Correction the Bias of Odds Ratio of Misclassified Smoking and Myocardial Infarction in Shahid Modares Hospital

In many Epidemiological studies, the purpose is to investigate the relation between some diseases and their probable risk factors. Misclassification of exposure in casecontrol studies can lead to bias estimate of the Odds ratio and distort the relationship between exposure and outcome. The purpose of this manuscript is to identify the possible misclassification of exposure to smoking in relation to Myocardial Infarction (MI) and to correct the estimated Odds ratio. This study includes two parts, main study and validation study. The main study was a case-control which compared 509 MI patients to 428 controls, for main study status of exposure was measured with misclassification. Then in validation study, subsamples include 105 cases and 95 controls were selected randomly from the main study groups which the exact exposure to smoke was evaluated. The data from the latter study were used to estimate the misclassification parameters and correct the Odds ratio employing matrix method. In main study the relation of exposure and outcome was not significant =1.25 (95%CI: 0.959-1.628, p=.051). Estimation of misclassification parameters for case and control group was (SE=0.97, SP=0.774) and (SE=1, SP=0.731) respectively. After validation of the main study results and correction of misclassification error, the Odds ratio was estimated at 1.329 (95%CI: 1.027- 1.72, p=0.015) and the relation of exposure of smoke and MI was significant. In case-control studies negative impact of exposure misclassification can cause misleading about the relationship between exposure and outcome variables. In such studies using a method such as the matrix method in order to adjust the bias due to misclassification is necessary.