One year ago, if you had asked Berdi Safford, MD, a family physician in northwest Wash-ington, whether she wanted to embark on a quality improvement (QI) project in her busy family practice, she probably would have burst into laughter — and then into tears.

“Anything that sounded like ‘continuous quality improvement’ seemed so overwhelming and involved so much change that there was no way I could do it,” she says, “and so it never got done.”

But today, quality improvement is being done in Safford's practice. In fact Family Care Network, the group without walls for which she is medical director, is now in the throes of an ambitious quality improvement initiative headed by the Institute for Healthcare Improvement (IHI), an organization dedicated to using the principles of quality improvement to identify best practices in the delivery of health care and the design of medical practices. The project is part of IHI's “Breakthrough Series” on chronic disease care — a 13-month quality-improvement initiative that involves approximately 30 organizations nationwide. Half are focused on congestive heart failure; half are focused on diabetes. By next spring, Family Care Network hopes to demonstrate that the principles of quality improvement can change physician and staff behavior and improve outcomes for patients with diabetes at all 17 Family Care Network sites. It's hardly a small task, and “the goals are set so high that you can't get there by just trying harder,” says Safford. “We're talking about practice redesign. This is not business as usual.”

What gives Family Care Network some hope that the group's goals are indeed achievable is one simple idea: Big change begins with small steps. “When you break it down into tiny little pieces, it's more workable,” says Safford. And with that, the network is tackling the gargantuan task of improving diabetes care through a rapid series of “baby steps.”

Step 1: Start where you are

When Family Care Network became part of the IHI diabetes project earlier this year, the physicians assumed they were starting from a better position than the average primary care group. Many of them were already using flow sheets in their patient charts, were familiar with the American Diabetes Association's (ADA) clinical guidelines, and had been paying attention to data about their diabetes care for at least one health plan, which was collecting the data to meet HEDIS (Health Plan Employer Data and Information Set) reporting requirements.

But a pre-project self-assessment graded the network “in the C range — much to our surprise and disappointment,” says Safford. And the health plan partnering with Family Care Network in the IHI project fared even worse: “We ended up in the D range,” says Chelle Moat, MD, MPH, medical director for quality management for Northwest Washing-ton Medical Bureau, which has its history in traditional insurance, not managed care.

The lesson? There's always room for improvement. By honestly assessing your present performance, says Moat, “you can look at places where you might focus your energy to try to get better.”

Step 2: Understand the problem

Diabetes affects 15.7 million people, or 5.9 percent of the U.S. population, according to data from the Centers for Disease Control and Prevention. Roughly one-third of diabetes cases remain undiagnosed. The disease is the leading cause of new cases of blindness in adults ages 20 to 74; it is the leading cause of end-stage renal disease; it accounts for more than half of lower-limb amputations in the United States; and it contributed to 193,140 deaths in 1996, making it the seventh leading cause of death that year. In addition, diabetes accounts for an estimated 14 percent of health care costs, or $98 billion in direct and indirect costs, according to the ADA.

What makes diabetes so difficult to manage effectively, says Safford, is a set of three problems: One, patients can “get by” for long periods of time without anyone discovering their conditions. Two, most health care delivery systems are set up for acute, reactive care, not proactive care of patients with chronic conditions. And three, “effective management of diabetes requires enormous behavioral changes on the part of patients, providers and systems,” says Safford.

Difficult as these obstacles may be, research has shown that significant improvements can be achieved through four strategies:

• Glycemic control,

• Intensive self-management support (diet, exercise, self-monitoring of blood glucose, foot care),

• Cardiac risk reduction (smoking, aspirin, blood pressure, lipids),

• Screening and management of complications (retinopathy, nephropathy, neuropathy).

