Evidence and harm: time for reflection

Excerpt

The three authors of the paper A scoping review of unintended harm associated with public health interventions: Towards a typology and an understanding of underlying factors in this issue (Allen-Scott et al. 2014) have provided an extremely useful and insightful service to those in public health who are concerned with the effect of public health interventions on populations. All too often, practitioners of public health have been more concerned with the evidence of effectiveness of interventions than with the potential adverse effect of the interventions (Fuller and Potvin 2012; Kelly and Moore 2012). This is hardly surprising because the scientific training of most practitioners and researchers in public health is designed chiefly to assess the effectiveness of the intervention itself (McQueen and Jones 2007). Assessing effect is the gold standard of institutional efforts such as those of Cochrane (1972), the Community Guide and many other such efforts (Briss et al. 2000; NICE 2008, 2009). It is also a common expectation of individual research efforts (Simoes et al. 2012, Jackson and Dickinson 2011). However, in almost all cases, the rigor expected of scientific research far exceeds the reality of what is obtainable in complex population -based interventions. In reality, most public health interventions are messy, multivariate, loaded with possibilities for error and thus a major underlying problem of such efforts is the common finding of insufficient evidence of effectiveness. While this may not be such a startling outcome for one trained in the social sciences, it is often puzzling for those expecting the rigor of the vaunted RCT. As I have discussed elsewhere there are no simple, quick, or easy solutions to this common outcome (McQueen 2013). There are radical approaches, one could simply only report evidence that is sufficient, a rather draconian measure and ethically questionable. A “softer” solution is to report in depth why evidence could not be found and offer strategies for resolving the situation to arrive at an appropriate ‘better’ answer. Another option that suggests the precautionary principle is to make it clear that there is no evidence that any harm has been done by carrying out the interventions with insufficient evidence and continue to explore the area of insufficient evidence. Clearly, another alternative is to make use of the typologies suggested by Allen-Scott, Hatfield and McIntyre (2014). …