Real-world Data Research Articles

Enhancing Criminal Analysis through Multi-Model Integration: Addressing Challenges and Ensuring Ethical Implementation

Abstract: Law enforcement agencies face numerous challenges in criminal analysis, from data quality issues to ethical concerns. To address these challenges, this research proposes a novel approach: integrating multiple modelling techniques into a unified framework. By combining regression analysis, decision trees, support vector machines, neural networks, and ensemble methods, this approach aims to provide more robust and accurate solutions. The research seeks to develop a framework for integration, evaluate effectiveness using real-world data, and explore ethical implications. Ultimately, the goal is to advance criminal analysis, promote multi-model integration in law enforcement, and ensure ethical implementation for justice.

Assessing the Impact of Seasonal Decomposition on the Time Series Analysis Accuracy: A Comprehensive Study

Abstract: Time series forecasting plays a pivotal role in decision-making across various domains, ranging from finance to healthcare and weather prediction. The accurate prediction of future values in a time series is vital for informed planning and resource allocation. The aim of this study is to explore whether the utilization of seasonal decomposition techniques, such as classical decomposition, X-12-ARIMA, and seasonal decomposition of time series (STL), can improve the effectiveness of time series forecasting models by separating the data into its distinct components, including trend and seasonality. We conduct a comprehensive analysis using real-world time series data, employing popular forecasting models like ARIMA, exponential smoothing, and machine learning-based approaches. By comparing the forecasting accuracy of these models with and without the application of seasonal decomposition techniques, we provide empirical evidence to support the hypothesis. Our research results provide valuable insights into the tangible ad- vantages of integrating seasonal decomposition techniques in time series forecasting, potentially contributing to enhanced decision support systems across diverse application domains.

Artificial Intelligence and Machine Learning in Hepatocellular Carcinoma Screening, Diagnosis and Treatment - A Comprehensive Systematic Review

Background: Medical image analysis plays a crucial role in the screening, monitoring, diagnosis, and prognosis of diseases. Hepatocellular Carcinoma (HCC) often remain asymptomatic in their early stages. Timely identification and intervention are crucial to prevent the progression to decompensated liver diseases and advanced-stage HCC, minimizing morbidity and mortality. Methodology: This study examined twenty relevant research articles that met the inclusion criteria. Utilizing the PRISMA criteria, a systematic search was conducted on PubMed/MEDLINE and Google Scholar for studies on HCC screening employing Artificial Intelligence (AI) and Machine Learning (ML). The search focused on the keywords "artificial intelligence" and "hepatocellular carcinoma," with inclusion criteria specifying studies in the English language published in and after 2020. Exclusions were made for histology, animal research, and investigations conducted before 2020. Titles and abstracts were thoroughly reviewed, and any discrepancies were discussed. Results: The comprehensive review reveals the transformative impact of AI and ML on HCC screening, diagnosis, and therapy. ML models demonstrated effectiveness in early HCC diagnosis, distinguishing hepatic lesions, predicting treatment responses, and assessing recurrence risks across various techniques. Integration of mass spectrometry-based technologies, advanced imaging, and real-world data significantly improved diagnostic accuracy and clinical decision support. AI models exhibited diagnostic expertise with potential applications in therapy suggestions, personalized surveillance, and prognostic evaluations. However, further validation and seamless integration into clinical practice are essential for realizing their full potential. Conclusion: This systematic study underscores the progress made in the application of AI and ML in HCC screening, diagnosis, and treatment. Numerous studies highlight the capability of AI and ML systems to enhance hepatocellular carcinoma diagnosis, predict outcomes, and optimize therapy. The enduring and versatile nature of these technologies points towards a revolutionary future in personalized and efficient HCC management.

Evaluating the accuracy of a state-of-the-art large language model for prediction of admissions from the emergency room.

