System and method for identifying complex patients, forecasting outcomes and planning for post discharge care

A method, comprising:

• receiving, by a system comprising a processor, patient information for patients currently admitted to a hospital from one or more electronic data sources prior to discharge of the patients, the patient information comprising clinical information about the patients and non-clinical information about the patients;
• applying, by the system, different care outcomes forecasting models to different subsets of parameters extracted from the patient information to predict care outcomes for the patients prior to the discharge of the patients, including discharge times, discharge destinations and their likelihoods, readmission risks and safety risks, wherein the different care outcomes forecasting models comprise machine learning models trained to predict the care outcomes based on learned correlations between the care outcomes and the different subsets of the parameters;
• and
• providing, by the system, care outcomes information identifying the care outcomes predicted for the patients to one or more care providers provides patient care outcomes information to one or more care providers via a graphical user interface displayed at one or more devices to facilitate managing and coordinating inpatient and post-discharge care for the respective patients;
• determining, by the system, whether the discharge destinations predicted for the patients violate one or more defined discharge rules;
• generating, by the system, an alert in response to a determination that a discharge destination violates the one or more defined discharge rules prior to discharge of the patients;
• providing, by the system, the alert via the graphical user interface;
• receiving, by the system, feedback information via the graphical user interface regarding patient needs and barriers associated with placing the patients at their discharge destinations;
• and
• retraining, by the system, one or more of the care outcomes forecasting models based on the user feedback to generate updated care outcomes forecasting models with improved accuracy;
• and
• applying, by the system, the one or more updated care outcomes forecasting models to at least some of the different subsets of parameters extracted from the patient information to predict updated care outcomes for the patients prior to the discharge of the patients.

A non-transitory machine-readable storage medium, comprising executable instructions that, when executed by a processor, facilitate performance of operations, comprising:

• receiving patient information for patients currently admitted to a hospital from one or more electronic data sources prior to discharge of the patients, the patient information comprising clinical information about the patients and non-clinical information about the patients;
• applying different care outcomes forecasting models to different subsets of parameters extracted from the patient information to predict care outcomes for the patients prior to the discharge of the patients, including discharge times, discharge destinations and their likelihoods, readmission risks and safety risks, wherein the different care outcomes forecasting models comprise machine learning models trained to predict the care outcomes based on learned correlations between the care outcomes and the different subsets of the parameters;
• providing care outcomes information identifying the care outcomes predicted for the patients to one or more care providers via a graphical user interface displayed at one or more devices to facilitate managing and coordinating inpatient and post-discharge care for the respective patients;
• determining, by the system, whether the discharge destinations predicted for the patients violate one or more defined discharge rules;
• generating, by the system, an alert in response to a determination that a discharge destination violates the one or more defined discharge rules prior to discharge of the patients;
• providing, by the system, the alert via the graphical user interface;
• receiving, by the system, feedback information via the graphical user interface regarding patient needs and barriers associated with placing the patients at their discharge destinations;
• and
• retraining, by the system, one or more of the care outcomes forecasting models based on the user feedback to generate updated care outcomes forecasting models with improved accuracy;
• and
• applying, by the system, the one or more updated care outcomes forecasting models to at least some of the different subsets of parameters extracted from the patient information to predict updated care outcomes for the patients prior to the discharge of the patients.