Last week, a wide-ranging interview on data in health care took place between Dr. Donald Berwick and Colin Hill of GNS Healthcare. Dr. Berwick and Hill got together in the Cambridge, Mass. office of the Institute for Healthcare Improvement, a health care reform organization founded by Dr. Berwick, to discuss data issues related to O’Reilly’s upcoming Strata Rx conference.
Berwick returned to IHI after his year as administrator of Centers for Medicare & Medicaid Services. Throughout these changes he has maintained his stalwart advocacy for better patient care, a campaign that has always been based on a society’s and a profession’s moral responsibility. Even an IHI course for the “Patient Safety Executive” program puts “Building a just culture” on its agenda.
Among the topics Berwick and and Hill look at in these videos are the importance of transparency or “turning on the lights,” ways of learning from the health provider system itself as well as from clinical trials, types of personalized medicine, the impediments to collecting useful data that can improve care, exemplary organizations that deliver better healthcare, and how long change will take.
The full video appears below.
Highlights from the conversation include:
- Introducing the speakers. [Discussed at the 0:00 mark]
- Basics: The triple aim, and what progress we are making toward it. At least we are talking of change now. [Discussed at the 0:37 mark]
- Impediments to quality and cost: Data collection and digitization are necessary but not sufficient. Data allows you to answer questions you never could before. Current systems don’t support that kind of data collection, but we are moving toward systems that do: variability, patient trajectories, comparisons of providers, accountability. [Discussed at the 4:50 mark]
- NUKA in Anchorage, Alaska has dramatically reduced hospitalizations and improved patient outcomes while lowering costs and getting high patient satisfaction ratings. They track patients and intervene before problems arise. [Discussed at the 10:46 mark]
- Feedback loops help on many levels: determining whether care is helping a patient, and determining whether the organization is using the practices it defined as successful. NUKA treats patients as people who own the system, and patients are given “tremendous voice.” [Discussed at the 13:10 mark]
- There are many philosophies of improvement: a learning approach, a market approach, an accountability approach, etc. But all depend on collecting data. “Turning the lights on” is essential. Not everything can be measured: some aspects of quality are less structured. But data allows one to perform experiments off-line. It also permits individualization of care. [Discussed at the 14:40 mark]
- Learning from the health care system itself. Collecting data in the field as an alternative or complement to randomized clinical trials. A randomized trial that “neatens up the world” may remove the variables you need to know. Trials are just one tool, and our data can swamp what we learn from trials. (It is interesting to note that the Institute of Medicine has recently described a learning health care system for nurses.) [Discussed at the 19:28 mark]
- Who are winners and losers in the competitive landscape created by big data? Information threatens the providers who do not perform as well. We will undergo a period of “combat and push-back,” and they have some validity: the interpretations we put on data sometimes really do have problems. [Discussed at the 20:53 mark]
- What do you do with the patient who is not typical–whose response to treatments does not conform to the “standard of care”? We depend a lot on experimentation and serendipity now, but data will allow us to act much faster. [Discussed at the 25:33 mark]
- How much can we accomplish by just doing what we know is right, and how much do we need new knowledge? We can confidently say that things we can correct right now are causing 34% waste. And this is not the limit: other countries spend at least a third less than we do, and of course they also have waste they can eliminate. [Discussed at the 27:35 mark]
- Personalized medicine: we will be able differentiate different types of a disease such as breast cancer. Whenever we give a treatment, we are implicitly doing an experiment on the patient. But capturing the data on what works will help us make better decisions next time. [Discussed at the 31:51 mark]
- Genetic data and other “omics”: having little effect yet, but our expectations of these advances will be fulfilled over time. [Discussed at the 35:01 mark]
- How long will it take to widely implement quality and cost improvements? Several institutions have achieved tremendous improvements over 5 to 7 years, but people making money off of the current system are resisting change. Bundled payments are just a start, because they treat a single episode. The real advance will be population-based payments, where we really support people’s health. Kaiser and Atrius Health are successful models. The country is at a turning point, because we cannot continue paying 18% of the country’s income on health care. We can either shift costs to individuals and reduce health care, or be smarter about providing care. The former would save money, but at the cost of our health and our “moral compass.” The public just believes what we have taught them: that more is better. [Discussed at the 37:49 mark]
- Where did Berwick get the feeling of moral imperative that he brings to health care? [Discussed at the 46:00 mark]