HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:
Guest Access | Sign In via User Name/Password

Electronic Letters to:

CRITICAL CARE MEDICINE: Yen-Yuan Chen, Alfred F. Connors, Jr, and Allan Garland Effect of Decisions to Withhold Life Support on Prolonged Survival Chest 2008; 133: 1312-1318 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

Electronic letters published:

Study has poor face validity despite sophisticated analytic techniques

Samuel M. Brown   (3 July 2008)

Study has poor face validity despite sophisticated analytic techniques 3 July 2008
Samuel M. Brown, Physician University of Utah Health Sciences Center Send letter to journal: Re: Study has poor face validity despite sophisticated analytic techniques Samuel.Brown{at}hsc.utah.edu Samuel M. Brown	The authors should be congratulated for the sophistication of their analysis. Propensity-score matching is indeed a useful tool for evaluating the impact of a non-randomizable intervention in an observational cohort. The fundamental question they hoped to address is an interesting and important one. Unfortunately, their hypothesis has poor face validity, and in the face of this problem, such sophisticated statistical techniques merely distract from the underlying problem. Fundamentally the authors have confused absolute medical futility, a useful interpretive framework for evaluating CPR in the setting of septic shock resulting in cardiac arrest, with improbability of a meaningful recovery after prolongation of life by heroic medical therapy. The assumption that renal replacement therapy (or to a lesser extent mechanical ventilation) is the equivalent of CPR in terms of futility represents an insuperable flaw in the paper. The former two therapies are generally recognized as prolonging life, even in critically ill patients with limited prospects of meaningful recovery. Withholding these therapies would almost certainly decrease the length of life, even if refusal to perform CPR had a negligible effect on survival. If this dataset could be reanalyzed with refusal of CPR--controlling for withholding of non-futile intensive therapies like dialysis or mechanical ventilation--as the independent predictor, it might provide meaningful information about the study's stated hypothesis. Alternatively, attempts to analyze quality of life in patients who declined physician proposals to limit life-prolonging intensive therapies would be another useful line of investigation, as would an attempt to compare "meaningful" survival in groups who receive and do not receive life support. The published analysis unfortunately provides minimal if any ground for causal inference on either of these fronts. The title of the accompanying editorial ascribes too much explanatory power to the study as currently published.