Some critical remarks on the initial detection of lung ailments using clinical data and chest radiography

An important area of health care is the problem of detecting lung ailments such as tuberculosis, pneumonia and lung cancer for individuals making their initial visit to a medical institution. Clinical data involving symptoms and signs has been the centre of interest for diagnosing patients. Important decisions sometimes need to be made before the availability of the results of further tests. However, besides clinical data, the chest radiograph is also available. This study uses logistic regression models to critically reconsider the use of age, loss of appetite, loss of weight and cough for predicting respiratory disease type. The main result of this study is that the probability of misclassifying the three disease type is large, and that good model fitting does not guarantee correct diagnosis. As a viable substitute, a graphical method of detection with an 85% chance of correct classification based on information extracted from the chest radiograph images is proposed