One area where insurance payers could expand the use of machine learning to further lower healthcare costs, improve patient experience, and improve patient outcome would be in their drug formularies.
As it stands, many drug formularies reflect cost-cutting, bulk price, and often exclusive deals with pharmaceutical companies and pharmacies rather than patient outcomes for each drug.
It is not uncommon to see a drug formulary that lists only one drug brand or its generic equivalent in a given drug/treatment category. This is an ineffective practice as not all patients respond well to any given drug or treatment. Those that don’t will require additional drugs or treatments which ultimately drives up the total costs in treating those patients, and the premium costs for nearly everyone else. Such also prolongs the patient’s suffering.
“Machine learning is used in drug design,” said Gray. “But I’m unaware of any insurers using it to improve efficiencies in drug formularies or costs in total per patient treatment.”
I’ve worked with many health insurers and have yet to see any of them use machine learning to improve the drug formulary money pit either. Here’s hoping the industry will begin that work soon.
Precision medicine will succeed or fail in large part according to the accessibility of drugs and treatments each patient needs. One-size-fits-all formularies will kill precision medicine before it even gets off the ground. And those same formularies will continue unabated as a money drain for insurers.
Payers, patients and providers will all see many benefits from widespread use of big data and machine learning across the entire healthcare ecosystem.