|
Risk Factors For Cardiovascular Diseases Among Iranian Women: Reanalysis Of A Case-Control Study With Bayesian Post Stratification Method
|
|
| Author:
|
SHARIFI N, FARAROUEI M, HASANZADEH J, AKBARZADE M
|
| Abstract:
|
Case-control studies are commonly used in epidemiology due to ease of handling and fast outcomes. However,
insufficient matching and non-representative sampling are important problems with a significant effect (bias) on study
results The goal of this study was to reanalyze the data from a case-control study to adjust control selection bias in the
estimation of the associations of several demographic and pregnancy related factors (multiple pregnancy, menopause,
and OCP consumption) with cardiovascular diseases (CVDs). To achieve this, census data was used as external
(population based) data.In the present study, the distribution of the study variables among case and controls, and index
variables (age and education) among cases, controls and a random sample of the source population are compared. Also,
results of logistic regression analysis (LRA) are compared with those obtained from Bayesian post stratification analysis
(BPSA). Results of LRA indicated significant associations between gravidity (OR=2.43,95% CI: 1.46 to 4.04),
hypertension (OR=2.47,95% CI: 1.68 to 3.64), hyperlipidemia (OR=1.87,95% CI: 1.27 to 2.75) and CVD. According to
the results from BPSA, the only significant association was found between hypertension (OR= 1.67, 95% CI: 1.07 to
2.62) and CVD. In addition, both methods of analysis revealed no significant associations between two important study
variables i.e. menopause (OR=0.94, 95% CI: 0.66 to 1.35) and OCP consumption (OR=1.24, 95% CI: 0.83 to 1.86) and
CVDs. The results of BPSA showed that frequent pregnancy, menopause and using OCP have no effect on CVDs,
whereas LRA found a significant association between the number of pregnancies and CVDs. The results suggest that at
least a part of the observed contradiction between the results of case-control studies are possibly due to selection bias.
|
| Keyword:
|
Women, CVD, Bayesian post stratification analysis, case-control design, selection bias Abbreviations :BPSA: Bayesian post stratification, CVD: cardiovascular disease, CHD: Coronary Heart Disease, MI: Myocardial infarction, OCP: oral contraceptive pills, PBRS: population base random sample, LRA: logistic regression analysis, MLRA: multiple logistic regression analysis
|
| EOI:
|
-
|
| DOI:
|
-
|
| Download:
|
Request For Article
|
|
|