Enriching Data Science and Health Care Education: Application and Impact of Synthetic Data Sets Through the Health Gym Project.

Large-scale medical data sets are vital for hands-on education in health data science but are often inaccessible due to privacy concerns. Addressing this gap, we developed the Health Gym project, a free and open-source platform designed to generate synthetic health data sets applicable to various areas of data science education, including machine learning, data visualization, and traditional statistical models. Initially, we generated 3 synthetic data sets for sepsis, acute hypotension, and antiretroviral therapy for HIV infection. This paper discusses the educational applications of Health Gym's synthetic data sets. We illustrate this through their use in postgraduate health data science courses delivered by the University of New South Wales, Australia, and a Datathon event, involving academics, students, clinicians, and local health district professionals. We also include adaptable worked examples using our synthetic data sets, designed to enrich hands-on tutorial and workshop experiences. Although we highlight the potential of these data sets in advancing data science education and health care artificial intelligence, we also emphasize the need for continued research into the inherent limitations of synthetic data.

Prescription opioid use among people with opioid dependence and concurrent benzodiazepine and gabapentinoid exposure: An analysis of overdose and all-cause mortality.

BACKGROUND: Studies investigating mortality risk associated with use of opioid analgesics, benzodiazepines, gabapentinoids, and opioid agonist treatment (OAT) among people with opioid dependence (PWOD) are lacking. This study addresses this gap using a cohort of 37,994 PWOD initiating opioid analgesics between July 2003 and July 2018 in New South Wales, Australia. METHODS: Linked administrative records provided data on dispensings, sociodemographics, clinical characteristics, OAT, and mortality. Cox proportional hazards models assessed associations between time-varying measures of individual and concurrent medicine use and OAT with all-cause mortality, accidental opioid overdose, non-drug induced accidents, and non-drug-induced suicide. Opioid analgesic dose effects, expressed as oral morphine equivalents (OMEs) per day, were also examined. OUTCOMES: During the study period, 3167 individuals died. Compared with no use, all medicines of interest were associated with increased accidental opioid overdose risk; hazard ratios (HR) ranged from 1.33 (95 % CI: 1.05-1.68) for opioid analgesic use to 6.10 (95 % CI: 4.11-9.06) for opioid analgesic, benzodiazepine and gabapentinoid use. Benzodiazepine use was associated with increased non-drug-induced accidents and non-drug-induced suicides. For all-cause mortality, all combinations of benzodiazepines and gabapentinoids with opioid analgesics were associated with increased risk (aHRs ranged from 1.35 to 2.73). For most medicines/medicine combinations, all-cause mortality risk was reduced when in OAT compared to out of OAT. Higher opioid analgesic doses were associated with increased all-cause mortality (e.g., 90-199 mg vs 1-49 mg OME per day: HR 1.90 [95 % CI: 1.52-2.40]). INTERPRETATION: The increased mortality risk associated with benzodiazepines and gabapentinoids among PWOD appear to be reduced when engaged in OAT. A greater focus on encouraging OAT engagement, providing overdose prevention education, and access and coverage of overdose antidotes is necessary to minimise the unintended consequences of medicines use in this population.

Sexual and drug use risk behaviour trajectories among people treated for recent HCV infection: the REACT study.

