Bayesiae Estimation for Modelling Congestive Heart Failure Deaths and Using Lorenz Curve

摘要

Background: Coronary Heart Disease (CHD) and specifically Congestive Heart Failure (CHF) are among the most common causes of death for both men and women throughout the world. Several studies have indicated that women sustaining Acute Myocardial Infarction (AMI) have similar mortality than men. Although some studies state that CHF affects women in greater numbers and the short-term outcomes for women are worse. Aim: The aim of this study is to use/evaluate Lorenz Curve and associated Gini Index to assess whether gender is an independent factor related to in-hospital mortality due to CHF in a geographically defined population. Statistical Method: Smoothed age-specific rates are obtained by employing Bayesian methods and Markov Chain and Monte Carlo (MCMC) techniques. Bayesian MCMC method was used to calculate the estimates. A Lorenz curve for the age specific rates is constructed by plotting empirical cumulative distribution of CHF deaths against cumulative distribution of incidence. Application: A total of 20,856 patients were treated with various cardiovascular disorders during a ten-year period from 1991-2002 at the Hamad General Hospital in Doha, Qatar. Of the patients, 3712 patients (2476 males and 1236 females) were diagnosed with CHF, and hence were studied to evaluate whether female gender was an independent predictor for poor prognosis. Results: The Lorenz curve obtained for the age specific rates showed variation in the incidence of CHF by gender. The present study shows that women are likely to have CHF similar to their male counterpart. Also the incidence rate for CHF seems to increase with age. The Lorenz curve showed that the age inequality was comparable among male and female patients. This fact was also proved by the Gini index being 0.15 for males and 0.17 for females. Conclusion: It was found that the Bayesian approach is a promising technique for the estimation of mortality rates for small geographic areas such as the State of Qatar.