For example, in the United Kingdom Prospective Diabetes Study, which involved 5,102 patients diagnosed with type 2 diabetes, a 1 percent reduction in HbA_1c levels resulted in a 17 percent reduction in all-cause mortality, an 18 percent reduction in myocardial infarction, a 15 percent reduction in stroke, a 35 percent reduction in microvascular complications and an 18 percent reduction in cataract extraction.¹

Step 3: Aim for something

The adage “aim for nothing and you'll hit it every time” is apropos of quality improvement and chronic disease care. One of the first assignments for the Family Care Network QI team, then, was to articulate its aim, or mission, which was “to help create successful partnerships between providers and patients in order to manage diabetes proactively and, thereby, improve clinical outcomes.” Based on that aim, the team defined its measures and goals for diabetes:

• Patient-specific self-management plans are documented in 60 percent of patient charts;

• A diabetes flowchart system that functions as a reminder for care is current in 80 percent of patient charts;

• Patients have improved glycemic control (90 percent of patients have two or more HbA_1c measures per year; 70 percent of patients have current HbA_1c levels of 7 percent or lower; no more than 15 percent of patients have HbA_1c levels of 8.5 percent or higher);

• Diabetic screening intervals and cardiac indicators are met (70 percent of patients have had a documented retinal exam in the past 12 months; 90 percent of most recent blood-pressure readings are 130/85 or better).

By tracking these measures each month, the network can see whether its performance is improving.

Step 4: Understand the ideal system of care

“Every system is perfectly designed to get the results it produces,” says Donald M. Berwick, MD, IHI president.

It follows, then, that if you want to get better results, you have to create a better system — not overnight but through a series of “baby steps” (see Step 5). That ideal system, according to IHI, has at least three themes:

Patient-centered care. If you were to take any family doctor off the street and ask him or her, “Do you practice patient-centered care?” the answer would be a resounding “ of course.” But according to IHI, truly patient-centered care is something most physicians haven't even contemplated.

In the current paradigm, says Safford, a patient's visit to the doctor for diabetes care is a lot like a trip to the principal's office. Its focus is correction, not empowerment, and “it ends up with the physician telling the patient what needs to change — the theme being, doctor knows best,” she says.

But patient-centered care means giving patients information and responsibility and getting them more involved in their own care. “The reality, at least what the studies tell us, is that the old approach, where we tell patients what to do and kind of beat up on them when they don't, doesn't work,” says Steven Alexander, MD, medical director for Lynden Family Medicine, another Family Care Network practice. “Patients at some point have to make the connection that they are responsible and they can change things and make their lives better.”

One way physicians can encourage this is by working with patients to set self-management goals that the patients find meaningful and doable. For example, “If you have a patient who eats five Big Macs a day, who smokes two packs of cigarettes a day, who is disabled and sits on the couch all day, and his goal is to walk to the mailbox three times a week, by golly that's his goal,” says Safford.

Helping the patient reach that one small goal may be the first step in helping the patient change his life.

Evidence-based care. Numerous studies have shown that physician practice patterns vary widely, even where highly effective, cost-efficient treatments have been identified (for example, prescribing beta blockers after a heart attack). Implementing evidence-based clinical guidelines, aka practice policies, clinical pathways, etc., can help. When Safford was exposed to the most recent ADA guidelines at an IHI learning session, she discovered “some of my diabetes thinking wasn't up to date even though I thought it was. For example, the revised goal for target blood pressure is less than 130/85, and most of us have been using 140/90. That's a big difference,” she explains. “I also learned to be more aggressive with the presence of microalbumin in the urine. I wasn't being aggressive enough in the [ACE inhibitor] doses.”

Jody Fox, RN, a member of the core QI team and triage nurse at Family Health, one of the Family Care Network sites, believes evidence-based clinical guidelines can also be useful to non-physician staff members by giving them a good overview of the disease and helping them understand why flow sheets, data collection and other tools for improving care are so important.

The ADA guidelines, updated in January 1999, can be accessed online at www.diabetes.org/DiabetesCare/Supplement199/S1.htm. In addition, the AAFP and ADA have partnered to create a set of guidelines titled “The Benefits and Risks of Controlling Blood Glucose Levels in Patients with Type 2 Diabetes Mellitus” (item #A928), which you can get a copy of by calling the AAFP at 800-944-0000.