INTRODUCTION: Exploration of sexual and drug use behaviours following treatment for recent hepatitis C virus (HCV) is limited. This analysis modelled behavioural trajectories following treatment for recent HCV and assessed reinfection. METHODS: Participants treated for recent HCV in an international trial (enrolled 2017-2019) were followed at 3-monthly intervals for up to 2 years to assess longitudinal behaviours. Population-averaged changes were assessed using generalized estimating equations. Distinct behavioural trajectories were identified using group-based trajectory modelling. HCV reinfection incidence was calculated using person-years (PY) of observation. RESULTS: During the follow-up of 212 participants (84% gay and bisexual men [GBM]; 69% HIV; 26% current injecting drug use [IDU]), behavioural trajectories for IDU and stimulant use (past month) did not change. However, population-averaged decreases in the likelihood of daily IDU (adjusted odds ratio [AOR] 0.83; 95% CI 0.72, 0.95) and opioid use (AOR 0.84; 95% CI 0.75, 0.93) were observed. Among GBM, behavioural trajectories for chemsex did not change. Population-averaged decreases in condomless anal intercourse with casual male partners (CAI-CMP) (AOR 0.95; 95% CI 0.90, 0.99) and group-sex (AOR 0.86; 95% CI 0.80, 0.93) were observed, but masked distinct trajectories. While a proportion had a decreased probability of CAI-CMP (23%) and group-sex (59%) post-treatment, a substantial proportion retained a high probability of these behaviours. High HCV reinfection incidence was observed for the sustained high probability IDU (33.0/100 PY; 95% CI 17.7, 61.3) and chemsex (23.3/100 PY; 95% CI 14.5, 37.5) trajectories. CONCLUSIONS: Limited sexual and drug use behavioural change was observed following treatment for recent HCV, supporting access to surveillance and (re)treatment.

Generating synthetic clinical data that capture class imbalanced distributions with generative adversarial networks: Example using antiretroviral therapy for HIV.

OBJECTIVE: Clinical data's confidential nature often limits the development of machine learning models in healthcare. Generative adversarial networks (GANs) can synthesise realistic datasets, but suffer from mode collapse, resulting in low diversity and bias towards majority demographics and common clinical practices. This work proposes an extension to the classic GAN framework that includes a variational autoencoder (VAE) and an external memory mechanism to overcome these limitations and generate synthetic data accurately describing imbalanced class distributions commonly found in clinical variables. METHODS: The proposed method generated a synthetic dataset related to antiretroviral therapy for human immunodeficiency virus (ART for HIV). We evaluated it based on five metrics: (1) accurately representing imbalanced class distribution; (2) the realism of the individual variables; (3) the realism among variables; (4) patient disclosure risk; and (5) the utility of the generated dataset for developing downstream machine learning models. RESULTS: The proposed method overcomes the issue of mode collapse and generates a synthetic dataset that accurately describes imbalanced class distributions commonly found in clinical variables. The generated data has a patient disclosure risk of 0.095%, lower than the 9% threshold stated by Health Canada and the European Medicines Agency, making it suitable for distribution to the research community with high security. The generated data also has high utility, indicating the potential of the proposed method to enable the development of downstream machine learning algorithms for healthcare applications using synthetic data. CONCLUSION: Our proposed extension to the classic GAN framework, which includes a VAE and an external memory mechanism, represents a promising approach towards generating synthetic data that accurately describe imbalanced class distributions commonly found in clinical variables. This method overcomes the limitations of GANs and creates more realistic datasets with higher patient cohort diversity, facilitating the development of downstream machine learning algorithms for healthcare applications.

Increasing national trend of direct-acting antiviral discontinuation among people treated for HCV 2016-2021.

BACKGROUND: Direct-acting antiviral (DAA) treatment discontinuation may negatively impact HCV elimination efforts. In Australia, DAA therapy is pharmacy dispensed, generally in 4-week amounts, with the approved duration (8-24 wk) and volume dispensed reported in pharmaceutical administrative data. This analysis assessed national HCV treatment discontinuation. METHODS: Individuals commencing DAAs between 2016 and 2021 were assessed for treatment discontinuation. Individuals with a single dispensation of their entire treatment course were excluded. Treatment discontinuation was defined as ≥4 weeks of approved treatment duration not dispensed. Factors associated with treatment discontinuation were assessed using Cox regression. Factors associated with retreatment following treatment discontinuation were assessed using logistic regression. RESULTS: Of 95,275 individuals who were treated, 88,986 were included in the analysis of whom 7532 (9%) discontinued treatment. Treatment discontinuation increased from 6% in the first half of 2016 to 15% in 2021. Longer treatment durations (vs. 8 wk) were associated with increased discontinuation risk (12 wk: adjusted HR = 3.23; 95% CI: 2.90, 3.59; p < 0.001, 16-24 wk: adjusted HR = 6.29; 95% CI: 5.55, 7.14; p < 0.001). Of individuals discontinuing treatment, 24% were retreated. Early discontinuation (4 wk treatment dispensed) increased the likelihood of retreatment (adjusted OR = 3.91; 95% CI: 3.44, 4.44; p < 0.001). Those with early discontinuation of glecaprevir/pibrentasvir 8 weeks (vs. sofosbuvir/velpatasvir 12 wk) had a lower likelihood of retreatment (adjusted OR = 0.62; 95% CI: 0.49, 0.79; p < 0.001). Initial treatment discontinuation was associated with an increased risk of retreatment discontinuation (adjusted HR = 4.41; 3.85, 5.05; p < 0.001). CONCLUSIONS: DAA treatment discontinuation increased over time corresponding to increasing treatment uptake through primary care among people who inject drugs. The use of simplified, short-duration therapies may reduce treatment discontinuation. Access to adherence support and retreatment will be essential for HCV elimination.

