Failures in this BMJ article:
Used unreliable survey data
Treatment was self-directed, self-selected, self-administered with no confirmation and no supervision; this was not a randomized controlled blinded study
Never administered fish oil
Never qualified product potency, purity, absorption
No dose-response relationship can be reported
Never quantified compliance, effectiveness
Never measured n3 index (ratio of n3 to n6)
Reported Risk/Hazard ratios are small and conflicting
Questionable funding by [injection] maniac
Unusual disease model that proved questionable/unreliable