Artificial intelligence (AI) and large language models (LLMs) can play a critical role in emergency room operations by augmenting decision-making about patient admission. However, there are no studies for LLMs using real-world data and scenarios, in comparison to and being informed by traditional supervised machine learning (ML) models. We evaluated the performance of GPT-4 for predicting patient admissions from emergency department (ED) visits. We compared performance to traditional ML models both naively and when informed by few-shot examples and/or numerical probabilities. We conducted a retrospective study using electronic health records across 7 NYC hospitals. We trained Bio-Clinical-BERT and XGBoost (XGB) models on unstructured and structured data, respectively, and created an ensemble model reflecting ML performance. We then assessed GPT-4 capabilities in many scenarios: through Zero-shot, Few-shot with and without retrieval-augmented generation (RAG), and with and without ML numerical probabilities. The Ensemble ML model achieved an area under the receiver operating characteristic curve (AUC) of 0.88, an area under the precision-recall curve (AUPRC) of 0.72 and an accuracy of 82.9%. The naïve GPT-4's performance (0.79 AUC, 0.48 AUPRC, and 77.5% accuracy) showed substantial improvement when given limited, relevant data to learn from (ie, RAG) and underlying ML probabilities (0.87 AUC, 0.71 AUPRC, and 83.1% accuracy). Interestingly, RAG alone boosted performance to near peak levels (0.82 AUC, 0.56 AUPRC, and 81.3% accuracy). The naïve LLM had limited performance but showed significant improvement in predicting ED admissions when supplemented with real-world examples to learn from, particularly through RAG, and/or numerical probabilities from traditional ML models. Its peak performance, although slightly lower than the pure ML model, is noteworthy given its potential for providing reasoning behind predictions. Further refinement of LLMs with real-world data is necessary for successful integration as decision-support tools in care settings.

Cost-effectiveness of sotagliflozin for the treatment of patients with diabetes and recent worsening heart failure.

Aim: To assesses the cost-effectiveness of sotagliflozin for the treatment of patients hospitalized with heart failure and comorbid diabetes. Materials & methods: Ade novo cost-effectiveness model with a Markov structure was created for patients hospitalized for heart failure with comorbid diabetes. Outcomes of interest included hospital readmissions, emergency department visits and all-cause mortality measured over a 30-year time horizon. Baseline event frequencies were derived from published real-world data studies; sotagliflozin's efficacy was estimated from SOLOIST-WHF. Health benefits were calculated quality-adjusted life years (QALYs). Costs included pharmaceutical costs, rehospitalization, emergency room visits and adverse events. Economic value was measured using the incremental cost-effectiveness ratio (ICER). Results: Sotagliflozin use decreased annualized rehospitalization rates by 34.5% (0.228 vs 0.348, difference: -0.120), annualized emergency department visits by 40.0% (0.091 vs 0.153, difference: -0.061) and annualized mortality by 18.0% (0.298 vs 0.363, difference: -0.065) relative to standard of care, resulting in a net gain in QAYs of 0.425 for sotagliflozin versus standard of care. Incremental costs using sotagliflozin increased by $19,374 over a 30-year time horizon of the patient, driven largely by increased pharmaceutical cost. Estimated ICER for sotagliflozin relative to standard of care was $45,596 per QALY. Conclusion: Sotagliflozin is a cost-effective addition to standard of care for patients hospitalized with heart failure and comorbid diabetes.

Real-life use of ropeg-interferon α2b in polycythemia vera: patient selection and clinical outcomes.

