Business
Engineering better care: A new approach
■ A group of researchers is investigating what lessons engineering and manufacturing have for physicians on how to run their practices. Can what works for blue collars help white coats?
By Tyler Chin — Posted Oct. 4, 2004
- WITH THIS STORY:
- » Manufacturing a new practice
- » External links
- » Related content
What principles of engineering and manufacturing can help office-based physicians practice better medicine, improve efficiency and boost their incomes?
Someone is about to answer that question. On Jan. 2, 2005, Indiana's Purdue University will launch the Regenstrief Center for Healthcare Engineering, which will research and apply engineering, management, science and information technology to re-engineer health care delivery and quality. It will be working with the Regenstrief Institute, a medical informatics research entity affiliated with Indiana University School of Medicine, and the Indiana University Medical Group, a 138-doctor primary care practice in Indianapolis.
Purdue, which has schools of nursing, pharmacy and health science but not medicine, expects to partner with other physician and health care entities, said Joseph Pekny, center interim director.
"We very much want the physician perspective, and we very much expect to be apprentice to the physician perspective," Pekny said. He also is a professor of chemical engineering and director of the e-Enterprise Center at Discovery Park, an interdisciplinary campus at Purdue.
The Regenstrief Foundation is providing $1 million annually for three years in start-up funding. The center also anticipates receiving federal grants and corporate and private funding.
For a physician's perspective, AMNews interviewed Thomas Inui, MD, president and CEO of the Regenstrief Institute. Dr. Inui, an internist, is also associate dean for health care research and professor of medicine at IU School of Medicine.
Question: How will principles of engineering be adapted and applied to re-engineer health care?
Answer: There are, in fact, a set of questions about organizational efficiency and quality that are recognizable in health care and also recognizable, for example, in manufacturing or service industries.
Engineering methods such as supply-chain theory and modeling could be applied to some classical questions in medicine like, "How long is it appropriate to go between visits if an individual has a chronic disease like diabetes?" "What is the maximally efficient and appropriate interval between visits?" is an example of a question that could be approached using engineering methods.
Another one is if someone is taking warfarin or a drug that needs adjustment from time to time, how often should that person be seen to keep them safe? Neither under- nor over-coagulated but [without wasting] the individual's or the physician's time and clinical resources in seeing them unnecessarily. Those are the kinds of questions that I think engineering could quite appropriately address in the practice of medicine.
Q: Endocrinologists generally see diabetics in the office once every three months. Isn't that an evidence-based guideline?
A: No. It's a habit. It's a question that hasn't been submitted systematically to research. There's a variety of research that could address that. It's called operations research, and it lives in engineering and other fields.
Q: What is operations research? It sort of conjures an image of a loading dock.
A: On a loading dock, the main aim is to get the freight on or off the trucks so [haulers] can deliver [goods quickly], spend a minimum amount of time on the dock or whatever.
In medical care, the aim [of operations research] is to keep patients in optimum health status and the methods that can be applied [to do that]. How do you do that? A mechanism for keeping patients in optimum health status is [the patient] coming to the clinic to see the doctor. In some sense the patient coming to the clinic to see the doctor is the loading dock, and everything that is going on there [involves] the doctor putting stuff on and taking stuff off.
You can use the same methods. The question is focusing on the critical process and deciding how to use the resources that you need efficiently in that critical process.
Q: How would that model work in the example of a patient with diabetes?
A: An interesting question for the science of medicine as well as practice would then be why [are diabetics seen quarterly]? If you ask physicians, the answer would be something like this: "Well, I feel as though I want to check in with patients to see how their control has been once every three months or four times a year because actually the pattern of control over that period of time might portend the need to adjust their drug regimen, diet or exercise."
Then I might say, "Is coming to see you necessary to know how the patient's control is going?" They would have to say, "No, not really." We have patients filling out blood sugar home checks on paper and mailing it to us that would allow me to do that. Or today we have systems that have Internet-based portals that patients can use to record their control. That [information] gets communicated to the physician, and the physician might call up and talk to patients about adjustments.
