Why this research matters

Predictive models & clinical scores help clinicians assess patients and make decisions

They range from simple 'scores' through to advanced statistical tools. 

Predictive models are not new! The first simple scores were created in the 1960s, and nowadays healthcare professionals use multiple types of predictive models daily. 

Patients can benefit from faster, more objective treatment recommendations when healthcare professionals use these tools. 

Models & scores work by simplifying complex data

Predictive models & scores can help us understand difficult clinical problems - for example: 'is surgery the right thing to do for this patient?'

They work by combining particular pieces of healthcare data (eg. test results, diagnoses, and the person's personal characteristics), and simplifying them into a prediction (eg. 'this patient is at high risk of suffering a bad outcome if we do surgery'). 

Models & scores aren't magic ✨

The thing is: models/scores can never fully represent reality. To some extent, the predictions we make will always be wrong. In many cases, the model is right often enough to be helpful. But if you're a patient who gets the wrong answer, this may affect the care you receive.