Two-level analysis of risk factors for ischemic cardiovascular disease—A community-based study from Henan, China

Abstract

Background: Ischemic cardiovascular disease (ICVD) has high incidence and high mortality worldwide. The studies of its risk factors were mostly concentrated on an individual level, and there are scarce studies on the two levels of risk factors which include individual and regional levels. Methods: The data were obtained from a community-based study in 4 cities and 6 counties of Henan, China. Risk factors were initially screened by one-way analysis of variance or chi-square test. Then, they were reanalyzed using a two-level logistic regression model to construct a personal disease risk prediction model. Results: A two-level ICVD risk prediction model comprised 11 variables: age, body mass index (BMI), family history of hypertension, marital status, salt intake, smoking, moderate recreational physical activities, alcohol intake, and education at the individual level. Among the unalterable risk factors, for each additional unit of age and family history of hypertension, the risk of ICVD increased by 1.08 and 1.07 units [b95% confidence interval (95%CI): 0.99-1.16, 0.97-1.17, both p < 0.0001], respectively. Among the modifiable risk factors, the ICVD risk increases by 0.67, 0.27, and 0.28 units for each additional unit of BMI, marital status, and education (b95%CI: 0.60-0.74, p < 0.0001; b95%CI: 0.14-0.40, p = 0.0012, 0.18-0.37, p = 0.0001). Conclusions:The two-level ICVD risk model can predict that the risk of one person for ICVD will be lower if one is younger, thinner, and well-educated without a family history of hypertension. Overall, the twolevel ICVD risk prediction model gets a better fitting effect than the single-level model.

Prevalence, awareness, treatment and control of dyslipidaemia and their determinants: results from a population-based survey of 60 283 residents in eastern China

“…While most previous studies have evaluated the potential influencing factors of dyslipidaemia using a traditional logistic regression model,10–13 fewer studies have used a multilevel statistical model. The two-level logistic regression model helps retain high-level features of the data and provides more stable regression coefficient 25. Moreover, our study employed standardised protocols and instruments and used a structured questionnaire that adhered to the standards set by the Chinese Center for Disease Control and Prevention.…”

“…The two-level logistic regression model helps retain high-level features of the data and provides more stable regression coefficient. 25 Moreover, our study employed standardised protocols and instruments and used a structured questionnaire that adhered to the standards set by the Chinese Center for Disease Control and Prevention. To ensure the authenticity and reliability of the data, the Nanjing Centers for Disease Control and Prevention implemented a strict quality control plan before, during and after the investigation.…”

“…A generalised linear mixed model was used to establish a two-level logistic regression model. First, a two-level logistic regression 25 null model without explanatory variables was fitted, and the intraclass correlation coefficient (ICC) was used to test whether the data had a hierarchical structure. The formula for ICC is ICC=var(V0j)/[var(V0j)+(π 2 /3)], where var(V0j) represents the variance of the random intercept (variance of level 2).…”

“…The two-level logistic regression model helps retain high-level features of the data and provides more stable regression coefficient. 25 Moreover, our study employed standardised protocols and instruments and used a structured questionnaire that adhered to the standards set by the Chinese Center for Disease Control and Prevention.…”