Author
Dr. Nigel SB Rawson, Ph.D.
Introduction
Advocates of increased regulation of new pharmaceuticals frequently cite the alarming statistic that 100,000 or more hospitalized patients in the United States die each year from adverse drug reactions (ADRs). Since the population of Canada is about one-tenth the size of the United States’ population, advocates of increased drug regulation in Canada argue that 10,000 hospitalized Canadian patients also probably die from ADRs annually.
Objective
The purpose of this article is to critically review the analyses on which these statistics are based, and to examine the actual empirical evidence for the rate of fatal ADRs from studies of hospital records in Canada.
Results
The original source of the American statistic on deaths due to ADRs was a meta-analysis. This is a method originally designed for the synthesis of randomized clinical trials, which are often sufficiently similar to allow them to be aggregated. However, the meta-analysis used observational studies which can vary widely so that they are frequently too different to combine. Moreover, 75% of the studies on which the statistic is based were published between 1964 and 1976, which challenges the relevancy of the findings because the practice of medicine has changed markedly since that time. More importantly, the incidence of fatal ADRs reported in the studies decreased steeply between 1964 and 1995 so that the estimated number of deaths due to ADRs in hospitalized patients in the United States based only on studies published after 1987 is 13,250. An analysis of actual hospital discharge and mortality data from Ontario for 1992/93 to 1996/97 found that an ADR (fatal or not) was recorded in 1.2% of the hospitalizations and only 0.05% of the hospitalizations had a fatal outcome. Assuming these rates were the same in other provinces, approximately 33,500 hospitalizations involve an ADR annually in Canada and 1,400 of these lead to death. The Canadian Adverse Events Study reviewed patient charts from admissions in 2000 to four randomly selected hospitals in five provinces and found that 3.1% of hospitalized patients experienced any type of ADR from which one can estimate that the number of deaths due to an ADR in hospital is 3,600.
Conclusions
The frequency of fatal ADRs should be accurately monitored and even low rates of occurrence deserve serious attention. However, public discussion of drug safety and regulatory policies to address drug safety should be informed by rigorous analyses of empirical evidence.
