RESUMEN

The standard approach to regression modeling for cause-specific hazards with prospective competing risks data specifies separate models for each failure type. An alternative proposed by Lunn and McNeil (1995) assumes the cause-specific hazards are proportional across causes. This may be more efficient than the standard approach, and allows the comparison of covariate effects across causes. In this paper, we extend Lunn and McNeil (1995) to nested case-control studies, accommodating scenarios with additional matching and non-proportionality. We also consider the case where data for different causes are obtained from different studies conducted in the same cohort. It is demonstrated that while only modest gains in efficiency are possible in full cohort analyses, substantial gains may be attained in nested case-control analyses for failure types that are relatively rare. Extensive simulation studies are conducted and real data analyses are provided using the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO) study.

RESUMEN

Objective: To examine the potential difference in survival and risk of death between asymptomatic and symptomatic SARS-CoV-2 patients, controlled by age and gender for all the attendance in hospitals of Khyber Pakhtunkhwa (KP), Pakistan. Methods: In this retrospective study, the medical records of 6273 SARS-CoV-2 patients admitted to almost all hospitals in Khyber Pakhtunkhwa during the first wave of the coronavirus outbreak from March to June 2020 were analysed. The effects of gender, age, and being symptomatic on the survival of SARS-CoV-2 patients were assessed using cure-survival models as opposed to the conventional Cox proportional hazards model. Results: The prevalence of initially symptomatic patients was 55.8%, and the overall mortality rate was 11.8%. The fitted cure-survival models suggest that age affects the probability of death (incidence) but not the short-term survival time of patients (latency); symptomatic patients have a higher risk of death than their asymptomatic counterparts, but the survival time of symptomatic patients is longer on average; gender has no significant effect on the probability of death and survival time. Conclusion: The available data and statistical results suggest that asymptomatic and young patients are generally less susceptible to initial infection with SARS-CoV-2 and therefore have a lower risk of death. Our regression models show that uncured asymptomatic patients generally have poorer short-term survival than their uncured symptomatic counterparts. The association between gender and survival outcome was not significant.

RESUMEN

BACKGROUND: Several epidemiologic studies have revealed a higher risk of cancer in patients with diabetes mellitus (DM) relative to the general population. To investigate whether the use of acarbose was associated with higher/lower risk of new-onset cancers. METHOD: We conducted a retrospective cohort study, using a population-based National Health Insurance Research Database of Taiwan. Both inpatients and outpatients with newly onset DM diagnosed between 2000 and 2012 were collected. The Adapted Diabetes Complications Severity Index (aDCSI) was used to adjust the severity of DM. The Cox proportional hazards regression model was used to estimate the hazard ratio (HR) of disease. RESULTS: A total of 22 502 patients with newly diagnosed DM were enrolled. The Cox proportional hazards regression model indicating acarbose was neutral for risk for gastroenterological malignancies, when compared to the acarbose non-acarbose users group. However, when gastric cancer was focused, acarbose-user group had significantly lowered HR than non-acarbose users group (p = 0.003). After adjusted for age, sex, cancer-related comorbidity, severity of DM, and co-administered drugs, the HR of gastric cancer risk was 0.43 (95% CI = 0.25-0.74) for acarbose-user patients. CONCLUSION: This long-term population-based study demonstrated that acarbose might be associated with lowered risk of new-onset gastric cancer in diabetic patients after adjusting the severity of DM.

Asunto(s)

