Predictive analytics in healthcare promises to significantly influence various stakeholder processes. In general, hospitals could benefit from more accurate predictive analysis, among other things, from more pronounced monitoring of quality indicators, or from more precise planning of accommodation capacities or from an increased level of supply optimization , etc. Insurance companies could increase their push for sustainable growth and higher performance. The medical community could provide more personalized patient-centered care guided by clinical decision support, while patients could receive higher quality care and better price transparency (quote {van2016randomized}). Health governments should therefore organize health plans in such a way as to pay particular attention to this patient population characterized by increased home care while avoiding additional and costly hospital admissions. Inherent in the cradle-to-grave care coverage plan, the collection and exchange of data deserves as much attention as the organization of the care itself. Say no to plagiarism. Get a tailor-made essay on "Why Violent Video Games Shouldn't Be Banned"? Get an Original Essay Hospital readmission (admission to the hospital within 30 days of discharge) is harmful to both patients and caregivers. While it is sometimes unavoidable, it is frequent and often associated with higher costs. Modern standards of care require effective discharge planning, including transfer of information at discharge, patient and parent education, and coordination of care after discharge. Analyzing hospital readmissions continues to be challenging due to the multitude of influencing factors (e.g. seasonal variations) and is considered a critical metric of healthcare quality and cost cite{stiglic2014readmission}. Based on the cite{srivastava2013pediatric} report, the readmission rate within 30 days is 19.6%, 34.0% within 90 days and 56.1% within one year of discharge. According to the Institute for Healthcare Improvement, of the 5 million hospital readmissions in the United States, approximately 76% are preventable, generating an annual cost of approximately $25 billion cite{srivastava2013pediatric}. The potential benefits of accurate models for predicting readmission risk have led to many types of research based on patient data incorporated into electronic health records (EHRs) cite {saunders2015impact, stiglic2015comprehensible}. However, all of these approaches attempt to quantify the risk of readmission upon patient discharge, but do not attempt to answer the very important question: What diagnoses might be involved in readmission? Highly accurate models capable of answering this question would provide not only an indicator of readmission risk, but also an assessment of the risk of specific complications (diagnosis or symptoms) upon subsequent hospitalization. These models could provide valuable decision support to doctors at the time of discharge (they could decide whether additional monitoring or testing is needed for a specific patient) and push analytical models from a predictive role towards a prescriptive role in healthcare decision support. of diagnoses/symptoms with which a patient is likely to be readmitted, we use the Predictive Clustering Trees cite{blockeel1998top, vens2008decision, kocev2013tree} (PCT) framework. PCTs generalize decision tree models. They look for homogeneous groups of observations to which it is possible to associate a.