Associations of Breast Arterial Calcifications with Cardiovascular Disease.

Cardiovascular diseases (CVD), including coronary artery disease (CAD), continue to be the leading cause of global mortality among women. While traditional CVD/CAD prevention tools play a significant role in reducing morbidity and mortality among both men and women, current tools for preventing CVD/CAD rely on traditional risk factor-based algorithms that often underestimate CVD/CAD risk in women compared with men. In recent years, some studies have suggested that breast arterial calcifications (BAC), which are benign calcifications seen in mammograms, may be linked to CVD/CAD. Considering that millions of women older than 40 years undergo annual screening mammography for breast cancer as a regular activity, innovative risk prediction factors for CVD/CAD involving mammographic data could offer a gender-specific and convenient solution. Such factors that may be independent of, or complementary to, current risk models without extra cost or radiation exposure are worthy of detailed investigation. This review aims to discuss relevant studies examining the association between BAC and CVD/CAD and highlights some of the issues related to previous studies' design such as sample size, population types, method of assessing BAC and CVD/CAD, definition of cardiovascular events, and other confounding factors. The work may also offer insights for future CVD risk prediction research directions using routine mammograms and radiomic features other than BAC such as breast density and macrocalcifications.

National trends in retreatment of HCV due to reinfection or treatment failure in Australia.

BACKGROUND & AIMS: Population-level uptake of direct-acting antiviral (DAA) treatment for hepatitis C virus (HCV) infection, including retreatment, can be estimated through administrative pharmaceutical dispensation data. However, the reasons for retreatment are not captured in these data. We developed a machine learning model to classify retreatments as reinfection or treatment failure at a national level. METHODS: Retreatment data from the REACH-C cohort (n = 10,843 treated with DAAs; n = 320 retreatments with known reason), were used to train a random forest model. Nested cross validation was undertaken to assess model performance and to optimise hyperparameters. The model was applied to data on DAA retreatment dispensed during 2016-2021 in Australia, to identify the reason for retreatment (treatment failure or reinfection). RESULTS: Average predictive accuracy, precision, sensitivity, specificity and F1-score for the model were 96.3%, 96.5%, 96.3%, 96.3% and 96.3%, respectively. Nationally, 95,272 individuals initiated DAAs, with treatment uptake declining from 32,454 in 2016 to 6,566 in 2021. Of those treated, 6,980 (7%) were retreated. Our model classified 51.8% (95% CI 46.7-53.6%; n = 3,614) of cases as reinfection and 48.2% (95% CI 46.4-53.3%; n = 3,366) as treatment failure. Retreatment for reinfection increased steadily over the study period from 14 in 2016 to 1,092 in 2020, stabilising in 2021. Retreatment for treatment failure increased from 73 in 2016 to 1,077 in 2019, then declined to 515 in 2021. Among individuals retreated for treatment failure, 50% had discontinued initial treatment. CONCLUSIONS: We used a novel methodology with high classification accuracy to evaluate DAA retreatment patterns at a national level. Increases in retreatment uptake for treatment failure corresponded to the availability of pangenotypic and salvage regimens. Increasing retreatment uptake for reinfection likely reflects increasing reinfection incidence. IMPACT AND IMPLICATIONS: This study used machine learning methodologies to analyse national administrative data and characterise trends in HCV retreatment due to reinfection and treatment failure. Retreatment for reinfection increased over time, reflecting increasing numbers of people at risk for reinfection following HCV cure. Increased retreatment for treatment failure corresponded to the availability of pangenotypic and salvage DAA regimens. The findings of this study can be used by public health agencies and policy makers to guide and assess HCV elimination strategies, while the novel methodology for monitoring trends in HCV retreatment has the potential to be used in other settings, and health conditions.