RESUMEN

BACKGROUND: American Indian/Alaska Natives (AI/ANs) disproportionately suffer from diabetes compared to non-Hispanic whites (NHW). In 2013, 69% of end-stage kidney disease (ESKD) in AI/ANs was caused by diabetes (ESKD-D) but accounts for only 44% of ESKD diagnoses in the overall USA population. Moreover, the diagnosis of diabetes and ESKD-D may be significantly related to social determinants of health. The purpose of this study was to conduct a survival analysis of AI/ANs and NHWs diagnosed with ESKD-D nationally and by Indian Health Service region and correlate the survival analysis to the Area Deprivation Index® (ADI®). METHODS: This manuscript reports a retrospective cohort analysis of 2021 United States Renal Data System data. Eligible patient records were AI/AN and NHWs with diabetes as the primary cause of ESKD and started dialysis on January 1, 2014, or later. RESULTS: A total of 81,862 patient records were included in this analysis, of which 1798 (2.2%) were AI/AN. AI/ANs survive longer, with an 18.4% decrease in risk of death compared to NHW. However, AI/ANs are diagnosed with ESKD-D and start dialysis earlier than NHWs. ADI® variables became significant as ADI® ratings increased, meaning persons with greater social disadvantage had worse survival outcomes. CONCLUSIONS: The findings reveal that AI/ANs have better survival outcomes than NWH, explained in part by initiating dialysis earlier than NHW. Additional research is needed to explore factors (e.g., social determinants; cultural; physiologic) that contribute to earlier diagnosis of ESKD-D in AI/ANs and the impact of prolonged dialysis on quality of life of those with ESKD-D.

RESUMEN

During patient follow-up in a randomized trial, some deaths may occur. Where death (or noncardiovascular death) is not part of an outcome of interest it is termed a competing risk. Conventional analyses (eg, Cox proportional hazards model) handle death similarly to other censored follow-up. Patients still alive are unrealistically assumed to be representative of those who died. The Fine and Gray model has been used to handle competing risks, but is often used inappropriately and can be misleading. We propose an alternative multiple imputation approach that plausibly accounts for the fact that patients who die tend also to be at high risk for the (unobserved) outcome of interest. This provides a logical framework for exploring the impact of a competing risk, recognizing that there is no unique solution. We illustrate these issues in 3 cardiovascular trials and in simulation studies. We conclude with practical recommendations for handling competing risks in future trials.

Asunto(s)

RESUMEN

BACKGROUND: Non-Hodgkin lymphoma (NHL) has been identified as a significant contributor to the cancer burden. This study investigates the incidence, mortality, and survival trends of NHL cancer in Brunei Darussalam from 2011 to 2020. METHODS: This is a registry-based retrospective study using de-identified data from the Brunei Darussalam Cancer Registry on patients diagnosed with NHL from 2011 to 2020 based on the ICD-10 codes C82-86. Statistical methods include descriptive statistics, age-specific and age-standardised incidence (ASIR) and mortality rates (ASMR), and joinpoint regression for trend analysis. Survival analysis was conducted using Kaplan-Meier plots, log-rank test, and Cox Proportional Hazards regression. RESULTS: From 2011 to 2020, 330 patients were diagnosed with NHL. The majority of patients were males (51.8%) and of Malay descent (82.7%). The age group most diagnosed was 55-74 years (42.3%), with a mean age at diagnosis being 55.1 years. The ASIRs were 12.12 for males and 10.39 per 100,000 for females; ASMRs were 6.11 for males and 4.76 per 100,000 for females. Diffuse large B-cell lymphoma was the most prevalent subtype, accounting for 39.1% of cases. The overall 5-year survival rate was 61.2%, with lower rates observed in older patients and those diagnosed at distant metastasis stage. Furthermore, older age and advanced stage diagnosis significantly increased mortality risk. NHL incidence and mortality rates in Brunei Darussalam remain stable over the period of 10 years, but highlights significant disparities in gender and age. CONCLUSIONS: The findings emphasize the importance of early detection and tailored treatments, especially for high-risk groups, in managing NHL's burden. These insights underline the need for focused healthcare strategies and continued research to address NHL's challenges.

Asunto(s)