There are several dimensions in which one could say checking in from time to time is necessary. One would be how good is the metabolic control. Others would be how often does the patient need reinforcement and education in their behaviors, and how often do they need prescriptions rewritten or refilled. Another would be how often do they need to be seen to have preventive care because it's not just about glucose. It's about the whole raft of stuff that has do with diabetes complications, which are mainly vascular in impact.
Vascular disease prevention, then, is not only glucose control. It's blood pressure control, it's smoking cessation if they smoke, and sensible diet, sensible exercise -- a whole series of dimensions in which you need to assess from time to time a patient's status in order to do what physicians do. Which is to work across that whole spectrum of issues and provide optimal information and support to patients.
You say, "Are there some patients who are really easy to control and every time you see them everything is hunky-dory?" At the other end of the spectrum you say, "Are there patients whose diabetes or other condition make it tough for them to stay in good control? Whose diabetes might be brittle and really waxes and wanes?"
Yes, a physician would say that whole spectrum of patients really exists in a population out there. So are there ways in which you can use science to figure out where you are on the spectrum with a patient and start to make recommendations that differ from patient to patient and how often you see them?
Well, yes, I guess you could. Maybe you could let some people go [to the office] once a year or something like that. For other people, maybe you have to see them once a month to really optimally control [their condition]. What would happen to the efficiency of your operation if you stratify people like that and individualize the follow-up? Well, I guess on the whole we have to find out. That's a good question for operations research, but there is the potential -- without injuring anybody's health and maybe improving the health of the people on the brittle end of the spectrum -- that you could use fewer resources to see more people.
Q: Aren't supply-chain management and just-in-time manufacturing more applicable to hospitals than office-based doctors?
A: Yes, I'd say in general that the larger the institution and the more complex the set of commodities or resources that are being used for patient care, the more robust the methods would be [applied] in improving efficiency. On the other hand, you can use similar kind of methods even for an individual physician's office.
For example, many offices are moving to so-called open-access appointment making. How would you decide how much of your schedule to leave open? Whether you need the same open block on Monday as you need on Wednesday? How would you decide how to staff your office with, let's say, nursing assistant coverage for a heavy versus light period?
Actually, you could just arbitrarily say you're going to have two hours every day. But it will probably turn out that there's a kind of pent-up utilization on Monday. That's often the case, and there's just-before-the-weekend utilization on Friday, especially with pediatrics. Except that the Friday [utilization] differs. There isn't as much in the summer as there is in winter.
Would there be a way to study, even in a single office situation, what the nursing assistance is that you need and how much of the schedule would it be appropriate to leave for open-access, same-day appointment by the day of the week through the year by the season of the year? Yes, there would be. The availability of open schedule and the question of how much [time] to make available when is an understandable example of just-in-time resource management.
Q: Have you seen those principles used elsewhere in health care? If so, where?
A: I have seen these kinds of principles used, largely within the Veterans Administration health care system, and I think used to good effect. The VA system is the country's largest integrated delivery system, even bigger than Kaiser [Permanente] because it's nationwide and present in all parts of the nation. It has been singled out by Institute of Medicine studies as the system in North America that's made the most rapid and substantial progress in improving quality of health care.
Part of the reason [for that] is that they have all the prerequisites, like [an] EMR and various reporting systems. But part of the reason also is they also subject themselves to questions like those we've been discussing to operations research within the VA system.
Q: Since this is being done already at the VA, is there a need for the Regenstrief Center for Healthcare Engineering?
A: Oh, yeah. In North America we have an incredibly diverse health care sector. The VA is a good example of an integrated -- and relative to the rest of the health care sector, a standardized -- system from every point of view: information systems, human resources policy, equipment, contracting for supplies and so forth.
Most of health care in North America is not standardized and not integrated. Instead, it operates in this largely fee-for-service environment without a lot of the infrastructure that is built into an integrated delivery system.
The question for health care at large is [whether] we can apply this kind of research approach to the diverse health care provision that dominates U.S. health care. What are the generalizable lessons that could be applied across the diversity of environments in which people get care?