Current issues of ACP Journal Club are published in Annals of Internal Medicine

Quality Improvement

Systematic practice-based interventions are better than conferences for improving professional practice

ACP J Club. 1996 Jan-Feb;124:22. doi:10.7326/ACPJC-1996-124-1-022

Source Citation

Davis DA, Thomson MA, Oxman AD, Haynes RB. Changing physician performance. A systematic review of the effect of continuing medical education strategies. JAMA. 1995 Sep 6;274:700-5.



To determine the effectiveness of educational activities designed to improve professional practice and patient outcomes.

Data Sources

Trials were collected using bibliographic databases (MEDLINE, ERIC, NTIS, CINAHL, HEALTHLINE, and EMBASE), hand searches of selected journals, review of bibliographies of relevant studies, and contact with experts.

Study Selection

Studies were selected if they were randomized controlled trials or trials with alternate allocation; if the study intervention was replicable and directed at changing physician behavior or patient outcomes; if the study population consisted of health professionals, most of whom were practicing physicians or residents; and if objective measurements of performance and health care outcomes were included.

Data Extraction

Data were extracted on type of intervention (educational materials, formal continuing medical education [CME] programs, outreach visits, local opinion leaders or educational influentials, patient-mediated interventions [e.g., educational materials or reminders], audit and feedback, and reminders); complexity of interventions (single, combination of 2 strategies, or multifaceted); outcome measures and their domains (disease prevention and health promotion, disease management, clinical management, resource utilization, prescribing); effectiveness of interventions (positive, negative, inconclusive, and mixed); types of physicians studied; and settings.

Main Results

99 studies that evaluated 160 programs met criteria. 51 studies included internists, 35 included family physicians, and 34 included residents. 75 studies took place in outpatient settings. 99 interventions (62%) showed an improvement in ≥ 1 major outcome. 70% showed an improvement in physician behavior, and 48% showed an improvement in health care outcomes. Interventions using 2 or more strategies were more frequently effective than were interventions using only 1 strategy. Most interventions showed improvements in performance or outcomes, including positive effects on prescribing (11 of 14 studies), prevention and screening (30 of 36 studies), resource utilization (17 of 24 studies), general medical management (32 of 58 studies), lifestyle or educational counseling (10 of 18 studies), and clinical management (2 of 4 studies). Among single strategies tested in more than a few interventions, reminders were the most effective, followed by patient-mediated strategies. Outreach and opinion leaders were also effective. Other strategies, including traditional CME, were less effective.


Individualized, practice-based interventions are more effective at changing physician behaviors and patient outcomes than are traditional, didactic, large group interventions.

Source of funding: Not stated.

For article reprint: Dr. D.A. Davis, Office of Continuing Education, Faculty of Medicine, University of Toronto, 150 College Street, Toronto, Ontario M5S 1A8, Canada. FAX 416-971-2200.


Interventions improve primary care processes but not necessarily outcomes

No one who has slept through hours of traditional CME lectures on a sub-specialized topic would argue with the findings of Davis and colleagues. Davidoff (1) has also questioned the efficacy of this mode of education. The most effective educational strategies are closely intertwined with clinical practice.

The reason that traditional CME is becoming irrelevant has to do with what our practices are becoming. The "new" practice of primary care is evolving away from an industrial model that rewards production of patient visits. The patient-centered approach acknowledges that what patients actually want is less computed tomography and more answers and education from their providers. Some patients are getting their health care information directly from the Internet. Today, providers need patient care information just to stay ahead of their patients! This re-engineering can take many lessons from the evolution of the information economy out of a manufacturing base.

Yano and colleagues defined access in terms of the ease with which patients could speak to or get an appointment with their care providers. In fact, a nurse's telephone call can be exactly what the patient means by "access." We will have to drop the 1950s model of a patient traveling to meet face-to-face with a physician as the sine qua non of health care. The new model of health care and CME is still being formed.

We can be sure of one thing: The paradigm of providers memorizing protocols can now be safely buried along with bleeding as a treatment for illness. We do not need to memorize protocols, we need to be reminded when to use them. Computer reminders were a common and successful method of achieving the goals of each of these articles. Another source of studies of the effects of information services and utilization management on patient care is the Columbia Registry of Information and Utilization Management Trials (2).

Davis and colleagues specifically excluded "administrative interventions," such as computer-based patient records or the problem-based format for records. According to Weed's (3) then-futuristic, computer-based, problem-oriented system, medical records should guide and teach. Barnett (4) argues that the most powerful use of information technology in undergraduate medical education will come when computer-based patient records are tightly integrated into educational resources and the student can learn at the instant the information is needed. The same can be said of postgraduate education. Perhaps how-to sessions on computer-based, problem-oriented patient records now should be a topic in primary care CME. The health care education equivalent of just-in-time manufacturing may be just around the corner.

As we continue to study the best ways to implement these changes, the humanistic aspect cannot be lost or made secondary. As managed care and economic realities force computers into patient care, we should be studying how best to bring them into the examination room without sacrificing the quality of the physician-patient relationship. Perhaps what we do best in providing health care is humanistic information processing. If so, we may see computerization continue to improve what we do and, eventually, improve outcomes.

Bruce Slater, MD
George Washington UniversityHerndon, Virginia, USA


1. Davidoff F. ACP Observer. 1993 Oct:17.

2. Balas EA, Stockman MG, Mitchell JA, et al. The Columbia Registry of Information and Utilization Management Trials. J Am Med Informatics Assoc. 1995;2:307-15.

3. Weed L, ed. Medical Records, Medical Education and Patient Care: The Problem-Oriented Record as a Basic Tool. Chicago: Year Book Medical Publishers; 1969.

4. Barnett GO. Information technology and medical education. JAMA. 1995;2:285-91.