Quantifying the impact of survivor treatment bias in observational studies
Bibliographic record
Abstract
RATIONALE: Observational cohort studies are frequently used to measure the impact of therapies on the time to a particular outcome. Treatment often has a time-variant nature since it is frequently initiated at varying times during a patient's follow-up. Studies in the medical literature frequently ignore the time-dependent nature of treatment exposure. Survivor treatment bias can arise when the time dependent nature of treatment exposure is ignored since patients who survived to receive treatment may be healthier than patients who died prior to receipt of treatment. AIMS AND OBJECTIVES: The objective of the current study was to explicitly quantify the magnitude of survivor-treatment bias. METHODS: Monte Carlo simulations using parameters obtained from an analysis of patients admitted to hospital with a diagnosis of acute myocardial infarction in Ontario, Canada. RESULTS AND CONCLUSIONS: When the true treatment was null (hazard ratio of 1), estimated treatment effects varied from a 4% reduction in mortality to a reduction in mortality of 27% when the time varying nature of the treatment was ignored. Furthermore, survivor-treatment bias increased as the time required foe exposed patients to receive treatment increased. Similarly, survivor treatment bias was amplified as exposure was defined to be exposure at any time prior to mortality compared to exposure within a fixed time interval starting at the time origin. Ignoring the time-dependent nature of treatment results in overly optimistic estimates of treatment effects. Depending on the period required for patients to initiate therapy, treatments with no effect on survival can appear to be strongly associated with improved survival. The current study is the first to explicitly quantify the magnitude of bias that results from ignoring the time-varying nature of treatment exposure in survival studies.
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
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 | 0.352 | 0.706 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.004 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.001 | 0.006 |
| Scholarly communication | 0.004 | 0.004 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.005 | 0.004 |
| Insufficient payload (model declined to judge) | 0.002 | 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, unvalidatedMachine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
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".