CliqueFluxNet: Unveiling EHR Insights with Stochastic Edge Fluxing and Maximal Clique Utilisation Using Graph Neural Networks.

Electronic Health Records (EHRs) play a crucial role in shaping predictive are models, yet they encounter challenges such as significant data gaps and class imbalances. Traditional Graph Neural Network (GNN) approaches have limitations in fully leveraging neighbourhood data or demanding intensive computational requirements for regularisation. To address this challenge, we introduce CliqueFluxNet, a novel framework that innovatively constructs a patient similarity graph to maximise cliques, thereby highlighting strong inter-patient connections. At the heart of CliqueFluxNet lies its stochastic edge fluxing strategy - a dynamic process involving random edge addition and removal during training. This strategy aims to enhance the model's generalisability and mitigate overfitting. Our empirical analysis, conducted on MIMIC-III and eICU datasets, focuses on the tasks of mortality and readmission prediction. It demonstrates significant progress in representation learning, particularly in scenarios with limited data availability. Qualitative assessments further underscore CliqueFluxNet's effectiveness in extracting meaningful EHR representations, solidifying its potential for advancing GNN applications in healthcare analytics.

Comparative effectiveness of sodium-glucose cotransporter-2 inhibitors for new-onset gastric cancer and gastric diseases in patients with type 2 diabetes mellitus: a population-based cohort study.

OBJECTIVE: To compare the risks of gastric cancer and other gastric diseases in patients with type-2 diabetes mellitus (T2DM) exposed to sodium-glucose cotransporter 2 inhibitors (SGLT2I), dipeptidyl peptidase-4 inhibitors (DPP4I) or glucagon-like peptide-1 receptor agonists (GLP1a). DESIGN: This was a population-based cohort study of prospectively collected data on patients with T2DM prescribed SGLT2I, DPP4I or GLP1a between January 1st 2015 and December 31st 2020 from Hong Kong. The outcomes were new-onset gastric cancer, peptic ulcer (PU), acute gastritis, non-acute gastritis, and gastroesophageal reflux disease (GERD). Propensity score matching (1:1) using the nearest neighbour search was performed, and multivariable Cox regression was applied. A three-arm comparison between SGLT2I, DPP4I and GLP1a was conducted using propensity scores with inverse probability of treatment weighting. RESULTS: A total of 62,858 patients (median age: 62.2 years old [SD: 12.8]; 55.93% males; SGLT2I: n = 23,442; DPP4I: n = 39,416) were included. In the matched cohort, the incidence of gastric cancer was lower in SGLT2I (Incidence rate per 1000 person-year, IR: 0.32; 95% confidence interval, CI 0.23-0.43) than in DPP4I (IR per 1000 person-year: 1.22; CI 1.03-1.42) users. Multivariable Cox regression found that SGLT2I use was associated with lower risks of gastric cancer (HR 0.30; 95% CI 0.19-0.48), PU, acute gastritis, non-acute gastritis, and GERD (p < 0.05) compared to DPP4I use. In the three-arm analysis, GLP1a use was associated with higher risks of gastric cancer and GERD compared to SGLT2I use. CONCLUSIONS: The use of SGLT2I was associated with lower risks of new-onset gastric cancer, PU, acute gastritis, non-acute gastritis, and GERD after matching and adjustments compared to DPP4I use. SGLT2I use was associated with lower risks of GERD and gastric cancer compared to GLP1a use.

Continuous patient state attention model for addressing irregularity in electronic health records.

BACKGROUND: Irregular time series (ITS) are common in healthcare as patient data is recorded in an electronic health record (EHR) system as per clinical guidelines/requirements but not for research and depends on a patient's health status. Due to irregularity, it is challenging to develop machine learning techniques to uncover vast intelligence hidden in EHR big data, without losing performance on downstream patient outcome prediction tasks. METHODS: In this paper, we propose Perceiver, a cross-attention-based transformer variant that is computationally efficient and can handle long sequences of time series in healthcare. We further develop continuous patient state attention models, using Perceiver and transformer to deal with ITS in EHR. The continuous patient state models utilise neural ordinary differential equations to learn patient health dynamics, i.e., patient health trajectory from observed irregular time steps, which enables them to sample patient state at any time. RESULTS: The proposed models' performance on in-hospital mortality prediction task on PhysioNet-2012 challenge and MIMIC-III datasets is examined. Perceiver model either outperforms or performs at par with baselines, and reduces computations by about nine times when compared to the transformer model, with no significant loss of performance. Experiments to examine irregularity in healthcare reveal that continuous patient state models outperform baselines. Moreover, the predictive uncertainty of the model is used to refer extremely uncertain cases to clinicians, which enhances the model's performance. Code is publicly available and verified at https://codeocean.com/capsule/4587224 . CONCLUSIONS: Perceiver presents a computationally efficient potential alternative for processing long sequences of time series in healthcare, and the continuous patient state attention models outperform the traditional and advanced techniques to handle irregularity in the time series. Moreover, the predictive uncertainty of the model helps in the development of transparent and trustworthy systems, which can be utilised as per the availability of clinicians.