Is self-rated health associated with cardiovascular risk factors and disease in a low-income setting? A cross-sectional study from the Amazon Basin of Brazil.

OBJECTIVE: Prior studies have suggested that self-rated health may be a useful indicator of cardiovascular disease. Consequently, we aimed to assess the relationship between self-rated health, cardiovascular risk factors and subclinical cardiac disease in the Amazon Basin.
DESIGN: Cross-sectional study. SETTING, PARTICIPANTS AND
INTERVENTIONS: In participants from the Amazon Basin of Brazil we obtained self-rated health according to a Visual Analogue Scale, ranging from 0 (poor) to 100 (excellent). We performed questionnaires, physical examination and echocardiography. Logistic and linear regression models were applied to assess self-rated health, cardiac risk factors and cardiac disease by echocardiography. Multivariable models were mutually adjusted for other cardiovascular risk factors, clinical and socioeconomic data, and known cardiac disease. OUTCOME MEASURES: Cardiovascular risk factors and subclincial cardiac disease by echocardiography.
RESULTS: A total of 574 participants (mean age 41 years, 61% female) provided information on self-rated health (mean 75±21 (IQR 60-90) points). Self-rated health (per 10-point increase) was negatively associated with hypertension (OR 0.87 (95% CI 0.78 to 0.97), p=0.01), hypercholesterolaemia (OR 0.89 (95%CI 0.80 to 0.99), p=0.04) and positively with healthy diet (OR 1.13 (95%CI 1.04 to 1.24), p=0.004). Sex modified these associations (p-interaction <0.05) such that higher self-rated health was associated with healthy diet and physical activity in men, and lower odds of hypertension and hypercholesterolaemia in women. No relationship was found with left ventricular ejection fraction <45% (OR 0.97 (95% CI 0.77 to 1.23), p=0.8), left ventricular hypertrophy (OR 0.97 (95% CI 0.76 to 1.24), p=0.81) or diastolic dysfunction (OR 1.09 (95% CI 0.85 to 1.40), p=0.51).
CONCLUSION: Self-rated health was positively associated with health parameters in the Amazon Basin, but not with subclinical cardiac disease by echocardiography. Our findings are of hypothesis generating nature and future studies should aim to determine whether assessment of self-rated health may be useful for screening related to policy-making or lifestyle interventions. TRIAL REGISTRATION NUMBER: Clinicaltrials.gov: NCT04445103; Post-results. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

This is the first study to examine self-rated health in a rural part of the Amazon Basin of Brazil using an internationally recognised questionnaire, EQ-5D-5L.

We applied a state-of-the-art echocardiographic imaging protocol to identify underlying cardiovascular disease.

Self-reported health behaviour could be subject to social and cultural biases.

Because no standard values of the EQ-5D-5L health instrument have been published for Brazil, it is not possible to compare our findings with other populations.

The study design was cross-sectional.

Introduction

Cardiovascular disease is the leading cause of mortality worldwide and accounts for more than 31% of all deaths and 8% of public hospitalisations in Brazil.1 2 Since the 1960s, Brazil has experienced a transition in health behaviour and cardiovascular risk factors, where tobacco consumption has declined and obesity has increased.2 Approximately 35% of Brazilian adults suffer from hypertension,3 the prevalence of diabetes mellitus is rising4 and a high proportion of adults do not practise recommended levels of physical activity.2 Differences in perception of risk factors and variability in access to healthcare unequivocally affect health behaviour and the lifetime risk of cardiovascular disease. In this regard, self-rated health is widely used as a health indicator in various populations,5 is strongly associated with cardiovascular morbidity6 7 and provides prognostic information on mortality.8 Self-rated health and cardiovascular risk factors are also both influenced by sex.9 10 Throughout the last decades, assessment of self-rated health has become increasingly important and is often used for healthcare surveillance and in policy-making.

