When developing clinical prediction models, it can be challenging to balance between global models that are valid for all patients and personalized models tailored to individuals or potentially unknown subgroups. To aid such decisions, we propose a diagnostic tool for contrasting global regression models and patient-specific (local) regression models. The core utility of this tool is to identify where and for whom a global model may be inadequate. We focus on regression models and specifically suggest a localized regression approach that identifies regions in the predictor space where patients are not well represented by the global model. As localization becomes challenging when dealing with many predictors, we propose modeling in a dimension-reduced latent representation obtained from an autoencoder. Using such a neural network architecture for dimension reduction enables learning a latent representation simultaneously optimized for both good data reconstruction and for revealing local outcome-related associations suitable for robust localized regression. We illustrate the proposed approach with a clinical study involving patients with chronic obstructive pulmonary disease. Our findings indicate that the global model is adequate for most patients but that indeed specific subgroups benefit from personalized models. We also demonstrate how to map these subgroup models back to the original predictors, providing insight into why the global model falls short for these groups. Thus, the principal application and diagnostic yield of our tool is the identification and characterization of patients or subgroups whose outcome associations deviate from the global model. Introduction In clinical research, conclusions about potential relationships between patient characteristics and outcomes often are based on regression models. More specifically, there might not be just some random variability across the parameters of patients, e.g. as considered in regression modeling with random effects (Pinheiro and Bates, 2000), but different regions in the space spanned by the patient characteristics might require different parameters. For example, the relation of some patient characteristics to the outcome might be more pronounced for older patients with high body weight, without having a corresponding pre-defined subgroup indicator. While sticking to a global model keeps interpretation simple and is beneficial in terms of statistical stability, it would at least be useful to have some diagnostic tool for judging the potential extent of deviations from the global model.

As regulatory agencies increasingly recognise real-world evidence as a complement to traditional clinical trial data, interest has grown in applying Bayesian methods across both interventional and observational research (Boulanger and Carlin (2021). A central objective in many clinical investigations is the delineation of patient subgroups that exhibit comparable disease-related characteristics (He, Belouali, Patricoski, Lehmann, Ball, Anagnostou, Kreimeyer, and Botsis (2023)). Electronic Health Records (EHR) have become an important resource for such phenotypic analyses (Hripcsak and Albers (2013)). Bayesian approaches to patient phenotyping in clinical observational studies have been limited by the computational challenges associated with applying the Markov Chain Monte Carlo (MCMC) approach to real-world data. Hubbard, Huang, Harton, Oganisian, Choi, Utidjian, Eneli, Bailey, and Chen (2019) proposed a Bayes latent class model that could be used in a general context for observational studies that use EHR data. They consider the common clinical context where gold-standard phenotype information, such as genetic and laboratory data, is not fully available. A general model of this form has high potential applicability for use in clinical decision support across disease areas for both primary and secondary clinical databases. Latent Class Analysis (LCA) is widely used when we want to identify patient phenotypes or subgroups given multivariate data (Lanza and Rhoades (2013)). A challenge in clinical LCA is the prevalence of mixed data, where we may have combinations of continuous, nominal, ordinal and count data.

Large language models (LLM) have achieved impressive performance on medical question-answering benchmarks. However, high benchmark accuracy does not imply that the performance generalizes to real-world clinical settings. Medical question-answering benchmarks rely on assumptions consistent with quantifying LLM performance but that may not hold in the open world of the clinic. Yet LLMs learn broad knowledge that can help the LLM generalize to practical conditions regardless of unrealistic assumptions in celebrated benchmarks. We seek to quantify how well LLM medical question-answering benchmark performance generalizes when benchmark assumptions are violated. Specifically, we present an adversarial method that we call MedFuzz (for medical fuzzing). MedFuzz attempts to modify benchmark questions in ways aimed at confounding the LLM. We demonstrate the approach by targeting strong assumptions about patient characteristics presented in the MedQA benchmark. Successful "attacks" modify a benchmark item in ways that would be unlikely to fool a medical expert but nonetheless "trick" the LLM into changing from a correct to an incorrect answer. Further, we present a permutation test technique that can ensure a successful attack is statistically significant. We show how to use performance on a "MedFuzzed" benchmark, as well as individual successful attacks. The methods show promise at providing insights into the ability of an LLM to operate robustly in more realistic settings.

