government agendas, especially during pandemic periods, in order to save lives 58 .
It should be noted that the study showed a worsening of lifestyles, with these changes being even more intense in the adult population with NCDs. Thus, a scenario of worsened indicators can be predicted, including increased mortality from NCDs, which may compromise the achievement of global goals such as reducing premature mortality by 30% by 2030 5 .
In this context, it is worth highlighting the contemporaneity of intersectoral policies for the prevention and control of NCDs in response to COVID-19. Prevention measures concerning NCDs should not be interrupted. Healthcare services, especially primary health care, must adapt to this new reality in order to support and manage the increase in risks and to continue providing health care for NCDs patients. Professionals in the field must be involved in planning strategies to respond to COVID-19, in such a way they can guarantee the right to health. Nevertheless, specific recommendations should be nationally and locally made available for patients with NCDs, their families, and their caregivers 32 . Furthermore, the COVID-19 pandemic and the NCD epidemic show the urgency in advancing stricter regulatory measures for tobacco, alcohol, and unhealthy food, promoting PA, and strengthening health systems during and after the pandemic 1 .
Among the limitations of the study, it is pointed out the fact that data were collected through the Web, which may not reach all population strata, considering that not everyone has access to this means of communication, which may lead to under-or overestimation of proportion of indicators. This limitation was minimized by sample calibration by applying post-stratification weights. Another issue refers to the fact that the questions are self-reported, and may be subject to memory bias. The fact that data collection took place at the beginning of the pandemic (3 rd month) may be subject to situational changes throughout the temporal evolution of the pandemic.
All in all, our findings suggest that adults with NCDs had their lifestyles more altered due to reduced PA, increased sedentary lifestyle, reduced consumption of healthy foods, and increased consumption of ultra-processed foods. Within this context, it is imperative to adopt public policies that foment health promotion actions as well as regulatory measures for the protection and prevention of NCDs and the expansion of primary health care for patients with NCDs. If urgent measures are not taken, there is a risk of not achieving the goals of Agenda 2030 -SDG concerning the reduction of NCDs.