LKS Next – Deep Learning Systems for Actionable Healthcare Knowledge Generation
LKS Next
Sector: Health
Business Case
Healthcare organisations are increasingly interested in the application of artificial intelligence (AI) to improve healthcare, reduce costs and improve the efficiency of healthcare services. Two of the cases in which this technological transition is to be applied are the care workload in A&E and the risk in patients with chronic heart failure pathology.
Objectives
Firstly, the aim is to create a model capable of predicting the hours with the highest emergency care workload. Secondly, to create a classification system that determines the risk percentage of each patient with chronic heart failure pathology.
Use case
The peak hour prediction model allows healthcare personnel to anticipate and reorganize the work schedule of their staff efficiently, assigning more workers to critical hours and fewer staff to the rest of the hours. The risk classification system is a decision support system for cardiologists, helping them to focus patient treatment.
Infrastructure
On Premise
Technology
Automatic or Deep Learning RPA
Data
The data used are: the case loads in the different emergency sectors and shifts of health centres in the Basque Country. Data on patients diagnosed with chronic heart failure; physical condition, blood tests, medication intake, etc.
Resources
Multidisciplinary team composed of data scientists and programmers.
Difficulties and learning
Treatment in A&E was influenced by factors external to those collected in the database. Regarding the risk classification model, the homogenisation of the raw variables of a clinical history into variables suitable for the model.
KPIs (business impact and metrics of the model)
The critical hours prediction model returns the next hours when A&E will have a high case load. As for the risk classification model, it returns a risk rating that categorises patients as very low, low, medium and high risk.
Funding
This project was funded by Hazitek, a business R&D support programme.
Collaborators, Partners
Kronikgune