To understand whether self-rated health in future studies may be used to screen for cardiac disease in low-income settings, we aimed to investigate the relationship with cardiovascular risk factors and disease in the general population from the Amazon Basin of Brazil. We hypothesised that higher self-rated health was associated with less cardiovascular risk factors and disease, and that these relationships were modified by sex.11

Methods

Study site

The study was conducted in the municipality of Cruzeiro do Sul, Acre (Northern Brazil; Amazon Basin). The prevalence for cardiovascular disease in Acre (5815 per 100 000 inhabitants) is below the average rate for Brazil (6025 per 100 000 inhabitants) and the region is considered to be one of the poorest in Brazil and has one of the lowest population densities.12 13

Patient and public involvement

Patients or the public were not involved in the study design, recruitment to and conduct of the study nor reporting of results. All patients were informed of the results from their own examinations conducted in the study. Data will be made available on reasonable request to the corresponding author.

Sampling

This cross-sectional study was conducted as a part of the Malaria Heart Study (ClinicalTrials.gov: NCT04445103). Participants from the general population were enrolled from June 2020 to December 2020. Through randomisation, we selected 10 local healthcare clinics from Cruzeiro do Sul, equally distributed between rural and urban areas. Local healthcare agents provided lists of persons associated with each clinic, who we invited to participate in the study (figure 1). We included persons ≥18 years old who completed the examination programme and responded to all questionnaires. Exclusion criteria were ongoing pregnancy, ongoing infection as assessed by examination of a medical doctor and presence of Plasmodium in peripheral blood smears. A total of 504 participants from the general population were included from healthcare clinics. As a part of the main study, we also examined patients diagnosed with uncomplicated malaria in healthcare clinics. This group of participants underwent a follow-up examination a median of 30 days later, when they had completed treatment and had no symptoms of malaria. According to the above-mentioned inclusion and exclusion criteria, a total of 70 participants from this group were eligible for inclusion (figure 1).

Figure 1

Flow chart. Inclusion of the study population in Cruzeiro do Sul, Acre.

Data collection

Two different questionnaires were administered by trained interviewers (ie, study personnel). These interviewers also filled out the questionnaires. The first was the EuroQol-5 Domain-5 Level (EQ-5D-5L) questionnaire, which is validated in Brazilian Portuguese (study registration no.: 28276).14 For the purpose of this study, we used data from theEuroQol Visual Analogue Scale, which provides a single estimate of self-rated health ranging from 0 to 100 points on a continuous scale. Zero represents the worst possible self-rated health and 100 represents ideal health. The second questionnaire was used to gather information about socioeconomic status, race, cardiovascular risk factors, known cardiac disease (prior myocardial infarction and heart failure) and current medications. Race was self-reported, and two persons did not answer this question. Afterwards, participants underwent a physical examination to measure height, weight and blood pressure. Fingerstick point-of-care blood draws were used to measure glucose levels and to obtain thick and thin blood slides. Giemsa stained thick and thin blood slides were analysed by two independent microscopists to detect Plasmodium. A medical doctor (PB) evaluated all patients. None of them displayed clinical signs or symptoms of heart disease (absence of shortness of breath, chest pain, swelling of legs and irregular heart rhythm). All data were quality controlled by AEH and PB on a daily basis.

Cardiovascular risk factors

We assessed seven different cardiovascular risk factors. Hypertension was defined as a physician diagnosis of hypertension or intake of antihypertensive medication, hypercholesterolaemia as a physician diagnosis of dyslipidaemia or intake of lipid-lowering medication, and diabetes as a physician diagnosis of diabetes or fasting blood glucose >126 mg/dL.15 Body mass index (BMI) was calculated as: body weight (kilograms)/height2 (metres), and obesity was defined as BMI ≥30 kg/m2. Participants were classified as smokers if they were current smokers or had previously smoked. A healthy diet was defined as intake of any quantity of vegetables with a main meal ≥3 times/week. Physical activity was defined as participation in any kind of physical activity, on a weekly basis, during leisure time. We did not apply any time limit or threshold.

Biochemistry

Field procedures

During examinations, we collected peripheral venous blood samples in citrate, EDTA and serum-separator tubes, which were cooled at 2°C–8°C. Citrate plasma was immediately separated by centrifugation (12 min, 3200 rpm) in a mobile laboratory and transferred to Eppendorf tubes.

