Bias is something that affects us all; its all-pervasive nature means it influences everything we do from supermarket choices through to study design. But what of perception? Bias might lead the result of our study to be skewed or flawed, but if our perception of where the issue lies is incorrect we may select the wrong thing to study. So, how wrong can our perception be? Well, very, according to IPSOS MORI and their annual Perils of Perception report for 2015.[1]

This report polls members of the public in 33 countries on their understanding of key issues affecting their nation. The results show a significant gap between perception and reality across a number of issues that specifically relate to society and health in Great Britain.

Our perception of the distribution of wealth was one of the most distorted views. When asked what proportion of wealth the top 1% of the population own the guess was 59%, more than twice the true figure, which is 23%.

On immigration the perception of Britons is that 25% of the population are immigrants, nearly double the actual figure of 13%.

We also know we have an ageing population, but perhaps not to the extent we believe. The estimate of the average age of the population was 51 years old, when it is in fact 40.

With regard to obesity we may be complacent; the average estimate of the proportion of people over the age of 20 who are overweight or obese was 44% when it is in fact 62%.
 
And before you think that the Great British Public are better or worse than elsewhere, we are not. Ranked 16^th of 33 countries in IPSOS MORI's provocatively titled “Index of Ignorance” we are firmly in mid-table. If you are reading the blog from either Ireland or South Korea (ranked 27^th and 28^th) you potentially have a better perception of issues affecting your nation than if in Mexico, India and Brazil who occupy the top 3 positions. But this is not an issue that is delineated along boundaries of low-, middle- or high-income countries in case you were to infer that from the results: New Zealand is ranked at number 5 and Belgium at number 7.

So this is all good fun, interesting stuff, but what does it mean? Well certainly not that we should quietly reassure ourselves that we would have been much closer to the real figure than most of the population. Bias and perception issues are at their most insidious when we fail to acknowledge that we may be subject to them.

These findings are in fact an endorsement of the way, as CLAHRCs, we structure what we do. By bringing together academics, patients and those involved in delivering care, we challenge each others perceptions of the issues related to service delivery. That way we can work collaboratively to solve that issues that are in fact real issues, rather than those which we perceive to be the issue.

-- Paul Bird, CLAHRC WM Head of Programme Delivery (Engagement)

In the early days of CLAHRCs, 'getting evidence into practice' was an important objective. We set about closing the T2 gap and used implementation science to get doctors to prescribe evidence-based care, dentists to use tooth protecting resins, and nurses to make regular observations. That is to say, we were concerned with how to make practitioners comply with standards over which they had complete jurisdiction. Theories of individual behaviour change were invoked, and rather then choose a theory on the basis of its impressive sounding title (e.g. prospect theory, social network theory), a framework was developed to identify barriers and facilitators of change.[1]

But practitioners increasingly follow the evidence when it is compelling and when the evidence-based standard is in their gift.[2] So, the big (and much more interesting) problem now is how to change the service in a generic way rather than simply to increase performance on a specific measure - we are becoming more concerned with draining the swamp than zapping individual mosquitoes.[3] In our CLAHRC we recently evaluated a compound (multi-component) intervention to improve home dialysis rates, having promulgated a guideline supporting improved access to such a service. We showed that agreement with the proposed change among stakeholders, an agreed implementation plan, managerial support, and product champions all facilitated the success of the intervention in taking West Midlands from the worst to the best performing region in England. However, the king of all intervention components was a financial incentive.[4] Fulop and colleagues have now published a similar multi-methods evaluation of an arguably even more complex intervention to improve access to acute stroke care.[5] The findings are very similar, save that we found more emphasis on financial incentives and also more problems in communication with patients; something that would perhaps not stand out in the hyper acute stroke context. The Fulop paper is an advance on ours in (at least) two respects. First, they compare and contrast across two regions/CLAHRCs and I always think controls should be used if possible; even one is better than none. Second, they illustrate the causal model with diagrams that make the theoretical framework they are using clear, a practice that is helpful in communicating the very real distinctions between the intervention as planned, its implementation/adaption, its upstream effects (e.g. staff knowledge/morale), its downstream effects (at the patient 'level'), and the context in which all takes place.[6] People muddle these concepts and hence fall over their feet, but Fulop and colleagues have shown themselves to be sure-footed!

-- Richard Lilford, CLAHRC WM Director

Genetically modified rice – called ‘golden rice’ – can increase yields and, since it produces beta-carotene, can prevent the sequelae of vitamin A deficiency that is common in those with a predominantly rice-based diet. For an interesting article on the controversy over use of this GM crop in Bangladesh, and its potential costs-benefit, please read Uttam Deb’s article from the Copenhagen Consensus Center.[1]

-- Richard Lilford, CLAHRC WM Director

Leave a comment

Consider an intervention or event that has occurred in a single site, such as one state in the US. The standard method for evaluation of the effects would be to compare aggregate outcomes across the intervention site and control sites. The synthetic control method is a refinement of this standard approach. Here pre-intervention data from controls is used to determine a weighted average of the controls that reproduces the pre-intervention trend in the intervention site.[1] In this way the outcome observed in the intervention sites is compared with the intervention that would have been expected in the intervention sites if it behaved in the same way as the synthetic control. The synthetic control can be considered the counter-factual outcome at the intervention site. The choice of controls is not left to the investigator and the difference in expected outcomes between the intervention and the control units is made explicit. The method reminded me of propensity scoring in clinical epidemiology, in the sense that both methods seek to provide an unbiased estimate of the treatment effect.

-- Richard Lilford, CLAHRC WM Director