The Health Gym: synthetic health-related datasets for the development of reinforcement learning algorithms.

In recent years, the machine learning research community has benefited tremendously from the availability of openly accessible benchmark datasets. Clinical data are usually not openly available due to their confidential nature. This has hampered the development of reproducible and generalisable machine learning applications in health care. Here we introduce the Health Gym - a growing collection of highly realistic synthetic medical datasets that can be freely accessed to prototype, evaluate, and compare machine learning algorithms, with a specific focus on reinforcement learning. The three synthetic datasets described in this paper present patient cohorts with acute hypotension and sepsis in the intensive care unit, and people with human immunodeficiency virus (HIV) receiving antiretroviral therapy. The datasets were created using a novel generative adversarial network (GAN). The distributions of variables, and correlations between variables and trends in variables over time in the synthetic datasets mirror those in the real datasets. Furthermore, the risk of sensitive information disclosure associated with the public distribution of the synthetic datasets is estimated to be very low.

Trends in opioid analgesic utilisation among people with a history of opioid dependence.

BACKGROUND: We aimed to characterise opioid analgesic utilisation over a 16-year period among a cohort of people with a history of opioid dependence, comparing rates of use in and out of opioid agonist treatment (OAT). METHODS: Retrospective cohort study in New South Wales, Australia, including 28,891 people with documented opioid dependence initiating opioid analgesics between July 2003 and December 2018. Linked administrative records provided data on prescription dispensings, sociodemographics, clinical characteristics, and OAT. Generalised estimating equation models estimated the incidence and adjusted incidence rate ratios (IRR) comparing periods in and out of OAT for the number of opioid analgesic dispensings (overall, for strong opioids, and the most commonly dispensed opioid types) and the amount dispensed in oral morphine equivalent milligrams (OME). RESULTS: At initiation, 43.7% of the cohort were enrolled in OAT. The most commonly initiated opioid was codeine (including combinations with paracetamol; 67.8%), and 49.6% of the cohort were dispensed a psychotropic medicine in the previous 90 days. Incidence of all opioid analgesic dispensings was higher during periods out of OAT compared to in OAT (5.8 v. 2.3 dispensings per person-year; IRR 0.39, 95% CI 0.38, 0.41), with findings similar when stratified by type. Being in OAT was associated with a lower OME amount dispensed compared to out of OAT (-57.7%, 95% CI-58.8, -56.7). CONCLUSIONS: People with opioid dependence had high rates of recent psychotropic medicine utilisation and current OAT enrolment at the time of opioid analgesic initiation. OAT was associated with a significant reduction in opioid analgesic dispensing.

Deep learning DCE-MRI parameter estimation: Application in pancreatic cancer.

