Ecological studyindependent cross-checkabstract checked★ key study I-CHM-106 standard review
Why rising cancer survival rates can mislead
Are increasing 5-year survival rates evidence of success against cancer?
Overview
Welch HG et al.. JAMA 2000 · PMID 10865276 · DOI 10.1001/jama.283.22.2975
Survival gains over time matched rising diagnosis rates, not falling death rates.
Does the study support the specific claim? Partly · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Check whether survival trends reflect real progress
- How was the study done?
Analysis of US cancer statistics 1950-1995
- How many people, animals or samples were studied?
20 tumour types
- What was actually found?
Survival changes tracked diagnosis, not deaths
- How large was the observed effect?
r=0.00 with mortality; r=+0.49 with incidence
- What do the results mean in everyday life?
Earlier diagnosis can make survival look longer without people living longer
- What are the study's limitations?
Ecological, old data
- What does it say about Elawady's specific claim?
Partly supports scepticism about survival statistics; does not show early diagnosis shortens life
- Do the results apply to Germany and Europe?
Moderate
- What can explicitly not be concluded from this study?
Cannot judge any specific treatment
Scientific in-depth review
1. Full reference
- Authors
- Welch HG et al.
- Journal
- JAMA
- Year
- 2000
- Identifier
- PMID 10865276 DOI 10.1001/jama.283.22.2975
- Verification
- verified – PubMed efetch 2026-10-26 (PublicationType and CommentsCorrections checked); Crossref 2026-10-26
- Retracted
- no
- Corrections
- none
- Extent checked
- abstract only
- Conflicts of interest and funding
US government support (non-PHS); COI not stated in abstract.
2. Research question
Do changes in 5-year survival over time track changes in cancer mortality?
3. Study design and population
- Study design
- Ecological study – Ecological analysis of SEER statistics, tumour as unit of analysis
- Population
US population (SEER), 20 most common solid tumours
- Country
USA
- Period
1950-1995
- Sample size
20 tumour types
- Inclusion and exclusion criteria
20 most common solid tumours
4. Methods and endpoints
- Exposure / intervention
Change in 5-year survival
- Comparator
Change in mortality and incidence
- Endpoints
Correlation of survival change with mortality and incidence change
- Statistical methods
Pearson and Spearman correlation
5. Actual results
- Results
5-year survival increased for all 20 tumour types (by 3% pancreas to 50% prostate); mortality fell for 12 and rose for 8; correlation of survival change with mortality change r=0.00 (Spearman -0.07); with incidence change r=+0.49 (Spearman +0.37).
- Absolute risks
not applicable
- Authors' conclusion
5-year survival is valid for comparing therapies in randomised trials, but changes over time bear little relation to mortality and mainly reflect changes in diagnosis.
6. Strengths
- Long period, population data
- Simple transparent method
7. Weaknesses
- Ecological; 20 data points
- Does not assess treatment effects directly
- Data end 1995
8. Risk of bias
- Orienting tool
- none applicable
- Type of assessment
- simplified assessment (abstract only)
- Full formal assessment
- no
- Overall risk of bias
- not assessable
- Rationale
Ecological analysis; not a test of treatment efficacy.
9. Assessment of the statistical interpretation
Correlation coefficients on 20 tumour types; no CIs in abstract.
10. Relevance to the specific claim
Supports a narrow part of C-CHM-13: population survival trends can be inflated by earlier diagnosis (lead-time) and overdiagnosis. It does not show that treatment is useless; the authors state survival is valid within randomised trials.
11. Independent contrary findings
Randomised trials with mortality endpoints (I-CHM-107; I-CHM-01, I-CHM-02).
12. Transferability to Germany and Europe
Applicable to European and Egyptian survival statistics in general.
13. Overall methodological judgement
Correct methodological caution about survival statistics, not evidence against treatment.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | not assessable |
| Precision of results | low |
| Directness | indirect |
| Reproducibility | replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
None.
15. Linked fact checks
- Oncology only buys time: 'survival rates' are meaningless, patients die during treatment, and early diagnosis only makes you 'die early'. – Studies with mixed or unclear results