Each year, the United States spends more than $9,000 per person on health care. Yet the world’s most costly health care system is neither its most effective nor efficient. Citizens of other developed countries enjoy better health outcomes and longer lives at a fraction of the price. According to the Institute of Medicine and the American Medical Association, around 30% of US health care dollars are wasted through misuse of resources. Retsef Levi, professor of operations management at MIT Sloan School of Management, isn’t sure he has all the answers to America’s health conundrum. But he thinks that he and his many colleagues at MIT are asking the right questions.
“If you want to change a system, you first need to understand how that system came into being,” says Levi, the J. Spencer Standish (1945) Professor of Operations Management at Sloan. “This country’s health system started to develop inorganically in the early 20th century focusing on care for sick people in hospitals. These hospitals delivered their services in exchange for fees. As a result, we have a system that focuses on treating the sick, instead of one that helps people to stay well.”
Born in Israel, Levi spent 11 years as an intelligence officer in the Israeli Defense Forces before earning degrees in mathematics and operations research. The latter discipline, he explains, is the science of leveraging data through models to inform decisions, particularly in a context of uncertainty. Levi and his colleagues use analytics and other quantitative methods to build data-driven models that help leaders and managers make decisions under the risks of an uncertain future.
Levi conducts research across a range of complex systems including supply chains, revenue management, and food safety. The uncertainties of health care, he says, are particularly complex—a unique web of technology, human behavior, politics, and culture. “Understanding the exact nature of the uncertainty is key,” he explains. “For example, I can’t predict exactly how many people will visit the emergency room at Mass General Hospital [MGH] on a specific hour and day in December. But I can build a model based on historical observations from the past three years on that day, and on similar days, and be able to get close to predicting.”
For Levi, the increasing availability of data opens up exciting opportunities in health care. Until recently, it was very difficult to predict when hospital patients were going to be discharged. Now, with big data from the new electronic medical records system at MGH and advanced analytics techniques, Levi’s MIT team and its MGH collaborators are able to predict daily discharges at an accuracy of over 90%. This enables providers to allocate resources more efficiently and substantially reduce patients’ wait time. The collaborative team also works on developing timely outpatient interventions and predictive risk models to reduce unnecessary and costly hospital admissions.
Levi believes the most commonly prescribed remedy for the US health care system—focusing merely on creating incentives that encourage hospitals to limit treatment and thus save money—is destined to fail. “You cannot change performance just by changing your pay structure,” he contends. “You need to design for performance, and then follow up with appropriate incentives.”
A co-leader in MIT Sloan’s Initiative for Health Systems Innovation, Levi is convinced that analytics coupled with human intelligence can help create a system designed to promote positive health outcomes and not just treat illness. “This requires a fundamental shift,” says Levi. “Not just in the US, but across the globe. At the moment, we don’t even have comprehensive metrics that can measure and help manage health outcomes.”
“I believe we’ve been asking the wrong questions about our health care system to this point,” says Levi. “Inefficient systems are not only costly, but are usually associated with bad outcomes. And our system is inefficient. Yet the only way to engage clinical teams to drive change is to focus on improving patient outcomes, which ultimately lead to lower cost and more efficient systems.”
Nicely written article! Professor Levi is sharing some critical insights about the origins of our healthcare infrastructure and metrics; these origins are not bad, but they didn’t foster incorporating the efficiency and outcomes measures we really need. That “…we don’t even have comprehensive metrics that can measure and help manage health outcomes” is wholly consistent with recent observations I made while I assessed how to measure impacts of implementing focused pre-emptive genomic risk screening. Every situation seems to have its own outcome measures, and those are limited by the quantity and quality of data available in a fragmented care and information landscape. Measures that could “work” across many areas and overcome the infrastructure deficiencies would be very valuable.
Here’s a diagram on incremental improvements in safety that lead to disruptive improvements in care: https://drive.google.com/file/d/15ExUap–D8lILbhMBL9S_2bWGqDCkrYc/view?usp=sharing
The problem with cost (upper right hand of diagram) is a lack of real free-market competition due to government policy. That’s where diet and exercise have a tremendous advantage – assuming the availability of ‘healthy’ food. One could easily argue that the government should subsidize and monitor that … particularly if it paid for all reactive health care.