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is an MRI technique for quantifying perfusion that can be used in clinical applications for classification of tumours and other types of diseases. Conventionally, the non-linear least squares (NLLS) methods is used for tracer-kinetic modelling of DCE data. However, despite promising results, NLLS suffers from long processing times (minutes-hours) and noisy parameter maps due to the non-convexity of the cost function. In this work, we investigated physics-informed deep neural networks for estimating physiological parameters from DCE-MRI signal-curves. Three voxel-wise temporal frameworks (FCN, LSTM, GRU) and two spatio-temporal frameworks (CNN, U-Net) were investigated. The accuracy and precision of parameter estimation by the temporal frameworks were evaluated in simulations. All networks showed higher precision than the NLLS. Specifically, the GRU showed to decrease the random error on ve by a factor of 4.8 with respect to the NLLS for noise (SD) of 1/20. The accuracy was better for the prediction of the ve parameter in all networks compared to the NLLS. The GRU and LSTM worked with arbitrary acquisition lengths. The GRU was selected for in vivo evaluation and compared to the spatio-temporal frameworks in 28 patients with pancreatic cancer. All neural network approaches showed less noisy parameter maps than the NLLS. The GRU had better test-retest repeatability than the NLLS for all three parameters and was able to detect one additional patient with significant changes in DCE parameters post chemo-radiotherapy. Although the U-Net and CNN had even better test-retest characteristics than the GRU, and were able to detect even more responders, they also showed potential systematic errors in the parameter maps. Therefore, we advise using our GRU framework for analysing DCE data.

Incorporating uncertainty in learning to defer algorithms for safe computer-aided diagnosis.

Deep neural networks are increasingly being used for computer-aided diagnosis, but erroneous diagnoses can be extremely costly for patients. We propose a learning to defer with uncertainty (LDU) algorithm which identifies patients for whom diagnostic uncertainty is high and defers them for evaluation by human experts. LDU was evaluated on the diagnosis of myocardial infarction (using discharge summaries), the diagnosis of any comorbidities (using structured data), and the diagnosis of pleural effusion and pneumothorax (using chest x-rays), and compared with 'learning to defer without uncertainty information' (LD) and 'direct triage by uncertainty' (DT) methods. LDU achieved the same F1 score as LD but deferred considerably fewer patients (e.g. 36% vs. 69% deferral rate for diagnosing pleural effusion with an F1 score of 0.96). Furthermore, even when many patients were assigned the wrong diagnosis with high confidence (e.g. for the diagnosis of any comorbidities) LDU achieved a 17% increase in F1 score, whereas DT was not applicable. Importantly, the weight of the defer loss in LDU can be easily adjusted to obtain the desired trade-off between diagnostic accuracy and deferral rate. In conclusion, LDU can readily augment any existing diagnostic network to reduce the risk of erroneous diagnoses in clinical practice.

Predicting cardiovascular risk from national administrative databases using a combined survival analysis and deep learning approach.

BACKGROUND: Machine learning-based risk prediction models may outperform traditional statistical models in large datasets with many variables, by identifying both novel predictors and the complex interactions between them. This study compared deep learning extensions of survival analysis models with Cox proportional hazards models for predicting cardiovascular disease (CVD) risk in national health administrative datasets. METHODS: Using individual person linkage of administrative datasets, we constructed a cohort of all New Zealanders aged 30-74 who interacted with public health services during 2012. After excluding people with prior CVD, we developed sex-specific deep learning and Cox proportional hazards models to estimate the risk of CVD events within 5 years. Models were compared based on the proportion of explained variance, model calibration and discrimination, and hazard ratios for predictor variables. RESULTS: First CVD events occurred in 61 927 of 2 164 872 people. Within the reference group, the largest hazard ratios estimated by the deep learning models were for tobacco use in women (2.04, 95% CI: 1.99, 2.10) and chronic obstructive pulmonary disease with acute lower respiratory infection in men (1.56, 95% CI: 1.50, 1.62). Other identified predictors (e.g. hypertension, chest pain, diabetes) aligned with current knowledge about CVD risk factors. Deep learning outperformed Cox proportional hazards models on the basis of proportion of explained variance (R2: 0.468 vs 0.425 in women and 0.383 vs 0.348 in men), calibration and discrimination (all P <0.0001). CONCLUSIONS: Deep learning extensions of survival analysis models can be applied to large health administrative datasets to derive interpretable CVD risk prediction equations that are more accurate than traditional Cox proportional hazards models.

A Machine Learning Approach to Predict the Added-Sugar Content of Packaged Foods.

