“When a payor starts to think not just about paying the bill, but also envisioning improvements in care, flows, and processes, a lot can happen.”
A growing share of health care spending is made by governments via social insurance programs. The power-of-the-purse gives governments huge potential to prod health care organizations to make improvements in the quality of care. Don Berwick, the former Administrator of the United States’ two biggest federally-funded health care programs, explains how it can be done.
As more governments in emerging markets adopt some form of social insurance, can you give them one big ‘Do’ and ‘Don’t’?
On the ‘Don’t’, stay as far away as you can from fee-for-service payment. I predict that social insurance systems structured around payment for population-based care will perform far better than ones replicating the broken fee-for-service model. For the ‘Do’, and this is related to the ‘Don’t’, increase flexibility in how funds are used. Too many systems are paying for the repair shop instead of working on the causes. For example, you can have a billion-dollar hospital but nobody working on clean water or food chain safety. Think holistically. The World Health Organization’s ‘Health in All Policies’ framework is very relevant to social insurance.
What did you learn from your time as Administrator of the U.S. Medicare and Medicaid programs about how a payor can drive quality in health care and improve outcomes?
A payor can do a lot to set policies that improve quality. When I was Administrator, we had a number of policy tools to improve quality that were mandated by the Affordable Care Act [the comprehensive U.S. health reform law passed in 2010]. First was support for improvement. We were able to directly fund with a budget of a billion dollars over two years many national efforts where hospitals came together to learn what they were each doing on important issues like unnecessary readmissions and hospital-acquired infections. The second theme was transparency. Because Medicare and Medicaid manage health care payments for a third of Americans, we had extensive information on claims that enabled us to find out who was doing an especially good job—and to share this information. The third theme was incentives at the organizational level. When we linked payments to improvements, which we were authorized to do, it generated a tremendous amount of activity in governance and leadership. For example, when we imposed small penalties on hospitals who performed in the lowest quartile on unnecessary readmissions of patients, we observed a lot of positive response.
What is the role of regulators and can you over-regulate?
Definitely, there is a risk of over-regulation and this can be toxic to improvement. If regulation is done wrong, it induces ‘gaming,’ meaning people will lie, or hide data, or find a way to get control of the information. They expend their energy in protecting themselves instead of trying to make improvements. If the regulator uses blame and shame or embarrassment as a tool, people will not set ambitious aims. That said, there are certain minimum standards that need to be met.
What kind of data is required to measure and manage the quality of care?
Measuring and managing are two different tasks. For regulatory purposes, one needs reliable information on the ‘big dots’, the major performance characteristics of care. Here, the mistake regulators can make is to measure too much instead of coming up with a relatively small control panel of highly important indices. In managing the quality of care, the challenge is different. Most improvement is local. Preventing infections may be a national agenda but every single hospital must work at it. To do that, they need locally-collected data and that needs to be gathered, made available, and used in much more rapid cycle times.
What are the common mistakes in data collection?
Measuring too much. Every measure adds cost to the system, so you have to be careful. Using measurement for blame or punishment or fear. Measuring parts instead of the whole. For instance, you can measure breast cancer screening but that doesn’t mean you have a great breast cancer care system. Another error is failing to represent the voice of the patient. Without their voice, your data system can be rather sterile, lacking heart and soul.
This article has been edited for clarity and conciseness.
Donald M. Berwick, MD, MPP, FRCP, KBE, is President Emeritus and Senior Fellow at the Institute for Healthcare Improvement (IHI), an organization he co-founded and led as President and CEO for 18 years. A leading U.S. authority on health care quality and improvement, he has authored or co-authored over 190 scientific articles and six books. From 2010–2011, Dr. Berwick served as Administrator of the Centers for Medicare and Medicaid Services. Other previous affiliations include Clinical Professor of Pediatrics and Health Care Policy at the Harvard Medical School and Professor of Health Policy and Management at the Harvard School of Public Health. Dr. Berwick currently serves as Lecturer in the Department of Health Care Policy at Harvard Medical School. Follow him on twitter: @donberwick