摘要：Analysis of the infl uence of risk factors aims to describe the unidirectional relationship between risk factors to an incident or a spesific disease, one of which is binary logistic regression. This analysis is applied to case of TIA because TIA is a warning that stroke will occur. This study was carried out to determine risk factors that aff ect case of TIA in Dr. Soetomo Regional Public Hospital Surabaya in 2012-2015 and the best binary logistic regression model. This study was an observational and case control study. Subjects were 90 inpatients at nerve division. Data were collected by observing the patient’s card status to get information of variables examined. Independent variables were hypertension, dyslipidemia, and diabetes mellitus. Result of simultaneous test showed that at least one variable that aff ected TIA (p = 0,000). Partial test showed that hypertension (p = 0.015; OR = 4.327), dyslipidemia (p = 0.000; OR = 10.455), and diabetes mellitus (p = 0.032; OR = 3.942) aff ected TIA (p 0.05). Binary logistic regression can be used to analyze risk factors of TIA in Dr. Soetomo Regional Public Hospital Surabaya in 2012–2015 with high prediction accuracy. Patients with dyslipidemia have the highest risk of TIA. Hypertension or diabetes mellitus was also signifi cant risk factor of TIA.
其他摘要：Analysis of the infl uence of risk factors aims to describe the unidirectional relationship between risk factors to an incident or a spesific disease, one of which is binary logistic regression. This analysis is applied to case of TIA because TIA is a warning that stroke will occur. This study was carried out to determine risk factors that aff ect case of TIA in Dr. Soetomo Regional Public Hospital Surabaya in 2012-2015 and the best binary logistic regression model. This study was an observational and case control study. Subjects were 90 inpatients at nerve division. Data were collected by observing the patient’s card status to get information of variables examined. Independent variables were hypertension, dyslipidemia, and diabetes mellitus. Result of simultaneous test showed that at least one variable that aff ected TIA (p = 0,000). Partial test showed that hypertension (p = 0.015; OR = 4.327), dyslipidemia (p = 0.000; OR = 10.455), and diabetes mellitus (p = 0.032; OR = 3.942) aff ected TIA (p < 0.05). This independent variables have contributed as much as 49% to TIA with prediction accuracy was 67%. Model obtained was fi t (p > 0.05). Binary logistic regression can be used to analyze risk factors of TIA in Dr. Soetomo Regional Public Hospital Surabaya in 2012–2015 with high prediction accuracy. Patients with dyslipidemia have the highest risk of TIA. Hypertension or diabetes mellitus was also signifi cant risk factor of TIA.
