Background: It is widely believed that providing doctors with guidelines will lead to more effective clinical practice and better patient care. However, different studies have shown contradictory results in quality improvement as a result of guideline implementation. The aim of this study was to compare family doctors' knowledge and self-reported care of type 2 diabetes patients with recommendation standards of the clinical practice guideline.
Methods: In April 2003 a survey was conducted among family doctors in Estonia. The structured questionnaire focused on the knowledge and self-reported behavior of doctors regarding the guideline of type 2 diabetes. The demographic and professional data of the respondents was also provided.
Results: Of the 354 questionnaires distributed, 163 were returned for a response rate of 46%. Seventy-six percent of the responded doctors stated that they had a copy of the guideline available while 24% reported that they did not. Eighty-three percent of the doctors considered it applicable and 79% reported using it in daily practice. The doctors tended to start treatment with medications and were satisfied with treatment outcomes at higher fasting blood glucose levels than the levels recommended in the guideline. Doctors' self-reported performance of the tests and examinations named in the guideline, which should be performed within a certain time limit, varied from overuse to underuse. Blood pressure, serum creatinine, eye examination and checking patients' ability to manage their diabetes were the best-followed items while glycosylated hemoglobin and weight reduction were the most poorly followed. Doctors' behavior was not related to the fact of whether they had the guideline available, whether they considered it applicable, or whether they actually used it.
Conclusion: Doctors' knowledge and self-reported behavior in patient follow-up of type 2 diabetes is very variable and is not related to the reported availability or usage of the guideline. Practice guidelines may be a useful source of information but they should not be overestimated.