Laboratory

Serum-separator tubes underwent centrifugation (10 min, 3000 rpm) to extract serum which was subsequently stored at −20°C in Eppendorf tubes. Laboratory analyses were performed at Citolab and Centro de Diagnósticos, Cruzeiro do Sul, Acre, Brazil. Using EDTA blood, a complete blood count with a differential was conducted (NX-350, Sysmex, Japan; Citolab), and reticulocytes were counted manually (Citolab).16 Citrate plasma was used to analyse coagulation parameters (Coagmaster 2.0, Wama Diagnóstica, Brazil; Citolab). Serum was used to measure creatinine, bilirubin and C reactive protein (Cobas c111, Roche Diagnostics, Switzerland; Citolab and Centro de Diagnósticos). Analyses of C reactive protein were only available in a subset of participants (n=436).

Echocardiography

A single medical doctor either performed or supervised all echocardiographic examinations (PB). Quality control was conducted on a frequent basis in a central imaging laboratory (Herlev-Gentofte Hospital, Denmark) by an investigator certified in echocardiography by the European Association of Cardiovascular Imaging. Examinations were performed bedside (Vivid-IQ, GE Healthcare, Norway), and images were stored offline for analysis in EchoPAC BT13 (V.203.82). Analyses were conducted by AW according to contemporary guidelines.17 Rheumatic heart disease was assessed by PB according to the World Heart Federation criteria.18 We assessed three categories of subclinical left ventricular (LV) cardiac disease, namely: (1) reduced contractile function defined as LV ejection fraction <45%, (2) LV hypertrophy defined as LV mass index >115 g/m2 for men and>95 g/m2 for women and (3) diastolic dysfunction determined according to existing guidelines.19 Classification of diastolic dysfunction involved assessment of early and late mitral inflow velocity, mitral annular early diastolic velocity, tricuspid regurgitation velocity and the left atrial volume index. Additional details are described in the online supplemental data methods.

Statistics

Baseline characteristics for the study population were stratified according to tertiles of self-rated health (cut-offs of 70 and 91 points) and sex. Due to the nature of the distribution, tertiles of self-rated health did not contain equal amounts of participants. P for trend was calculated using linear regression models and the Cuzick non-parametric test for trend.20 Differences between groups were compared using the χ2 test, Student’s t-test and the Wilcoxon rank-sum test, as appropriate. Histograms were conducted to display the distribution of self-rated health. In all statistical tests, self-rated health was treated as a continuous variable. Logistic regression models were conducted to examine the relationship between self-rated health and cardiovascular risk factors and disease. Multivariable models were adjusted for core variables: clinical data (age, sex and race), socioeconomic data (work, family income and living area), known cardiac disease (prior myocardial infarction, heart failure and rheumatic heart disease). Included variables were selected based on prior studies of self-rated health21–23 and were defined prior to commencing data analyses. In addition, all associations with cardiovascular risk factors were mutually adjusted for all other risk factors. Interactions with sex were also examined. Family income was log-transformed to provide a normal distribution. The relationship between self-rated health and (1) the sum of cardiac risk factors (hypertension, hypercholesterolaemia, diabetes, obesity and smoking) and (2) echocardiographic parameters were assessed in linear regression models, which were adjusted for the core variables. As this was a secondary study, no sample size calculation was conducted. All analyses were conducted in Stata V.14.2 (StataCorp) and RStudio V.1.3 (R, Vienna, Austria). Two-sided p values <0.05 were considered statistically significant.

Results

A total of 574 participants were assessed (mean age 41±15 years, 61% female). Mean self-rated health was 75±21 points (IQR 60–90 points) (figure 2A). Four participants (<1%) reported 0 points and 91 participants (16%) reported 100 points. The prevalences of cardiovascular risk factors were 20% for hypertension, 16% for hypercholesterolaemia, 6% for diabetes, 23% for obesity, 38% for current or prior smoking, 52% for unhealthy diet and 63% for absence of physical activity. Participants with lower self-rated health more frequently had all of the above risk factors and were older compared with participants with high self-rated health (p-trend <0.05; table 1). No differences were observed in socioeconomic characteristics, biochemistry or subclinical cardiac disease by echocardiography (reduced LV ejection fraction, hypertrophy, diastolic dysfunction) across tertiles of self-rated health (table 1).

