Rapid Responses to:

EDITORIALS: Nick Freemantle and Mel Calvert Composite and surrogate outcomes in randomised controlled trials BMJ 2007; 334: 756-757 [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Composite resources and value freedom

Grazyna T Adamiak   (11 May 2007)

Composite resources and value freedom 11 May 2007
Grazyna T Adamiak, PhD, MA, MPH Unemployed Send response to journal: Re: Composite resources and value freedom	In addition to the fact that hospital admission does not fulfil clinical criteria of specificity and sensitivity in order to work as an outcome measure or clinical end point, the use of the component ”hospital admission” in composite measures of outcome is misleading. In reality, the composite measures in cardiological RCT studies are questionable due to use of interventions such as admissions (implying access of inpatient beds) or revascuralisation/PTCA procedures as independent variables. The problem is that composite measures simultaneously use dependent and independent variables, confounding relations and correlations between them. Is the PTCA intervention a dependent or explanatory variable in this context? This means that composite end points are often based on assumption of availability of composite resources. This is not only misleading but also presupposing access to some kind of configurations and combinations of health care facilities and resources. Used in this way the composite end points have capacity to drive organisational changes in health care without being explicitly targeted under the cover of clinical freedom or autonomy and false impression of being “pure” medical measures. There is also a semantic problem in the notion of “composite end point” as it invokes semantic connotations that are not value free in opposite to the free of semantic and moral judgements notions such as cells or embryos. The problem of confounding due to clinically uncontrollable factors that the RCT is aimed to avoid is thus present in all RCTs using composite end points including interventions as components. As soon as clinicians use this kind of notions they leave their autonomous area of clinical medicine and move into the context of interventions, economic resources and events taking place in the real world as opposed to isolated conditions of an RCT, e.g. in university hospitals. Competing interests: None declared