Rui Fu

There is a lack of rigorous methodological development for descriptive epidemiology, where the goal is to describe and identify the most important associations with an outcome given a large set of potential predictors. This has often led to the Table 2 fallacy, where one presents the coefficient estimates for all covariates from a single multivariable regression model, which are often uninterpretable in a descriptive analysis. We argue that machine learning (ML) is a potential solution to this problem. We illustrate the power of ML with an example analysis identifying the most important predictors of alcohol abuse among sexual minority youth. The framework we propose for this analysis is as follows: (1) Identify a few ML methods for the analysis, (2) optimize the parameters using the whole data with a nested cross-validation approach, (3) rank the variables using variable importance scores, (4) present partial dependence plots (PDP) to illustrate the association between the important variables and the outcome, (5) and identify the strength of the interaction terms using the PDPs. We discuss the potential strengths and weaknesses of using ML methods for descriptive analysis and future directions for research. R codes to reproduce these analyses are provided, which we invite other researchers to use.

Background: Operations research techniques enable health care administrators to optimize resource allocation and to find solutions to staff and patient scheduling problems. We aimed to conduct the first systematic review of the international literature on the use of operations research for allocating deceased-donor kidneys. Methods: We searched the MEDLINE, EMBASE, and PubMed databases from inception to February 2023. Two reviewers independently screened the title/abstract and subsequently the full text of potentially eligible articles and abstracted the data. Quality assessment of the final set of studies was conducted using Subben’s checklist. Results: Of the 302 citations identified, 5 studies were included. These studies covered three themes, including (1) provider-facing decision aids to determine the timing of transplant for single or multiple patients; (2) system-level planning on kidney allocation based on blood type matching rules; and (3) patient-facilitated wait times estimation using incomplete information. Markov models, sequential stochastic assignment models, and queuing models were amongst the most used techniques. Although we found all included studies to meet Subben’s criteria, we believe the checklist in its current form lacks items to assess the validity of model inferences. As such, we ended this review with a set of practical recommendations. Conclusions: Our review demonstrated the utility of operations research techniques in assisting the system, healthcare providers, and patients in the transplantation process. More research is needed to reach a consensus on a model that can be used to support the decision-making of different stakeholders for efficient kidney allocation, with the ultimate goal of reducing the gap between kidney supply and demand and enhancing the population’s well-being.