[Variability in primary care drug prescription in 2003 in Castile-la Mancha, Spain]

Detecting variability in clinical practice is important, given that it may entail the underuse or overuse of resources. This study was aimed at studying the variability of drug prescription in Primary Care in Castile-La Mancha in terms of certain sociosanitary characteristics of each health district.

The prescribing of various therapeutic groups in 2003 was reviewed in 180 health districts, the following variables having been gathered: number of packages having been prescribed per 100 inhabitants, percentage of inhabitants over age 65 in the district, visit rate (visits/1,000 inhabitants), physicians/1,000 inhabitants and distance to the reference hospital.

The greatest degree of drug-prescribing variability was found regarding fibrates (42.95% variance factor) and H2-blockers (38.61%). The group showing the closest correlation between its prescription rate and the variables analyzed was that of the angiotensin-converting enzyme inhibitors (mean Spearman p: 0.719); antidepressants however showing a weak correlation (0.324). The percentage of inhabitants over age 65 and the visit rate are the variables arising more significantly in the multiple regression models constructed for each treatment group. The highest determination coefficient was found in the angiotensin enzyme-converting inhibitors (R2=0.761). The mean determination coefficient for all of the groups, weighted on the basis of the prescribing volume, was 0.492.

A high degree of drug-prescribing variability was found to exist according to the different districts. Generally speaking, the variables analyzed explain part of this variability.