Autoregressive forecasting is central to predictive control in diabetes and hemodynamic management, where different operating zones carry different clinical risks. Standard models trained with teacher forcing suffer from exposure bias, yielding unstable multi-step forecasts for closed-loop use. We introduce Soft-Token Trajectory Forecasting (SoTra), which propagates continuous probability distributions (``soft tokens'') to mitigate exposure bias and learn calibrated, uncertainty-aware trajectories. A risk-aware decoding module then minimizes expected clinical harm. In glucose forecasting, SoTra reduces average zone-based risk by 18\%; in blood-pressure forecasting, it lowers effective clinical risk by approximately 15\%. These improvements support its use in safety-critical predictive control.

Predictive modelling is vital to guide preventive efforts. Whilst large-scale prospective cohort studies and a diverse toolkit of available machine learning (ML) algorithms have facilitated such survival task efforts, choosing the best-performing algorithm remains challenging. Benchmarking studies to date focus on relatively small-scale datasets and it is unclear how well such findings translate to large datasets that combine omics and clinical features. We sought to benchmark eight distinct survival task implementations, ranging from linear to deep learning (DL) models, within the large-scale prospective cohort study UK Biobank (UKB). We compared discrimination and computational requirements across heterogenous predictor matrices and endpoints. Finally, we assessed how well different architectures scale with sample sizes ranging from n = 5,000 to n = 250,000 individuals. Our results show that discriminative performance across a multitude of metrices is dependent on endpoint frequency and predictor matrix properties, with very robust performance of (penalised) COX Proportional Hazards (COX-PH) models. Of note, there are certain scenarios which favour more complex frameworks, specifically if working with larger numbers of observations and relatively simple predictor matrices. The observed computational requirements were vastly different, and we provide solutions in cases where current implementations were impracticable. In conclusion, this work delineates how optimal model choice is dependent on a variety of factors, including sample size, endpoint frequency and predictor matrix properties, thus constituting an informative resource for researchers working on similar datasets. Furthermore, we showcase how linear models still display a highly effective and scalable platform to perform risk modelling at scale and suggest that those are reported alongside non-linear ML models.

Rising consumer awareness regarding risk management and implementation of big data solutions are driving the market for clinical risk grouping software. Scorecard & visualization tools, dashboard analytics, and risk reporting are the three product types of clinical risk grouping solutions. Scorecard & visualization tools segment dominated the market with the largest share due to its ability to predict payment processes accurately and project per-patient risk. The rising need to reduce healthcare costs through these two channels is expected to augment the growth of the segment during the forecast period. Hospitals, payers, ambulatory care centers, and long-term care centers, among others are the end-users of clinical risk grouping solutions, of which hospitals accounted for the largest market share in 2018.

An electronic health record dataset was used for this study with similar risk factor information used across all models. Nineteen different prediction techniques were applied including 12 families of machine learning models (such as neural networks) and seven statistical models (such as Cox proportional hazards models). It was found that the various models had similar population-level model performance (C-statistics of about 0.87 and similar calibration). However, the predictions for individual CVD risks varied widely between and within different types of machine learning and statistical models, especially in patients with higher CVD risks. Most of the machine learning models, tested in this study, do not take censoring into account by default (i.e., loss to follow-up over the 10 years).

In recent years, increasingly augmentation of health data, such as patient Electronic Health Records (EHR), are becoming readily available. This provides an unprecedented opportunity for knowledge discovery and data mining algorithms to dig insights from them, which can, later on, be helpful to the improvement of the quality of care delivery. Predictive modeling of clinical risk, including in-hospital mortality, hospital readmission, chronic disease onset, condition exacerbation, etc., from patient EHR, is one of the health data analytic problems that attract most of the interests. The reason is not only because the problem is important in clinical settings, but also there are challenges working with EHR such as sparsity, irregularity, temporality, etc. Different from applications in other domains such as computer vision and natural language processing, the labeled data samples in medicine (patients) are relatively limited, which creates lots of troubles for effective predictive model learning, especially for complicated models such as deep learning. In this paper, we propose MetaPred, a meta-learning for clinical risk prediction from longitudinal patient EHRs. In particular, in order to predict the target risk where there are limited data samples, we train a meta-learner from a set of related risk prediction tasks which learns how a good predictor is learned. The meta-learned can then be directly used in target risk prediction, and the limited available samples can be used for further fine-tuning the model performance. The effectiveness of MetaPred is tested on a real patient EHR repository from Oregon Health & Science University. We are able to demonstrate that with CNN and RNN as base predictors, MetaPred can achieve much better performance for predicting target risk with low resources comparing with the predictor trained on the limited samples available for this risk.

This studentship is funded by the EPSRC DTP covering fees and stipend (RCUK rate) Start date September 2018 for 3.5 years. Applicants must be from the UK/EU and have obtained (or be about to obtain) a minimum 2:1 Bachelors degree in a relevant subject area. Applications should be submitted online, select PhD Health Sciences on the application form. For information on how to apply for this project, please visit the Faculty of Biology, Medicine and Health Doctoral Academy website (View Website). Interviews will be held in Manchester in May 2018.