BACKGROUND: Dietary guidelines recommend limiting the intake of added sugars. However, despite the public health importance, most countries have not mandated the labeling of added-sugar content on packaged foods and beverages, making it difficult for consumers to avoid products with added sugar, and limiting the ability of policymakers to identify priority products for intervention. OBJECTIVE: The aim was to develop a machine learning approach for the prediction of added-sugar content in packaged products using available nutrient, ingredient, and food category information. METHODS: The added-sugar prediction algorithm was developed using k-nearest neighbors (KNN) and packaged food information from the US Label Insight dataset (n = 70,522). A synthetic dataset of Australian packaged products (n = 500) was used to assess validity and generalization. Performance metrics included the coefficient of determination (R2), mean absolute error (MAE), and Spearman rank correlation (ρ). To benchmark the KNN approach, the KNN approach was compared with an existing added-sugar prediction approach that relies on a series of manual steps. RESULTS: Compared with the existing added-sugar prediction approach, the KNN approach was similarly apt at explaining variation in added-sugar content (R2 = 0.96 vs. 0.97, respectively) and ranking products from highest to lowest in added-sugar content (ρ = 0.91 vs. 0.93, respectively), while less apt at minimizing absolute deviations between predicted and true values (MAE = 1.68 g vs. 1.26 g per 100 g or 100 mL, respectively). CONCLUSIONS: KNN can be used to predict added-sugar content in packaged products with a high degree of validity. Being automated, KNN can easily be applied to large datasets. Such predicted added-sugar levels can be used to monitor the food supply and inform interventions aimed at reducing added-sugar intake.

Using administrative data to predict cessation risk and identify novel predictors among new entrants to opioid agonist treatment.

BACKGROUND: Longer retention in opioid agonist treatment (OAT) is associated with improved treatment outcomes but 12-month retention rates are often low. Innovative approaches are needed to strengthen retention in OAT. We develop and compare traditional and deep learning-extensions of Cox regression to examine the potential for predicting time in OAT at individuals' first episode entry. METHODS: Retrospective cohort study in New South Wales, Australia including 16,576 people entering OAT for the first time between January 2006 and December 2017. We develop 12-month OAT cessation prediction models using traditional and deep learning-extensions of the Cox regression algorithm with predictors evaluated from linked administrative datasets. Proportion of explained variation, calibration, and discrimination are compared using 5 × 2 cross-validation. RESULTS: Twelve-month cessation rate was 58.4%. The largest hazard ratios for earlier cessation from the deep learning model were observed for treatment factors, including private dosing points (HR=1.54, 95% CI=1.49-1.60) and buprenorphine medication (HR=1.43, 95% CI=1.39-1.46). Diagnostic codes for homelessness (HR=1.09, 95% CI=1.04-1.13), outpatient treatment for drug use disorders (HR=1.10, 95% CI=1.06-1.15), and occupant of vehicle accident (HR=1.04, 95% CI=1.01-1.07) from past-year health service presentations were identified as significant predictors of retention. We observed no improvement in performance of the deep learning model over traditional Cox regression. CONCLUSIONS: Deep learning may be more useful in identifying novel risk factors of OAT retention from administrative data than evaluating individual-level risk. An increased focus on addressing structural issues at the population level and considering alternate models of care may be more effective at improving retention than delivering fully personalised OAT.

Psychotropic medicine prescribing and polypharmacy for people with dementia entering residential aged care: the influence of changing general practitioners.