Having good clinical guidelines in your possession is only half the battle, though. You must also translate those guidelines into practice, in part by making sure that the recommendations are well understood by staff members and that you have physician buy-in. “The reality is that staff members key off the physician to a great extent, and if it doesn't matter to the physician, it's not going to matter to the staff,” says Alexander. (For more information on identifying good clinical policies and implementing them, see the reading list.)

Population-based care. The goal of population-based care is, very simply, to satisfy the health care needs of all your patients, even the patient who hasn't set foot in your office for two years. This goes against the natural tendency of most health care systems, which is to focus only on the acute problem of the patient sitting in the exam room, never mind the problems lurking in the shadows. What population-based care requires is that you: 1. Identify who your patients are and what conditions they have (a patient registry) and 2. Monitor their care proactively.

For many practices, the challenge of developing a comprehensive and accurate patient registry may seem insuperable. The gradual computerization of practice is likely to make it easier, but even now it's possible to make a start. Family Care Network, for example, is working with its local hospital and other providers to create an Internet-based, communitywide registry, something they didn't think was possible just a few months ago. (In an upcoming issue of FPM, we will examine how the group's registries have evolved.)

The second part of population-based care is being proactive and keeping patients from falling through the cracks. Often, the problem with diabetes is that “we may think the endocrinologist is taking care of the patients, and the endocrinologist thinks we're taking care of them,” says Alexander. One solution is the use of flow sheets to track needed services for patients. (We'll discuss this issue, and Family Care Network's solution, in a later article in this series.)

Step 5: Think “baby steps.”

Quality improvement can often seem overwhelming to physicians and staff members because it feels like more work. In fact, says Alexander, “it is more work. And everybody feels like they're overworked already. That's the dilemma.”

The solution, says Safford, is to “start small with one doctor, one patient, one something. You learn what works and what doesn't, and you fix the problems before you implement the solution on a grand scale.” A common framework for quality improvement is the PDSA cycle:

• Plan what change needs to be made;

• Do it, preferably on a small scale;

• Study it to see whether it accomplished what you had hoped;

• Act on what you have learned.

According to what participants in the IHI project learned, the key is to avoid belaboring any part of the cycle. Rapid-cycle improvement involves simply coming up with a good idea, trying it, checking to see if it worked, and then expanding on it or coming up with a better idea. “This was all completely new for our group, and we realized we had a lot of work to do,” says Fox, “but it was also exciting because we were setting goals, trying new things and making changes right away.”

Often, an important question to ask of a new idea is, “Does it work, or is it just more paperwork?” says Fox. “If it's just more paperwork, we won't implement it.” (For more information on the principles of quality improvement, see the reading list.)

In the first five days after the core team attended its first IHI learning session, each member of the team was required to complete at least one PDSA cycle. Safford's assignment was to ask one patient about his or her diabetes goal. “So I came home, and the next patient with diabetes I saw was a woman I've cared for for a long time, with moderate success. She still smokes,” says Safford. “But I walked in and I sat down, smiled and said, ‘Hi, what is your goal for your visit with me today?’ And she didn't even think about the answer. It just spit right out: ‘Get out of here alive!’”

The patient's response not only confirmed the “principal's office” theory, but it provided Safford with a perfect opener to patient-centered care. “The bottom line of that 15-minute visit was we agreed we weren't even going to touch smoking because she's so full of guilt about the fact that she hasn't quit,” says Safford. “Instead, what she was going to focus on was taking a walk every day.”

That brief interaction was a simple but successful PDSA cycle. Safford tried a new technique (asking about self-management goals), and the patient's response taught her something. “I might ask it a little differently next time or might have more materials present when I do it again, but I started generalizing that to other office visits, not just around diabetes, and it was very enlightening.”

