Meta-analysisindependent cross-checkfull text checked★ key study I-CHM-07 in-depth review
Each month of delay in cancer treatment raises the risk of death
Mortality due to cancer treatment delay: systematic review and meta-analysis.
Overview
Hanna TP et al.. BMJ 2020 · PMID 33148535 · DOI 10.1136/bmj.m4087
For example, every 4-week delay of breast surgery was linked to 8% higher mortality.
Does the study support the specific claim? No · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Quantify harm from delay.
- How was the study done?
Meta-analysis.
- How many people, animals or samples were studied?
1.27 million patients.
- What was actually found?
Higher mortality with delay.
- How large was the observed effect?
HR 1.06-1.08 per 4 weeks (surgery).
- What do the results mean in everyday life?
Do not postpone recommended treatment.
- What are the study's limitations?
Observational, mostly US registry data; valid only for delays of about 4-16 weeks; not suitable for predicting an individual's risk.
- What does it say about Elawady's specific claim?
Safety.
- Do the results apply to Germany and Europe?
High.
- What can explicitly not be concluded from this study?
The effect of delays longer than 16 weeks or of refusing treatment altogether; the effect for every cancer type; proof of causation.
Scientific in-depth review
1. Full reference
- Authors
- Hanna TP et al.
- Journal
- BMJ
- Year
- 2020
- Identifier
- PMID 33148535 DOI 10.1136/bmj.m4087
- Verification
- verified – PubMed esummary + efetch 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
No project-specific funding (authors' personal fellowships/chairs: Ontario Institute for Cancer Research, UKRI ESRC, NIHR, Canada Research Chair). Competing interests: unrestricted research funding from Roche for an unrelated project (TPH); none other.
2. Research question
Mortality per 4-week treatment delay.
3. Study design and population
- Study design
- Meta-analysis – Systematic review and random-effects meta-analysis of retrospective observational studies (28 registry/population-based, 6 institutional)
- Population
Patients receiving curative, neoadjuvant or adjuvant surgery, systemic treatment or radiotherapy for bladder, breast, colon, rectum, lung, cervix, head and neck cancer
- Country
International; 25 of 34 studies from the USA (others Canada, Korea, Taiwan, China)
- Period
Medline, 1 Jan 2000 to 10 April 2020 (bladder 22 April 2020); underlying datasets 1988-2014
- Sample size
34 studies, 17 indications, 1,272,681 patients (study sizes 174-420,792)
- Inclusion and exclusion criteria
English-language full publications reporting HR for overall survival by treatment delay; only 'high validity' studies (prognostic factors age, stage, treatment, comorbidity/functional status described and balanced or adjusted; for colon, lung and cervix also analysis addressing the waiting-time paradox); prostate cancer and intentional delay after rectal neoadjuvant therapy excluded
4. Methods and endpoints
- Exposure / intervention
Treatment delay: diagnosis to first treatment, or end of one treatment to start of next (adjuvant/neoadjuvant intervals)
- Comparator
Shorter delay
- Endpoints
Overall survival
- Statistical methods
HRs converted to per-4-week delay assuming log-linearity; DerSimonian-Laird random-effects pooling; I2; publication bias not tested (too few studies); post hoc sensitivity analysis adding 12 studies lacking comorbidity adjustment
5. Actual results
- Results
Significant association in 13 of 17 indications. Surgery HR per 4 weeks 1.06-1.08 (e.g. head and neck 1.06, 1.04-1.08; breast 1.08, 1.03-1.13; colectomy 1.06, 1.01-1.12; lung 1.06, 0.93-1.19, not significant). Systemic: bladder neoadjuvant 1.24 (1.03-1.50), breast adjuvant 1.09 (1.07-1.11), breast neoadjuvant 1.28 (1.05-1.56), colorectal adjuvant 1.13 (1.09-1.17); NSCLC adjuvant 1.01 (0.99-1.04) and bladder adjuvant 1.04 (0.98-1.11) not significant. Radiotherapy: head and neck radical 1.09 (1.05-1.14), cervix adjuvant 1.23 (1.00-1.50), breast adjuvant after BCS 0.98 (0.88-1.09, single study). No high-validity data for 5 radiotherapy indications or cervical surgery. Sensitivity analysis changed little except breast neoadjuvant systemic treatment.
- Absolute risks
Not reported for patients; authors' model example: a 12-week surgical delay for all breast cancer patients for one year would cause ~1400 excess deaths in the UK, 6100 USA, 700 Canada, 500 Australia (assumes 12% mortality without delay).
