Identifying heart failure hospitalizations: can an automated algorithm become a gold standard?

I shall not today attempt further to define the kinds of material I understand to be embraced within that shorthand description [ " hard-core pornography " ]; and perhaps I could never succeed in intelligibly doing so. But I know it when I see it, and the motion picture involved in this case is not that. [Emphasis added.] —Justice Potter Stewart, concurring opinion in Jacobellis v. Ohio 378 US 184 (1964), about possible obscenity in The Lovers. T hus, I know it when I see it became a well-known colloquial expression in which one attempts to categorize an observation that perhaps is subjective or without clearly defined parameters. In this issue of Circulation: Heart Failure, this concept is central in evaluating classification of heart failure hospitalizations because Loehr et al 1 describe the comparison of an automated algorithm with a physician reviewer panel who presumably know it when they see it. The article is of particular relevance given the increasing importance of accurately assessing events, such as heart failure hospitalizations which, as the authors point out, are a crucial part of clinical research and epidemiology in heart failure. In clinical research, heart failure hospitalizations have become a component of the primary outcome of recent clinical trials to assess the efficacy of new therapies. Of course, it was not always so because early clinical heart failure trials focused on all-cause mortality. Therapies, such as converting enzyme inhibitors, β blockers, and aldosterone antagonists, became standard of care because of their demonstration of life-saving potential. 2–4 Because the field has matured in knowledge and treatment, it has become more difficult to prove efficacy in reducing mortality, and it has also become apparent that a large part of the burden of heart failure involves hospitalizations, particularly those related to the disease. Therapies may now be recognized as beneficial with a composite outcome, including heart failure hospitalization in which only this component of the outcome is favorable, such as in the Valsartan in Heart Failure Trial (Val-HeFT) with favorable adjudicated heart failure hospitalization results but neutral mortality. 5 The accurate identification of such events takes on even greater prominence because one considers that they are increasingly seen as sen-tinel events independent of other factors that influence patient prognosis. 6 Furthermore, heart failure hospitalizations are now serving as a metric in evaluating success of programs and, in the short term, even influence reimbursement schemas. In clinical …