This study examines the use of natural language processing (NLP) models to evaluate whether language patterns used by item writers in a medical licensure exam might contain evidence of biased or stereotypical language. This type of bias in item language choices can be particularly impactful for items in a medical licensure assessment, as it could pose a threat to content validity and defensibility of test score validity evidence. To the best of our knowledge, this is the first attempt using machine learning (ML) and NLP to explore language bias on a large item bank. Using a prediction algorithm trained on clusters of similar item stems, we demonstrate that our approach can be used to review large item banks for potential biased language or stereotypical patient characteristics in clinical science vignettes. The findings may guide the development of methods to address stereotypical language patterns found in test items and enable an efficient updating of those items, if needed, to reflect contemporary norms, thereby improving the evidence to support the validity of the test scores.

This article proposes a method for automated service selection to improve treatment efficacy and reduce re-hospitalization costs. A predictive model is developed using the National Home and Hospice Care Survey (NHHCS) dataset to quantify the effect of care services on the risk of re-hospitalization. By taking the patient's characteristics and other selected services into account, the model is able to indicate the overall effectiveness of a combination of services for a specific NHHCS patient. The developed model is incorporated in Monte-Carlo Tree Search (MCTS) to determine optimal combinations of services that minimize the risk of emergency re-hospitalization. MCTS serves as a risk minimization algorithm in this case, using the predictive model for guidance during the search. Using this method on the NHHCS dataset, a significant reduction in risk of re-hospitalization is observed compared to the original selections made by clinicians. An 11.89 percentage points risk reduction is achieved on average. Higher reductions of roughly 40 percentage points on average are observed for NHHCS patients in the highest risk categories. These results seem to indicate that there is enormous potential for improving service selection in the near future.

Tens of millions of women suffer from infertility worldwide each year. In vitro fertilization (IVF) is the best choice for many such patients. However, IVF is expensive, time-consuming, and both physically and emotionally demanding. The first question that a patient usually asks before the IVF is how likely she will conceive, given her basic medical examination information. This paper proposes three approaches to predict the cumulative pregnancy rate after multiple oocyte pickup cycles. Experiments on 11,190 patients showed that first clustering the patients into different groups and then building a support vector machine model for each group can achieve the best overall performance. Our model could be a quick and economic approach for reliably estimating the cumulative pregnancy rate for a patient, given only her basic medical examination information, well before starting the actual IVF procedure. The predictions can help the patient make optimal decisions on whether to use her own oocyte or donor oocyte, how many oocyte pickup cycles she may need, whether to use embryo frozen, etc. They will also reduce the patient's cost and time to pregnancy, and improve her quality of life.

We investigate the effect of the proportional hazards assumption on prognostic and predictive models of the survival time of patients suffering from amyotrophic lateral sclerosis (ALS). We theoretically compare the underlying model formulations of several variants of survival forests and implementations thereof, including random forests for survival, conditional inference forests, Ranger, and survival forests with $L_1$ splitting, with two novel variants, namely distributional and transformation survival forests. Theoretical considerations explain the low power of log-rank-based splitting in detecting patterns in non-proportional hazards situations in survival trees and corresponding forests. This limitation can potentially be overcome by the alternative split procedures suggested herein. We empirically investigated this effect using simulation experiments and a re-analysis of the PRO-ACT database of ALS survival, giving special emphasis to both prognostic and predictive models.