OBJECTIVE: To examine relationships between changing general practitioner after entering residential aged care and overall medicines prescribing (including polypharmacy) and that of psychotropic medicines in particular. DESIGN: Retrospective data linkage study. SETTING, PARTICIPANTS: 45 and Up Study participants in New South Wales with dementia who were PBS concession card holders and entered permanent residential aged care during January 2010 - June 2014 and were alive six months after entry. MAIN OUTCOME MEASURES: Inverse probability of treatment-weighted numbers of medicines dispensed to residents and proportions of residents dispensed antipsychotics, benzodiazepines, and antidepressants in the six months after residential care entry, by most frequent residential care GP category: usual (same as during two years preceding entry), known (another GP, but known to the resident), or new GP. RESULTS: Of 2250 new residents with dementia (mean age, 84.1 years; SD, 7.0 years; 1236 women [55%]), 625 most frequently saw their usual GPs (28%), 645 saw known GPs (29%), and 980 saw new GPs (44%). The increase in mean number of dispensed medicines after residential care entry was larger for residents with new GPs (+1.6 medicines; 95% CI, 1.4-1.9 medicines) than for those attended by their usual GPs (+0.7 medicines; 95% CI, 0.4-1.1 medicines; adjusted rate ratio, 2.42; 95% CI, 1.59-3.70). The odds of being dispensed antipsychotics (adjusted odds ratio [aOR], 1.59; 95% CI, 1.18-2.12) or benzodiazepines (aOR, 1.69; 95% CI, 1.25-2.30), but not antidepressants (aOR, 1.32; 95% CI, 0.98-1.77), were also higher for the new GP group. Differences between the known and usual GP groups were not statistically significant. CONCLUSIONS: Increases in medicine use and rates of psychotropic dispensing were higher for people with dementia who changed GP when they entered residential care. Facilitating continuity of GP care for new residents and more structured transfer of GP care may prevent potentially inappropriate initiation of psychotropic medicines.

Improved unsupervised physics-informed deep learning for intravoxel incoherent motion modeling and evaluation in pancreatic cancer patients.

PURPOSE: Earlier work showed that IVIM-NETorig , an unsupervised physics-informed deep neural network, was faster and more accurate than other state-of-the-art intravoxel-incoherent motion (IVIM) fitting approaches to diffusion-weighted imaging (DWI). This study presents a substantially improved version, IVIM-NEToptim , and characterizes its superior performance in pancreatic cancer patients. METHOD: In simulations (signal-to-noise ratio [SNR] = 20), the accuracy, independence, and consistency of IVIM-NET were evaluated for combinations of hyperparameters (fit S0, constraints, network architecture, number of hidden layers, dropout, batch normalization, learning rate), by calculating the normalized root-mean-square error (NRMSE), Spearman's ρ, and the coefficient of variation (CVNET ), respectively. The best performing network, IVIM-NEToptim was compared to least squares (LS) and a Bayesian approach at different SNRs. IVIM-NEToptim 's performance was evaluated in an independent dataset of 23 patients with pancreatic ductal adenocarcinoma. Fourteen of the patients received no treatment between two repeated scan sessions and nine received chemoradiotherapy between the repeated sessions. Intersession within-subject standard deviations (wSD) and treatment-induced changes were assessed. RESULTS: In simulations (SNR = 20), IVIM-NEToptim outperformed IVIM-NETorig in accuracy (NRMSE(D) = 0.177 vs 0.196; NMRSE(f) = 0.220 vs 0.267; NMRSE(D*) = 0.386 vs 0.393), independence (ρ(D*, f) = 0.22 vs 0.74), and consistency (CVNET (D) = 0.013 vs 0.104; CVNET (f) = 0.020 vs 0.054; CVNET (D*) = 0.036 vs 0.110). IVIM-NEToptim showed superior performance to the LS and Bayesian approaches at SNRs < 50. In vivo, IVIM-NEToptim showed significantly less noisy parameter maps with lower wSD for D and f than the alternatives. In the treated cohort, IVIM-NEToptim detected the most individual patients with significant parameter changes compared to day-to-day variations. CONCLUSION: IVIM-NEToptim is recommended for accurate, informative, and consistent IVIM fitting to DWI data.

Big data and predictive modelling for the opioid crisis: existing research and future potential.

