Inverse Optimization and its Applications in Measuring Clinical Pathway Concordance
Bibliographic record
Abstract
Inverse optimization aims to infer the parameters of an optimization problem given a set of observed decisions that are assumed to be optimal or minimally suboptimal. In this thesis, we propose novel inverse optimization techniques and present new applications. In particular, we use inverse optimization in developing a metric for measuring clinical pathway concordance. The resulting metric can help monitor variations in the healthcare system, identify bottlenecks, and provide data-driven evidence to inform health policy decisions. We first propose an inverse optimization-based framework for measuring clinical pathway concordance in diseases with simple patient journeys. We construct a network where the nodes represent the clinical activities, and patients accumulate costs as they visit the nodes and move through arcs. We use linear programming duality to formulate a two-phased inverse optimization problem to obtain arc costs that make the clinical pathways optimal for a shortest path problem on the network. The arc costs are the weights used in the concordance metric. We then provide an in-depth case study applying our inverse optimization-based model and metric to real patient data from stage III colon cancer patients. We also develop a rigorous framework to identify the sources of discordance in the patient population. Next, we propose an approach for learning arc costs in a hierarchical network that can be used to measure clinical pathway concordance for diseases that have complex patient journeys. The overall network comprises nested subnetworks defined over multiple levels in a hierarchical network structure. The hierarchical structure facilitates the modelling of complex pathways by dividing them into shorter sections and allowing us to incorporate crucial contextual information. We then apply our methodology to a large dataset of breast cancer patients. Finally, we extend the inverse optimization methodology by quantifying the uncertainty around the point estimates for unknown parameters of an optimization problem. Even though uncertainty is inherent in inverse optimization with noisy data, the quantification of this uncertainty has been treated very little in the literature. We propose one parametric and one non-parametric approach to estimate the unknown parameters and construct confidence regions for the point estimates.
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 | 0.010 | 0.048 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.001 | 0.004 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.002 | 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; 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".