Optimizing Management and Medication of Febrile Children in Out-of-hours Primary Care: CHILI Cluster RCT (CHILI)

A GP-parent information exchange tool in the form of an interactive booklet has the potential to provide parents with information about symptoms and fever management and consistent information during GP consultations. Thereby enhancing their self-management and providing them with safety net advice when they return home.

It is hypothesized that the use of such an interactive booklet during consultations for febrile children at GP out-of-hours centres will result in a reduced number of antibiotic prescriptions, improved parental satisfaction and reduced intention to re-consult.

The development of the interactive booklet concerned a three-stage process and is based on extensive qualitative work among parents, GPs and other professionals involved in childhood fever management. The booklet incorporates already existing information about fever, alarm symptoms, advice on use of medication and specific infectious diseases that frequently occur in childhood in combination with fever such as upper respiratory tract infections, and otitis media.

We will perform a cluster-randomised controlled trial at 20 GP out-of-hours centres in the Netherlands. GP out-of-hours centres will be stratified by size, to ensure equal distribution of size between the intervention and control group. The required number of clusters and participants was based on the following assumptions: (1) ICC of 0.01, (2) alpha of 0.05, power of 0.80, (3) proportion of antibiotic prescriptions in control group of 25% and a proportion of 19% in the intervention group (6% minimal clinical relevant difference) and (4) 10% loss to follow-up and 10% efficiency loss based on unequal cluster sizes. Based on a previous cohort study, we estimated to include 1000 children per cluster (GP out-of-hours centre) within six months, resulting in a need for 20 clusters and an effective sample size of 737 patients in the intervention and control group (1474 in total).

The booklet will be used during consultations with febrile children at the GP out-of-hours centres that are randomly allocated by computer to the intervention. The child's symptoms will determine which information and advice parents receive from the GP.

Statistical analysis will be performed based on intention to treat principle by performing multilevel logistic regression analysis using IBM SPSS version 21.0 and MLwiN software. We will determine independent factors associated with antibiotic prescriptions. The same will be done for secondary outcomes.

All data will be obtained, managed and monitored according to the guidelines of Good Clinical Practice.