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

Quality Improvement

Computer use enhances clinician task performance

ACP J Club. 1996 Mar-April;124:51. doi:10.7326/ACPJC-1996-124-2-051

Source Citation

Sullivan F, Mitchell E. Has general practitioner computing made a difference to patient care? A systematic review of published reports. BMJ. 1995 Sep 30;311:848-52.



To study the effect of computer use in primary care on physician-patient consultations, clinician performance, and patient care.

Data Sources

Searches were done in MEDLINE, BIDS, and GPlit using the terms computers in medicine, primary care, family practice, and medical informatics. Books, bibliographies, and conference proceedings and their references were also reviewed, as were references suggested by experts.

Study Selection

Studies were selected if they were about the effect of computers in primary care and if they reported using random allocation of study groups, adjusted or no baseline difference between groups, the unit of allocation, outcome measures, and completeness of follow-up. 30 studies were included.

Data Extraction

2 investigators independently reviewed and scored each study. Each methodologic criterion was given a score of 2, 1, or 0. Differences in scores were discussed by the investigators, and agreed-on scores were reached. Studies were classified according to the consultation process, clinician task performance, and patient outcome.

Main Results

6 studies examined the effect of computers on the consultation process. 4 of these reported that the consultation period was lengthened by 48 to 90 seconds when a computer was used, and 2 showed an increase in physician-initiated medical discussion and a decrease in patient-initiated social conversation. 21 studies examined clinician task performance. Of these, 3 studies showed an increase in vaccination rates of 8% to 18% when a computer was used. In 11 studies, other preventive tasks (including test ordering, physician and patient reminders, and screening procedures) increased by as much as 50% when a computer was used. 4 studies examined physician prescribing practices and reported a 13% to 30% reduction in prescribing costs, an increase in generic prescribing, and a decrease in prescribing time for both physicians and receptionists when these procedures were computerized. 2 studies that examined management of chronic disease showed an increase in the number of pulse, fundal, and blood glucose examinations. Of the 3 studies that measured patient outcome, 1 reported a mean decrease of 5 mm Hg in diastolic blood pressure in patients with moderate hypertension, and 2 studies that measured patient satisfaction detected no difference between patients whose care did and did not include computer use.


The introduction of computers in primary care settings has a positive effect on clinician task performance. The consultation process is lengthened by adding physician-initiated clinical information, but patient-initiated conversation is decreased. Few studies examined the effect of computing on patient outcome.

Source of funding: Clinical Resource and Audit Group of the Scottish Office Home and Health Department.

For article reprint: Dr. F. Sullivan, Department of General Practice, University of Glasgow, Woodside Health Centre, Glasgow, Scotland G20 7LR, UK. FAX 44-141-353-3402.


For many clinicians, computers are a mixed blessing. In fee-for-service settings, they increase the efficiency of submitting claims, billing, and tracking payments but are also perceived as a threat to clinicians' diagnostic and therapeutic roles. With the large capital outlays necessary to get started and the rapid evolution of computer hardware and software, one can understand why clinicians are skeptical of current electronic medical record systems.

The review by Sullivan and Mitchell rigorously catalogued the methods and results of previous trials that studied the effect of innovations in medical informatics on the processes and, in distressingly few cases, the clinical outcomes of care. These trials, although few in number, do show that computer reminders improved compliance with established guidelines, that electronic prescribing decreases costs, and that these systems can take longer to use than paper charts.

No medium is more portable or user friendly for recording medical information than the pen-and-paper chart, but paper is the most costly medium for storing information and the worst for accessing it. The potential benefits of electronic medical records go beyond merely reproducing the paper chart in electronic form. Computers can also do tasks that physicians do not relish or do well, such as watching for rare events and providing access to remote information sources or to experts. Electronic medical records also free up time so that clinicians can do tasks that computers will never do well, such as interacting with patients and families.

Future clinical trials similar to those reviewed by Sullivan and Mitchell must continue to show the potential developers, purchasers, and users of electronic medical systems which innovations work and what patient and societal benefits can be expected. Only then will individual clinicians and health care organizations embrace computers the way grocers, accountants, engineers, and mechanics have.

William M. Tierney, MD
Regenstrief Institute for Health CareIndianapolis, Indiana, USA