- Authors' conclusion
Even a four-week delay of cancer treatment is associated with increased mortality across surgical, systemic and radiotherapy indications for seven cancers; policies minimising system-level delays could improve population survival. Authors state results apply only to delays of about 4-16 weeks, should not be used beyond 16 weeks or for individual risk prediction, and patients should not start treatment before being medically fit.
6. Strengths
- Restriction to studies adjusted for major prognostic factors
- Large registry populations
- Common per-4-week metric across modalities
- Sensitivity analysis on validity criteria
- Discussion of waiting-time paradox and immortal-time bias
7. Weaknesses
- All primary studies retrospective observational: residual confounding by comorbidity/performance status (authors' main limitation)
- Single database (Medline) plus reference lists/Google; English only; no registered protocol reported
- Log-linear dose-response assumption; valid only for 4-16 weeks
- Publication bias not assessed; few studies per indication
- 25/34 studies from the USA (insurance/system factors)
- Immortal-time bias may bias toward null (authors)
8. Risk of bias
- Orienting tool
- AMSTAR 2
- 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
Overall unchanged (some_concerns). A formal AMSTAR 2 confidence rating would likely be 'low' because of no reported protocol and limited search; the direction of the association is consistent and supported by cancer-specific endpoints in 13 of 15 studies. Single-reviewer structured judgement from the full text; a complete formal assessment requires two independent reviewers and the review/study protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| Item 2 protocol (critical) | high | No a priori registered protocol reported; PRISMA followed. |
| Item 4 comprehensive search (critical) | some concerns | Medline only (Ovid) plus reference lists and Google; English only; abstracts excluded. |
| Item 5 duplicate selection | low | Two reviewers screened abstracts (Covidence). |
| Item 6 duplicate extraction | some concerns | Not described in main text. |
| Item 7 excluded studies (critical) | some concerns | Exclusion reasons counted in PRISMA flow (e.g. 100 not high validity); full list not in main text. |
| Item 9 RoB of primary studies (critical) | some concerns | Custom 'high validity' criteria on prognostic-factor control; not ROBINS-I/E; residual confounding remains. |
| Item 11 synthesis (critical) | some concerns | Random-effects pooling reasonable; log-linear conversion an assumption, partly supported by spline analyses in primary studies. |
| Item 13 RoB in interpretation (critical) | low | Authors discuss residual confounding, scope limits and individual non-applicability. |
| Item 15 publication bias (critical) | some concerns | Not assessed because of too few studies. |
| Item 16 COI | low | Declared. |
Full-text review (25 October 2026): Added search limits (Medline only, English), study composition (all retrospective, 25/34 US), full indication-specific HRs incl. non-significant ones, authors' scope limits (4-16 weeks, not individual prediction), modelled excess-death example, AMSTAR 2 item judgements; removed extrapolation to 'months on a diet trial' from statistical interpretation; precision and directness downgraded. Bias overall unchanged (some_concerns).
9. Assessment of the statistical interpretation
Per-4-week HRs of ~1.06-1.13 for most indications imply relative increases in mortality that compound over 4-16 weeks (e.g. +17% for 8 and +26% for 12 weeks of breast surgery delay under the log-linear assumption). Associations, not proven causal effects; authors explicitly warn against extrapolating beyond 16 weeks or to individual patients. The data concern system-level delays in patients who were eventually treated, not treatment refusal.
10. Relevance to the specific claim
Delaying treatment to 'try the system first' carries measurable mortality risk.
11. Independent contrary findings
12. Transferability to Germany and Europe
High.
13. Overall methodological judgement
Key evidence that system-level treatment delays of weeks are associated with higher mortality; for diet-first or refusal scenarios it gives directional, indirect support only.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | indirect |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Effect of delays beyond 16 weeks; causal effect after better control of confounding (e.g. landmark analyses); radiotherapy indications lacking high-validity data.
15. Linked fact checks
- Chemotherapy and tumour-removal surgery do not address root causes; the body will reactivate its 'defensive plan' elsewhere, so these treatments should be avoided. – Studies contradicting or limiting the claim
- Cancer patients should not take chemotherapy, radiotherapy or biological therapy (or should stop them) and follow the Tayebat system instead; 'we are against the word medicine'. – Studies contradicting or limiting the claim
- Cancer will respond to the Tayebat system; a patient with a malignant tumour should start the system now (diet as cancer treatment). – Studies contradicting or limiting the claim