Robyn E. Wootton

Who they are

Wootton is a genetic epidemiologist at the MRC Integrative Epidemiology Unit and the School of Experimental Psychology at the University of Bristol, the group that developed Mendelian randomisation as a practical method. George Davey Smith is a co-author on the 2018 paper, which is part of why it carries the weight it does.

The logic of the design is worth spelling out, because it is the answer to the problem the observational literature cannot solve. A cohort study measuring optimism against death has to decide which covariates to adjust for, and a covariate can be a confounder or a step on the causal pathway with no way to tell from the data. Mendelian randomisation sidesteps that by using genetic variants associated with the exposure as instruments: genotype is fixed at conception, so it cannot be caused by later illness or by socio-economic position in the way a questionnaire answer can. Running it bidirectionally tests both arrows at once.

Her team used genome-wide association summary data with participant numbers from 83,198 to 339,224, plus a follow-up in 337,112 UK Biobank participants, against eleven cardiometabolic outcomes including coronary artery disease and myocardial infarction. The result, verbatim: evidence of a causal effect of body mass index on subjective wellbeing was seen, with each one kilogram per square metre increase causing a 0.045 standard deviation reduction in wellbeing, on an interval from 0.084 to 0.006, P = 0.02; and no clear evidence of a causal effect was seen between subjective wellbeing and the other cardiometabolic health measures, in either direction.

Read that the right way round. The only causal arrow the study found ran from body mass index to lower wellbeing. The arrow everybody wants, from wellbeing to heart disease, was not there in either direction, in a well-powered test by the group that built the method.

The construct caveat is not a technicality. The vault searched in both standard and extended mode for Mendelian randomisation studies of optimism, dispositional optimism, the Life Orientation Test, subjective wellbeing, life satisfaction and the wellbeing spectrum against coronary artery disease, myocardial infarction, heart failure, stroke, longevity and all-cause mortality, and found no Mendelian randomisation study of dispositional optimism specifically. So her study is the nearest causal design on the nearest construct, and any parallel drawn from it must say so.

The reconciliation with the one positive genetic result is where this becomes an argument rather than a null. Rukh and colleagues, in 2022, instrumented a wellbeing-spectrum composite built from four traits, life satisfaction, positive affect, neuroticism and depressive symptoms, and found it causally associated with lower risk of heart failure and myocardial infarction. Of the composite's constituent traits, only depressive symptoms showed a positive causal association with either outcome, and the authors' own summary is that the association is mainly driven by depressive symptoms. Subjective wellbeing alone gives nothing; a composite containing depressive symptoms gives something its own authors attribute to the depressive symptoms. The genetic evidence points at the negative pole, exactly as the subscale evidence does.

Our read is the structured abstract and the results paragraph at the BMJ article page. The full body was not read, and nothing beyond those is attributed here.

Key papers

Wootton is a working genetic epidemiologist whose output is journal articles rather than books.

Wootton, R. E., Lawn, R. B., Millard, L. A. C., Davies, N. M., Taylor, A. E., Munafo, M. R., Timpson, N. J., Davis, O. S. P., Davey Smith, G., & Haworth, C. M. A. (2018). "Evaluation of the causal effects between subjective wellbeing and cardiometabolic health: mendelian randomisation study." BMJ 362, k3788, doi:10.1136/bmj.k3788. Structured abstract and results paragraph read in full for this article; full body not read. Author order confirmed independently.

Key ideas

Bidirectional Mendelian randomisation tests both causal directions at once, using genetic variants as instruments that cannot have been caused by later illness.

Body mass index causally lowers subjective wellbeing, at 0.045 standard deviations per unit.

There is no clear causal effect between subjective wellbeing and cardiometabolic health in either direction, across eleven outcomes.

The exposure is subjective wellbeing and not dispositional optimism. No Mendelian randomisation study of dispositional optimism exists.

Taken with the one positive genetic result, which its own authors attribute to depressive symptoms, the causal signal in this area sits at the negative pole.

---