Underreporting and selection bias of serious road traffic injuries in auto insurance claims and police reports in British Columbia, Canada
Bibliographic record
Abstract
Administrative datasets (police reports, insurance claims, medical records), form the basis for road safety research, but suffer from under-reporting and selection bias. Data linkage can provide a fuller picture of road traffic injuries and provide insight into dataset-specific biases. We examined the overlap of serious road traffic injuries involving motor vehicles reported in hospitalization records, police reports, and insurance claims in British Columbia, Canada (2015 – 2019) and assess selection bias within each injury dataset. We probabilistically linked police reports, insurance claims, and hospital admissions to a provincial population directory, identifying distinct persons and injuries across datasets. Injuries were linked to sociodemographic and geographic details from other government data including age, sex, low-income status, neighbourhood income and health authority. We analyzed serious injuries to drivers, cyclists and pedestrians. We assessed the proportion of injuries captured by a database (ascertainment rate) and assessed selection bias based on which sociodemographic groups were more likely to only be captured in hospital admissions. From 2015 to 2019, we estimated 57,097 motor vehicle-involved injuries (48,198 motor vehicle drivers, 2,641 cyclists, 6,258 pedestrians). Insurance claims had the highest ascertainment rate for drivers (95.7%), but lower for cyclists (83.3%) and pedestrians (76.5%). Police records and hospital admissions better captured cyclist and pedestrian injuries compared to driver injuries. Unlinked hospital admission injuries were more likely from low-income and remote populations. The underreporting highlights the need for improved injury data collection especially for pedestrian and cyclists, to better capture the full injury burden, particularly among marginalized sociodemographic groups.
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 | 0.019 | 0.061 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.007 | 0.017 |
| Science and technology studies | 0.005 | 0.002 |
| Scholarly communication | 0.004 | 0.001 |
| Open science | 0.004 | 0.004 |
| Research integrity | 0.001 | 0.001 |
| 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, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
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".