RESUMEN

Intubation for mechanical ventilation (MV) is one of the most common high-risk procedures performed in Intensive Care Units (ICUs). Early prediction of intubation may have a positive impact by providing timely alerts to clinicians and consequently avoiding high-risk late intubations. In this work, we propose a new machine learning method to predict the time to intubation during the first five days of ICU admission, based on the concept of cure survival models. Our approach combines classification and survival analysis, to effectively accommodate the fraction of patients not at risk of intubation, and provide a better estimate of time to intubation, for patients at risk. We tested our approach and compared it to other predictive models on a dataset collected from a secondary care hospital (AZ Groeninge, Kortrijk, Belgium) from 2015 to 2021, consisting of 3425 ICU stays. Furthermore, we utilised SHAP for feature importance analysis, extracting key insights into the relative significance of variables such as vital signs, blood gases, and patient characteristics in predicting intubation in ICU settings. The results corroborate that our approach improves the prediction of time to intubation in critically ill patients, by using routinely collected data within the first hours of admission in the ICU. Early warning of the need for intubation may be used to help clinicians predict the risk of intubation and rank patients according to their expected time to intubation.

Asunto(s)

RESUMEN

When individuals are released from prison, they typically enter a period of post confinement community supervision. While under community supervision, their behaviors are subject to special conditions requiring them to report to supervisors and prohibiting certain behaviors such as drug and alcohol use. Many supervisees are returned to prison because they violate those special conditions, or because they commit minor crimes that would not result in prison were they not being supervised. But others are returned to prison for serious new crimes. We distinguish the two as nuisance behaviors (the former) and pernicious behaviors (the latter). Our research applies competing events survival analysis to distinguish a structural model that accounts for nuisance behaviors from a structural model that accounts for pernicious behaviors. We demonstrate that returning offenders to prison for technical violations and minor crimes may reduce the incidence of major crimes because the occurrence of nuisance behaviors and pernicious behaviors are highly correlated. Our findings support the theory that nuisance behaviors signal the likelihood of pernicious behaviors.

RESUMEN

AIM: Kidney transplantation remains the preferred standard of care for patients with kidney failure. Most patients do not access this treatment and wide variations exist in which patients access transplantation. We sought to develop a model to estimate post-kidney transplant survival to inform more accurate comparisons of access to kidney transplantation. METHODS: Development and validation of prediction models using demographic and clinical data from the Australia and New Zealand Dialysis and Transplant Registry. Adult deceased donor kidney only transplant recipients between 2000 and 2020 were included. Cox proportional hazards regression methods were used with a primary outcome of patient survival. Models were evaluated using Harrell's C-statistic for discrimination, and calibration plots, predicted survival probabilities and Akaike Information Criterion for goodness-of-fit. RESULTS: The model development and validation cohorts included 11 302 participants. Most participants were male (62.8%) and Caucasian (79.2%). Glomerulonephritis was the most common cause of kidney disease (45.6%). The final model included recipient, donor, and transplant related variables. The model had good discrimination (C-statistic, 0.72; 95% confidence interval (CI) 0.70-0.74 in the development cohort, 0.70; 95% CI 0.67-0.73 in the validation cohort and 0.72; 95% CI 0.69-0.75 in the temporal cohort) and was well calibrated. CONCLUSION: We developed a statistical model that predicts post-kidney transplant survival in Australian kidney failure patients. This model will aid in assessing the suitability of kidney transplantation for patients with kidney failure. Survival estimates can be used to make more informed comparisons of access to transplantation between units to better measure equity of access to organ transplantation.

