Description

Death by liver cirrhosis continues to increase, given the increase in alcohol consumption rates, chronic hepatitis infections, and obesity-related liver disease. Notwithstanding the high mortality of this disease, liver diseases do not affect all sub-populations equally. The early detection of pathology is a determinant of patient outcomes, yet female patients appear to be marginalized when it comes to early diagnosis of liver pathology. The dataset comprises 584 patient records collected from the NorthEast of Andhra Pradesh, India. The prediction task is to determine whether a patient suffers from liver disease based on the information about several biochemical markers, including albumin and other enzymes required for metabolism.

Additional Information

This data set contains records of 416 patients diagnosed with liver disease and 167 patients without liver disease. This information is contained in the class label named 'Selector'(167 healthy vs 416 diseased patients). There are 10 variables per patient: age, gender, total Bilirubin, direct Bilirubin, total proteins, albumin, A/G ratio, SGPT, SGOT and Alkphos. Of the 583 patient records, 441 are male, and 142 are female.

The current dataset has been used to study

• differences in patients across US and Indian patients that suffer from liver diseases.
• gender-based disparities in predicting liver disease, as previous studies have found that biochemical markers do not have the same effectiveness for male and female patients.