Figure 2

Histograms of self-rated health. Distribution of self-rated health in the (A) entire study population (n=574), (B) in men (n=224) and (C) in women (n=350).

Table 1

Baseline clinical characteristics by tertiles of self-rated health

	Tertiles of self-rated health
	1st tertile (n=231)	2nd tertile (n=226)	3rd tertile (n=117)	P trend*
	0 to 70	71 to 90	91 to 100
Baseline
Age, years	46±16	38±13	39±15	<0.001
Female, %	154 (67%)	127 (56)	69 (59)	0.06
Self-reported race, %				0.51
White	33 (14)	24 (11)	20 (17)
Mixed	163 (71)	175 (77)	77 (66)
Black	32 (14)	26 (12)	18 (15)
Indigenous	2 (1)	1 (<1)	1 (1)
BMI, kg/m2	28±6	27±5	26±4	0.002
Abdominal circumference, cm	90±14	87±12	84±11	<0.001
Asthma, %	11 (5)	8 (4)	2 (2)	0.36
COPD, %	3 (1)	3 (1)	1 (1)	0.92
History of MI, %	2 (1)	2 (1)	1 (1)	1
Heart failure, %	3 (1)	2 (1)	0 (0)	0.47
Rheumatic heart disease, %	7 (3)	7 (3)	4 (3)	0.97
SBP, mmHg	134±20	131±20	131±19	0.29
DBP, mmHg	83±12	81±11	82±12	0.17
Risk factors
Hypertension, %	66 (29)	32 (14)	14 (12)	<0.001
Hypercholesterolaemia, %	52 (23)	23 (10)	14 (12)	<0.001
Diabetes, %	21 (9)	6 (3)	6 (5)	0.012
Obesity, %	68 (29)	45 (20)	20 (17)	0.012
Smoking, %	106 (46)	65 (29)	46 (39)	<0.001
Healthy diet, %	87 (38)	130 (58)	59 (50)	<0.001
Physical activity, %	64 (28)	94 (42)	53 (45)	<0.001
Socioeconomic status
Work status, %				0.09
Employed	77 (33)	98 (43)	53 (45)
Self-employed	20 (9)	23 (10)	9 (8)
Other	134 (58)	105 (47)	55 (47)
Family income, BRL	1250(800, 2000)	1500(1000, 3000)	1200(800, 2000)	0.11
Rural living area, %	92 (40)	78 (35)	55 (47)	0.08
Biochemistry
Blood sugar, mg/dL	110±74	100±27	110±49	0.1
Bilirubin, mg/dL	0.3(0.2, 0.5)	0.4(0.2, 0.5)	0.4(0.2, 0.5)	0.55
Platelets, x 109/L	229±76	240±67	234±66	0.28
Leucocytes, x 109/L	6.35±1.99	6.38±1.72	6.53±1.92	0.68
Reticulocytes, %	0.75±0.19	0.80±0.22	0.77±0.22	0.44
Haemoglobin, g/L	140±14	142±14	142±12	0.13
C reactive protein, mg/L	0(0, 0)	0(0, 0)	0(0, 0)	0.44
Creatinine, mg/dL	0.9±0.3	0.9±0.2	0.9±0.2	0.59
INR	1.02±0.12	1.01±0.10	1.02±0.10	0.3
Echocardiography
LV ejection fraction <45%, %	9 (4)	6 (3)	3 (3)	0.69
LV hypertrophy, %	9 (4)	4 (2)	4 (3)	0.39
Diastolic dysfunction, %	7 (3)	5 (2)	1 (1)	0.43
LV ejection fraction, %	57±6	57±5	58±5	0.48
LV mass index, g/m2	71±18	68±17	70±16	0.11
E/e’	7.3±2.6	6.7±2.1	6.9±2.3	0.014
E/A	1.2±0.5	1.3±0.4	1.3±0.4	0.003
Left atrial volume index, mL/m2	20±6	19±5	19±4	0.025
TR velocity, m/s	2.3±0.3	2.3±0.3	2.3±0.2	0.34

Normally distributed variables are displayed as mean±SD. Non-normally distributed variables are presented as median (IQR). Proportions are displayed as n (%).

