Validation of a Cyclic Algorithm to Proxy Number of Lines of Systemic Cancer Therapy Using Administrative Data
Bibliographic record
Abstract
PURPOSE Researchers are automating the process for identifying the number of lines of systemic cancer therapy received by patients. To date, algorithm development has involved manual modifications to predefined classification rules. In this study, we propose a supervised learning algorithm for determining the best-performing proxy for number of lines of therapy and validate this approach in four patient groups. MATERIALS AND METHODS We retrospectively analyzed BC Cancer pharmacy records from patients’ cancer diagnosis until end of follow-up (cohort-specific, 2014/2015). We created and validated a cyclic algorithm in patients with advanced cancers of varying histologies, diffuse large B-cell lymphoma, follicular lymphoma, and chronic lymphocytic leukemia. To assess internal and external validity, we used a split-sample approach for all analyses and considered lines of therapy identified through manual review as our criterion standard. We measured agreement using correlation coefficients, mean squared error, nonparametric hypothesis testing, and quantile-quantile plots. RESULTS Cohorts comprised 91 patients with advanced cancers, 121 with chronic lymphocytic leukemia, 440 with follicular lymphoma, and 679 with diffuse large B-cell lymphoma. Number of lines of therapy received and patients’ treatment period length varied substantially across cohorts. Despite these differences, our algorithm successfully identified a best-performing proxy for number of lines of therapy for each cohort, which was moderate to highly correlated with (within-sample: 0.73 ≤ Pearson correlation ≤ 0.84; out-of-sample: 0.52 ≤ Pearson correlation ≤ 0.76) and whose distribution did not significantly differ from the criterion standard within or out of sample ( P > .10). CONCLUSION Supervised learning is an ideal tool for generating a best-performing proxy that recognizes prescription drug patterns and approximates number of lines of therapy. Our cyclic approach can be used in jurisdictions with access to administrative pharmacy data.
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 | 0.035 | 0.083 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".