Asunto(s)

RESUMEN

PURPOSE: This study introduces a novel method for estimating the variance of life expectancy since diagnosis (LEC) and loss in life expectancy (LLE) for cancer patients within a relative survival framework in situations where life tables based on the entire general population are not accessible. LEC and LLE are useful summary measures of survival in population-based cancer studies, but require information on the mortality in the general population. Our method addresses the challenge of incorporating the uncertainty of expected mortality rates when using a sample from the general population. METHODS: To illustrate the approach, we estimated LEC and LLE for patients diagnosed with colon and breast cancer in Sweden. General population mortality rates were based on a random sample drawn from comparators of a matched cohort. Flexible parametric survival models were used to model the mortality among cancer patients and the mortality in the random sample from the general population. Based on the models, LEC and LLE together with their variances were estimated. The results were compared with those obtained using fixed expected mortality rates. RESULTS: By accounting for the uncertainty of expected mortality rates, the proposed method ensures more accurate estimates of variances and, therefore, confidence intervals of LEC and LLE for cancer patients. This is particularly valuable for older patients and some cancer types, where underestimation of the variance can be substantial when the entire general population data are not accessible. CONCLUSION: The method can be implemented using existing software, making it accessible for use in various cancer studies. The provided example of Stata code further facilitates its adoption.

Asunto(s)

RESUMEN

Objective: To compare the performance of radiomics-based machine learning survival models in predicting the prognosis of glioblastoma multiforme (GBM) patients. Methods: 131 GBM patients were included in our study. The traditional Cox proportional-hazards (CoxPH) model and four machine learning models (SurvivalTree, Random survival forest (RSF), DeepSurv, DeepHit) were constructed, and the performance of the five models was evaluated using the C-index. Results: After the screening, 1792 radiomics features were obtained. Seven radiomics features with the strongest relationship with prognosis were obtained following the application of the least absolute shrinkage and selection operator (LASSO) regression. The CoxPH model demonstrated that age (HR = 1.576, p = 0.037), Karnofsky performance status (KPS) score (HR = 1.890, p = 0.006), radiomics risk score (HR = 3.497, p = 0.001), and radiomics risk level (HR = 1.572, p = 0.043) were associated with poorer prognosis. The DeepSurv model performed the best among the five models, obtaining C-index of 0.882 and 0.732 for the training and test set, respectively. The performances of the other four models were lower: CoxPH (0.663 training set / 0.635 test set), SurvivalTree (0.702/0.655), RSF (0.735/0.667), DeepHit (0.608/0.560). Conclusion: This study confirmed the superior performance of deep learning algorithms based on radiomics relative to the traditional method in predicting the overall survival of GBM patients; specifically, the DeepSurv model showed the best predictive ability.

RESUMEN

Policies aimed at increasing employment among older people often focus on the statutory retirement age. Taking into account the characteristics of workers and work-related factors, we examine the impact of reaching the statutory retirement age on continuing employment. In addition to the use of survival trees, we propose a novel method to predict the probability of staying in employment based on an ensemble of survival trees. We focus on Poland as an example of a European country with a particularly low share of older workers in the labor force. Moreover, reform was carried out in Poland in 2017, lowering the previously raised pension eligibility age. Like other EU countries, pension eligibility in Poland starts after reaching the statutory retirement age. Our results suggest that the timing of retirement is determined by the statutory retirement age to a limited extent compared to other factors. In the case of women, a match of education and occupation, the employment sector, and holding a managerial position had a greater impact on continuing employment than reaching retirement age. In the case of men, the type of job contract had the greatest impact on continuing employment. Our findings indicate that the policies and initiatives aimed at extending working life should pay more attention to work-related factors and gender differences in employment.