Step 6: Commission your pilot group

Family Care Network has three pilot sites working on the IHI diabetes project, with a goal of rolling out the program to all 17 of the group's sites before the end of the year. In most practices, the initial pilot could be “as simple as taking one doctor in the office and saying ‘for this doctor's patients we're going to try this and see how it works.’ Get the doctor who's most interested, and try it,” says Safford.

Once you've identified your pilot group, conduct a training session so that individuals understand the goals of the project and can see the big picture (essentially, steps one through five, above). Safford and her colleagues held a mandatory training session for all physicians, clinical staff and office managers of the three pilot sites. At the end of the presentation, the pilots were required to have a staff meeting on the spot and begin planning.

What always happens is you get these great ideas, and then you go back and you get swallowed up in day-to-day practice. Nothing gets implemented,” says Safford. “So we charged them with immediately answering four questions.” [Note: Each of the following questions will be addressed in greater detail, including the pilots' answers, in upcoming issues of FPM.]

1. How are you going to identify your patients who have diabetes? This would involve creating some type of registry, from a simple paper-based system containing the names of a few identified patients to a communitywide, Internet-based registry. To begin identifying patients, practices may be able to pull the data out of their own computer systems or patient charts. Health plans can help by finding patients with diabetes based on claims or pharmacy data.

2. How are you going to be reminded of what services your patients with diabetes need? The answer here will probably entail the creation of a flow sheet that can be placed in patients' charts to serve as a reminder and record results. Perhaps the most difficult part of using a flow sheet is deciding who will own it. Can the physicians be relied on to enter the data? Can the nurses handle the additional work? Should the task be shared by several individuals? Practices will also need a system for reviewing flow sheets periodically and following up with patients who have missed needed services. Some health plans may be willing to help with this task, for example, by sending the practice quarterly reports listing which patients had eye exams, based on claims data.

3. How are you going to “preplan” diabetic visits? The idea behind preplanned visits is to have patients complete their lab work, eye checks, foot checks and any other necessary services prior to their checkups so that the family physician has full information on the patient's condition and can provide better care. Most practices can't give patients one-stop shopping and same-day results, so the challenge will be to prompt patients to complete those services prior to their checkups. Reminder letters or phone calls are among the strategies, and again health plans may be able to offer assistance.

4. How are you going to handle patient self-management goals? Will the nurses be in charge of initiating this conversation with patients, or will the physicians handle it? Will it happen in group sessions or one-on-one? How will you document patient self-management goals in the chart, and whose responsibility is it to follow up with patients regarding their goals?

The pilot groups had just 30 minutes to come up with their best first guess at answering these questions. Then, to hold everyone accountable, the core team handed out two 3 ×5 cards to everyone and asked them to write down one thing they were going to do differently when they went back to the office. They wrote it once for themselves and once to hand in. A week and a half later, the QI team contacted everyone to see how they had done. “Most of them agreed that the call was helpful, kind of spurring them forward,” says Safford. “And most people had really done something. They just needed some more encouragement to keep doing it.”

Your mission …

The pilot sites in Family Care Network's diabetes project have been working at a rapid pace on the four key questions listed above, with a goal of having solid answers by the end of the year. “The real test,” says Alexander, “is going to be whether we can do this with the pilots and improve care, then try to expand upon that, spread the gospel and implement it.”

In the coming issues of FPM, mostly in our Improving Patient Care department, we will report on the network's progress, including what worked and what didn't for each of the four questions. For practices even remotely interested in quality improvement, this is the perfect opportunity to take a baby step or two toward providing better diabetes care. Work with your staff and colleagues to answer the first question on the list: How are we going to identify our patients who have diabetes? Move on to questions two, three and four as you can. Then compare your answers to what Family Care Network offers as its solutions in upcoming issues of FPM.

“I think with this kind of thing if you're not really challenged or pushed, it isn't going to happen,” says Alexander. “There has to be something driving you to pursue it, or life's just too busy and you aren't going to get at it.”

So consider this your push and take a step toward improved patient care, then join us for the next article in this series.