*P for trend was calculated using linear regression models for normally distributed variables and Cuzick’s non-parametric test for trend for non-normally distributed variables.

A, late mitral inflow velocity; BMI, body mass index; COPD, chronic obstructive pulmonary disease; DBP, diastolic blood pressure; e', mitral annular early diastolic velocity; E, early mitral inflow velocity; INR, international normalised ratio; LV, left ventricular; MI, myocardial infarction; SBP, systolic blood pressure; TR, tricuspid regurgitation.

In unadjusted logistic regression models, better self-rated health was significantly associated with lower odds of all cardiovascular risk factors (p<0.05 for all; table 2). In adjusted models, self-rated health (per 10-point increase) was associated with lower odds of hypertension (OR 0.87 (95% CI 0.78 to 0.97), p=0.01), hypercholesterolaemia (OR 0.89 (95% CI 0.80 to 0.99), p=0.04) and higher odds of healthy diet (OR 1.13 (95% CI 1.04 to 1.24), p=0.004). In multivariable models, better self-rated health was also associated with the sum of cardiovascular risk factors (beta=−0.07 per 10-point increase (95% CI −0.10 to −0.03), p<0.001). The associations remained unchanged when we excluded participants recently treated for malaria (online supplemental table 1).

Table 2

Association between self-rated health (per 10 increase), cardiovascular risk factors and disease in the entire population (n=574)

	Unadjusted OR (95% CI)	P value	Adjusted OR(95% CI)*	P value	P interaction sex
Risk factors
Hypertension	0.77 (0.71 to 0.85)	<0.001	0.87 (0.78 to 0.97)	0.011	0.005
Hypercholesterolaemia	0.83 (0.75 to 0.91)	<0.001	0.89 (0.80 to 0.99)	0.044	0.29
Diabetes	0.84 (0.73 to 0.97)	0.021	1.02 (0.86 to 1.22)	0.80	0.17
Obesity	0.90 (0.82 to 0.98)	0.017	0.95 (0.86 to 1.05)	0.30	0.78
Smoking	0.86 (0.79 to 0.93)	<0.001	0.96 (0.87 to 1.05)	0.39	0.003
Heathy diet	1.11 (1.03 to 1.20)	0.008	1.13 (1.04 to 1.24)	0.004	0.002
Physical activity	1.16 (1.06 to 1.26)	0.001	1.09 (0.99 to 1.20)	0.079	<0.001
Subclinical cardiac disease
LV ejection fraction <45%	0.88 (0.73 to 1.08)	0.22	0.97 (0.77 to 1.23)	0.82	0.88
LV hypertrophy	0.87 (0.72 to 1.07)	0.19	0.97 (0.76 to 1.24)	0.81	0.31
Diastolic dysfunction	0.92 (0.75 to 1.15)	0.47	1.09 (0.85 to 1.40)	0.51	0.63

*Multivariable models were mutually adjusted for cardiovascular risk factors (hypertension, hypercholesterolaemia, diabetes, obesity, smoking, healthy diet, physical activity) in addition to age, sex, work, family income, living area (rural/urban) and prior heart disease.

LV, left ventricular.

Subclinical cardiac disease by echocardiography

No significant associations were found between self-rated health (per 10-point increase) and subclinical cardiac disease by echocardiography: LV ejection fraction <45% (OR 0.88 (95% CI 0.73 to 1.08), p=0.22), LV hypertrophy (OR 0.87 (95% CI 0.72 to 1.07), p=0.19) or diastolic dysfunction (OR 0.92 (95% CI 0.75 to 1.15), p=0.47) (table 2). No individual echocardiographic parameters (assessed continuously) were significantly associated with self-rated health in multivariable models (p>0.05 for all; table 3).

