Choice of a standard of proof in public health - how much evidence is enough to justify adoption of an intervention or policy change - is perhaps the central conceptual question in public policy on health inequalities. As a matter of values and priorities, there is no algorithm for doing this choice nor can it be made on scientific grounds. It is high time we foreground this point not only in discussions of public policy on health inequalities, but also in the training of researchers, practitioners and policy specialists. Understanding standards of proof must be recognized as a core professional competence, and incorporated into the relevant curricula.
Professor of Global Health Policy, Newcastle University, UK
Co-editor, Journal of Public Health
What Public Health Policy Can Learn from the Murders of Nicole Brown Simpson and Ron Goldman
For readers of tender years and no immediate access to Wikipedia, Nicole Brown Simpson was the ex-wife of former (US) football superstar O.J. Simpson, who in 1994 was accused of murdering her and a companion, Ron Goldman. Mr. Simpson was acquitted of criminal charges in the deaths, but later found liable in a civil suit brought by the victims’ families for wrongful death damages. This outcome is superficially perverse, but legally plausible in common law countries because the standard of proof in criminal proceedings – every element of the offence must be proved beyond a reasonable doubt – is more demanding than in civil proceedings, where the standard of proof refers to the balance of probabilities or a preponderance of the evidence. This reflects a judgement about the dangers of being wrong in different kinds of ways: it is considered more objectionable to convict an innocent defendant than to hold that the same defendant must pay damages to those who have been wronged, or their survivors.
In public health, application of an analogy with legal proceedings is nothing new. Forty years ago, environmental economist Talbot Page published a critically important article about a class of hazards he called environmental risks, like certain kinds of toxic discharges, with such characteristics as incomplete knowledge of mechanism, long latency periods and serious if not irreversible effects. Scientific research on these hazards, he pointed out, is organized around limiting ‘false positive’ results – for example, inferring a causal relationship where none in fact exists. In the case of criminal trials, limiting false positive results (that is, convictions of the innocent) is theoretically embodied in the principle of requiring proof beyond a reasonable doubt. However, Page argued that the approach of limiting false positives may be inappropriate in environmental policy because of the consequences of failing to take action to control a hazard that is later found to cause, say, cancer or birth defects with lifelong consequences. “In its extreme,” wrote Page, “the approach of limiting false positives requires positive evidence of ‘dead bodies’ before acting”.
Unfortunately, this is not an extreme. With a few exceptions, that has been the default position in public policy with regard to health inequalities and social determinants of health. As David Stuckler and Sanjay Basu have pointed out, “in countries where austerity is ascendant, we’re undergoing a massive and untested experiment on human health, and left to count the dead.” Exactly. The logic is that of industrial employers who insist that only findings from prospective epidemiological studies will justify regulating workplace processes that give their workers cancer. In 10 or 20 years, longitudinal studies of populations exposed to austerity may provide stronger evidence than is now available of the long-term health impacts of the attack on social provision and the targeting of poor communities that have been the centrepieces of post-2010 social policy in the UK, or the youth unemployment rates of 40-50 percent that are the legacies of the financial crisis and subsequent austerity in Italy, Spain and Greece. Vindication of what many of us have been saying all along will be cold comfort to the survivors.
Choice of a standard of proof – how much evidence is enough to justify adoption of an intervention or policy change – is perhaps the central conceptual question in public policy on health inequalities, and the contrasting positions can be dramatic. As noted in my previous posting, in 2010 Sir Michael Marmot and colleagues wrote: ‘It is hard to see how even ideologically driven commentators could think that having insufficient money to live on is irrelevant to health inequalities’. On the other hand, a senior fellow of the somewhat left-of-centre Brookings Institution recently wrote in the Journal of the American Medical Association that: “The evidence on social determinants of health is growing, but is still insufficient to convince many key decision makers. For instance, there is good research on the link between such housing problems as mould or substandard accommodations and health, and between family or social ‘toxic’ stress and long-term mental health and other patterns. But purported linkages between health and other social conditions, such as general poverty, lack reliable evidence. Much more basic research is needed to understand the key determinants.”
As in the legal system, the choice of a standard of proof in public health cannot be made on scientific grounds. There is no algorithm for doing so. It is a matter of values and priorities – in the first instance, those of the ‘key decision-makers’ referred to, in a decontextualized way, in the JAMA article. One must ask: do they really think that living in poverty or relying on low-wage, precarious employment are unrelated to the possibilities for living healthily? That such situations don’t involve toxic stress? (The work of Linda Tirado and Barbara Ehrenreich, among others, is eloquent on this point.) Sadly, this lack of comprehension is shared by many of our professional colleagues. The reply to an article I published on this issue in 2013 described me as “noisy” in the second paragraph – completely missing the analytical point, as do many others.
Forty years on from the appearance of Page’s landmark article, it is time to foreground this point not only in discussions of public policy on health inequalities, but also in the training of researchers, practitioners and policy specialists. Understanding standards of proof must be recognized as a core professional competence, and incorporated into the relevant curricula.
Author’s apology: a few of the hyperlinks in this posting may lead to articles that are behind a paywall. If you have trouble accessing these, please contact me and I will ensure that you obtain a copy.