Ropeginterferon-alfa2b (ropegIFNα2b) is a long-acting IFN formulation with broad FDA/EMA approval as a therapy of polycythemia vera (PV) with no symptomatic splenomegaly. There is currently lack of information on the real-world patient selection, including the impact of local reimbursement policies, and drug management, particularly: type/timing of screening and follow-up tests; absolute/relative contraindications to therapy; ropegIFNα2b dose and combinations with hydroxyurea. As a sub-analysis of the PV-ARC retrospective study (NCT06134102), we here report our monocenter experience with ropegIFNα2b in the period from January 2021, corresponding to drug availability outside clinical trial, and December 2023. Among the 149 patients with EMA/FDA indication, only 55 (36.9%) met the local reimbursement criteria and 18 (12.1%) received ropegIFNα2b. Thanks to appropriate screening, relative/absolute contraindications to ropegIFNα2b were detected and managed in a multidisciplinary manner. Efficacy and safety of ropegIFNα2b was confirmed, with 3 cases of early molecular response. General use of low ropegIFNα2b dose, with frequent need for hydroxyurea combinations, was noted. This real-world experience suggests a significant impact of local regulations on drug prescription and the need for greater real-world data collection on ropegIFNα2b in PV patients. Also, it describes appropriate multidisciplinary screening and monitoring procedures during ropegIFNα2b therapy.

Success rates of intensive aphasia therapy: real-world data from 448 patients between 2003 and 2020.

Aphasia is a devastating consequence after stroke, affecting millions of patients each year. Studies have shown that intensive speech and language therapy (SLT) is effective in the chronic phase of aphasia. Leveraging a large single-center cohort of persons with aphasia (PWA) including patients also in the subacute phase, we assessed treatment effects of intensive aphasia therapy in a real-world setting. Data were collected at the Aachen aphasia ward in Germany between 2003 and 2020. Immediate treatment responses across different language domains were assessed with the Aachen Aphasia Test (AAT) using single-case psychometrics, conducted before and after 6-7 weeks of intensive SLT (10 h per week, median (IQR) dosage = 68 (61-76)). We adjusted for spontaneous recovery in subacute patients. Differential treatment effects between subgroups of chronicity and predictors of therapy response were investigated. A total of 448 PWA were included (29% female, median (IQR) age = 54 (46-62) years, median (IQR) time post-onset = 11 (6-20) months) with 12% in the early subacute, 15% in the late subacute and 74% in the chronic phase of aphasia. The immediate responder rate was 59%. Significant improvements in all AAT subtests und subscales were observed hinting at broad effectiveness across language domains. The degree of therapy-induced improvement did not differ between the chronicity groups. Time post-onset, dosage of therapy and aphasia severity at the beginning of treatment were predictors of immediate treatment response. Intensive therapy protocols for aphasia after stroke are yielding substantial responder rates in a routine clinical setting including a wide range of patients.

Tackling school segregation with transportation network interventions: an agent-based modelling approach

We address the emerging challenge of school segregation within the context of free school choice systems. Households take into account both proximity and demographic composition when deciding on which schools to send their children to, potentially exacerbating residential segregation. This raises an important question: can we strategically intervene in transportation networks to enhance school access and mitigate segregation? In this paper, we propose a novel, network agent-based model to explore this question. Through simulations in both synthetic and real-world networks, we demonstrate that enhancing school accessibility via transportation network interventions can lead to a reduction in school segregation, under specific conditions. We introduce group-based network centrality measures and show that increasing the centrality of certain neighborhood nodes with respect to a transportation network can be an effective strategy for strategic interventions. We conduct experiments in two synthetic network environments, as well as in an environment based on real-world data from Amsterdam, the Netherlands. In both cases, we simulate a population of representative agents emulating real citizens’ schooling preferences, and we assume that agents belong to two different groups (e.g., based on migration background). We show that, under specific homophily regimes in the population, school segregation can be reduced by up to 35%. Our proposed framework provides the foundation to explore how citizens’ preferences, school capacity, and public transportation can shape patterns of urban segregation.