RESUMEN

Neural-network-based outcome predictions may enable further treatment personalization of patients with head and neck cancer. The development of neural networks can prove challenging when a limited number of cases is available. Therefore, we investigated whether multitask learning strategies, implemented through the simultaneous optimization of two distinct outcome objectives (multi-outcome) and combined with a tumor segmentation task, can lead to improved performance of convolutional neural networks (CNNs) and vision transformers (ViTs). Model training was conducted on two distinct multicenter datasets for the endpoints loco-regional control (LRC) and progression-free survival (PFS), respectively. The first dataset consisted of pre-treatment computed tomography (CT) imaging for 290 patients and the second dataset contained combined positron emission tomography (PET)/CT data of 224 patients. Discriminative performance was assessed by the concordance index (C-index). Risk stratification was evaluated using log-rank tests. Across both datasets, CNN and ViT model ensembles achieved similar results. Multitask approaches showed favorable performance in most investigations. Multi-outcome CNN models trained with segmentation loss were identified as the optimal strategy across cohorts. On the PET/CT dataset, an ensemble of multi-outcome CNNs trained with segmentation loss achieved the best discrimination (C-index: 0.29, 95% confidence interval (CI): 0.22-0.36) and successfully stratified patients into groups with low and high risk of disease progression (p=0.003). On the CT dataset, ensembles of multi-outcome CNNs and of single-outcome ViTs trained with segmentation loss performed best (C-index: 0.26 and 0.26, CI: 0.18-0.34 and 0.18-0.35, respectively), both with significant risk stratification for LRC in independent validation (p=0.002 and p=0.011). Further validation of the developed multitask-learning models is planned based on a prospective validation study, which has recently completed recruitment.

RESUMEN

OBJECTIVES: Between 1997 and 2021, the number of children looked after (CLA) in Wales, UK, increased steadily, with stark inequalities. We aimed to assess how deprivation and maternal and child perinatal characteristics influence the risk of becoming CLA in Wales. STUDY DESIGN: We constructed a prospective longitudinal cohort of children born in Wales between April 2006 and March 2021 (n = 395,610) using linked administrative records. METHODS: Survival models examined the risk of CLA from birth by small-area deprivation and maternal and child perinatal characteristics. Population attributable fractions quantify the potential impact of action on modifiable risk factors. RESULTS: Children from the most deprived fifth of the population were 3.4 times more likely to enter care than those in the least deprived (demographic adjusted hazard ratios [aHRs] 3.40, 95% confidence interval [CI] 3.08, 3.74). Maternal mental health problems in pregnancy (fully aHR, 2.03, 95% CI 1.88, 2.19) and behavioural factors, such as smoking (aHR 2.46, 95% CI 2.34-2.60), alcohol problems (aHR 2.35, 95% CI 1.70-3.23) and substance use in pregnancy (aHR 5.72, 95% CI 5.03-6.51), as well as child congenital anomalies (aHR 1.46, 95% CI 1.16-1.84), low birth weight (aHR 1.28, 95% CI 1.17, 1.39) and preterm birth (aHR 1.16, 95% CI 1.06, 1.26), were associated with higher risk of CLA status. The risk of CLA in the population may be reduced by 35% (95% CI 0.33, 0.38) if children in the two most deprived fifths of the population experienced the conditions of those in the least deprived. CONCLUSIONS: Deprivation and perinatal maternal health are important modifiable risk factors for children becoming CLA. Our analysis provides insight into the mechanisms of intergenerational transfer of disadvantage in a vulnerable section of the child population and identifies targets for public health action.

RESUMEN

BACKGROUND: The association of childhood adversities with mortality has rarely been explored, and even less studied is the question of whether any excess mortality may be potentially preventable. This study examined the association between specific childhood adversities and premature and potentially avoidable mortality (PPAM) in adulthood in a representative sample of the general population. Also, we examined whether the associations were potentially mediated by various adult socioeconomic, psychosocial, and behavioral factors. METHODS: The study used data from the National Population Health Survey (NPHS-1994) linked to the Canadian Vital Statistics Database (CVSD 1994-2014) available from Statistics Canada. The NPHS interview retrospectively assessed childhood exposure to prolonged hospitalization, parental divorce, prolonged parental unemployment, prolonged trauma, parental problematic substance use, physical abuse, and being sent away from home for doing something wrong. An existing definition of PPAM, consisting of causes of death considered preventable or treatable before age 75, was used. Competing cause survival models were used to examine the associations of specific childhood adversities with PPAM in adulthood among respondents aged 18 to 74 years (rounded n = 11,035). RESULTS: During the 20-year follow-up, 5.4% of the sample died prematurely of a cause that was considered potentially avoidable. Childhood adversities had a differential effect on mortality. Physical abuse (age-adjusted sub-hazard ratio; SHR 1.44; 95% CI 1.03, 2.00) and being sent away from home (age-adjusted SHR 2.26; 95% CI 1.43,3.57) were significantly associated with PPAM. The associations were attenuated when adjusted for adulthood factors, namely smoking, poor perceived health, depression, low perceived social support, and low income, consistent with possible mediating effects. Other adversities under study were not associated with PPAM. CONCLUSION: The findings imply that the psychological sequelae of childhood physical abuse and being sent away from home and subsequent uptake of adverse health behavior may lead to increased risk of potentially avoidable mortality. The potential mediators identified offer directions for future research to perform causal mediation analyses with suitable data and identify interventions aimed at preventing premature mortality due to potentially avoidable causes. Other forms of adversities, mostly related to household dysfunction, may not be determinants of the distal health outcome of mortality.

