Methodological issues in a cluster-randomized trial to prevent dementia by intensive vascular care

Abstract

Objectives: Description of methodological issues in a trial designed to evaluate if a multi-component intervention aimed at vascular risk factors can prevent dementia.

Design, setting and participants: Multi-center, open, cluster-randomized controlled clinical trial (preDIVA) including 3535 non-demented subjects aged 70-78, executed in primary practice and coordinated from one academic hospital. General practices are randomized to standard care or intensive vascular care.

Intervention: Vascular care consists of 4-monthly visits to a practice nurse who monitors all cardiovascular risk factors. Hypertension, hypercholesterolemia, overweight, lack of physical exercise and diabetes are strictly controlled according to a protocol and treated in a way, tailored to the characteristics of individual participants.

Measurements: Primary outcomes are incident dementia and disability; secondary outcomes are mortality, vascular events (stroke, myocardial infarction, peripheral vascular disease), cognitive decline and depression.

Results: Between May 2006 and February 2009, 3535 subjects from 115 general practices have been included. The clusters have an average size of 31 (SD 22, range 2-114). 1658 Patients from 52 practices were randomized to the standard care condition and 1877 patients in 63 practices to the vascular care condition.

Discussion: When designing a cluster-randomized trial, clustering of patient data within GP practices leads to a loss of power. This should be adjusted for in the power calculation. Since intensive vascular care will probably lead to a reduction in cardiovascular mortality, the competing risks of mortality and dementia should be taken into account.