Target trial emulation of carfilzomib safety among patients with relapsed/refractory multiple myeloma using a nationwide observational data in Korea

PurposeCarfilzomib, commonly used for relapsed/refractory multiple myeloma (RRMM), has been associated with various adverse events in randomized controlled trials (RCTs). However, real-world safety data for a more diverse population are needed, as carfilzomib received expedited approval. This study aimed to evaluate carfilzomib’s safety in Korea by comparing new users of KRd (carfilzomib, lenalidomide, and dexamethasone) to Rd (lenalidomide and dexamethasone) using a nationwide administrative claims database.MethodsThe retrospective cohort study utilized target trial emulation, focusing on adverse events in various organ systems similar to the ASPIRE trial.ResultsThis study included 4,580 RRMM patients between 2007 and 2020, and the KRd group showed significantly higher risks of hematologic adverse events (anemia, neutropenia, thrombocytopenia) and some non-hematologic adverse events (cough, hypokalemia, constipation, hypertension, heart failure) compared to the Rd group. Among non-hematologic adverse events, cardiovascular events (heart failure [HR 2.04; 95% CI 1.24–3.35], hypertension [HR 1.58; 95% CI 1.15–2.17]) had the highest risk in the KRd group.ConclusionThe safety profile of carfilzomib in Korean patients was similar to previous RCTs. Therefore, caution should be exercised when using carfilzomib in Asian individuals with RRMM due to the increased risk of cardiovascular adverse events.

Transformer-Powered Surrogates Close the ICF Simulation-Experiment Gap with Extremely Limited Data

Abstract Recent advances in machine learning, specifically transformer architecture, have led to significant advancements in commercial domains. These powerful models have demonstrated superior capability to learn complex relationships and often generalize better to new data and problems. This paper presents a novel transformer-powered approach for enhancing prediction accuracy in multi-modal output scenarios, where sparse experimental data is supplemented with simulation data. The proposed approach integrates transformer-based architecture with a novel graph-based hyper-parameter optimization technique. The resulting system not only effectively reduces simulation bias, but also achieves superior prediction accuracy compared to the prior method. We demonstrate the efficacy of our approach on inertial confinement fusion experiments, where only 10 shots of real-world data are available, as well as synthetic versions of these experiments.

Balancing Data Acquisition Benefits and Ordering Costs for Predictive Supplier Selection and Order Allocation

The strategic selection of suppliers and the allocation of orders across multiple periods have long been recognized as critical aspects influencing company expenditure and resilience. Leveraging the enhanced predictive capabilities afforded by machine learning models, direct lookahead models—linear programming models that optimize future decisions based on forecasts generated by external predictive modules—have emerged as viable alternatives to traditional deterministic and stochastic programming methodologies to solve related problems. However, despite these advancements, approaches implementing direct lookahead models typically lack mechanisms for updating forecasts over time. Yet, in practice, suppliers often exhibit dynamic behaviours, and failing to update forecasts can lead to suboptimal decision-making. This study introduces a novel approach based on parametrized direct lookahead models to address this gap. The approach explicitly addresses the hidden trade-offs associated with incorporating forecast updates. Recognizing that forecasts can only be updated by acquiring new data and that the primary means of acquiring supplier-related data is through order allocation, this study investigates the trade-offs between data acquisition benefits and order allocation costs. An experimental design utilizing real-world automotive sector data is employed to assess the potential of the proposed approach against various benchmarks. These benchmarks include decision scenarios representing perfect foresight, no data acquisition benefits, and consistently positive benefits. Empirical findings demonstrate that the proposed approach achieves performance levels comparable to those of decision-makers with perfect foresight while consistently outperforming benchmarks not balancing order allocation costs and data acquisition benefits.