Asunto(s)

RESUMEN

BACKGROUND: Individual-level social capital prevents cognitive decline. However, a few studies have focused on the effects of community-level social capital on dementia. Therefore, we investigated the association between community-level social capital and dementia onset based on longitudinal study data on older adults in Japan. METHODS: We used longitudinal data from the Japan Gerontological Evaluation Study, obtained over nine years (2010-2019). In total, 35,921 physically and cognitively independent individuals (16,848 males and 19,073 females) aged ≥65 years and nested within 308 communities in seven municipalities participated in the study. Dementia onset was assessed using the public long-term care insurance registration. Social capital was assessed using three dimensions: civic participation, social cohesion, and reciprocity. We performed a two-level multilevel survival analysis stratified by sex, calculated hazard ratios (HRs), and 95% confidence intervals (CIs). RESULTS: During the follow-up, 6245 (17.4%) dementia onset cases were identified. The cumulative incidence of dementia was 16.2% in males and 18.4% in females. After adjusting for covariates, individual-level civic participation was associated with a lower incidence of dementia in both males and females (HR, 0.84; 95% CI, 0.77-0.92; HR, 0.78; 95% CI, 0.73-0.84). Community-level civic participation and social cohesion were associated with a lower incidence of dementia among females (HR, 0.96; 95% CI, 0.93-0.99; HR, 0.93; 95% CI, 0.88-0.98) and cross-level interaction on social cohesion among females (HR, 0.95; 95% CI, 0.90-0.99). CONCLUSIONS: Living in a community with high civic participation and social cohesion is associated with a lower incidence of dementia among older females. Therefore, promoting civic participation and social cohesion in the community may be a useful population-based strategy to delay or prevent the onset of dementia.

Asunto(s)

RESUMEN

Developing an efficient computational scheme for high-dimensional Bayesian variable selection in generalised linear models and survival models has always been a challenging problem due to the absence of closed-form solutions to the marginal likelihood. The Reversible Jump Markov Chain Monte Carlo (RJMCMC) approach can be employed to jointly sample models and coefficients, but the effective design of the trans-dimensional jumps of RJMCMC can be challenging, making it hard to implement. Alternatively, the marginal likelihood can be derived conditional on latent variables using a data-augmentation scheme (e.g., Pólya-gamma data augmentation for logistic regression) or using other estimation methods. However, suitable data-augmentation schemes are not available for every generalised linear model and survival model, and estimating the marginal likelihood using a Laplace approximation or a correlated pseudo-marginal method can be computationally expensive. In this paper, three main contributions are presented. Firstly, we present an extended Point-wise implementation of Adaptive Random Neighbourhood Informed proposal (PARNI) to efficiently sample models directly from the marginal posterior distributions of generalised linear models and survival models. Secondly, in light of the recently proposed approximate Laplace approximation, we describe an efficient and accurate estimation method for marginal likelihood that involves adaptive parameters. Additionally, we describe a new method to adapt the algorithmic tuning parameters of the PARNI proposal by replacing Rao-Blackwellised estimates with the combination of a warm-start estimate and the ergodic average. We present numerous numerical results from simulated data and eight high-dimensional genetic mapping data-sets to showcase the efficiency of the novel PARNI proposal compared with the baseline add-delete-swap proposal.