RESUMEN

Objective: Explainable machine learning (XAI) was introduced in this study to improve the interpretability, explainability and transparency of the modelling results. The survex package in R was used to interpret and compare two survival models - the Cox proportional hazards regression (coxph) model and the random survival forest (rfsrc) model - and to estimate overall survival (OS) and its determinants in heart failure (HF) patients using these models. Methods: We selected 1159 HF patients hospitalised at the First Affiliated Hospital of Kunming Medical University. First, the performance of the two models was investigated using the C-index, the integrated C/D AUC, and the integrated Brier score. Second, a global explanation of the whole cohort was carried out using the time-dependent variable importance and the partial dependence survival profile. Finally, the SurvSHAP(t) and SurvLIME plots and the ceteris paribus survival profile were used to obtain a local explanation for each patient. Results: By comparing the C-index, the C/D AUC, and the Brier score, this study showed that the model performance of rfsrc was better than coxph. The global explanation of the whole cohort suggests that the C-reactive protein, lg BNP (brain natriuretic peptide), estimated glomerular filtration rate, albumin, age and blood chloride were significant unfavourable predictors of OS in HF patients in both the cxoph and the rfsrc models. By including individual patients in the model, we can provide a local explanation for each patient, which guides the clinician in individualising the patient's treatment. Conclusion: By comparison, we conclude that the model performance of rfsrc is better than that of coxph. These two predictive models, which address not only the whole population but also selected patients, can help clinicians personalise the treatment of each HF patient according to his or her specific situation.

RESUMEN

Background: Certain viral infections have been linked to the development of neurodegenerative diseases. This study aimed to investigate the association between cytomegalovirus (CMV) infection and five neurodegenerative diseases, spinal muscular atrophy (SMA) and related syndromes, Parkinson's disease (PD), Alzheimer's disease (AD), multiple sclerosis (MS), and disorders of the autonomic nervous system (DANS). Methods: This prospective cohort included white British individuals who underwent CMV testing in the UK Biobank from January 1, 2006 to December 31, 2021. A Cox proportional hazard model was utilized to estimate the future risk of developing five neurodegenerative diseases in individuals with or without CMV infection, adjusted for batch effect, age, sex, and Townsend deprivation index in Model 1, and additionally for type 2 diabetes, cancer, osteoporosis, vitamin D, monocyte count and leukocyte count in Model 2. Bidirectional Mendelian randomization was employed to validate the potential causal relationship between CMV infection and PD. Findings: A total of 8346 individuals, consisting of 4620 females (55.4%) and 3726 males (44.6%) who were white British at an average age of 56.74 (8.11), were included in this study. The results showed that CMV infection did not affect the risk of developing AD (model 1: HR [95% CI] = 1.01 [0.57, 1.81], P = 0.965; model 2: HR = 1.00 [0.56, 1.79], P = 0.999), SMA and related syndromes (model 1: HR = 3.57 [0.64, 19.80], P = 0.146; model 2: HR = 3.52 [0.63, 19.61], P = 0.152), MS (model 1: HR = 1.16 [0.45, 2.97], P = 0.756; model 2: HR = 1.16 [0.45, 2.97], P = 0.761) and DANS (model 1: HR = 0.65 [0.16, 2.66], P = 0.552; model 2: HR = 0.65 [0.16, 2.64], P = 0.543). Interestingly, it was found that participants who were CMV seronegative had a higher risk of developing PD compared to those who were seropositive (model 1: HR = 2.37 [1.25, 4.51], P = 0.009; model 2: HR = 2.39 [1.25, 4.54], P = 0.008) after excluding deceased individuals. This association was notably stronger in males (model 1: HR = 3.16 [1.42, 7.07], P = 0.005; model 2: HR = 3.41 [1.50, 7.71], P = 0.003), but no significant difference was observed in the female subgroup (model 1: HR = 1.28 [0.40, 4.07], P = 0.679; model 2: HR = 1.27 [0.40, 4.06], P = 0.684). However, a bidirectional Mendelian randomization analysis did not find a genetic association between CMV infection and PD. Interpretation: The study found that males who did not have a CMV infection were at a higher risk of developing PD. The findings provided a new viewpoint on the risk factors for PD and may potentially influence public health approaches for the disease. Funding: National Natural Science Foundation of China (81873776), Natural Science Foundation of Guangdong Province, China (2021A1515011681, 2023A1515010495).

