Cirrhosis - a chronic liver condition - most commonly presents with life-threatening complications (e.g. hepatic encephalopathy); this gives patients with cirrhosis rates of readmission as high as 10-50% within 7-90 days after initial discharge, with 27.1% of these readmissions found to have been preventable. The use of electronic health record (EHR) data such as lab tests and vitals is currently being used to predict a patient's risk of readmission, but while the predictive power of EHR-based machine learning models excels at predicting the risk of readmission for a single patient, these models fail to predict readmission possibility when integrating data from multiple patients. Shi et. al. attempted to create a META-DES framework that can take into account patient variability and disease presentation variability to better predict the risk of readmission. They performed a retrospective study on a cohort of 3307 cirrhosis-afflicted individuals - varying in complications demographic factors, and labratory test results - with at least one record of readmission in Chongqing, China. Potential predictors of readmission after 14- or 30-days were evaluated with the Mann-Whitney U test, the chi-square test, or the Fischer exact test and predictors with a P value under 0.05 were kept as possible predictors for the DES framework to select from when deciding the best combination of predictors. The DES framework was trained on a variety of different subsets and conditions, giving DES the ability to fill in the gaps when faced with scarce EHR information as well as adapt to fluctuations in a patient's health information. DES first created a large pool of classifiers by turning the possible combinations of side effects and comorbidities of cirrhosis into binary variables (allowing wider capture of different possible patient profiles), and then created a meta classifier that can assess the capability of each classifier as a predictor. The dynamic classifier theory of selection was used when developing the DES framework to choose the classifiers within specific data subsets that had the most accurate predictions. Of the cohort studied, 12.8% of patients were readmitted after 14 days while 26.6% of patients were ere readmitted after 30 days. When assessing for accurate predictability for readmission after 14 days, the framework had a maximal accuracy of 0.783; for the accurate predictability of readmission after 30 days, the maximal accuracy was 0.683. Labratory test results at both discharge and readmission had a strong correlation with the anticipated clinical outcomes for patients with cirrhosis, or in other words, there was a statistically significant difference (P values less than 0.001) between the labratory results of those with cirrhosis who were readmitted and those who were not.