RESUMEN

BACKGROUND: . At present, the prognostic prediction in advanced oral cavity squamous cell carcinoma (OCSCC) is based on the tumor-node-metastasis (TNM) staging system, and the most used imaging modality in these patients is magnetic resonance image (MRI). With the aim to improve the prediction, we developed an MRI-based radiomic signature as a prognostic marker for overall survival (OS) in OCSCC patients and compared it with published gene expression signatures for prognosis of OS in head and neck cancer patients, replicated herein on our OCSCC dataset. METHODS: For each patient, 1072 radiomic features were extracted from T1 and T2-weighted MRI (T1w and T2w). Features selection was performed, and an optimal set of five of them was used to fit a Cox proportional hazard regression model for OS. The radiomic signature was developed on a multi-centric locally advanced OCSCC retrospective dataset (n = 123) and validated on a prospective cohort (n = 108). RESULTS: The performance of the signature was evaluated in terms of C-index (0.68 (IQR 0.66-0.70)), hazard ratio (HR 2.64 (95% CI 1.62-4.31)), and high/low risk group stratification (log-rank p < 0.001, Kaplan-Meier curves). When tested on a multi-centric prospective cohort (n = 108), the signature had a C-index of 0.62 (IQR 0.58-0.64) and outperformed the clinical and pathologic TNM stage and six out of seven gene expression prognostic signatures. In addition, the significant difference of the radiomic signature between stages III and IVa/b in patients receiving surgery suggests a potential association of MRI features with the pathologic stage. CONCLUSIONS: Overall, the present study suggests that MRI signatures, containing non-invasive and cost-effective remarkable information, could be exploited as prognostic tools.

RESUMEN

Background: Breast cancer (BC) survival prediction can be a helpful tool for identifying important factors selecting the effective treatment reducing mortality rates. This study aims to predict the time-related survival probability of BC patients in different molecular subtypes over 30 years of follow-up. Materials and methods: This study retrospectively analyzed 3580 patients diagnosed with invasive breast cancer (BC) from 1991 to 2021 in the Cancer Research Center of Shahid Beheshti University of Medical Science. The dataset contained 18 predictor variables and two dependent variables, which referred to the survival status of patients and the time patients survived from diagnosis. Feature importance was performed using the random forest algorithm to identify significant prognostic factors. Time-to-event deep-learning-based models, including Nnet-survival, DeepHit, DeepSurve, NMLTR and Cox-time, were developed using a grid search approach with all variables initially and then with only the most important variables selected from feature importance. The performance metrics used to determine the best-performing model were C-index and IBS. Additionally, the dataset was clustered based on molecular receptor status (i.e., luminal A, luminal B, HER2-enriched, and triple-negative), and the best-performing prediction model was used to estimate survival probability for each molecular subtype. Results: The random forest method identified tumor state, age at diagnosis, and lymph node status as the best subset of variables for predicting breast cancer (BC) survival probabilities. All models yielded very close performance, with Nnet-survival (C-index=0.77, IBS=0.13) slightly higher using all 18 variables or the three most important variables. The results showed that the Luminal A had the highest predicted BC survival probabilities, while triple-negative and HER2-enriched had the lowest predicted survival probabilities over time. Additionally, the luminal B subtype followed a similar trend as luminal A for the first five years, after which the predicted survival probability decreased steadily in 10- and 15-year intervals. Conclusion: This study provides valuable insight into the survival probability of patients based on their molecular receptor status, particularly for HER2-positive patients. This information can be used by healthcare providers to make informed decisions regarding the appropriateness of medical interventions for high-risk patients. Future clinical trials should further explore the response of different molecular subtypes to treatment in order to optimize the efficacy of breast cancer treatments.