RESUMEN

BACKGROUND: Predicting an individual's risk of death from COVID-19 is essential for planning and optimising resources. However, since the real-world mortality rate is relatively low, particularly in places like Hong Kong, this makes building an accurate prediction model difficult due to the imbalanced nature of the dataset. This study introduces an innovative application of graph convolutional networks (GCNs) to predict COVID-19 patient survival using a highly imbalanced dataset. Unlike traditional models, GCNs leverage structural relationships within the data, enhancing predictive accuracy and robustness. By integrating demographic and laboratory data into a GCN framework, our approach addresses class imbalance and demonstrates significant improvements in prediction accuracy. METHODS: The cohort included all consecutive positive COVID-19 patients fulfilling study criteria admitted to 42 public hospitals in Hong Kong between January 23 and December 31, 2020 (n = 7,606). We proposed the population-based graph convolutional neural network (GCN) model which took blood test results, age and sex as inputs to predict the survival outcomes. Furthermore, we compared our proposed model to the Cox Proportional Hazard (CPH) model, conventional machine learning models, and oversampling machine learning models. Additionally, a subgroup analysis was performed on the test set in order to acquire a deeper understanding of the relationship between each patient node and its neighbours, revealing possible underlying causes of the inaccurate predictions. RESULTS: The GCN model was the top-performing model, with an AUC of 0.944, considerably outperforming all other models (p < 0.05), including the oversampled CPH model (0.708), linear regression (0.877), Linear Discriminant Analysis (0.860), K-nearest neighbours (0.834), Gaussian predictor (0.745) and support vector machine (0.847). With Kaplan-Meier estimates, the GCN model demonstrated good discriminability between low- and high-risk individuals (p < 0.0001). Based on subanalysis using the weighted-in score, although the GCN model was able to discriminate well between different predicted groups, the separation was inadequate between false negative (FN) and true negative (TN) groups. CONCLUSION: The GCN model considerably outperformed all other machine learning methods and baseline CPH models. Thus, when applied to this imbalanced COVID survival dataset, adopting a population graph representation may be an approach to achieving good prediction.

Asunto(s)

RESUMEN

Objective: Several studies have investigated the correlation between blood lipids and homocysteine, but no clear conclusions have been defined yet. Therefore, we utilized data from National Health and Nutrition Examination Survey (NHANES) to explore the correlation between serum homocysteine (Hcy) levels and hyperlipidemia, which is determined by the levels of total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglycerides (TG). We believe this study can provide a scientific basis for the prevention and treatment of lipid abnormalities. Methods: The data used in this study were sourced from NHANES 1999-2006, linked with National Death Index mortality data from January 1999 to December 2019. We employed logistic regression to assess the associations between Hcy levels and the presence of hyperlipidemia. Additionally, survival analysis using Kaplan-Meier estimate and Cox proportional hazards regression model was conducted to evaluate the associations between Hcy levels and all-cause mortality in the hyperlipidemia population. Results: (1) A total of 13,661 subjects were included in the study. There were statistically significant differences in Hcy levels across different groups based on gender, age, race, marital status, education level, hypertension status, diabetes status, and Body Mass Index (BMI) (P < 0.05). (2) In the overall population, hyperhomocysteinemia (HHcy) was associated with an increased risk of high-TC hyperlipidemia (P < 0.05). Subgroup analysis by gender showed that HHcy in females was associated with an increased risk of dyslipidemia (OR = 1.30, 95% CI: 1.07-1.59, P < 0.05) and high-LDL-C hyperlipidemia (OR = 1.30, 95% CI: 1.00-1.68, P < 0.05). In addition, subgroup analysis by age revealed that HHcy in middle-aged people was associated with an increased risk of high-TC hyperlipidemia (OR = 1.21, 95% CI: 1.03-1.41, P < 0.05) and high-LDL-C hyperlipidemia (OR = 1.23, 95% CI: 1.06-1.43, P < 0.05). (3) HHcy was consistently associated with an increased mortality risk in the hyperlipidemia population (HR = 1.49, 95% CI: 1.35-1.65, P < 0.05). Conclusion: There was positive correlation between Hcy levels and the presence of hyperlipidemia. In the overall population, HHcy was associated with an increased risk of high-TC hyperlipidemia. Among females, HHcy is linked to an increased risk of dyslipidemia and high-LDL-C hyperlipidemia. In middle-aged people, HHcy was associated with an elevated risk of high-TC hyperlipidemia and high-LDL-C hyperlipidemia. In addition, HHcy increased the all-cause mortality rate in hyperlipidemia patients.