A need exists to accurately estimate overdose risk and improve understanding of how to deliver treatments and interventions in people with opioid use disorder in a way that reduces such risk. We consider opportunities for predictive analytics and routinely collected administrative data to evaluate how overdose could be reduced among people with opioid use disorder. Specifically, we summarise global trends in opioid use and overdoses; describe the use of big data in research into opioid overdose; consider the potential for predictive modelling, including machine learning, for prevention and monitoring of opioid overdoses; and outline the challenges and risks relating to the use of big data and machine learning in reducing harms that are related to opioid use. Future research for improving the coverage and provision of existing interventions, treatments, and resources for opioid use disorder requires collaboration of multiple agencies. Predictive modelling could transport the concept of stratified medicine to public health through novel methods, such as predictive modelling and emulated trials for evaluating diagnoses and prognoses of opioid use disorder, predicting treatment response, and providing targeted treatment recommendations.

The effect of person, treatment and prescriber characteristics on retention in opioid agonist treatment: a 15-year retrospective cohort study.

BACKGROUND AND AIMS: There is limited evidence on the relationship between retention in opioid agonist treatment for opioid dependence and characteristics of treatment prescribers. This study estimated retention in buprenorphine and methadone treatment and its relationship with person, treatment and prescriber characteristics. DESIGN: Retrospective longitudinal study. SETTING: New South Wales, Australia. PARTICIPANTS: People entering the opioid agonist treatment programme for the first time between August 2001 and December 2015. MEASUREMENTS: Time in opioid agonist treatment (primary outcome) was modelled using a generalized estimating equation model to estimate associations with person, treatment and prescriber characteristics. FINDINGS: The impact of medication type on opioid agonist treatment retention reduced over time; the risk of leaving treatment when on buprenorphine compared with methadone was higher among those who entered treatment earlier [e.g. 2001-03: odds ratio (OR) = 1.59, 95% confidence interval (CI) = 1.45-1.75] and lowest among those who entered most recently (2013-15: OR = 1.23, 95% CI = 1.11-1.36). In adjusted analyses, risk of leaving was reduced among people whose prescriber had longer tenure of prescribing (e.g. 3 versus 8 years: OR = 0.94, 95% CI = 0.93-0.95) compared with prescribers with shorter tenure. Aboriginal and Torres Strait Islander people, being of younger age, past-year psychosis disorder and having been convicted of more criminal charges in the year prior to treatment entry were associated with increased risk of leaving treatment. CONCLUSION: In New South Wales, Australia, retention in buprenorphine treatment for opioid dependence, compared with methadone, has improved over time since its introduction in 2001. Opioid agonist treatment retention is affected not only by characteristics of the person and his or her treatment, but also of the prescriber, with those of longer prescribing tenure associated with increased retention of people in opioid agonist treatment.

Medications used disproportionately during pregnancy: Priorities for research on the risks and benefits of medications when used during pregnancy.

PURPOSE: To identify medications used disproportionately more or less among pregnant women relative to women of childbearing age. METHODS: Medication use among pregnant women in New South Wales, Australia was identified using linked perinatal and pharmaceutical dispensing data from 2006 to 2012. Medication use in women of childbearing age (including pregnant women) was identified using pharmaceutical dispensing data for a 10% random sample of the Australian population. Pregnant social security beneficiaries (n = 111 612) were age-matched (1:3) to female social security beneficiaries in the 10% sample. For each medication, the risk it was dispensed during pregnancy relative to being dispensed during an equivalent time period among matched controls was computed. Medications were mapped to Australian pregnancy risk categories. RESULTS: Of the 181 included medications, 35 were statistically significantly more commonly dispensed to pregnant women than control women. Of these, 23 are categorised as posing no increased risk to the foetus. Among medications suspected of causing harm or having insufficient safety data, the strongest associations were observed for hydralazine, ondansetron, dalteparin sodium and ranitidine. Use was less likely during pregnancy than control periods for 127 medications, with the strongest associations observed for hormonal contraceptives and progestogens. CONCLUSIONS: Most medications found to be used disproportionately more by pregnant women are indicated for pregnancy-related problems. A large number of medications were used disproportionately less among pregnant women, where avoidance of some of these medications may pose a greater risk of harm. For many other medications avoided during pregnancy, current data are insufficient to inform this risk-benefit assessment.