Oncology · Medical Literature
[Correspondence] Representativeness of the Cancer Survival Index in Colombia – Authors' reply
Abstract
We welcome the comments from Oscar Ramirez and colleagues, regarding the hospital-based surveillance system (VIGICANCER1) that includes 55–60% of Colombian children diagnosed with cancer. As VIGICANCER is not a population-based cancer registry (PBCR), it was not invited to contribute data to the CONCORD programme for the global surveillance of cancer survival. CONCORD-4 included data on 2725 children diagnosed with cancer during 2000–22 from four established PBCRs (Bucaramanga, Cali, Manizales, and Pasto), which covered about 8% of the national population of children in 2015.
Clinical relevance
Does VIGICANCER, Colombia's hospital-based childhood cancer surveillance system, meet the population-based registry criteria required for inclusion in the CONCORD-4 global childhood cancer survival programme, and how representative are the four Colombian population-based registries included in CONCORD-4 of the national childhood population? __SNAPSHOT__ {"studyDesign":"Correspondence (Authors' reply) — not a primary research study; commentary clarifying methodological and representativeness issues raised by external correspondents regarding a previously published global surveillance analysis (CONCORD-4)","setting":"Colombia; national-level cancer surveillance infrastructure, including four established population-based cancer registries (PBCRs): Bucaramanga, Cali, Manizales, and Pasto","sampleSize":"n=2,725 children diagnosed with cancer during 2000–2022, drawn from four Colombian PBCRs included in CONCORD-4; VIGICANCER (the hospital-based system under discussion) covers approximately 55–60% of Colombian children diagnosed with cancer nationally but was not included in CONCORD-4","population":"Children (aged 0–14 years) diagnosed with any cancer in Colombia; CONCORD-4 Colombian registries collectively covered approximately 8% of the national population of children in 2015","intervention":"Participation of population-based cancer registries in the CONCORD-4 global childhood cancer survival surveillance programme","comparator":"VIGICANCER (hospital-based surveillance system) vs. population-based cancer registries (PBCRs) — the correspondence clarifies the structural and methodological distinction between these two registry types","primaryOutcome":"Eligibility and representativeness of Colombian cancer surveillance data for inclusion in CONCORD-4; no new primary survival outcome data are reported in this correspondence","followUp":"Diagnosis period 2000–2022 (for the Colombian CONCORD-4 data cited); follow-up duration for individual patients not reported in this correspondence","fundingSource":"Not reported for this correspondence; the parent CONCORD-4 study was funded by Cancer Research UK, Institut National du Cancer, St Jude Children's Research Hospital, National Cancer Institute, and Dell Technologies"}
Key findings
This is an authors' reply correspondence, not a primary research paper; no new effect sizes, p-values, or survival statistics are reported. The authors clarify three key factual points: (1) VIGICANCER is a hospital-based surveillance system—not a population-based cancer registry (PBCR)—and therefore did not meet CONCORD-4 eligibility criteria, which require PBCR status. (2) CONCORD-4 included data on 2,725 children diagnosed with cancer during 2000–2022 from four established Colombian PBCRs (Bucaramanga, Cali, Manizales, and Pasto). (3) These four registries collectively covered approximately 8% of the national population of Colombian children in 2015, indicating a substantial representativeness gap relative to the national childhood population. VIGICANCER, by contrast, currently captures approximately 55–60% of Colombian children diagnosed with cancer through 27 institutions across 10 cities, though via a hospital-based rather than population-based architecture. No quantitative survival outcomes are introduced or compared in this correspondence.
Practice implications
This correspondence establishes a principled methodological distinction between hospital-based and population-based cancer surveillance: only the latter can produce unbiased, denominator-defined survival estimates free from referral and selection bias, and CONCORD's requirement for PBCR status reflects this scientific standard. The four Colombian PBCRs included in CONCORD-4 covered only ~8% of the national childhood population, a significant limitation acknowledged implicitly by the authors, meaning Colombia's CONCORD-4 survival estimates may not be nationally representative and likely reflect urban, more resourced settings. VIGICANCER's broader geographic reach (~55–60% of national childhood cancer cases via 27 institutions across 10 cities) offers complementary clinical and programmatic data but cannot substitute for population-based denominators required for unbiased net survival estimation. Clinicians and policymakers interpreting Colombian data within CONCORD-4 should be aware that reported figures reflect a geographically and demographically restricted subset of the Colombian childhood cancer population. This correspondence does not establish or change any clinical management recommendations; its importance is methodological and epidemiological. Note: only the abstract and contextual literature were available for analysis. __LIMITATIONS__ 1. This is a brief correspondence, not a primary study; no new data, statistical analyses, or survival outcomes are presented. 2. The four Colombian PBCRs in CONCORD-4 covered only ~8% of the national childhood population in 2015, substantially limiting the national representativeness of derived survival estimates. 3. VIGICANCER, despite broader operational reach (~55–60% of national cases), is explicitly hospital-based and cannot generate population-based denominators; its exclusion from CONCORD-4 is methodologically defensible but leaves a data gap on the majority of Colombian children with cancer. 4. The extended diagnostic period cited (2000–2022) is unusual for CONCORD-4 (originally designed through 2019) and may reflect a CONCORD-4 extension or updated data cycle not fully detailed in this correspondence. 5. No survival statistics, confidence intervals, or comparative outcome data are reported, precluding any quantitative clinical inference. 6. Generalizability of the four PBCR estimates to rural, indigenous, or lower-resourced Colombian populations is uncertain and not addressed. __BOTTOM_LINE_CATEGORY__ low_certainty __BOTTOM_LINE__ This authors' reply correspondence clarifies that VIGICANCER's hospital-based architecture disqualifies it from CONCORD-4 inclusion, which requires population-based cancer registry data to ensure unbiased survival estimation. The four Colombian PBCRs included in CONCORD-4 covered only ~8% of the national childhood population, meaning Colombian survival estimates in that programme should be interpreted as geographically restricted rather than nationally representative. Clinicians and policymakers should treat Colombia's CONCORD-4 figures as provisional benchmarks pending broader population-based coverage, and should recognize VIGICANCER's complementary but structurally distinct role in national childhood cancer surveillance.