Risk assessmentindependent cross-checkfull text checked★ key study I-EGG-15 in-depth review
EFSA: Salmonella in poultry flocks
Salmonella control in poultry flocks and its public health impact.
Overview
EFSA BIOHAZ Panel. EFSA J 2019 · PMID 32626222 · DOI 10.2903/j.efsa.2019.5596
Overall, laying-hen flocks kept outside cages tend to have less Salmonella, but studies conflict; crowding, large farms and stress clearly increase it.
Does the study support the specific claim? Yes · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Contributory factors and control options for Salmonella in EU poultry
- How was the study done?
EFSA scientific opinion with attribution modelling and risk-factor reviews
- How many people, animals or samples were studied?
Not applicable; EU poultry flocks
- What was actually found?
Evidence 'points to' lower Salmonella in non-cage than cage flocks (e.g. Austria 34.4% vs 7.9% of flocks), but some studies found the opposite and the cause is unclear. Crowding, large farms and stress conclusively increase Salmonella. A 1% target would cut layer-related human cases by 53.4% (about 254,400 cases).
- How large was the observed effect?
Attribution modelling; reviews
- What do the results mean in everyday life?
Better housing and hygiene reduce Salmonella.
- What are the study's limitations?
Housing studies mostly older than the 2012 cage ban; confounded by flock size and density; incomplete national data; model estimates with wide uncertainty
- What does it say about Elawady's specific claim?
Supports the direction (cages worse on average) but not the specific '44% vs 3.2%' figures or a large, consistent difference.
- Do the results apply to Germany and Europe?
Directly EU.
- What can explicitly not be concluded from this study?
No egg-level risk numbers for consumers; cannot separate the cage effect from flock size and stocking density.
Scientific in-depth review
1. Full reference
- Authors
- EFSA BIOHAZ Panel
- Journal
- EFSA J
- Year
- 2019
- Identifier
- PMID 32626222 DOI 10.2903/j.efsa.2019.5596
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
EFSA scientific opinion requested by the European Commission; prepared by the BIOHAZ Panel and working group under EFSA's declarations-of-interest policy; no specific conflicts reported in the full text
2. Research question
Contributory factors and control options for Salmonella in EU poultry
3. Study design and population
- Study design
- Risk assessment – EFSA BIOHAZ Panel scientific opinion (mandate from the European Commission): literature reviews on risk factors (ToR 3-5), descriptive analysis of EU monitoring data 2014-2016 (and MS housing data 2016-2017), and Bayesian Salmonella source attribution model (EFSA_SAM) for 23 Member States
- Population
EU poultry flocks
- Country
EU (23 MSs in attribution)
- Period
Monitoring data 2014-2016 (attribution baseline 2016; housing-system data 2016-2017); opinion adopted 2019
- Sample size
Attribution: 28 of 944 serovars covering 65,087 (85%) of 76,284 reported human cases; housing data from 19 countries (6 usable for mixed model); literature reviews (no meta-analysis because most studies did not report OR/RR)
- Inclusion and exclusion criteria
Not applicable
4. Methods and endpoints
- Exposure / intervention
Housing system (conventional cage, enriched cage, barn, free-range, organic), outdoor access, stocking density, farm/flock size, stress; target serovars and flock prevalence targets
- Comparator
Cage vs non-cage; conventional vs enriched cages; 2016 situation vs 1% target scenario
- Endpoints
Salmonella occurrence/persistence in laying hen and broiler flocks; attributed human salmonellosis true cases (corrected for under-reporting)
- Statistical methods
Bayesian source attribution with 95% credible intervals; multivariate mixed model on MS housing data; qualitative weight-of-evidence synthesis with structured uncertainty tables
5. Actual results
- Results
Achieving a 1% target in laying hens would reduce layer-associated human salmonellosis true cases by 254,400 (CrI95 98,540-602,700), i.e. 53.4% (CrI95 39.1-65.7) of layer-associated and 6.2% of all estimated 4.08 (2.22-7.39) million true cases. Of 2.19 million estimated S. Enteritidis cases, 40.9% attributed to broilers, 33.2% pigs, 19.8% layers, 6.1% turkeys. Housing: 'Conflicting evidence ... Overall, evidence points to a lower occurrence in non-cage systems compared to cage systems'; whether due to housing itself, furniture, break in infection cycle or reduced stocking density is unclear. Examples cited: Austrian baseline survey 34.4% of cage vs 7.9% of non-cage flocks positive (barn 15.3%, free-range 7.0%, organic 1.4%); French study 30.9% caged vs 7.9% on-floor; UK non-cage OR 0.14 (0.04-0.49); some studies found the reverse (e.g. Romanian retail eggs: Salmonella in albumen of 33% organic, 28% free-range, 25% aviary, 14% cage eggs). MS data 2016-2017: Netherlands 4.5% cage vs 0.68% non-cage S. Enteritidis-positive flocks; Czech Republic and Italy similar. Outdoor access and conventional vs enriched cages: inconclusive. Conclusive evidence that higher stocking density, larger farms and stress increase Salmonella occurrence, persistence and spread in laying hens. The figures '44% vs 3.2%' do not appear in this opinion.