Table 3

Self-rated health (per 10 point increase) and echocardiographic parameters in the entire population (n=574)

	Unadjusted beta (95% CI)	P value	Adjusted beta (95% CI)*	P value
Echocardiography
Left ventricular ejection fraction	0.04 (−0.16 to 0.25)	0.67	0.04 (−0.17 to 0.25)	0.71
Left ventricular mass index	−0.46 (−1.12 to 0.21)	0.18	0.12 (−0.46 to 0.70)	0.69
e’	0.40 (0.25 to 0.54)	<0.001	0.06 (−0.04 to 0.15)	0.23
E/e’	−0.16 (−0.25 to −0.07)	0.001	0.01 (−0.07 to 0.09)	0.76
E/A	0.03 (0.02 to 0.05)	<0.001	0.01 (−0.01 to 0.01)	0.99
Left atrial volume index	−0.26 (−0.46 to −0.06)	0.012	−0.05 (−0.22 to 0.13)	0.61
Tricuspid regurgitation velocity	−0.01(−0.02 to −0.01)	0.21	0.01 (−0.01 to 0.1)	0.95

*Multivariable models were adjusted for age, sex, work, family income, living area (rural/urban) and prior heart disease.

A, late mitral inflow velocity; e’, mitral annular early diastolic velocity; E, early mitral inflow velocity.

Interactions with sex

Self-rated health was higher in men than in women (77 vs 73 points) but the difference was not statistically significant (p=0.09) (figure 2B–C). In general, women had higher BMI, lower income, less frequently smoked and were less physically active compared with men (p<0.05 for all; online supplemental table 2). Sex modified the associations with hypertension, smoking, healthy diet and physical activity, but not subclinical cardiac disease by echocardiography (table 2). Unadjusted associations with cardiovascular risk factors, stratified by sex, are presented in figure 3. For men, higher self-rated health (per 10-point increase) yielded greater odds of a healthy diet (adjusted OR 1.33 (95% CI 1.12 to 1.59), p=0.002) and physical activity (adjusted OR 1.24 (95% CI 1.03 to 1.50), p=0.02). For women, higher self-rated health (per 10-point increase) was associated with lower odds of hypertension (adjusted OR 0.85 (95% CI 0.74 to 0.97), p=0.016) and hypercholesterolaemia (adjusted OR 0.87 (95% CI 0.76 to 0.99), p=0.046). The associations remained unchanged when we excluded participants from the malaria group (online supplemental table 2 and figure 1).

Figure 3

Forest plot. Association between self-rated health (per 10-point increase) and cardiovascular risk factors stratified by sex. *Indicates that the association persisted to be significant in multivariable models.

Discussion

This study has two principal findings. First, in a sample of the general population from the Amazon Basin, we found that self-rated health was significantly associated with cardiovascular risk factors and that these associations were modified by sex. Second, self-rated health was not associated with cardiac disease assessed by echocardiography. These findings indicate that in a low-income setting, self-rated health may to some extent provide information on cardiac risk profiles.

Self-rated health has previously been related to cardiovascular disease in various observational studies.24–26 Higher self-rated health is related to a lower burden of cardiovascular risk factors (hypertension, hypercholesterolaemia, diabetes, obesity), associations that persist after accounting for sociodemographic characteristics and baseline cardiac disease. Proposed mechanisms involve (1) chronic elevation of inflammatory cytokines (‘immune-activated sickness’),27 (2) a poorly balanced activation of the autonomous nervous system and (3) glucose levels.28 Furthermore, self-rated health has been linked to subclinical cardiac alterations, for example, elevated coronary artery calcium score,24 cardiac biomarkers29 and reduced right ventricular function.30 We found no associations with left or right ventricular echocardiographic parameters, possibly because our sample was derived from an overall healthy general population, participants were young (mean age 41 years) and echocardiographic alterations may possibly occur later in the cascade of cardiac pathology compared with elevated calcium scores and biomarkers. Another potential reason could be low statistical power due to the limited size of the study population.

