Applying statistical techniques to lower health care costs

I have blogged earlier about the use of statistical techniques to identify high intensity locations of road accidents and malaria incidence and then intervene with greater focus to address the problem.

In a recent New Yorker article (gated), Atul Gawande points to the application of the same techniques - focus on the most expensive patients ("super-utilizers") to lower health care costs. A physician in Camden, New Jersey, Jeffrey Brenner, found that one per cent of patients are responsible for thirty per cent of medical costs. These patients were subjected to an intensive treatment regime to both treat and undergo prevention regimes,

"Brenner’s team, which includes a nurse practitioner and a social worker, make regular home visits and phone calls to check in about new and existing complaints, unfilled prescriptions, and other complications that could land them back in the hospital. They help apply for disability insurance and fill out paperwork for state-run housing where their medication can be overseen. They encourage these super-utilizers to improve their lives with steps like quitting smoking, cooking more, joining Alcoholics Anonymous—even going to church."

The results have been dramatic. Brenner found that his first 36 patients saw a 40% reduction in average monthly hospital and emergency room visits, and 56% reduction in average hospital bills—savings. A similar intervention in Atlantic City found that after 12 months, the first 1200 patients had 40% fewer emergency-room visits and hospital admissions, 25% fewer surgical procedures, and 25% reduction in treatment costs. Gawande writes,

"An important idea is getting its test run in America: the creation of intensive outpatient care to target hot spots, and thereby reduce over-all health-care costs. But, if it works, hospitals will lose revenue and some will have to close. Medical companies and specialists profiting from the excess of scans and procedures will get squeezed. This will provoke retaliation, counter-campaigns, intense lobbying for Washington to obstruct reform."

In this context, a recently released report (pdf here) by the British Government's Behavioural Insight Team draws on insights from behavioural science to show that significant health improvements can be achieved with a series of nudge-based interventions. These nudges can be applied to get people to give up smoking and drinking, get them to exercise more and turn up for routine periodic check-ups and so on. Atleast some of the same techniques could be applied to these "super-utilizers".