- Absolute risks
Modelled: 254,400 (CrI95 98,540-602,700) fewer layer-associated true human cases per year under a 1% target vs 2016, out of about 4.08 million estimated true cases in 23 MSs
- Authors' conclusion
Retain S. Enteritidis, S. Typhimurium (incl. monophasic) and S. Infantis as targets; an all-serovar target would be more effective. A 1% layer target would cut layer-associated cases by about half. Evidence on cage vs non-cage conflicts but overall points to lower occurrence in non-cage systems; causal mechanism unclear; outdoor access and enriched vs conventional cages inconclusive; stocking density, farm size and stress conclusively increase Salmonella.
6. Strengths
- Authoritative EU assessment with explicit uncertainty analysis
- Combines literature, EU monitoring data and attribution modelling
- Transparent about conflicting evidence and data gaps
7. Weaknesses
- Housing evidence largely from studies before the 2012 EU ban on conventional cages
- Cage effect confounded by flock size, stocking density, cleaning and infection history (EFSA states the cause is unclear)
- No meta-analysis; heterogeneous studies; operator sampling up to 10-fold less sensitive than official sampling
- MS housing data incomplete (only 3 countries had cage and non-cage data for the mixed model)
- Attribution model excludes 5 MSs and depends on under-reporting multipliers
8. Risk of bias
- Orienting tool
- No standard tool fits (regulatory risk assessment / expert opinion); structured judgement of validity threats
- Type of assessment
- structured assessment based on the full text by a single reviewer (no formal double review)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Single-reviewer structured judgement based on the full text; not a formal assessment (would need two independent reviewers). Overall changed from not_assessable to some_concerns because the full text allows judging the underlying evidence.
| Domain | Judgement | Rationale |
|---|---|---|
| Evidence identification | low | Documented systematic searches (Web of Science, PubMed, Scopus updates) with predefined risk-factor categories. |
| Primary-study validity (housing comparison) | high | Mostly observational, older, heterogeneous studies with confounding by flock size, density and hygiene; conflicting directions. |
| Data quality (monitoring) | some concerns | Incomplete serotyping and housing reporting, sampling sensitivity differences between operators and officials. |
| Modelling (attribution) | some concerns | Bayesian model with uncertainty propagated; wide credible intervals; assumptions on under-reporting and excluded MSs. |
| Interpretation | low | Conclusions cautiously worded and consistent with the evidence; uncertainty tables provided. |
Full-text review (25 October 2026): Added EFSA wording ('conflicting evidence', mechanism unclear), example prevalences (Austria 34.4% vs 7.9%), counter-examples, attribution details with CrIs; noted 44% vs 3.2% not in the opinion; consistency set to mixed; structured validity domains; bias overall changed from not_assessable to some_concerns.
9. Assessment of the statistical interpretation
Attribution results are modelled estimates with wide credible intervals (e.g. 98,540-602,700). Cage vs non-cage evidence is qualitative weight of evidence; prevalence contrasts come from individual studies and are partly confounded.
10. Relevance to the specific claim
C-EGG-09 (supports cage effect).
11. Independent contrary findings
Not applicable.
12. Transferability to Germany and Europe
Directly EU.
13. Overall methodological judgement
Authoritative EU opinion supporting the direction of the cage claim (lower Salmonella in non-cage laying hen flocks overall) but describing the evidence as conflicting, mechanism unclear and older than the 2012 conventional-cage ban; it does not contain the '44% vs 3.2%' figures and does not support 'far more common' as a general statement.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Enriched vs conventional cages; current post-2012 EU data by housing; egg-level (not flock-level) contamination by housing system.
15. Linked fact checks
- Salmonella is far more common in caged egg production than in cage-free systems (44% vs 3.2%). – Studies supporting the claim