Study yields system for citing QI successes

Key characteristics common to high performers’

While many studies have been designed to evaluate outcomes and identify appropriate standards of care, far fewer have as their prime mover the establishment of a system to measure those standards. But such was clearly the case in a recent study conducted by researchers at the Yale University School of Medicine in New Haven, CT.

"We knew we would follow up this study with some quantitative, randomly sampled hospital-based study," notes Elizabeth H. Bradley, PhD, MBA, assistant professor in the department of epidemiology and public health. "We also knew that, in the long run, we wanted measures of which pieces of quality make a difference to outcomes and which don’t."

Quality has many components, but it is not often measured carefully, Bradley notes. "Usually, what is measured is whether quality is on’ or off.’ On a larger scale, we wanted a really good classification system to measure what we thought were the key characteristics of QI efforts. What has previously been done has been to look at things with a very blunt instrument — how many people were trained, whether docs were involved, and so forth."

Ostensibly, the study examined the use of beta-blockers after myocardial infarction.1 Interviews were conducted with physicians, nurses, quality managers, and administrative staff at hospitals in eight different states. In addition to being asked about the usage of beta-blockers, they also were asked questions about data monitoring and feedback, physician leadership, and sustaining change. They were encouraged to expand their answers with specific anecdotes.

"The best way to come up with a classification system is through qualitatively open-ended interviews with hospital staff," explains Bradley. "And the goal of this study was to develop a classification system."

What emerged, then, was not only a picture of how often these hospitals used beta-blockers following myocardial infarction, but what characteristics were identified that were common to the "high performers."

"In order to be a high performer, a hospital had to have had 65% of discharged AMI [acute myocardial infarction] patients prescribed beta-blockers," Bradley explained. "We never expect it to be 100%; some patients have contraindications."

Six key characteristics emerged that were common to broad-based efforts at quality improvement:

  • goals of the efforts;
  • administrative support;
  • support among clinicians;
  • design and implementation of improvement initiatives;
  • use of data;
  • modification of variables.

Of the six, however, four were not found in hospitals with less or no improvement: Shared goals for improvement; substantial administrative support; strong physician leadership advocating beta-blocker use; and use of credible data feedback. These four factors were further analyzed to develop what Bradley and her colleagues describe as the "taxonomy" of factors. (See table, below)

Classification System Taxonomy
The taxonomy developed by Yale University researchers in their beta-blocker study reveals six characteristics common to quality improvement efforts. The following system, according to Elizabeth H. Bradley, PhD, MBA, assistant professor in the department of epidemiology and public health, may be employed as a quality checklist for any number of medical conditions:
Factors Dimensions Concepts
Goals Content Improve patient care; maintain financial position; enhance reputation
Specificity Explicit performance targets set vs. more vague goals of improving generally
Challenge Zero defects vs. more lenient goals
Sharedness Widespread sharing of and agreement with goals vs. conflicting/unsupported goals

Administrative support Philosophy Supportive of quality improvement vs. indifferent, vs. negative
Resources Human/technical resource availability; training in quality improvement techniques

Clinical support Physicians Presence of leaders; level of engagement; ability to lead change; supportive vs. detract ing from performance improvement efforts
Nursing staff Presence of nurturing leaders; facilitators; support for performance improvement efforts
Ancillary staff Involvement of necessary ancillary and support staff
Performance improvement 
Initiative type Enhancing adherence to existing system; redesigning existing system
Implementation style Methods used to increase adherence to standards; consequences of compliance or noncompliance; focus on improvement vs. faultfinding; participation vs. autocracy; degree of teamwork

Use of data Validity Awareness and credibility of research evidence; source and perceived quality of the data on current practice
Timeliness Frequency of reports providing data on current practice
Benchmarking Use of data from comparable sites, groups, or physician peers to aid in interpretation of hospital or physician performance data

Contextual factors Hospital size Capacity; staffed beds
System affiliation Part of health system vs. independent
Ownership type Nonprofit; for-profit; government-owned
Financial constraints Degree of competition for cardiac care; financial strength of organization
Organizational turbulence Turnover of senior administrative/clinical staff; mergers/acquisitions; unionization
Source: Bradley EH, Holmboe ES, Mattera JA, et al. A qualitative study of increasing ß-blocker use after myocardial infarction: Why do some hospitals succeed? JAMA 2001; 285:2,604-2,611.

Why the discrepancy between the original six factors and the final four? "We first tried to come up with six factors," notes Bradley. "Once we had them, we wanted to see which of these factors were actually common in high-performing hospitals." She notes that it is possible that in a larger-scale study the relative importance of these six factors may vary.

Bradley says this classification system can be used to provide concrete guidance to help plan effective interventions in other areas of medicine. "With any medical condition you are looking at, you would take the taxonomy almost like a check list: Is there really administrative support? Have we set goals? Do we have strong physician leadership? Do we have a data feedback system that people consider valid? It’s general enough to do that."

The strongest take-home message for quality managers from this study, says Bradley, may be the recognition of the limits of the actual technology of improvement. "Flowcharts, Delphi methods, all of those things that Berwick and Demming taught us, are incredibly important, but what may ultimately separate us are those softer’ cultural things; we need that as a foundation."


1. Bradley EH, Holmboe ES, Mattera JA, et al. A qualitative study of increasing ß-blocker use after myocardial infarction: Why do some hospitals succeed? JAMA 2001; 285:2,604-2,611.

Need more information?

For more information on this system, contact:

Elizabeth H. Bradley, PhD, MBA, Assistant Professor, Department of Epidemiology and Public Health, Yale School of Medicine, 60 College St., New Haven, CT 06520-8034. Telephone: (203) 785-2937.