Question 1 (Logistic Regression). Data for this study were collected on subjects who presented to the emergency departments at 6 New England hospitals with a primary complaint of chest pain,

shortness of breath, or other cardiac symptoms. We are interested in determining the relationship between in-hospital mortality (MORT: 1=in-hospital death, 0=alive) and the patient having a history of high blood pressure (HXHIBP: 0=no, 1=yes). la. How many patients die while in the hospital? What proportion of patients died while in the hospital? (1 point) 1b. We will fit a logistic model to look at the association between in-hospital mortality (MORT)and history of high blood pressure (HXHIBP). Write the equation corresponding to this logistic regression model (in terms of log odds). (1 point) 1c. Fit the logistic regression model with in-hospital mortality as the outcome and history of high blood pressure as the predictor of interest. Report and interpret the odds ratio and p-value for history high blood pressure. (1 point) 1d. Add age (AGE), ST segment elevation on presenting ECG (STEL: 0=no elevation, 1=STelevated), and chest pain as primary presentation symptom (SX1CP: 0=not chest pain, 1=chestpain) to the model above. In a table, report the OR and p-value for these 4 variables. Commenton which variables are associated with mortality. (1 point) 1f. Is there evidence of confounding of the relationship between history of high BP and mortality by age, ST elevation and chest pain? Briefly justify your response. (1 point) 1g. Currently the coefficient for age represents a 1-year change in age. Please calculate the OR for a 10-year increase in age (using output from 1d). Interpret the OR. (1 point)