By Howard White, CEO
As my family and friends know, I’ve been a runner all my life. Now, after a knee replacement at the end of last year, I am working on getting back some of the speed I have lost over the years. I am using evidence to do so.
Studies show that to run faster we take longer steps, not more steps per minute. I tried this out myself last week. My daughter Laila and I went out to run 200m at different speeds. It was absolutely true: my cadence (steps per minute) was actually a little bit lower in my fastest run compared to my slowest. But my step length increased by over 60 percent from 1.19m to 1.90m per step.
But here’s the thing. Studies also show that people slow down as they get older because their stride length gets shorter, whereas cadence hardly goes down at all. That’s not been true for me. My cadence now is around 155-165 steps per minute, compared to 180-185 in my prime. There may be some easy gains in speed from bringing my cadence back up.
So how to increase cadence? One way is ‘running to the beat’ by playing music at the desired cadence. Search ‘180 bpm’ on YouTube or Amazon Music and there’s a playlist for you.
And does running to the beat work? Yes. I looked at three studies, such as this one. They all found that it is possible to retrain cadence to have people run faster or slower than their pre-study natural cadence. Two of the studies disaggregated by gender, you can find the studies here and here. They both report that running to the beat is more effective for women then men. So, echoing the opening of Trish Greenhalgh’s book, How To Read a Paper, I want to know ‘Will it Work for Me?’. I suspect not. Coaches across the decades have tried to improve my running form to no avail.
The finding that running to the beat improves cadence is not transferable to me, because the effect is less with men than women and because of my general lack of coordination.
Do the findings apply to me?
Gender differences in effects are not restricted to running to the beat. They occur across all sectors. But they are frequently ignored.
It is often said that the evidence-based approach has made most progress in medicine because how a human body reacts to a drug is a universal, and so generalisable, finding. This is simply incorrect. There are many treatments for which effects are different for men than women, for example the benefits of aspirin for cardiovascular disease are greater for men than women. And there are sound physiological reasons to expect women to react differently to men. The established differences in the effects of drugs by gender and race are the basis of the rise of personalised medicine.
Yet, a review of 86 clinical trials published in prominent medical journals found that of the 50 studies including both men and women, women are under-represented, accounting for just over a third of those enrolled in the trial. And three-quarters of the studies did not report effects disaggregated by gender. Another review focused on long-acting antipsychotic drugs has similar findings. Of the 40 included studies just over one-third (36%) of patients were women and only 6 out of 40, that is 15%, reported results separately for men and women.
This bias in research is highlighted in Caroline Criado Perez’s book Invisible Women: Exposing Data Bias in a World Designed for Men. Because of these biases, women are: 47% more likely to be injured in a car accident, regularly prescribed drugs which will, at best, have no benefit and may do harm, and use appliances like cell phones which are too big for them.
Introducing the middle-range theory approach
Why is any of this relevant to Campbell?
Campbell reviews summarise the relevant global evidence on a policy or practice. This global body of evidence is intended to inform decisionmaking by national and local government, as well as in non-governmental agencies. But they always want to know, ‘Do these findings apply to me?’ Or even, ‘But these findings don’t apply to me because…’.
In comes middle-range theory. Campbell is a consortium member of the Centre of Excellence for Development Impact and Learning (CEDIL). CEDIL has a programme of work exploring the use of middle-range theory to inform the use of evaluation findings and the design of evidence-based programmes.
Middle-range theory sits between general theory, which is too abstract to be of practical use, and the nitty gritty details of a project specific theory of change. The approach focuses on the causal mechanisms by which a general category of interventions is expected to work.
CEDIL has just published a discussion paper, and accompanying methods brief, authored by a team led by the philosopher Nancy Cartwright, a member of CEDIL’s Intellectual Leadership Team. Nancy lays out a 10- step approach for developing a middle-range theory which includes looking at the barriers and facilitators – which she calls derailers and supporters – for the various causal mechanisms to operate as well as safeguards which may be in place against the derailers.
An equity extension
Though I was closely involved in Nancy’s work, I realise now that we perhaps need make an 11th step explicit. That is, to ensure that we have an equity-augmented middle-range theory. An equity-augmented theory explicitly addresses heterogeneity in how causal process may operate on account of gender, race, class, location, disability or other aspects of socio-economic status. Whilst these factors may be captured as barriers or moderators, they may be overlooked unless there is the explicit intent to address them.
One of the examples used by Nancy and her colleagues is a conditional cash transfer for parents to send their children to school. It is readily apparent that the causal process by which this incentive payment encourages the action may have a gender differential if households place less value on education for girls, or even are actively opposed to it or live in an area where it is discouraged. Similarly, as the paper points out, the process won’t work if there is no school available, which may well be the case for remote communities. Or parents of ethnic minorities may not send their children either because they are subject to teasing and bullying, or they do not speak the language used in the school and school textbooks.
Some years ago I worked with our Editor in Chief Vivian Welch and others, to produce an equity extension to PRISMA which recommends incorporating analysis of disadvantaged groups into reviews. Vivian is also Co-convenor of Campbell’s equity group, run jointly with Cochrane. The aim of this group is to ‘encourage authors of Campbell and Cochrane reviews to include explicit descriptions of the effect of the interventions not only on the whole population but to describe their effect upon the disadvantaged and/or their ability to reduce socioeconomic inequalities.’ In 2018, members of this equity group assessed sex/gender analysis in Campbell and Cochrane reviews. Only 20 out of 137 Campbell reviews published up to September 2017 conducted an analysis across sex/gender.
To further promote equity analysis, we shall be embarking on an equity review of all Campbell systematic reviews published to date. That is, do they report disaggregated effects by ethnicity, education or other dimensions of social disadvantage? And does their discussion of how the intervention is meant to work discuss possible heterogeneity by gender or other factors? For a sample of these reviews, we shall examine the included studies to see if they have such an analysis which has been ignored by the review authors. If you would like to be involved, or financially support this work, please get in touch: info@campbellcollaboration.org.
Now, I’m off to do some cadence-increasing drills.
A longer version of this blog can be found on my running blog.