RESUMEN

Evidence linking greenness to all-site and site-specific cancers remains limited, and the complex role of air pollution in this pathway is unclear. We aimed to fill these gaps by using a large cohort in southern China. A total of 654,115 individuals were recruited from 2009 to 2015 and followed-up until December 2020. We calculated the normalized difference vegetation index (NDVI) in a 500-m buffer around the participants' residences to represent the greenness exposure. Cox proportional-hazards models were used to evaluate the impact of greenness on the risk of all-site and site-specific cancer mortality. Additionally, we assessed both the mediation and interaction roles of air pollution (i.e., PM2.5, PM10, and NO2) in the greenness-cancer association through a causal mediation analysis using a four-way decomposition method. Among the 577,643 participants, 10,088 cancer deaths were recorded. We found a 10% (95% CI: 5-16%) reduction in all-site cancer mortality when the NDVI increased from the lowest to the highest quartile. When stratified by cancer type, our estimates suggested 18% (95% CI: 8-27%) and 51% (95% CI: 16-71%) reductions in mortality due to respiratory system cancer and brain and nervous system cancer, respectively. For the above protective effect, a large proportion could be explained by the mediation (all-site cancer: 1.0-27.7%; respiratory system cancer: 1.2-32.3%; brain and nervous system cancer: 3.6-109.1%) and negative interaction (all-site cancer: 2.1-25.7%; respiratory system cancer: 2.0-25.7%; brain and nervous system cancer: not significant) effects of air pollution. We found that particulate matter (i.e., PM2.5 and PM10) had a stronger causal mediation effect (25.0-109.1%) than NO2 (1.0-3.6%), while NO2 had a stronger interaction effect (25.7%) than particulate matter (2.0-2.8%). In summary, greenness was significantly beneficial in reducing the mortality of all-site, respiratory system, and brain and nervous system cancer in southern China, with the impact being modulated and mediated by air pollution.

Asunto(s)

RESUMEN

BACKGROUND: Severe heart failure (HF) has a higher mortality during vulnerable period while targeted predictive tools, especially based on drug exposures, to accurately assess its prognoses remain largely unexplored. Therefore, this study aimed to utilize drug information as the main predictor to develop and validate survival models for severe HF patients during this period. METHODS: We extracted severe HF patients from the MIMIC-IV database (as training and internal validation cohorts) as well as from the MIMIC-III database and local hospital (as external validation cohorts). Three algorithms, including Cox proportional hazards model (CoxPH), random survival forest (RSF), and deep learning survival prediction (DeepSurv), were applied to incorporate the parameters (partial hospitalization information and exposure durations of drugs) for constructing survival prediction models. The model performance was assessed mainly using area under the receiver operator characteristic curve (AUC), brier score (BS), and decision curve analysis (DCA). The model interpretability was determined by the permutation importance and Shapley additive explanations values. RESULTS: A total of 11,590 patients were included in this study. Among the 3 models, the CoxPH model ultimately included 10 variables, while RSF and DeepSurv models incorporated 24 variables, respectively. All of the 3 models achieved respectable performance metrics while the DeepSurv model exhibited the highest AUC values and relatively lower BS among these models. The DCA also verified that the DeepSurv model had the best clinical practicality. CONCLUSIONS: The survival prediction tools established in this study can be applied to severe HF patients during vulnerable period by mainly inputting drug treatment duration, thus contributing to optimal clinical decisions prospectively.

Asunto(s)