RESUMEN

PURPOSE: To identify 5-year survival prognostic variables in patients with colorectal cancer (CRC) and to propose a survival prognostic score that also takes into account changes over time in the patient's health-related quality of life (HRQoL) status. METHODS: Prospective observational cohort study of CRC patients. We collected data from their diagnosis, intervention, and at 1, 2, 3, and 5 years following the index intervention, also collecting HRQoL data using the EuroQol-5D-5L (EQ-5D-5L), European Organization for Research and Treatment of Cancer's Quality of Life Questionnaire-Core 30 (EORTC-QLQ-C30), and Hospital Anxiety and Depression Scale (HADS) questionnaires. Multivariate Cox proportional models were used. RESULTS: We found predictors of mortality over the 5-year follow-up to be being older; being male; having a higher TNM stage; having a higher lymph node ratio; having a result of CRC surgery classified as R1 or R2; invasion of neighboring organs; having a higher score on the Charlson comorbidity index; having an ASA IV; and having worse scores, worse quality of life, on the EORTC and EQ-5D questionnaires, as compared to those with higher scores in each of those questionnaires respectively. CONCLUSIONS: These results allow preventive and controlling measures to be established on long-term follow-up of these patients, based on a few easily measurable variables. IMPLICATIONS FOR CANCER SURVIVORS: Patients with colorectal cancer should be monitored more closely depending on the severity of their disease and comorbidities as well as the perceived health-related quality of life, and preventive measures should be established to prevent adverse outcomes and therefore to ensure that better treatment is received. TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT02488161.

Asunto(s)

RESUMEN

This article focuses on shared frailty models for correlated failure times, as well as joint frailty models for the simultaneous analysis of recurrent events (eg, appearance of new cancerous lesions or hospital readmissions) and a major terminal event (typically, death). As extensions of the Cox model, these joint models usually assume a frailty proportional hazards model for each of the recurrent and terminal event processes. In order to extend these models beyond the proportional hazards assumption, our proposal is to replace these proportional hazards models with generalized survival models, for which the survival function is modeled as a linear predictor through a link function. Depending on the link function considered, these can be reduced to proportional hazards, proportional odds, additive hazards, or probit models. We first consider a fully parametric framework for the time and covariate effects. For proportional and additive hazards models, our approach also allows the use of smooth functions for baseline hazard functions and time-varying coefficients. The dependence between recurrent and terminal event processes is modeled by conditioning on a shared frailty acting differently on the two processes. Parameter estimates are provided using the maximum (penalized) likelihood method, implemented in the R package frailtypack (function GenfrailtyPenal). We perform simulation studies to assess the method, which is also illustrated on real datasets.

Asunto(s)

RESUMEN

BACKGROUND: In decision modeling with time-to-event data, there are a variety of parametric models that can be used to extrapolate the survival function. Each model implies a different hazard function, and in situations in which there is moderate censoring, this can result in quite different survival projections. External information such as expert opinion on long-term survival can more accurately characterize the uncertainty in these extrapolations. OBJECTIVE: We present a general and easily implementable approach to incorporate various types of expert opinions into parametric survival models, focusing on opinions about survival at various landmark time points. METHODS: Expert opinion is incorporated into parametric survival models using Bayesian and frequentist approaches. In the Bayesian method, expert opinion is included through a loss function and in the frequentist approach by penalizing the likelihood function, although in both cases the core approach is the same. The issue of aggregating multiple expert opinions is also considered. RESULTS: We apply this method to data from a leukemia trial and use previously elicited expert opinion on survival probabilities for that particular trial population at years 4 and 5 to inform our analysis. We take a robust approach to modeling expert opinion by using pooled distributions and fit a broad class of parametric models to the data. We also assess statistical goodness of fit of the models to both the observed data and expert opinion. CONCLUSIONS: Expert opinions can be implemented in a straightforward manner using this novel approach; however, more work is required on the correct elicitation of these quantities. HIGHLIGHTS: Presentation of a novel and open-source method to incorporate expert opinion into decision modeling.Extends upon earlier work in that expert opinion can be incorporated into a wide range of parametric models.Provides methodological guidance for directly including expert opinion in decision modeling, which is a research focus area in NICE TSD 21.1.

