Requiring physicians to respond to computerized reminders increased physician compliance with preventive care protocols
ACP J Club. 1993 Sept-Oct;119:61. doi:10.7326/ACPJC-1993-119-2-061
Litzelman DK, Dittus RS, Miller ME, Tierney WM. Requiring physicians to respond to computerized reminders improves their compliance with preventive care protocols. J Gen Intern Med. 1993 Jun;8:311-7.
To evaluate the effectiveness of requiring physicians to respond to computer-generated reminders for improving preventive health care delivery.
Randomized controlled trial with a 6-month intervention period.
University-affiliated, primary care, general internal medicine practice.
31 faculty internists, 145 residents, and 5407 patients (mean age 58 y, 81% women) who had ≥ 1 scheduled visits to their physician during the study period and who were eligible for ≥ 1 of 3 cancer screening tools (fecal occult blood testing [FOBT], mammography, and Papanicolaou [Pap] testing).
The unit of randomization was a half-day practice session (4 separate practices each meeting for 8 half-day sessions per week resulting in 32 practice sessions). 16 practice sessions (with 15 faculty internists, 77 residents, and 2827 patients) were allocated to computer-generated reminders to which the participants were required to circle 1 of 4 responses. The other 16 practice sessions (with 16 faculty internists, 68 residents, and 2580 patients) were allocated to the computer-generated reminder only.
Main outcome measure
Physician compliance with the 3 study tests.
During the 6-month study, 8541 visits were made by 5407 patients who were eligible for ≥ 1 of 3 cancer screening tests (3552 [66%] for FOBT, 2919 [54%] for Pap testing, and 2654 [49%] for mammography). Intervention physicians complied more frequently than control physicians with all reminders combined (46% vs 38%, 95% CI for the absolute difference 2% to 12%) and separately for reminders of FOBT (61% vs 49%, CI for the difference 5% to 20%) and mammography (54% vs 47%, CI for the difference 0% to 13%) but not for reminders of Pap testing (21% vs 18%, CI for the difference -1% to 7%). When residents were analyzed separately, intervention residents were more compliant than control residents with all reminders combined and separately for FOBT and mammography but not Pap testing. No difference was found between intervention and control faculty for all tests combined or any single test. The effect of the intervention for residents but not faculty was greatest for patients ≥ 70 years old.
Compared with using computer-generated reminders only, requiring physicians to respond to a computer-generated reminder increased compliance with the delivery of preventive health care to patients.
Sources of funding: Health Research Service Administration and Agency for Health Care Policy and Research.
For article reprint: Dr. D.K. Litzelman, Regenstrief Institute for Health Care, 5th Floor, 1001 West Tenth Street, Indianapolis, IN 46202, USA. FAX 317-630-2667.
Physicians readily agree that health promotion and early disease detection play a vital role in patient management. Increasing public awareness and declining costs of preventive services have encouraged physicians to increase the use of preventive procedures. Since 1972, the age-adjusted mortality caused by strokes has declined by more than 50% because of the early detection and treatment of hypertension, and since 1950, cervical cancer mortality has declined more than 73% because of the increasing use of the Pap smear (1). Yet, physicians still underuse preventive procedures that have proven benefit and widespread support from the medical community.
Previous research has shown the potential role of computer-prompting systems in improving physician compliance with preventive protocols. Advocates for computer prompting cite cost-effectiveness, efficiency in reminding physicians about multiple events, and acceptability to physicians and office staff as reasons for considering computers (2). The articles by Litzelman and Tape and their colleagues address the role of computer prompting to enhance patient receipt of preventive procedures. Both studies were done in university hospitals. The study by Litzelman and colleagues was done to enhance physician compliance with computer-generated reminders that had been available to physicians for 14 years before the start of the study. Despite long-term access to a computer-prompting system, physician compliance with the reminders was less than 50%. In the study by Tape and colleagues, computer-generated reminders were introduced for the study. Unlike the former study, however, physicians were not required to respond to the reminders.
One reason for Tape and colleagues' negative findings was the inability of the study to detect less than a 30% change in the performance of any preventive procedure. The authors suggested that smaller changes may not be clinically significant. Many studies, however, have reported smaller changes in compliance. Tape and colleagues, however, did discuss several important caveats to the evaluation of studies of physician office structure: 1) the correlation between faculty attitudes about preventive procedures and compliance rates of the residents they precept; and 2) the avoidance of computers by some physicians in order to not appear "incompetent." Computer literacy may also partially explain why the intervention with a computer printout had better results than the intervention with a monitor display only.
Litzelman and colleagues showed an 8% absolute improvement when requiring physicians to act on prompts. It may not be practical, however, for busy clinicians to respond to every computer prompt on every patient. For example, the authors also reported the most common reason for physician noncompliance was lack of time. Neither study could improve physician compliance with Pap smears, where time usage may also be an important deterrent.
Other options have been effective in improving physician compliance. Harris and colleagues (3) found that a nurse-initiated prompting system increased the overall delivery of 7 preventive services and that this compared favorably with a computer-prompting system after adjustment for the number of patients enrolled. Gann and colleagues (4) described the positive influence of flow sheets, of patient-initiated mammography scheduling, and of characteristics of the mammography center itself on mammography use. These findings all suggest that physicians have a number of options to improve their use of preventive procedures, including paper systems, nurse-initiated prompting, and even computers. Computers may have a valuable role in physician office structure, but other effective alternatives exist to remind physicians about preventive health care.
James E. Shaw, MD
University of North CarolinaChapel Hill, North Carolina, USA