Developing an Efficient Model for Online Grocery Order Fulfillment

Due to the convenience of online grocery apps and home delivery, online grocery shopping has become popular in recent years. Globally, consumer behavior has significantly changed the consumption and purchase patterns of online grocery shopping. This study aimed to develop an efficient model for online grocery order fulfillment that both reduces costs and increases supply chain efficiency and sustainability. This study first aimed to develop the current picking model by adopting real-world data from a store in Riyadh, Saudi Arabia. Subsequently, four proposed models were developed to improve the efficiency and sustainability of the online grocery order fulfillment process. The results show a significant improvement in all models over the current picking model. The percentage improvements in fulfillment time per product are as follows: single order picking—8.33%; batch order picking—6.78%; zone order picking—3.08%; and hybrid order picking—13.20%, which combines zone and batch order picking. Retailers and online grocery apps could adopt these models to increase efficiency and sustainability. Also, these models have great potential for future research and improvement by optimizing product placement, in addition to picking methods and picking routes, which are the focus of this study.

An Assessment of Semaglutide Safety Based on Real World Data: From Popularity to Spontaneous Reporting in EudraVigilance Database

Some glucagon-like peptide-1 receptor agonists (GLP-1 RAs), first used in the treatment of type 2 diabetes mellitus (T2DM), have been approved for the treatment of obesity in patients with or without T2DM (liraglutide—LIR, semaglutide—SEM, and tirzepatide—TIR). Social media had an important influence on the off-label use of GLP-1 RAs for obesity, especially for SEM. We analyzed the Google queries related to SEM to assess people’s interest in this drug. We also investigated the occurrence of adverse drug reactions (ADRs) by searching the EudraVigilance database (EV) for Individual Case Safety Reports (ICSRs) that reported SEM as the suspected drug and performed a descriptive and a disproportionality analysis. The data obtained for SEM were compared to other GLP-1 RAs. SEM had the highest proportions of searches on Google associated with the term “weight loss” and presented the lowest number of severe ADRs, but it also had the highest number of ICSRs reported in EV. Even though no unexpected safety issues have been reported for it until now, SEM has a hi3gh tendency for overdose reports. The most frequent off-label use was reported for SEM and TIR. In order to lower the risks of ADRs, the off-label use should be reduced and carefully monitored.

Validation of algorithms in studies based on routinely collected health data: general principles.

Clinicians, researchers, regulators, and other decision-makers increasingly rely on evidence from real-world data (RWD), including data routinely accumulating in health and administrative databases. RWD studies often rely on algorithms to operationalize variable definitions. An algorithm is a combination of codes or concepts used to identify persons with a specific health condition or characteristic. Establishing the validity of algorithms is a prerequisite for generating valid study findings that can ultimately inform evidence-based health care. This paper aims to systematize terminology, methods, and practical considerations relevant to the conduct of validation studies of RWD-based algorithms. We discuss measures of algorithm accuracy; gold/reference standard; study size; prioritizing accuracy measures; algorithm portability; and implication for interpretation. Information bias is common in epidemiologic studies, underscoring the importance of transparency in decisions regarding choice and prioritizing measures of algorithm validity. The validity of an algorithm should be judged in the context of a data source, and one size does not fit all. Prioritizing validity measures within a given data source depends on the role of a given variable in the analysis (eligibility criterion, exposure, outcome or covariate). Validation work should be part of routine maintenance of RWD sources.

Predictive Modelling on Competitor Analysis Performance by using Generalised Linear Models and Machine Learning Approach

Competitive analysis in digital and technology is trending in the business field. However, the field of digital and technology in the business world is vast and challenging to analyse. The purpose of this research is first to identify the success factor which represent the company performance. Second, is to identify the significant services provided by the company to their business user. Then, based on the first and second objectives, a predictive modelling is developed to produce the best solution to their business user. The research is implementing a case study from Telecommunication Company and using data science life cycle methodology. The statistical modelling that is used to develop the competitor’s analysis model is generalised linear model (GLM) which integrated with machine learning approach. Furthermore, the synthetic data set is created by using Gamma Distribution, Gaussian Distribution and Poisson Distribution due to some data from the case study is confidential. The synthetic data set is based on existing real data which are from Telecommunication Company sentiment analysis data, were used to investigate the performance of the proposed model. The machine learning technique is used to get the accuracy of the significant GLM which has been developed. The accuracy is tested by using the error rates of the machine learning technique which are Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and R-squared. This research discovered that the business solution to the significant service for the business user and discovered the best statistical model to be used for the business solution. The results show that the Gumbel distribution is the best fit model for the synthetic dataset where the values of RMSE is 1.0574, MAE is 0.9168 and R-squared is 0.3994, and the significant success factor that has been identified by using the GLM is advertising success factor. The model developed can be improved with another type of data set and different sizes of data. Hence, further studies and real-world data are required for better validation.

