Decision Support and Predictive Analytics

Ci4CC Decision Support and Predictive Analytics Initiative 

Ci4CC Decision Support & Predictive Analytics initiative brings together the oncology community with academia and industry leaders in data science, machine learning and artificial intelligence, to identify opportunities for application of these technologies, and form partnerships to develop products that take predictive analytics studies further than research and into actual practice. This initiative will lead and support studies on the actual impact and effectiveness of such systems, as well as establishment of best practices and the standard path to successful pilots, implementation and adoption.  This initiative will focus on:

  • Identifying, understanding, validating and optimizing what is already being used as decision support tools for oncology population. 
  • Establishing a standard path and best practices to bringing existing prediction models that have shown great promise in retrospective studies, to real world.
  • Requirements for successful implementation of decision support tools, such as technology and infrastructure requirements, user engagement and adoption, ownership, legal, training and support. 
  • Identifying new opportunities and developing new predictive models for clinical decision support using state of the art technology. 

As cancer centers collect huge amounts of data through EMR systems, imaging devices, genomic tests, wearables, and patient reported outcomes platforms, and develop more sophisticated big data analytics capabilities, they are starting to move from basic descriptive analytics towards the realm of predictive analytics and technology-enabled real-time decision support. Application of Machine learning and deep learning in healthcare are rapidly becoming some of the most discussed, perhaps most hyped topics in healthcare analytics. This may be just one of the steps along the journey to healthcare analytics maturity, but it actually represents a huge leap forward for early adopters. Machine learning is a well-studied discipline with a long history of success in many industries. Oncology research and care can learn valuable lessons from this previous success to jumpstart the utility of predictive analytics for improving cancer research, care and treatment. 

Instead of simply presenting raw, fragmented data about past events to a clinician, predictive analytics estimate the likelihood of a future outcome based on patterns in historical data. This allows clinicians to receive alerts about potential adverse events ahead of time, and therefore make more informed, timely decisions. The importance of being one step ahead of events is obvious in cancer care, where a patient’s life might depend on a quick reaction and a finely-tuned sense of when something is going wrong. However, not every decision support tool is about alerts and rapid response. Machine learning systems can recommend optimal treatment decisions based on many similar past patients and outcomes. This enables clinicians join forces in an easier more systematic fashion, and benefit from the cumulative knowledge and experience of past and present colleagues. 

High-value use cases for predictive analytics exist throughout oncology ecosystem. The opportunity that currently exists for cancer centers is to define what “predictive analytics” means to them and how can it be used most effectively to make improvements. Predictions are most useful when the knowledge can be translated into action. Adoption, the willingness to intervene, is the key to harnessing the power of historical and real-time data. Importantly, to best gauge efficacy and value, the predictions, the end users, and the interventions must be integrated within the same system and workflow.

Initiative Lead:  Nasim Eftekhari




Like to join the initiative?  Please email Nasim here