Adherence to prescription guidelines in general practice: inter-doctor variation and trends over time.

Background: Evidence-based working has become increasingly important in medical care. A growing number of clinical guidelines has been implemented aiming at improving quality of care and decreasing inter-doctor variation. Objectives: To describe trends in adherence to prescription guidelines of the Dutch College of General Practitioners (NHG) and to study inter-doctor variation over a 5-year period (2003–2007). Methods: Data were used from routine electronic medical records from 80 representative general practices in the Netherlands (LINH), including around two million diagnosis- coded prescriptions per year. These data were combined with electronic data from the NHG-guideline database. Per diagnosis (ICPC-coded) we determined the percentage of prescriptions that included a drug that recommended for that particular diagnosis. To determine inter-doctor variation 95% confidence-intervals were calculated using multilevel analyses. Multilevel analyses were also used to test trends in guideline adherence over time. Results: A trend in guideline adherence (either upwards or downwards) was found for a minority of diagnoses. Adherence levels varied considerably between diagnoses. For example, while GPs chose a recommended drug in <5% of the prescriptions for acute bronchitis and incontinence, they chose >85% of recommended drugs for enuresis and constipation. Guidelines for the respiratory system generally have low adherence. Inter-doctor variation was large for all diagnoses, but constant over time. Conclusions: Guideline adherence is constant over time, but varies across guidelines and GPs. Efforts to rationale prescribing should focus on a selection of diseases with a low level of guideline adherence combined with risks for patient safety such as diseases of the respiratory system.