Search

Abstract Background Chronic conditions place a considerable burden on modern healthcare systems. Within New Zealand and worldwide cardiovascular disease (CVD) affects a significant proportion of the population and it is the leading cause of death. Like other chronic diseases, the course of cardiovascular disease is usually prolonged and its management necessarily long-term. Despite being highly effective in reducing CVD risk, non-adherence to long-term medication continues to be a longstanding challenge in healthcare delivery. The study investigates the benefits of integrating patient history and assesses the contribution of explicitly temporal models to medication adherence prediction in the context of lipid-lowering therapy. Methods Data from a CVD risk assessment tool is linked to routinely collected national and regional data sets including pharmaceutical dispensing, hospitalisation, lab test results and deaths. The study extracts a sub-cohort from 564,180 patients who had primary CVD risk assessment for analysis. Based on community pharmaceutical dispensing record, proportion of days covered (PDC) $$\ge$$ ≥ 80 is used as the threshold for adherence. Two years (8 quarters) of patient history before their CVD risk assessment is used as the observation window to predict patient adherence in the subsequent 5 years (20 quarters). The predictive performance of temporal deep learning models long short-term memory (LSTM) and simple recurrent neural networks (Simple RNN) are compared against non-temporal models multilayer perceptron (MLP), ridge classifier (RC) and logistic regression (LR). Further, the study investigates the effect of lengthening the observation window on the task of adherence prediction. Results Temporal models that use sequential data outperform non-temporal models, with LSTM producing the best predictive performance achieving a ROC AUC of 0.805. A performance gap is observed between models that can discover non-linear interactions between predictor variables and their linear counter parts, with neural network (NN) based models significantly outperforming linear models. Additionally, the predictive advantage of temporal models become more pronounced when the length of the observation window is increased. Conclusion The findings of the study provide evidence that using deep temporal models to integrate patient history in adherence prediction is advantageous. In particular, the RNN architecture LSTM significantly outperforms all other model comparators.

Abstract Individuals diagnosed with colorectal adenomas with high-risk features during screening colonoscopy have increased risk for the development of subsequent adenomas and colorectal cancer. While US guidelines recommend surveillance colonoscopy at 3 years in this high-risk population, surveillance uptake is suboptimal. To inform future interventions to improve surveillance uptake, we sought to assess surveillance rates and identify facilitators of uptake in a large integrated health system. We utilized a cohort of patients with a diagnosis of ≥ 1 tubular adenoma (TA) with high-risk features (TA ≥ 1 cm, TA with villous features, TA with high-grade dysplasia, or ≥ 3 TA of any size) on colonoscopy between 2013 and 2016. Surveillance colonoscopy completion within 3.5 years of diagnosis of an adenoma with high-risk features was our primary outcome. We evaluated surveillance uptake over time and utilized logistic regression to detect factors associated with completion of surveillance colonoscopy. The final cohort was comprised of 405 patients. 172 (42.5%) patients successfully completed surveillance colonoscopy by 3.5 years. Use of a patient reminder (telephone, electronic message, or letter) for due surveillance (adjusted odds = 1.9; 95%CI = 1.2–2.8) and having ≥ 1 gastroenterology (GI) visit after diagnosis of an adenoma with high-risk features (adjusted odds = 2.6; 95%CI = 1.6–4.2) significantly predicted surveillance colonoscopy completion at 3.5 years. For patients diagnosed with adenomas with high-risk features, surveillance colonoscopy uptake is suboptimal and frequently occurs after the 3-year surveillance recommendation. Patient reminders and visitation with GI after index colonoscopy are associated with timely surveillance completion. Our findings highlight potential health system interventions to increase timely surveillance uptake for patients diagnosed with adenomas with high-risk features.