RESUMEN

The study aimed to assess the impact of changes in blood pressure on cardiovascular events in the Chinese population. It enrolled 33 179 Chinese participants aged ≥35 years (57.1% women) without CVD at baseline. BP status was defined according to the 2017 ACC/AHA hypertension guidelines. The type of BP change was defined as change in BP status from baseline to the end of follow-up, included stable BP <130/80, <130/80 to ≥130/80, ≥130/80 to <130/80 mm Hg, persistent BP ≥130/80 mm Hg. The hazard ratio (HR) of incident CVD by change in BP category was estimated using Cox proportional hazards and Fine-Gray models. During median follow-up of 3.17 years, 2023 CVD events occurred. Compared with BP <120/80, 120-129/<80 mm Hg at baseline (HR = 1.66, 95% CI: 1.09-2.53), 130-139/80-89 mm Hg (HR = 1.35, 95% CI: 0.94-1.95), and ≥140/90 mm Hg (HR = 2.46, 95% CI: 1.78-3.40) were risk factors for CVD. Compared with the group with stable BP <130/80 mm Hg, the risk of CVD was 1.88 (95% CI: 1.40-2.53) in the group with persistent BP ≥130/80 mm Hg and 1.40 (95% CI: 1.01-1.94) in the group of BP decreased to <130/80 mm Hg. These results showed that BP 120-129/<80, 130-139/80-89, and ≥140/90 mm Hg were associated with a high risk of CVD. Over time, persistent BP ≥130/80 mm Hg increased the risk of CVD, but a return to <130/80 mm Hg from hypertension decreased the risk of CVD.

RESUMEN

INTRODUCTION: The national burden of gastric cancer (GC) is high in Georgia, which is determined by its high mortality and low survival. The study aimed to estimate the effect of age at diagnosis on the prognosis of GC patients diagnosed between 2015 and 2020 in Georgia. MATERIALS AND METHODS: We obtained data for the study from the national population-based cancer registry. All patients 15 years of age or older, diagnosed during 2015-2020 with invasive GC (site codes C16.0 to C16.9, International Classification of Diseases for Oncology), were eligible for inclusion in the analysis. We produced survival curves using the Kaplan-Meier method, and the log-rank test was used to compare survival between groups. Hazard ratios (HR) were estimated using univariate Cox proportional models and multivariate Cox proportional hazard models. The endpoint of the study was overall survival (OS). The level of statistical significance of the study findings was estimated using p-values and 95% confidence intervals (CI). A p-value<0.05 was considered statistically significant.  Results: A total of 1,828 gastric cancer cases were included in the statistical analysis. The average age of patients was 65 years. The bivariate Cox's regression analysis demonstrated that the risk of gastric cancer mortality increased gradually with the age of cancer patients. The HR and 95% CI were as follows: 1.5 (1.1-1.8) and 2.1 (1.5-2.5) in the 46-65 years and >65 years groups, respectively, with the <46 years group as a reference. Moreover, multivariable Cox's regression analysis proved that age is an independent risk factor for GC mortality (HR = 1.4; 95% CI = 1.2-1.8; p<.001).  Conclusion: We found that age at diagnosis was a significant predictor of the worse survival of GC patients diagnosed between 2015 and 2020 in Georgia.

RESUMEN

BACKGROUND: Temporal lobe (TL) white matter (WM) injuries are often seen early after radiotherapy (RT) in nasopharyngeal carcinoma patients (NPCs), which fail to fully recover in later stages, exhibiting a "non-complete recovery pattern". Herein, we explored the correlation between non-complete recovery WM injuries and TL necrosis (TLN), identifying dosimetric predictors for TLN-related high-risk WM injuries. METHODS: We longitudinally examined 161 NPCs and 19 healthy controls employing multi-shell diffusion MRI. Automated fiber-tract quantification quantified diffusion metrics within TL WM tract segments. ANOVA identified non-complete recovery WM tract segments one-year post-RT. Cox regression models discerned TLN risk factors utilizing non-complete recovery diffusion metrics. Normal tissue complication probability (NTCP) models and dose-response analysis further scrutinized RT-related toxicity to high-risk WM tract segments. RESULTS: Seven TL WM tract segments exhibited a "non-complete recovery pattern". Cox regression analysis identified mean diffusivity of the left uncinate fasciculus segment 1, neurite density index (NDI) of the left cingulum hippocampus segment 1, and NDI of the right inferior longitudinal fasciculus segment 1 as TLN risk predictors (hazard ratios [HRs] with confidence interval [CIs]: 1.45 [1.17-1.81], 1.07 [1.00-1.15], and 1.15 [1.03-1.30], respectively; all P-values < 0.05). In NTCP models, D10cc.L, D20cc.L and D10cc.R demonstrated superior performance, with TD50 of 37.22 Gy, 24.96 Gy and 37.28 Gy, respectively. CONCLUSIONS: Our findings underscore the significance of the "non-complete recovery pattern" in TL WM tract segment injuries during TLN development. Understanding TLN-related high-risk WM tract segments and their tolerance doses could facilitate early intervention in TLN and improve RT protocols.

