Predictive data analysis models that use machine learning algorithms can help to quickly identify patterns in clinical data, allowing researchers to gain new insights into how to improve the various steps from raw data to clinical decisions. Austrian researchers have now developed a software toolset which supports the rapid setup of predictive modelling solutions for a variety of healthcare-related applications.
By Dieter Hayn
Due to the ever-increasing amount of data generated in healthcare each day, healthcare professionals are more and more challenged by information overload. New tools, such as predictive data analysis models that use machine learning algorithms, can help to identify patterns in clinical data more efficiently. Requirements for predictive models in health and care are similar in many ways to applications in other domains. However, there are also various challenges that are specific to health and care settings.
Healthcare data – both a complex and sensitive issue
A recent paper by Dieter Hayn and colleagues from the AIT Austrian Institute of Technology, published in the journal it – Information Technology, describes both healthcare specific requirements and how these were addressed in their Predictive Analytics Toolset for Healthcare (PATH). So far, the toolset’s applications include the prediction of delirium in hospital patients, hospital re-admissions as well as the optimal amount of blood to be transfused during surgery.
“Health and care data are extremely complex, and they are sensitive in terms of privacy aspects. Developing machine learning tools from such data is a complex process. However, deploying such models in routine care is even more difficult. With our work we want to ease the implementation of machine learning tools in hospitals, to finally provide measurable benefits for the patients”, says Dieter Hayn, senior scientist at the AIT.
Read the original article here for free: