Many clinical trials lasting a few months or even a very few years may incur in a statistical error called the “immortal bias”, which may not have been made clear before the 21st century, even though it is relatively obvious. In the case of continuous therapies, it is not an intervention of a few months or of a couple of years that might change significantly the long term prognosis of a disease that lasts 30 years. That should have been obvious to all involved. It may sound like a complex statistical discussion, but in fact it is not.
Tobias Derfuss and Ludwig Kappos in JAMA (2012;308(3):290-291) went into this discussion and led to the papers by Donald A. Redelmeier and Sheldon M. Singh (Ann Intern Med. 2001 May 15;134(10):955-62, Survival in Academy Award-winning actors and actresses), and MP Sylvestre, E Huszti E and JA Hanley. (Ann Intern Med. 2006 Sep 5;145(5):361-3; discussion 392, Do Oscar winners live longer than less successful peers? A reanalysis of the evidence).
The 2001 study looked at 762 Academy of Motion Picture Arts and Sciences actors nominated for an academy award in a leading or a supporting role, and 887 matched controls in the same film, born in the same era. The median duration of follow-up from birth was 66 years. There were 772 deaths. Life expectancy was 3.9 years longer for Academy Award winners than for other, less recognized performers. This difference was equal to a 28% relative reduction in death rates.
The follow-up paper, a reanalysis, concluded that this was an artifact, and when this was compensated the advantage was less than a year, not significant. According to the authors and the discussion, many of the clinical trials of the kind accepted by the FDA lead to this kind of bias, which needs to be compensated. This is, also, quite obvious. It is as if the Oscar awardees and patients who undergo this or that treatment, were chosen from birth to live longer or be healthier with regard to this or that disease. It is quite obvious that only the period after intervention is relevant to the therapeutic or statistical question being posed.
In the words of Silvestre and colleagues: “The statistical method used to derive this difference gave winners an unfair advantage because it credited an Oscar winner’s years of life before winning toward survival subsequent to winning. When the authors of the second article reanalyzed the data using methods that avoided this “immortal time” bias, the survival advantage was closer to 1 year and was not statistically significant. The type of bias in Redelmeier and Singh’s study is not limited to longevity comparisons of persons who reach different ranks within their profession; it can, and often does, occur in nonexperimental studies of life- or time-extending benefits of medical interventions. The current authors suggest ways in which researchers and readers may avoid and recognize this bias.”
Prof. Dr. Paulo Bittencourt, PhD, FAAN
figure from Google