Could preferential prescribing and referral behaviour of physicians explain the elevated thrombosis risk found to be...

Recently published case-control studies on third versus second generation oral contraceptives showed a slightly increased risk for venous thromboembolism (VTE) and led to a discussion about biases or confounders external to the study such as, preferential prescribing of third generation pills to women at higher risk and differential diagnostic behaviour.

An interview survey with 102 physicians was performed in 2 weeks of December 1995 in Germany and 1209 of their patients were included in a retrospective cohort analysis of drug utilization and risk markers.

German physicians previously preferred to prescribe third generation pills whenever an increased risk was perceived, i.e. risk for VTE and for arterial thrombosis (risk factors, personal or family history of cardiovascular diseases). Almost all predefined risk scenarios were associated with an increased attention to non-specific venous symptoms and patients were more frequently sent for intensive diagnostic search for venous thromboses. The behaviour was not directly dependent on the type of OC; however the third generation pills were indirectly associated with higher risk in the pill-taking women. The physicians' view was confirmed by the analysis of the cohorts of patients: preferential prescribing was also found in this data set, although it was not as impressive as in the physicians responses.

The possibility cannot be excluded that preferential prescribing and differential diagnostic behaviour alone could have explained the roughly two-fold increased odds ratios from case-control studies recently published, had this information been available among the study variables.