Asunto(s)

RESUMEN

BACKGROUND: There hasn't been research done on the connection between serum anion gap (AG) levels and long-, medium-, and short-term all-cause mortality in congestive heart failure (CHF) patients. This study aims to investigate the association between serum anion gap levels and all-cause mortality in CHF patients after adjusting for other covariates. METHODS: For each patient, we gather demographic information, comorbidities, laboratory results, vital signs, and scoring data using the ICU (Intensive Care Unit) Admission Scoring System from the MIMIC-III database. The connection between baseline AG and long-, medium-, and short-term all-cause mortality in critically ill congestive heart failure patients was investigated using Kaplan-Meier survival curves, subgroup analysis, restricted cubic spline, and Cox proportional risk analysis. RESULTS: 4840 patients with congestive heart failure in total were included in this study. With a mean age of 72.5 years, these patients had a gender split of 2567 males and 2273 females. After adjusting for other covariates, a multiple regression analysis revealed that, in critically ill patients with congestive heart failure, all-cause mortality increased significantly with rising AG levels. In the fully adjusted model, we discovered that AG levels were strongly correlated with 4-year, 365-day, 90-day, and 30-day all-cause mortality in congestive heart failure patients with HRs (95% CI) of 1.06 (1.04, 1.08); 1.08 (1.05, 1.10); and 1.08 (1.05, 1.11) (p-value < 0.05). Our subgroup analysis's findings demonstrated a high level of consistency and reliability. K-M survival curves demonstrate that high serum AG levels are associated with a lower survival probability. CONCLUSION: Our research showed the association between CHF patients' all-cause mortality and anion gap levels was non-linear. Elevated anion gap levels are associated with an increased risk of long-, medium-, and short-term all-cause death in patients with congestive heart failure. Continuous monitoring of changes in AG levels may have a clinical predictive role.

Asunto(s)

RESUMEN

OBJECTIVE: Survival analysis is widely utilized in healthcare to predict the timing of disease onset. Traditional methods of survival analysis are usually based on Cox Proportional Hazards model and assume proportional risk for all subjects. However, this assumption is rarely true for most diseases, as the underlying factors have complex, non-linear, and time-varying relationships. This concern is especially relevant for pregnancy, where the risk for pregnancy-related complications, such as preeclampsia, varies across gestation. Recently, deep learning survival models have shown promise in addressing the limitations of classical models, as the novel models allow for non-proportional risk handling, capturing nonlinear relationships, and navigating complex temporal dynamics. METHODS: We present a methodology to model the temporal risk of preeclampsia during pregnancy and investigate the associated clinical risk factors. We utilized a retrospective dataset including 66,425 pregnant individuals who delivered in two tertiary care centers from 2015 to 2023. We modeled the preeclampsia risk by modifying DeepHit, a deep survival model, which leverages neural network architecture to capture time-varying relationships between covariates in pregnancy. We applied time series k-means clustering to DeepHit's normalized output and investigated interpretability using Shapley values. RESULTS: We demonstrate that DeepHit can effectively handle high-dimensional data and evolving risk hazards over time with performance similar to the Cox Proportional Hazards model, achieving an area under the curve (AUC) of 0.78 for both models. The deep survival model outperformed traditional methodology by identifying time-varied risk trajectories for preeclampsia, providing insights for early and individualized intervention. K-means clustering resulted in patients delineating into low-risk, early-onset, and late-onset preeclampsia groups-notably, each of those has distinct risk factors. CONCLUSION: This work demonstrates a novel application of deep survival analysis in time-varying prediction of preeclampsia risk. Our results highlight the advantage of deep survival models compared to Cox Proportional Hazards models in providing personalized risk trajectory and demonstrating the potential of deep survival models to generate interpretable and meaningful clinical applications in medicine.