Asunto(s)

RESUMEN

In many applications of hierarchical models, there is often interest in evaluating the inherent heterogeneity in view of observed data. When the underlying hypothesis involves parameters resting on the boundary of their support space such as variances and mixture proportions, it is a usual practice to entertain testing procedures that rely on common heterogeneity assumptions. Such procedures, albeit omnibus for general alternatives, may entail a substantial loss of power for specific alternatives such as heterogeneity varying with covariates. We introduce a novel and flexible approach that uses covariate information to improve the power to detect heterogeneity, without imposing unnecessary restrictions. With continuous covariates, the approach does not impose a regression model relating heterogeneity parameters to covariates or rely on arbitrary discretizations. Instead, a scanning approach requiring continuous dichotomizations of the covariates is proposed. Empirical processes resulting from these dichotomizations are then used to construct the test statistics, with limiting null distributions shown to be functionals of tight random processes. We illustrate our proposals and results on a popular class of two-component mixture models, followed by simulation studies and applications to two real datasets in cancer and caries research.

Asunto(s)

RESUMEN

The confounding between fixed effects and (spatial) random effects in a regression setup is termed spatial confounding. This topic continues to gain attention and has been studied extensively in recent years, given that failure to account for this may lead to a suboptimal inference. To mitigate this, a variety of projection-based approaches under the class of restricted spatial models are available in the context of generalized linear mixed models. However, these projection approaches cannot be directly extended to the spatial survival context via frailty models due to dimension incompatibility between the fixed and spatial random effects. In this work, we introduce a two-step approach to handle this, which involves (i) projecting the design matrix to the dimension of the spatial effect (via dimension reduction) and (ii) assuring that the random effect is orthogonal to this new design matrix (confounding alleviation). Under a fully Bayesian paradigm, we conduct fast estimation and inference using integrated nested Laplace approximation. Both simulation studies and application to a motivating data evaluating respiratory cancer survival in the US state of California reveal the advantages of our proposal in terms of model performance and confounding alleviation, compared to alternatives.

Asunto(s)

RESUMEN

BACKGROUND: There are situations when we need to model multiple time-scales in survival analysis. A usual approach in this setting would involve fitting Cox or Poisson models to a time-split dataset. However, this leads to large datasets and can be computationally intensive when model fitting, especially if interest lies in displaying how the estimated hazard rate or survival change along multiple time-scales continuously. METHODS: We propose to use flexible parametric survival models on the log hazard scale as an alternative method when modelling data with multiple time-scales. By choosing one of the time-scales as reference, and rewriting other time-scales as a function of this reference time-scale, users can avoid time-splitting of the data. RESULT: Through case-studies we demonstrate the usefulness of this method and provide examples of graphical representations of estimated hazard rates and survival proportions. The model gives nearly identical results to using a Poisson model, without requiring time-splitting. CONCLUSION: Flexible parametric survival models are a powerful tool for modelling multiple time-scales. This method does not require splitting the data into small time-intervals, and therefore saves time, helps avoid technological limitations and reduces room for error.

Asunto(s)

RESUMEN

Introduction: Limbic-predominant age-related TAR DNA-binding protein 43 (TDP-43) encephalopathy (LATE) is a recently defined neurodegenerative disease. Currently, there is no effective way to make a prognosis of time to stage-specific future conversions at an individual level. Methods: After using the Kaplan-Meier estimation and log-rank test to confirm the heterogeneity of LATE progression, we developed a deep learning-based approach to assess the stage-specific probabilities of time to LATE conversions for different subjects. Results: Our approach could accurately estimate the disease incidence and transition to next stages: the concordance index was at least 82% and the integrated Brier score was less than 0.14. Moreover, we identified the top 10 important predictors for each disease conversion scenario to help explain the estimation results, which were clinicopathologically meaningful and most were also statistically significant. Discussion: Our study has the potential to provide individualized assessment for future time courses of LATE conversions years before their actual occurrence.