Poor quality data from diet studies may dilute better quality evidenceBMJ 2016; 354 doi: https://doi.org/10.1136/bmj.i4219 (Published 15 August 2016) Cite this as: BMJ 2016;354:i4219
- George D Henderson, research associate
- Human Potential Centre, Auckland University of Technology, Auckland, New Zealand
The hypothesis behind Farvid and colleagues’ study,1 that cancer risk is modified early in life by micronutrients and other factors, is plausible. However, the data collection method—nostalgic re-creation of adults’ childhood diet—invites respondents to invent data.
I vaguely remember vegetables I ate in adolescence, in meals my mother cooked, but I can’t accurately recall fruit consumption. Yet the exclusion criteria …