Effectiveness, safety and discontinuation rates of bictegravir/emtricitabine/tenofovir alafenamide (BIC/FTC/TAF) in people with HIV using real-world data: a systematic review and meta-analysis.

The use of bictegravir/emtricitabine/tenofovir alafenamide (BIC/FTC/TAF) is based on the results of robust clinical trials. To assess the effectiveness and safety of BIC/FTC/TAF in treatment-naïve (TN) and treatment-experienced (TE) people with HIV using available real-world cohort studies. Systematic review and meta-analysis of publications and communications identified via Boolean search in Medline, PubMed and Embase, and conference abstracts reporting retrospective real-world use of BIC/FTC/TAF, published until 31 January 2024. The primary endpoint was the proportion of TN and TE people with HIV with viral load (VL) < 50 copies/mL at 48 weeks while on treatment. Of the 38 identified publications and conference abstracts, for the present analysis we included 12 publications (comprising 792 TN and 6732 TE individuals). For the three publications including 507 TN participants reporting the primary outcome, VL suppression was 97% [95% confidence intervals (CI): 89-100]. For the nine publications including 4946 TE participants reporting the primary outcome, VL suppression was 95% (95% CI: 94-96), with suppression >93% in all studies. Total discontinuations at 48 weeks in TE individuals were 3% (95% CI: 2-5), 1% (95% CI: 0-2) due to side effects. A total of four publications with 151 TE individuals with previous presence of M184V substitution were identified, reporting a suppression rate at 48 weeks of 95% (95% CI: 88-100). Real-world studies demonstrate low discontinuation rates and high rates of virologic suppression in individuals treated with BIC/FTC/TAF, both TN and TE with and without previous detection of M184V substitution.

How power shapes behavior: Evidence from physicians.

Power-the asymmetric control of valued resources-affects most human interactions. Although power is challenging to study with real-world data, a distinctive dataset allowed us to do so within the critical context of doctor-patient relationships. Using 1.5 million quasi-random assignments in US military emergency departments, we examined how power differentials between doctor and patient (measured by using differences in military ranks) affect physician behavior. Our findings indicate that power confers nontrivial advantages: "High-power" patients (who outrank their physician) receive more resources and have better outcomes than equivalently ranked "low-power" patients. Patient promotions even increase physician effort. Furthermore, low-power patients suffer if their physician concurrently cares for a high-power patient. Doctor-patient concordance on race and sex also matters. Overall, power-driven variation in behavior can harm the most vulnerable populations in health care settings.

Longitudinal Changes in Diagnostic Accuracy of a Differential Diagnosis List Developed by an AI-Based Symptom Checker: Retrospective Observational Study.

