*Geek Box: Control of Bias in the Lyon Diet-Heart Study
The Lyon Diet-Heart Study was a secondary prevention trial of a Mediterranean diet in subjects whom had already experienced a first cardiovascular event, specifically a myocardial infarction. The intervention group had a 73% reduction in risk of cardiovascular mortality or a non-fatal second MI after 5-years. The trial was designed as randomised and single blind.
In recruitment, the investigators used a two-step process where participants signed an original consent form to take part but without knowing that they were signing up to a dietary intervention, and following the consent to participate were then randomised to the intervention or control diet. At this point, subjects in the intervention group only signed a second consent form agreeing to change their diet to the intervention Med diet. This meant that the participants were unaware that they were taking part in a comparative trial.
Further, the physicians assigned to the intervention group were also unaware that their patients were taking part in a comparative trial. This was to avoid ‘physician bias’, whereby the physician may alter the treatment for their patients if they knew they were participating in a comparative intervention study. Regarding the control group, they did not have any dietary evaluation performed until the final year of the 5-year follow-up period in order not to influence their behaviour, as a potential limitation to nutrition intervention studies where the control group are aware they are participating in a trial is that they change/improve their diets.
The LDHS is an example that, notwithstanding methodological issues for nutrition interventions to comply with biomedical standards of double-blinding and placebo control, steps can be taken in the context of free-living, single-blind interventions to minimise the potential for bias toward the intervention group, or to prevent dramatic changes in the control group that undermine the intervention.