A confounder debiasing method for RCT-like comparability enables Machine Learning-based personalization of survival benefit in living donor liver transplantation
Bibliographic record
Abstract
Abstract Many clinical questions in medicine cannot be answered through randomized controlled trials (RCTs) due to ethical or feasibility constraints. In such cases, observational data is often the only available resource for evaluating treatment effects. To address this challenge, we have developed Decision Path Similarity Matching (DPSM), a novel machine learning (ML)-based algorithm that simulates RCT-like conditions to debias observational data. In this study, we apply DPSM to the clinical question of living donor liver transplantation (LDLT) versus deceased donor liver transplantation (DDLT), helping to identify which patients benefit most from LDLT. DPSM leverages decision paths from a Random Forest classifier to perform accurate, one-to-one matching between LDLT and DDLT recipients, minimizing confounding while retaining interpretability. Using data from the Scientific Registry of Transplant Recipients (SRTR), including 4,473 LDLT and 68,108 DDLT patients transplanted between 2002 and 2023, we trained independent Random Survival Forest (RSF) models on the matched cohorts to predict post-transplant survival. DPSM successfully reduced confounding associations between the two groups as shown by a decrease in area under the receiver operating characteristic (AUROC) from 0.82 to 0.51. Subsequently, RSF (C-index ldlt =0.67, C-index ddlt =0.74) outperformed the traditional Cox model (C-index ldlt =0.57, C-index ddlt =0.65). The predicted 10-year mean survival gain was 10.3% (SD = 5.7%). In conclusion, DPSM provides an effective approach for creating RCT-like comparability from observational data, enabling personalized survival predictions. By leveraging real-world data where RCTs are impractical, this method offers clinicians a tool for transitioning from population-level evidence to more nuanced, personalization.
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 | 0.153 | 0.372 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.006 | 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".