Benefit-risk assessment on farmed salmon

Scope

Purpose

The purpose of the assessment is to evaluate the recommendation given by Hites et al that people should not eat farmed salmon too often because of the pollutant concentrations.

Boundaries

• Pollutants: Toxaphene, dieldrin, PCB
• Nutrients: Omega-3 fatty acids
• Health impacts:
  • Total cancer caused by pollutants
  • Cardiovascular deaths and the impact of omega-3 fatty acids
• Population: the population of Western Europe (European Economic Area as of beginning of 2004)
• Current situation (year 2004)

Scenarios

• Recommendation for not to eat farmed salmon too often (yes/no)
• Setting up new stricter regulations about pollutants in fish feed (yes/no)

Intended users

• The primary users are public health authorities and decision-makers involved in giving food recommendations.
• The secondary user group is the general public. Quantitative estimates are offered to increase understanding of the magnitudes of the related issues.

Participants

The assessment is restricted to a group of environmental health researchers. See participant list in Farmed salmon (project).

Definition

Benefit-risk assessment on farmed salmon: To the assessment page | To the Analytica model
Decision variables: Variable:Recommendation for consumption of farmed salmon | Variable:Pollutant concentration limits for fish feed
Indicators: Variable:Pollutant health risk due to the consumption of salmon | Variable:Net health effects due to the consumption of salmon
Other variables: Variable:Persistent pollutant concentrations in fish feed | Variable:Persistent pollutant concentrations in salmon | Variable:Salmon intake in the population of the Western Europe | Variable:Exposure to persistent pollutants due to salmon in the population of the Western Europe | Variable:Dose-response function of persistent pollutants | Variable:Omega-3 content in salmon | Variable:Omega-3 intake due to salmon in the population of the Western Europe | Variable:Dose-response function of cardiovascular effects of omega-3 fatty acids | Variable:Total mortality in the Western Europe | Variable:Cardiovascular mortality in the Western Europe | Variable:Cardiovascular effects of omega-3 in salmon in the Western Europe

Analyses

• Value of information analyses
  • Outcome: Net health effect; Decision: Recommend restricted farmed salmon consumption?; Variables tested: all
  • Outcome: Net health effect; Decision: Lower limits for pollutants in fish feed?; Variables tested: all
• Importance analyses
  • Outcome: Mortality by recommendation; Variables tested: all
  • Outcome: Mortality by feed regulation; Variables tested: all

Result

Results

The main results and conclusions have been published by Tuomisto et al.

Figure. Benefits and risks (number of avoided deaths per year) of eating farmed salmon in Western Europe.

Table: Net health impacts (avoided deaths per year) of eating farmed salmon in Western Europe. The results include the health impacts of the total health impact of omega-3 intake from salmon.

Statistics or fractile	Business as usual	Recommend restrictions	Stricker rules for feed	Both
Mean	31062	22929	31428	23296
SD	22650	17105	22842	17317
0.01	576	404	653	460
0.025	1782	1286	1847	1356
0.05	3632	2697	3766	2803
0.25	14547	10590	14780	10793
0.5 (Median)	25187	18504	25479	18829
0.75	43064	31526	43628	32091
0.95	76141	56697	77010	57393
0.975	88267	66397	89040	67249
0.99	101832	79328	102202	80143

Table: Net health impacts (avoided deaths per year) of eating farmed salmon in Western Europe, compared with the business-as-usual scenario. The results include the health impacts of the total health impact of omega-3 intake from salmon.

Statistics or fractile	Business as usual	Recommend restrictions	Stricter rules for feed	Both
Mean	0	-8133	366	-7766
SD	0	9703	982	9637
0.01	0	-45895	-2335	-45557
0.025	0	-35624	-1662	-35171
0.05	0	-27745	-1155	-27041
0.25	0	-10929	-60	-10553
0.5 (Median)	0	-4876	264	-4560
0.75	0	-1567	820	-1307
0.95	0	-112	2117	147
0.975	0	-10	2635	459
0.99	0	14	3180	995

Figure. Value of information (VOI) calculated as avoided deaths per year in Western Europe. The first two bars are for the decision about recommending restrictions to salmon consumption. No single variable was important enough to gain any VOI; therefore, only the uncertain variable emphasized by Hites is shown. The other bars are for the decision about setting stricter limits for fish feed.

Conclusions

• Pollutant risk is much smaller than the net health benefit of farmed salmon
• Scientific uncertainties related to recommendations are unimportant
• Some scientific and political uncertainties related to feed limits are important