Artificial intelligence (AI) symptom checker models should be trained using real-world patient data to improve their diagnostic accuracy. Given that AI-based symptom checkers are currently used in clinical practice, their performance should improve over time. However, longitudinal evaluations of the diagnostic accuracy of these symptom checkers are limited. This study aimed to assess the longitudinal changes in the accuracy of differential diagnosis lists created by an AI-based symptom checker used in the real world. This was a single-center, retrospective, observational study. Patients who visited an outpatient clinic without an appointment between May 1, 2019, and April 30, 2022, and who were admitted to a community hospital in Japan within 30 days of their index visit were considered eligible. We only included patients who underwent an AI-based symptom checkup at the index visit, and the diagnosis was finally confirmed during follow-up. Final diagnoses were categorized as common or uncommon, and all cases were categorized as typical or atypical. The primary outcome measure was the accuracy of the differential diagnosis list created by the AI-based symptom checker, defined as the final diagnosis in a list of 10 differential diagnoses created by the symptom checker. To assess the change in the symptom checker's diagnostic accuracy over 3 years, we used a chi-square test to compare the primary outcome over 3 periods: from May 1, 2019, to April 30, 2020 (first year); from May 1, 2020, to April 30, 2021 (second year); and from May 1, 2021, to April 30, 2022 (third year). A total of 381 patients were included. Common diseases comprised 257 (67.5%) cases, and typical presentations were observed in 298 (78.2%) cases. Overall, the accuracy of the differential diagnosis list created by the AI-based symptom checker was 172 (45.1%), which did not differ across the 3 years (first year: 97/219, 44.3%; second year: 32/72, 44.4%; and third year: 43/90, 47.7%; P=.85). The accuracy of the differential diagnosis list created by the symptom checker was low in those with uncommon diseases (30/124, 24.2%) and atypical presentations (12/83, 14.5%). In the multivariate logistic regression model, common disease (P<.001; odds ratio 4.13, 95% CI 2.50-6.98) and typical presentation (P<.001; odds ratio 6.92, 95% CI 3.62-14.2) were significantly associated with the accuracy of the differential diagnosis list created by the symptom checker. A 3-year longitudinal survey of the diagnostic accuracy of differential diagnosis lists developed by an AI-based symptom checker, which has been implemented in real-world clinical practice settings, showed no improvement over time. Uncommon diseases and atypical presentations were independently associated with a lower diagnostic accuracy. In the future, symptom checkers should be trained to recognize uncommon conditions.

Mining Multimorbidity Trajectories and Co-Medication Effects from Patient Data to Predict Post–Hip Fracture Outcomes

Hip fractures have profound impacts on patients’ conditions and quality of life, even when they receive therapeutic treatments. Many patients face the risk of poor prognosis, physical impairment, and even mortality, especially older patients. Accurate patient outcome estimates after an initial fracture are critical to physicians’ decision-making and patient management. Effective predictions might benefit from analyses of patients’ multimorbidity trajectories and medication usages. If adequately modeled and analyzed, they could help identify patients at higher risk of recurrent fractures or mortality. Most analytics methods overlook the onset, co-occurrence, and temporal sequence of distinct chronic diseases in the trajectory, and they also seldom consider the combined effects of different medications. To support effective predictions, we develop a novel deep learning–based method that uses a cross-attention mechanism to model patient progression by obtaining “contextual information” from multimorbidity trajectories. This method also incorporates a nested self-attention network that captures the combined effects of distinct medications by learning the interactions among medications and how dosages might influence post-fracture outcomes. A real-world patient data set is used to evaluate the proposed method, relative to six benchmark methods. The comparative results indicate that our method consistently outperforms all the benchmarks in precision, recall, F-measures, and area under the curve. The proposed method is generalizable and can be implemented as a decision support system to identify patients at greater risk of recurrent hip fractures or mortality, which should help clinical decision-making and patient management.

Avoiding Delays in Reporting Time-to-Event Randomized Trials: Calendar Backstops and Other Approaches.

New oncology therapies that extend patients' lives beyond initial expectations and improving later-line treatments can lead to complications in clinical trial design and conduct. In particular, for trials with event-based analyses, the time to observe all the protocol-specified events can exceed the finite follow-up of a clinical trial or can lead to much delayed release of outcome data. With the advent of multiple classes of oncology therapies leading to much longer survival than in the past, this issue in clinical trial design and conduct has become increasingly important in recent years. We propose a straightforward prespecified backstop rule for trials with a time-to-event analysis and evaluate the impact of the rule with both simulated and real-world trial data. We then provide recommendations for implementing the rule across a range of oncology clinical trial settings.