Objective: To assess the evidence for the effectiveness of clinical practice guidelines (CPGs) in improving patient outcomes in primary care.
Data sources: A search of the MEDLINE, HEALTHPLAN, CINAHL and FAMLI databases was conducted to identify studies published between Jan. 1, 1980, and Dec. 31, 1995, concerning the use of guidelines in primary medical care. The keywords used in the search were "clinical guidelines," "primary care," "clinical care," "intervention," "randomized controlled trial" and "effectiveness."
Study selection: Studies of the use of CPGs were selected if they involved a randomized experimental or quasi-experimental method, concerned primary care, were related to clinical care and examined patient outcomes. Of 91 trials of CPGs identified through the search, 13 met the criteria for inclusion in the critical appraisal.
Data extraction: The following data were extracted, when possible, from the 13 trials: country and setting, number of physicians, number of patients (and the proportion followed to completion), length of follow-up, study method (including random assignment method), type of intervention, medical condition treated and effect on patient outcomes (including clinical and statistical significance, with confidence intervals).
Data synthesis: The most common conditions studied were hypertension (7 studies), asthma (2 studies) and cigarette smoking (2 studies). Four of the studies followed nationally developed guidelines, and 9 used locally developed guidelines. Six studies involved computerized or automated reminder systems, whereas the others relied on small-group workshops and education sessions. Only 5 of the 13 trials (38%) produced statistically significant results.
Conclusion: There is very little evidence that the use of CPGs improves patient outcomes in primary medical care, but most studies published to date have used older guidelines and methods, which may have been insensitive to small changes in outcomes. Research is needed to determine whether the newer, evidence-based CPGs have an effect on patient outcomes.