How patterns can power better treatment
Artificial intelligence can play a key role in improving outcomes in healthcare
By Zia Chishti
The goal of healthcare is to improve the quality and increase the length of people’s lives. AI has some great opportunities to help in this field – but many are misunderstood or mis-sold. There is a lot of hype around diagnostics, for instance. Today, with only a small handful of exceptions, many applications of AI in diagnostics have fallen flat.
How can healthcare harness artificial intelligence?
Where AI can really help generate better health outcomes is by matching patients to clinicians more efficiently. Home visits, for example, can be critical to a patient’s care but are all too often neglected. This is because educating patients that home visits are good for them is difficult, even though they are key to reducing morbidity and mortality rates. The education process is important – and who is trying to do the educating? Other humans; the insurer talking to the insured. Creating an interaction between those two people that increases the likelihood that the patient accepts a home visit can be an incredibly powerful thing. It helps ensure patients adhere to their treatments and often flags serious health issues before they become any worse.
Afiniti’s AI platform identifies subtle and valuable patterns in data and in human interaction. We match individual agents working for healthcare providers such as United Healthcare in the United States to individual patients, creating interactions that increase the number of home visits, which help to deliver better health outcomes.
How does Afiniti’s AI work?
Afiniti’s AI already has a great track record in business, working closely with leading British companies such as Sky and Virgin Media to help improve the interactions they have with millions of customers. Rather than customer calls being automatically assigned to agents in the order they come in, the algorithm analyses the behavioral traits of both agents and callers based on data such as prior interactions each party has had. What does this mean in practice? A better conversation that increases the likelihood that the customer receives the best TV and internet package for them and their family.
It is a similar principle in healthcare, albeit with more at stake. We can predict which pairings between insurers and insureds will generate the best dialogue that is more likely to mean that the patient will agree to a home visit from a clinician.
What is the role of humans in rolling out more AI-enabled technologies?
Firstly, we need to distinguish between data analytics and AI. Data collection, usage and analysis need to be regulated, particularly around issues of privacy. But there is a train of thought which says that government should regulate AI. This is wrong and represents a fundamental misunderstanding of what AI is. It is maths – and you cannot regulate maths.
Secondly, we need to recognise that AI is oriented around specific use cases. The era of the automaton is not upon us, despite the hyperbole surrounding the technology. Businesses, including healthcare providers, need to understand how specific applications of AI will help directly help them and their customers. Don’t believe all the hype!
Zia Chishti is chairman of the board and CEO at Afiniti.
Original article appears here: https://www.newstatesman.com/spotlight/healthcare/2019/06/how-patterns-can-power-better-treatment