Asunto(s)

RESUMEN

Epidemiologic studies frequently use risk ratios to quantify associations between exposures and binary outcomes. When the data are physically stored at multiple data partners, it can be challenging to perform individual-level analysis if data cannot be pooled centrally due to privacy constraints. Existing methods either require multiple file transfers between each data partner and an analysis center (e.g., distributed regression) or only provide approximate estimation of the risk ratio (e.g., meta-analysis). Here we develop a practical method that requires a single transfer of eight summary-level quantities from each data partner. Our approach leverages an existing risk-set method and software originally developed for Cox regression. Sharing only summary-level information, the proposed method provides risk ratio estimates and confidence intervals identical to those that would be provided - if individual-level data were pooled - by the modified Poisson regression. We justify the method theoretically, confirm its performance using simulated data, and implement it in a distributed analysis of COVID-19 data from the U.S. Food and Drug Administration's Sentinel System.

RESUMEN

PURPOSE: Several predictors have been identified for mental sickness absence, but those for recurrences are not well-understood. This study assesses recurrence rates for long-term mental sickness absence (LTMSA) within subgroups of common mental disorders (CMDs) and identifies predictors of recurrent LTMSA. METHODS: This historical prospective cohort study used routinely collected data from 16,310 employees obtained from a nationally operating Dutch occupational health service (ArboNed). Total follow-up duration was 23,334 person-years. Overall recurrence rates were assessed using Kaplan-Meier estimators. Recurrence rates within subgroups of CMDs were calculated using person-years. Univariable and multivariable Cox proportional hazards models were used to identify predictors. RESULTS: 15.6% of employees experienced a recurrent LTMSA episode within three years after fully returning to work after a previous LTMSA episode. Highest recurrence rates for LTMSA were observed after a previous LTMSA episode due to mood or anxiety disorders. Mood or anxiety disorders and shorter previous episode duration were predictors of recurrent LTMSA. No associations were found for age, gender, company size, full-time equivalent and job tenure. CONCLUSION: Employees should be monitored adequately after they fully returned to work after LTMSA. It is recommended to monitor high-risk employees (i.e. employees with mood or anxiety disorders and short LTMSA episode) more intensively, also beyond full return to work. Moreover, diagnosis of anxiety and depressive symptoms should be given a higher priority in occupational healthcare.

RESUMEN

BACKGROUND: Robust and practical prognosis prediction models for hepatocellular carcinoma (HCC) patients play crucial roles in personalized precision medicine. MATERIAL AND METHODS: We recruited two independent HCC cohorts (discovery cohort and validation cohort), totally consisting of 222 HCC patients undergone surgical resection. We quantified the expressions of immune-related proteins (CD8, CD68, CD163, PD-1 and PD-L1) in paired HCC tissues and non-tumor liver tissues from these HCC patients using immunohistochemistry (mIHC) assays. We constructed the HCC prognosis prediction model using five different machine learning methods based on the patients in the discovery cohort, such as Cox proportional hazards (CoxPH). RESULTS: We identified 19 features that were associated with overall survival of HCC patients in the discovery cohort (p < 0.1), such as immune-related features CD68+ and CD8+ cell infiltration. We constructed five HCC prognosis prediction models using five different machine learning methods. Among the five different machine learning models, the CoxPH model achieved the best performance (area under the curve [AUC], 0.839; C-index, 0.779). According to the risk score from CoxPH model, we divided HCC patients into high-risk group/low-risk group. In both discovery cohort and validation cohort, the patients in low-risk group showed longer overall survival compared with those in high-risk group (p = 1.8 × 10-7 and 3.4 × 10-5, respectively). Moreover, our novel scoring system efficiently predicted the 6, 12, and 18 months survival rate of HCC patients with AUC >0.75 in both discovery cohort and validation cohort. In addition, we found that the scoring system could also distinguish the patients with high/low risks of relapse in both discovery cohort and validation cohort (p = 0.00015 and 0.00012). CONCLUSION: The novel CoxPH-based risk scoring model on clinical, laboratory-testing and immune-related features showed high prediction efficiencies for overall survival and recurrence of HCCs undergone surgical resection. Our results may be helpful to optimize clinical follow-up or therapeutic interventions.

Asunto(s)