Does PELOD measure organ dysfunction…and is organ function a valid surrogate for death?

Excerpt

The last 2 decades have seen considerable improvements in the delivery of care to critically ill children, such that paediatric intensive care unit (PICU) mortality rates are now typically below 10%, and indeed often closer to 5% [1, 2]. Paradoxically, this may hinder evaluation of new therapies aimed at further improving survival, because the number of patients required for an adequately powered clinical trial with mortality as the endpoint may now be unfeasibly large. Several strategies exist that may allow for a reduction in study size without loss of power. These include: (1) screening out extremely low- and high-risk patients, (where treatment effects may be less pronounced), (2) employing stratified randomisation based on mortality risk at trial entry [3], and (3) estimating treatment effects after covariate adjustment [4]. Identifying a suitable screening/stratification tool poses several challenges. It must be reliable, relatively simple, calculable within a short time frame, and should not reflect local treatment preferences, such as choice of monitoring tool or a particular ancillary therapy. At face value, contemporary mortality risk tools appear to fulfil these criteria, although valid criticisms of this approach have been voiced [5]. …