Scientists try to design experiments that will yield maximal information. For instance, given the available evidence and a limitation on the number of variables that can be observed simultaneously, it may be more informative to intervene on variable $X$ and observe the response of variable $Y$ than to intervene on $X$ and observe $Z$ ; in other situations, the opposite may be true. Scientists must often make these decisions without primary data. To address this problem, in previous work, we created software for annotating aggregate statistics in the literature and deriving consistent causal explanations, expressed as causal graphs. This meta-analytic pipeline is useful not only for synthesizing evidence but also for planning experiments: one can use it strategically to select experiments that could further eliminate causal graphs from consideration. In this paper, we introduce interpretable policies for selecting experiments in the context of piecemeal causal discovery, a common setting in biological sciences in which each experiment can measure not an entire system but rather a strict subset of its variables. The limits of this piecemeal approach are only beginning to be fully characterized, with crucial theoretical work published recently. With simulations, we show that our experiment-selection policies identify causal structures more efficiently than random experiment selection. Unlike methods that require primary data, our meta-analytic approach offers a flexible alternative for those seeking to incorporate qualitative domain knowledge into their search for causal mechanisms. We also present a method that categorizes hypotheses with respect to their utility for identifying a system’s causal structure. Although this categorization is usually infeasible to perform manually, it is critical for conducting research efficiently.

Globally, lung cancer is responsible for nearly one in five cancer deaths. The National Lung Screening Trial (NLST) demonstrated the efficacy of low-dose computed tomography (LDCT) to identify early-stage disease, setting the basis for widespread implementation of lung cancer screening programs. However, the specificity of LDCT lung cancer screening is suboptimal, with a significant false positive rate. Representing this imaging-based screening process as a sequential decision making problem, we combined multiple machine learning-based methods to learn a partially-observable Markov decision process that simultaneously optimizes lung cancer detection while enhancing test specificity. Using NLST data, we trained a dynamic Bayesian network as an observational model and used inverse reinforcement learning to discover a rewards function based on experts' decisions. Our resultant predictive model decreased the false positive rate while maintaining a high true positive rate at a level comparable to human experts. Our model also detected a number of lung cancers earlier.

The purpose of the current study was to assess the prevalence of cyst formation at the brain-tumor interface in olfactory neuroblastoma. We used the UCLA patient-based Pathology and Radiology Head and Neck Database (UPP&R HAND) to identify the largest patient cohort reported to date with imaging and pathology data. Eighteen of thirty-one patients (58.1%) had evidence of intracranial extension on MRI, while four (22.0%) demonstrated cyst formation at the brain–tumor interface. The extent of intracranial extension was by far the strongest predictor for intracranial cyst formation, regardless of Hyams tumor grade, using a binary logistics regression model (p = 0.002) and ROC curve analysis (AUC 94.6%). Cyst formation at the brain-tumor interface was an uncommon imaging finding, and tends to occur with a larger component of intracranial tumor extension.

Predictive analytics in health is a complex, transdisciplinary field requiring collaboration across diverse scientific and stakeholder groups. Pilot implementation of participatory research to foster team science in predictive analytics through a partnered-symposium and funding competition. In total, 85 stakeholders were engaged across diverse translational domains, with a significant increase in perceived importance of early inclusion of patients and communities in research. Participatory research approaches may be an effective model for engaging broad stakeholders in predictive analytics.

To plan experiments, a biologist needs to evaluate a growing set of empirical findings and hypothetical assertions from diverse fields that use increasingly complex techniques. To address this problem, we operationalized principles (e.g., convergence and consistency) that biologists use to test causal relations and evaluate experimental evidence. With the framework we derived, we then created a free, open-source web application that allows biologists to create research maps, graph-based representations of empirical evidence and hypothetical assertions found in research articles, reviews, and other sources. With our ResearchMaps web application, biologists can systematically reason through the research that is most important to them, as well as evaluate and plan experiments with a breadth and precision that are unlikely without such a tool.

Cybersecurity research often describes people as understanding internet security in terms of metaphorical mental models (e.g., disease risk, physical security risk, or criminal behavior risk). However, little research has directly evaluated if this is an accurate or productive framework. To assess this question, two experiments asked participants to respond to a statistical reasoning task framed in one of four different contexts (cybersecurity, plus the above alternative models). Each context was also presented using either percentages or natural frequencies, and these tasks were followed by a behavioral likelihood rating. As in previous research, consistent use of natural frequencies promoted correct Bayesian reasoning. There was little indication, however, that any of the alternative mental models generated consistently better understanding or reasoning over the actual cybersecurity context. There was some evidence that different models had some effects on patterns of responses, including the behavioral likelihood ratings, but these effects were small, as compared to the effect of the numerical format manipulation. This points to a need to improve the content of actual internet security warnings, rather than working to change the models users have of warnings.