Lack of effect of a poster-based intervention to reduce the number of blood culture samples collected

Objective: To reduce the number of blood culture samples collected.

Patients and method: We performed a cluster randomized controlled trial in adult acute care, and subacute care and rehabilitation wards in a university hospital in France. A poster associating an image of eyes looking at the reader with a summary of blood culture sampling guidelines was displayed in hospital wards in the intervention group. The incidence rate of blood cultures per 1000 days during pre- and post-intervention periods was calculated.

Results: Thirty-one wards participated in the study. The median difference in blood cultures/1000 days between periods was -1.863 [-11.941; 1.007] in the intervention group and -5.824 [-14.763; -2.217] in the control group (P=0.27).

Conclusion: The intervention did not show the expected effect, possibly due to the choice of blood cultures as a target of good practice, but also to confounding factors such as the stringent policy of decreasing unnecessary costly testing.