Doctors need feedback to make QI work

Information about guidelines isn’t enough

Collecting data and monitoring outcomes is an expensive, time-consuming proposition. But all that effort may be worthless if you don’t have an effective system to provide feedback and suggestions of corrective action to physicians.

That is the conclusion of a randomized, controlled study of quality assurance cycles in 16 group practices in Boston.1 And it is echoed by leaders in the outcomes field.

"Just knowing you’re going to be measured on something is not enough," says Lee J. Hargraves, PhD, survey scientist at The Picker Institute in Boston and project director of the Quality Differences Project. "It was the performance feedback that led to gains in improvement."

Not all types of feedback are equally accepted or effective. In the study, which was conducted from 1978 to 1983 but was only recently analyzed in detail, Ambulatory Care Medical Audit Demonstration Project researchers coordinated quality assurance cycles and collected data from medical records.

Physicians appreciated the review and support from an external research organization. In fact, 89% would have liked to continue the project, which was only funded for a five-year period.1 "There is an idea that people are negative about getting feedback, and that’s not true at all," says Hargraves.

Look out for negative profiling

Hargraves cautions that some types of physician profiling, which rate physicians or result in report card scores, can carry a negative connotation. "That sets up a tone of looking for bad apples rather than quality improvement," he says.

Based on this and other research, outcomes experts offer these recommendations when setting up a quality improvement program:

Education about guidelines isn’t enough to produce changes in patient care.

"There’s no magic bullet," says Carol Horowitz, MD, MPH, assistant professor of health policy and medicine at Mount Sinai Medical Center in New York City. "It’s hard to change physicians’ behavior."

Prior research shows that clear, well-constructed guidelines, active educational forums, individual performance data, and continuing feedback are all important to improving performance,2 she says.

Horowitz and colleagues are comparing the success of "academic detailing," in which physicians receive one-on-one instruction about guidelines, along with feedback and support, and continuous quality improvement projects. The guidelines dealt with the detection and follow-up for patients with depression and use of appropriate medication for hypertension.

While the data have not yet been fully analyzed and published, Horowitz notes that complex factors impact quality improvement. "I don’t think anyone is going to tell you if you do these five things everything is going to work," she says.

For example, you need to consider how receptive physicians are to change when you design your program. Those who have frequently dealt with managed care organizations may be more receptive to guidelines than those who are accustomed to making autonomous decisions, she says.

In addition to learning that their performance differs from their peers, physicians need guidance about what that means and how they should change. That is the goal of "Breakthrough Series" of quality improvement efforts and educational forums developed by the Institute for Healthcare Improvement in Boston.

"If you give physicians feedback about their cesarean section rate, a physician might look at the data and say, ‘The C-section rate for our practice is too high,’" says Maureen Bisognano, the institute’s executive vice president and chief operating officer. "They might sincerely desire to improve that. But that doesn’t tell them what to do differently."

Provide physician involvement in each stage.

Physicians need input throughout the quality improvement project, from selecting the guidelines and criteria to discussing methods of measurement and feedback, outcomes experts say.

"You can’t say, ‘This is what we need to do’ if they don’t believe it needs to be fixed," says Horowitz.

Customize your quality improvement

Quality improvement should be tailored to the specific institutional culture and needs. In other words, it must be customized for your group practice, Horowitz advises.

Moreover, organizational changes may help physicians alter their practices to provide care consistent with guidelines. For example, if you want to increase smoking-cessation counseling, you may want to add a counselor to provide patient education and support to the physician, says Horowitz.

Make sure that guidelines and interventions are evidence-based.

If you believe that a certain change in practice will improve patient care and outcomes, you better be able to prove it. Physicians are more likely to adhere if guidelines are evidence-based, says Horowitz.

Your corrective action is also more likely to produce results if it has been demonstrated to be effective elsewhere.

The Institute for Healthcare Improvement searches for practices that have demonstrated unusual success in various target areas, such as cesarean rates. The institute brings the experts together with physicians who have signed on to an improvement project. (See editor’s note below for more information on the Breakthrough Series program.)

In another example, the institute found that physicians who stay up all night delivering babies are more likely to perform cesareans the next day with their laboring patients. That suggests a change in physician scheduling as a corrective action to reduce cesarean rates, Bisognano says.

Expect significant variation in adherence to guidelines.

Performance measurement of physicians doesn’t produce tidy "good" and "bad" categories. The Ambulatory Care Medical Audit Demonstration Project found variation that related to guidelines as well as patients.

"Someone can be very good at following up on low hemocrits and not as good on cancer screening," says Hargraves.

Some patients will refuse the screening or have contraindications to the standard treatment. No matter how dedicated to following the guidelines, physicians will never attain 100% adherence, says Horowitz.

"Patients won’t take a pill because their brother took it and got really sick," says Horowitz. Physicians may be unable to persuade them that the medication is best for them and may need to rely on a secondary choice, she says.

[Editor’s note: Participation in a 15-month-long collaborative as a part of the Breakthrough Series, which includes meetings with experts, costs $12,000. Results of the improvement efforts are presented at a two-day congress, which is open to nonparticipants and costs $695. For more information, contact the Institute for Healthcare Improvement, 135 Francis St., Boston, MA 02215. Telephone: (617) 424-4800.]


1. Palmer RH, Hargraves LJ. Quality measurement and improvement among primary care practitioners: Results from a multisite, randomized, controlled trial. Med Care 1996; 34:S1-S113.

2. Horowitz CR, et al. Conducting a randomized controlled trial of CQI and academic detailing to implement clinical guidelines. Journal on Quality Improvement 1996; 22:734-750.