Selection Bias in Real-World Data Studies Used to Support Health Technology Assessments: A Case Study in Metastatic Cancer
Bibliographic record
Abstract
Real-world evidence has been increasingly used to support evaluations of emerging therapies. These investigations are often conducted in settings that may not be representative of the underlying population. The purpose of this investigation was to empirically quantify the magnitude of this selection bias. Individuals diagnosed with solid metastatic cancer in Alberta, Canada, between 2010–2019 were identified using the provincial cancer registry for 13 common metastatic sites. Two outcomes used to support oncology reimbursement decisions were examined: the proportion of individuals who initiated systemic therapy and median overall survival (OS). These outcomes were assessed in the entire population and in a subset of individuals who were referred to a medical oncologist. Among the 23,152 individuals in the entire population, 40.8% (95% CI: 40.2–41.4) initiated systemic therapy, and the median OS from diagnosis was 5.4 months (95% CI: 5.3–5.6). Among those who were referred to a medical oncologist (n = 13,372; 57.8%), 67.4% (95% CI: 66.6–68.2) initiated systemic therapy, and the median OS from diagnosis was 11.2 months (95% CI: 10.9–11.5). The magnitude of bias varied by cancer site where lower referral rates were associated with greater bias. Non-referral is an important source of selection bias in real-world investigations. Studies that rely on limited-catchment real-world data should be interpreted with caution, particularly in metastatic cancer settings.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Direct model labels (unvalidated)
Per-model category and study-design labels from the labeling rounds. They are machine output, unvalidated, and the disagreement between models ships as data. No study design here is MEDLINE-validated yet.
| Model arm | Categories | Study design | Confidence |
|---|---|---|---|
| gpt | Metaresearch Domain: Methods · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Observational | high |
| grok | Metaresearch Domain: Methods · Genre: Empirical About the Canadian research system: no · About a Canadian topic: yes | Observational | high |
| opus | Metaresearch Domain: Methods · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Observational | high |
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.579 | 0.740 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.004 |
| Bibliometrics | 0.007 | 0.011 |
| Science and technology studies | 0.004 | 0.008 |
| Scholarly communication | 0.008 | 0.007 |
| Open science | 0.005 | 0.006 |
| Research integrity | 0.006 | 0.004 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedLabeled directly by 3 models reading the full record.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".