Oncology · Medical Literature
[Correspondence] Representativeness of the Cancer Survival Index in Colombia
Abstract
We read with interest the Article1 by Claudia Allemani and colleagues, which provides a landmark contribution to global cancer surveillance. Nevertheless, we have concerns regarding the representativeness of the Cancer Survival Index reported for Colombia: 72·7% for all cancers and 81·2% for the six WHO tracer cancers.
Clinical relevance
Do the Cancer Survival Index (CSI) figures reported for Colombia in the CONCORD-4 study (72·7% for all cancers combined; 81·2% for the six WHO tracer cancers) accurately represent the cancer survival experience of the Colombian national population, given the sub-national and geographically limited coverage of the cancer registries contributing to that estimate? __SNAPSHOT__ {"studyDesign":"Published correspondence (letter to the editor) critiquing methodological representativeness of a landmark global cancer surveillance analysis (CONCORD-4); not a primary study","setting":"Colombia; critique directed at the CONCORD-4 programme, which derived the Colombian CSI from sub-nationally representative population-based cancer registries (PBCRs)","sampleSize":"Not reported in this correspondence; the index figures (CSI 72·7% all cancers; CSI 81·2% for WHO tracer cancers) originate from the CONCORD-4 article by Allemani et al. (Lancet, 2026)","population":"Children and adolescents with cancer in Colombia, as captured by participating PBCRs and the VIGICANCER surveillance network; the correspondence questions whether these sources adequately represent the full national paediatric and general cancer population","intervention":"Not applicable (observational surveillance correspondence); the exposure of interest is sub-national registry coverage and its effect on the validity of a national-level Cancer Survival Index","comparator":"Not applicable; the correspondence contrasts the CONCORD-4 CSI against independent data from VIGICANCER and Colombian PBCRs to highlight potential discordance","primaryOutcome":"Representativeness of the Colombian CSI (all-cancers and WHO tracer cancer 5-year survival estimates) as derived from CONCORD-4 registry data","followUp":"Not reported","fundingSource":"Not reported in the abstract; VIGICANCER has received support from the My Child Matters Program and the Keira Grace Foundation per prior publications"}
Key findings
The correspondence raises concern that the CONCORD-4 CSI figures for Colombia — 72·7% (all cancers) and 81·2% (six WHO tracer cancers) — may overestimate true national cancer survival because the contributing registries cover only a small fraction of the Colombian population. In CONCORD-3, four participating Colombian PBCRs together covered approximately 9% of the total national population, a figure that improved upon the 6·9% covered in CONCORD-2 but remained substantially below the coverage achieved in most high-income countries (e.g., 85·6% in the USA, 21·7% in France). The VIGICANCER network, developed specifically to address this surveillance gap for childhood cancer, currently operates across 27 institutions in 10 Colombian cities and is estimated to capture approximately 55%–60% of all childhood cancer cases diagnosed nationally; nonetheless, VIGICANCER itself acknowledges that its lack of complete national coverage may affect representativeness. Prior real-world data from VIGICANCER and the Cali Population-based Cancer Registry recorded 5-year overall survival figures of approximately 51·7%–52·0% for children and 39·4%–44·0% for adolescents — substantially lower than the CONCORD-4 national CSI figures — underscoring the concern that registry-captured populations may reflect higher-resourced, urban care settings rather than the country as a whole. A well-documented discrepancy exists between survival estimates derived from population-based sources and those from administrative or hospital-based systems (e.g., the Colombian Ministry of Health's SIVIGILA, which showed >50% case underreporting compared to VIGICANCER in Cali). No new statistics, p-values, confidence intervals, or effect sizes are reported in this correspondence beyond the two CSI figures attributed to Allemani et al. (2026).
Practice implications
This correspondence does not present new empirical data but raises an important epidemiological validity concern: national-level cancer survival indices derived from sub-nationally representative registries may systematically overestimate true population-wide survival if the covered areas disproportionately include urban, better-resourced centres with higher-quality care and more complete case ascertainment. For Colombia specifically, the available evidence suggests that PBCRs historically covering approximately 9% of the national population are concentrated in cities such as Cali, Pasto, Manizales, and Bucaramanga — settings unlikely to reflect outcomes in rural or lower-resourced regions. The broader concern applies to all LMIC country estimates in CONCORD-4: survival indices must be interpreted with explicit acknowledgement of registry coverage denominators. The CONCORD-4 figures for Colombia should not be used for national benchmarking, health system performance evaluation, or policy target-setting without adjustment for or transparent disclosure of the partial population coverage. The population inadequately represented includes patients in rural Colombia, those receiving care through adult oncology services (particularly adolescents), and those in regions without an accredited PBCR. Clinicians and policymakers referencing these figures should treat them as registry-catchment estimates rather than as national statistics. Only the abstract of the correspondence was available for this summary; full text may contain additional data, referenced sources, or recommendations not captured here. __LIMITATIONS__ This is a correspondence letter — the shortest and lowest-evidential form of peer-reviewed publication; no primary data, statistical analyses, or formal methodology are presented.,Only the abstract was available; the full text may contain supporting data, quantitative arguments, or referenced local registry data not reflected in this summary.,The magnitude of potential bias in the CONCORD-4 Colombian CSI is not quantified; the correspondence identifies the direction of concern (likely overestimation) but provides no adjusted alternative estimate.,VIGICANCER itself acknowledges incomplete national coverage (approximately 55%–60% of childhood cases, with gradual urban expansion over the past decade), limiting its use as a definitive national comparator.,The critique is specific to childhood cancer surveillance context (the VIGICANCER authors' domain), but the CSI figures cited apply to all ages; the degree to which sub-national registry bias affects adult cancer estimates is not separately addressed.,No assessment is provided of whether the CONCORD-4 methodology included any sensitivity analysis, weighting, or correction for partial population coverage in Colombia.,Geographic and socioeconomic heterogeneity within Colombia (including the well-documented GINI inequality coefficient of 0·542 in 2021 and a poverty rate of approximately 39% post-pandemic) means that sub-national registry data from larger cities may not be generalisable to the national population.,Publication as correspondence means formal peer review of the specific claims made may be limited compared to a full research article. __BOTTOM_LINE_CATEGORY__ low_certainty __BOTTOM_LINE__ The CONCORD-4 Cancer Survival Index figures for Colombia (72·7% all-cancers; 81·2% WHO tracer cancers) should be interpreted cautiously, as they are derived from registries covering only a small fraction of the Colombian population and likely reflect higher-resourced urban settings rather than the full national experience. Real-world survival data from VIGICANCER and Colombian PBCRs suggest substantially lower outcomes in practice. Clinicians and public health officials should treat these figures as sub-national benchmarks rather than validated national statistics, and advocacy for expanded PBCR coverage across Colombia remains an urgent priority for meaningful cancer control planning.