Study design and participants
We obtained data from the 2021 nationwide STEPS survey conducted by the NCD Research Center in Iran. The detailed design protocol and methodology of the STEPS study 2021 have been described elsewhere19. Briefly, the World Health Organization introduced the STEPS studies according to its framework for NCD risk factors surveillance, assessment, and reporting. In Iran, the STEPS national surveys were first implemented in 2005 with eight follow-up rounds, with the latest survey being carried out in 2021.
The STEPS 2021 study utilized a systematic cluster classification method to calculate the sample size. After using a proportion-to-population size method, 3176 clusters were calculated to conduct this national survey, with 10 participants in each cluster (reduced to 9 due to COVID-19 restrictions in some clusters). The survey collected demographic, lifestyle behaviors, and metabolic information through three steps: (1) questionnaires (n = 27,874) (2) physical (n = 27,745) and (3) laboratory measurements (n = 18,119).
In this study, we included adults ≥ 18 years with self-reported hypertension, SBP ≥ 140, DBP ≥ 90, or antihypertensive medication use, who resided in rural and urban regions of 31 provinces in Iran (n = 9561). Individuals with severe mental or physical disabilities unable to answer the questionnaires or participate in anthropometry measurements; individuals who refused to provide laboratory samples; and pregnant women were excluded from the study. As illustrated in Fig. 1, after excluding cases with missing values, 9388 patients were included in the final analysis.
Flow chart of the study participants selection. BMI: Body Mass index.
Questionnaires
The STEPS questionnaire is primarily based on the standard tool created by World Health Organization, specifically utilizing the most recent version (version 3.2)20. This questionnaire has consisted of two main sections: core questions and expanded questions. This survey incorporated all core questions along with a majority of the expanded questions, which cover areas such as demographics, dietary habits, physical activity, and tobacco usage.
Physical measurements
Weight was recorded using a calibrated digital scale (Inofit), which was adjusted with a 5 kg reference weight each time it was moved. Height was assessed using a standard meter stick while the individual stood straight against a wall, ensuring that their heels, hips, and the back of their head were aligned with the wall. Blood pressure readings were taken in three separate rounds from the brachial artery, with each round separated by a three-minute interval. This process started after a 15-min rest period in a seated position, utilizing standard Beurer sphygmomanometers. The final blood pressure value was determined by averaging the second and third measurements.
Variables
We adopted five lifestyle factors, namely, physical activity, BMI, alcohol consumption, tobacco use, and dietary habits, to create a healthy lifestyle behavior score. Each lifestyle factor was classified into different levels based on a predetermined criteria, and patients were assigned points for each level. The full description of the criteria and scoring are outlined in Supplementary Table S1. In brief, physical activity was defined as total recreational and travel time in minutes per week and scored in tertiles; BMI was categorized with ≤ 24.9, 25–29.9, and ≥ 30.0 kg/m2 values21; alcohol consumption and tobacco use were classified based on individuals’ history of use as current, former, or never; and dietary habit items such as, mean fruits and vegetables intake were calculated from servings/day * days per week divided by 7, dairy products, and other nutritional variables including dairy type (based on fat content), red meat, fish, processed meat, sugar-sweetened beverages, salt intake, high-salt processed foods, main meal, snack, breakfast, whole grains, nuts, and nutrition facts label were scored based on frequency of consumption determined by expert opinion. Then we calculated a healthy lifestyle behavior score from the sum of all five lifestyle factor points for each patient. Using the tertile system, the healthy lifestyle behavior score (minimum score 0 to maximum 9) was divided into three groups of poor, moderate, and good lifestyle behavior for the first to third tertile, respectively. We then examined the relationship of each lifestyle behavior category and the trend of scores from 0 to 9 with the study outcomes.
The study outcomes consisted of uncontrolled hypertension, SBP, and DBP levels. Hypertension was defined as SBP ≥ 140 mmHg or DBP ≥ 90 mmHg, self-reported, diagnosed during physical assessment, or using antihypertensive medication. Patients diagnosed by a physician or using medication were classified as the aware group. Additionally, we defined controlled hypertension as SBP < 140 mmHg and DBP < 90 mmHg1.
Ethical approval
All individuals gave informed consent to the study conditions before participation, and the Research Council and Medical Ethics Committee at the Endocrine and Metabolism Research Institute approved the 2021 STEPS study protocol (IR.TUMS.EMRI.REC.1403.071). All research was performed in accordance with the Declaration of Helsinki.
Statistical analysis
The weighting process was done based on the age, sex, and area stratifications at the provincial level of the 2016 Iranian census, which the detailed process is described elsewhere19,22, and the nonresponse rate was considered in statistical analyses. After that it was cleared of unusual values. Descriptive statistics for categorical variables were presented as proportions and confidence intervals (CI), and numerical variables were described in terms of weighted means and corresponding CI. The chi-square test was used to compare all categorical variables pairwise (e.g., sex and lifestyle index). In addition, an independent t-test was used to show the difference between the means of the numerical variables.
Logistic regression was performed to assess the association between healthy lifestyle as an independent variable (in three levels: poor, moderate, and good) and uncontrolled hypertension as an outcome. First, the univariable model was run to illustrate the crude odds ratio (OR). Subsequently, the age- and sex-adjusted model and the fully adjusted model were created to control for the effects of potential confounders. In addition, the linear regression model was performed to examine the relation between a healthy lifestyle and mean SBP and DBP. As mentioned above, crude and adjusted models were created to control for confounding factors. A stepwise regression model with backward approach was used to select variables for the adjusted models. Therefore, all potential covariates were entered in model firstly, and then, variables with p-values under 0.20 were retained in the model and selected as covariates for adjustment23. The final covariates in the fully adjusted models were: age, sex, education, hypertension medication use, hypertension status, comorbidity, and marital status. To evaluate multicollinearity between the independent variables, the variance inflation factor criteria were tested in all models. The average variance inflation factor was below 5, indicating that there was no collinearity between the variables24.
Cases with missing values were dropped, and all analyses were performed using the complete cases strategy, as missing data did not significantly alter the results under sensitivity analysis with amputated data. Stata version 18 was used to analyze the data. The level of confidence was set at 95%, and p < 0.05 was assumed as a statistically significant.