Dr Anna Bergström, PhD, MSc Epidemiology

Dr Anna Bergström reads the methods section first and the abstract last, a habit she picked up in an epidemiology department and has never seen a good reason to drop. In most fields the two agree. In this one they frequently do not, and the gap between what a study did and what its summary says it showed is where nearly every misleading claim about the arts and health begins.

She is a public health researcher with a PhD and an MSc in epidemiology, and her work is in evaluation methods rather than in the arts. She came to this site through the back door: not as an enthusiast for creative health, but as somebody who kept encountering the same handful of figures in policy documents and could not find the studies underneath them. Her role here is deliberately narrow and, on a site like this one, load bearing. She checks the arithmetic.

In practice that means she goes through every efficacy claim on the site and confirms that the design, the sample size, the effect size, the confidence interval and the certainty rating are described as the source describes them, in the source’s own units. She is the reason a gait speed result is quoted in metres per minute here rather than converted into a livelier looking number, the reason a nested count of studies inside reviews is not flattened into two impressive figures side by side, and the reason a subgroup finding at one timepoint is never reported as a whole sample result. Her hardest rule is also the simplest: an uncontrolled before and after evaluation is not a trial, is never called one on this site, and does not become one by being cited in a government report.

She owns the three tier claim ladder set out in the editorial policy and signs off the pages where the evidence is weakest or most contested, including the methods explainer that the rest of the site links to.

The point she asks to have made whenever the tiers are explained is the one most easily lost. “No adequately powered controlled evidence” and “it does not work” are different statements, and treating them as the same is a mistake in the opposite direction from the one this site usually corrects. A Tier 3 label means the question has not been answered properly yet. Plenty of things people find genuinely valuable sit there, and some of them will move up when somebody finally runs the study.