Importantly, women had somewhat lower self-rated health than men, and the relationship with cardiovascular risk factors was further modified by sex. Both findings are in line with previously published data.31–33 While the mechanisms for this remain unknown, women may be particularly sensitive to chronic health conditions, thus affecting self-rated health.34 Recent studies have demonstrated that the prevalence of cardiovascular disease is higher in women, emphasising that an appraisal of sex differences is necessary to obtain maximum benefit of lifestyle interventions for the prevention of cardiac disease.35

Throughout the last decades, quality of life has been used as a tool to measure outcome of healthcare interventions and guide healthcare policy making. Although self-rated health represents a generic measure that encompasses many dimensions of health, and as such, has limited sensitivity to address specific health issues, it is considered a reliable measure to compare health in different populations and to evaluate disease burden.36 Because classic risk tools for cardiovascular disease do not capture social determinants, it has even been argued that self-rated health, in addition to classic risk factors, may be more useful for cardiovascular risk prediction. The EuroQol visual analogue scale constitutes a widely used tool for this purpose.14 In the Amazon Basin, the average self-rated health score was 75 points, which is lower compared with other studies from Brazil, where average scores of 78–84 points have been reported.9 37 Notably, none of these studies were conducted in Northern Brazil, and the assessed populations were younger than our sample. In addition, differences in cultural, regional and disease patterns may partake in understanding this difference, and further explain why general life expectancy in the Amazon Basin is below the national average in Brazil.38

Self-rated health relies on patient-centred care, which integrates the patient’s environment, values and preferences, hence making it meaningful to the patient and the treating clinician. It is a reproducible and consistent measure across different populations and geographical regions,36 and it may potentially complement well-established risk scoring models for cardiovascular disease.39 Because self-rated health is easily obtained, it can help to facilitate risk assessment strategies. This is particularly important in areas such as the Amazon Basin where access to healthcare is highly variable and often limited. Considering the close relationship we found with several cardiovascular risk factors, self-rated health could be obtained by non-medical personnel and enable screening of remote communities. Consequently, selected individuals, that is, persons with low self-rated health and no known cardiovascular risk factors, could be referred for risk factor optimisation in healthcare facilities. Furthermore, it could be used as a measure for the effect of primary healthcare prevention strategies, similar to what has been reported previously.40 Whether self-rated health is linked to clinical outcomes in the Amazon Basin, and if improvement in self-rated health could improve prognosis, should be explored in future studies.

Strengths and limitations

Socioeconomic status is perceived to be associated with self-rated health and cardiovascular risk factors,41 42 and despite our multivariable adjustment, residual confounding may still exist. Interestingly, parameters of socioeconomic status did not vary significantly across tertiles of self-reported health (table 1), indicating that this relationship may differ in this region. Health-related behaviour, including healthy diet and physical activity, was self-reported and this could be associated with bias. Furthermore, it is a limitation that the questionnaire for health behaviour has not been validated in other studies or settings. We adjusted our models for cardiac disease at baseline in an attempt to limit reverse causation; however, some effect may persist. To reduce bias, we had a clear and predefined hypothesis prior to commencing data analyses and a rigorous design for the sequence of questionnaires. To increase the sample size, we included a subgroup of participants recently treated for malaria (n=70). As this group was derived from the same population, had completed treatment for malaria and all associations remained significant when excluded, we do not believe its inclusion affects the generalisability of our results. Because no standard data values of the EQ-5D-5L have been published in Brazil, we did not apply data from the five dimensions of quality of life in this study, nor calculate index scores. Data from this study represent an important first step in establishing EQ-5D-5L index values for the rural parts of the Amazon basin. Reference values for the EQ-5D-3L9 have been published, but cross-walk datasets are not available. To avoid the inclusion of white coat hypertension, we defined hypertension based on prior physician diagnosis and/or intake of anti-hypertensive medication. While the generalisability of our findings to other regions in the world may be disputed, the Amazon Basin covers eight other countries in addition to Brazil. Hence, our findings are likely to be applicable to populations in these areas or to populations who share similar environment and culture.

Conclusion

Self-rated health was positively associated with a healthy lifestyle, and this relationship was modified by sex. Conversely, self-rated health was not associated with cardiac disease by echocardiography. On a hypothesis-generating basis, healthcare policies could potentially use self-rated health for screening or as a target to improve health behaviour. Nevertheless, this should be investigated in future validation studies.