Clinical trials involving multiple treatments utilize randomization of the treatment assignments to enable the evaluation of treatment efficacies in an unbiased manner. Such evaluation is performed in post hoc studies that usually use supervised-learning methods that rely on large amounts of data collected in a randomized fashion. That approach often proves to be suboptimal in that some participants may suffer and even die as a result of having not received the most appropriate treatments during the trial. Reinforcement-learning methods improve the situation by making it possible to learn the treatment efficacies dynamically during the course of the trial, and to adapt treatment assignments accordingly. Recent efforts using \textit{multi-arm bandits}, a type of reinforcement-learning methods, have focused on maximizing clinical outcomes for a population that was assumed to be homogeneous. However, those approaches have failed to account for the variability among participants that is becoming increasingly evident as a result of recent clinical-trial-based studies. We present a contextual-bandit-based online treatment optimization algorithm that, in choosing treatments for new participants in the study, takes into account not only the maximization of the clinical outcomes but also the patient characteristics. We evaluated our algorithm using a real clinical trial dataset from the International Stroke Trial. The results of our retrospective analysis indicate that the proposed approach performs significantly better than either a random assignment of treatments (the current gold standard) or a multi-arm-bandit-based approach, providing substantial gains in the percentage of participants who are assigned the most suitable treatments. The contextual-bandit and multi-arm bandit approaches provide 72.63% and 64.34% gains, respectively, compared to a random assignment.

Over the past decades, both critical care and cancer care have improved substantially. Due to increased cancer-specific survival, we hypothesized that both the number of cancer patients admitted to the ICU and overall survival have increased since the millennium change. MIMIC-III, a freely accessible critical care database of Beth Israel Deaconess Medical Center, Boston, USA was used to retrospectively study trends and outcomes of cancer patients admitted to the ICU between 2002 and 2011. Multiple logistic regression analysis was performed to adjust for confounders of 28-day and 1-year mortality. Out of 41,468 unique ICU admissions, 1,100 hemato-oncologic, 3,953 oncologic and 49 patients with both a hematological and solid malignancy were analyzed. Hematological patients had higher critical illness scores than non-cancer patients, while oncologic patients had similar APACHE-III and SOFA-scores compared to non-cancer patients. In the univariate analysis, cancer was strongly associated with mortality (OR= 2.74, 95%CI: 2.56, 2.94). Over the 10-year study period, 28-day mortality of cancer patients decreased by 30%. This trend persisted after adjustment for covariates, with cancer patients having significantly higher mortality (OR=2.63, 95%CI: 2.38, 2.88). Between 2002 and 2011, both the adjusted odds of 28-day mortality and the adjusted odds of 1-year mortality for cancer patients decreased by 6% (95%CI: 4%, 9%). Having cancer was the strongest single predictor of 1-year mortality in the multivariate model (OR=4.47, 95%CI: 4.11, 4.84).

The odds of responding well to "intensifying" antidiabetic regimens with an additional antihyperglycemic and of avoiding episodes of severe hypoglycemia could be increased by promising approaches in big data, machine learning, and real-time informatics, according to recent presentations at the American Diabetes Association (ADA) 78th Scientific Sessions, Orlando, Florida. The decision to add a glucagon-like peptide-1 receptor agonist (GLP-1 RA) to basal insulin and other oral antihyperglycemic agents that have failed to adequately control a patient's type 2 diabetes (T2DM) could be better informed, for example, with analysis of a range of patient characteristics including the other medications and dosages, and the severity and duration of diabetic symptoms and of concurrent conditions. Big-data algorithms might be used to consider these multiple parameters, and to possibly identify optimal patient characteristics for the new drug therapy, according to Esther Zimmermann, PhD, Novo Nordisk, Søborg, Denmark. "Machine learning is a new tool used for the analysis of big data that has the potential to identify trends and predict outcomes," Zimmermann explained, in describing her study. "The aim of this study was to use machine learning for extensive analysis of big, complex to data to, one, characterize patients on basal insulin for whom a GLP-1 RA was additionally prescribed and, two, identify predictors of 1 percent (or greater) reduction in A1c in (those) patients."