Subdistribution-Based Imputation for Deep Survival Analysis with Competing Events

2021 ◽  
Author(s):  
Shekoufeh Gorgi Zadeh ◽  
Charlotte Behning ◽  
Matthias Schmid
Keyword(s):  
Survival Analysis ◽  
Survival Data ◽  
Cox Regression ◽  
Classical Model ◽  
Event Analysis ◽  
Survival Times ◽  
Subdistribution Hazard ◽  
Time To Event Data ◽  
Competing Events ◽  
Real World Datasets

Abstract With the popularity of deep neural networks (DNNs) in recent years, many researchers have proposed DNNs for the analysis of survival data (time-to-event data). These networks learn the distribution of survival times directly from the predictor variables without making strong assumptions on the underlying stochastic process. In survival analysis, it is common to observe several types of events, also called competing events. The occurrences of these competing events are usually not independent of one another and have to be incorporated in the modeling process in addition to censoring. In classical survival analysis, a popular method to incorporate competing events is the subdistribution hazard model, which is usually fitted using weighted Cox regression. In the DNN framework, only few architectures have been proposed to model the distribution of time to a specific event in a competing events situation. These architectures are characterized by a separate subnetwork/pathway per event, leading to large networks with huge amounts of parameters that may become difficult to train. In this work, we propose a novel imputation strategy for data preprocessing that incorporates the subdistribution weights derived from the classical model. With this, it is no longer necessary to add multiple subnetworks to the DNN to handle competing events. Our experiments on synthetic and real-world datasets show that DNNs with multiple subnetworks per event can simply be replaced by a DNN designed for a single-event analysis without loss in accuracy.

scholarly journals The what, when and how of orthopaedic registers: an introduction into register-based research

2019 ◽  
Vol 4 (6) ◽  
pp. 337-343 ◽  
Author(s):  
Claus Varnum ◽  
Alma Bečić Pedersen ◽  
Per Hviid Gundtoft ◽  
Søren Overgaard
Keyword(s):  
Survival Analysis ◽  
Selection Bias ◽  
Survival Data ◽  
Cox Regression ◽  
Research Question ◽  
Cox Regression Analysis ◽  
Information Bias ◽  
Kaplan Meier ◽  
Routinely Collected Health Data ◽  
Competing Events

Establishment of orthopaedic registers started in 1975 and many registers have been initiated since. The main purpose of registers is to collect information on patients, implants and procedures in order to monitor and improve the outcome of the specific procedure. Data validity reflects the quality of the registered data and consists of four major aspects: coverage of the register, registration completeness of procedures/patients, registration completeness of variables included in the register and accuracy of registered variables. Survival analysis is often used in register studies to estimate the incidence of an outcome. The most commonly used survival analysis is the Kaplan–Meier survival curves, which present the proportion of patients who have not experienced the defined event (e.g. death or revision of a prosthesis) in relation to the time. Depending on the research question, competing events can be taken into account by using the cumulative incidence function. Cox regression analysis is used to compare survival data for different groups taking differences between groups into account. When interpreting the results from observational register-based studies a number of factors including selection bias, information bias, chance and confounding have to be taken into account. In observational register-based studies selection bias is related to, for example, absence of complete follow-up of the patients, whereas information bias is related to, for example, misclassification of exposure (e.g. risk factor of interest) or/and outcome. The REporting of studies Conducted using Observational Routinely-collected Data guidelines should be used for studies based on routinely-collected health data including orthopaedic registers. Linkage between orthopaedic registers, other clinical quality databases and administrative health registers may be of value when performing orthopaedic register-based research. Cite this article: EFORT Open Rev 2019;4 DOI: 10.1302/2058-5241.4.180097

Factors Influencing the Lifetime of Telecommunication and Information Technology Standards

2010 ◽  
pp. 242-259
Author(s):  
Knut Blind
Keyword(s):  
Survival Analysis ◽  
New Technologies ◽  
Cox Regression ◽  
Medical Science ◽  
International Standards ◽  
Control Mechanisms ◽  
Survival Times ◽  
Scientific Publications ◽  
The One ◽  
Standard Documents

Besides these very conceptual or theoretical approaches to deal with standards dynamics, several case study analyses exist, which focus on the standard maintenance and succession (Egyedi, Loeffen 2002) in order to answer the question how to deal with heritage relations between standards and on standard integrity (Egyedi, Hudson 2005) and in order to discuss control mechanisms that safeguard the integrity of (de facto) standards. This paper adds an additional methodological dimension to the analysis of the dynamics of standards by a strong focus on the life times of standards. The contribution of this paper to the emerging research on the dynamics of standards is twofold. First, the descriptive presentation of life times of standards focusing both on average publication years and survival times reflects on the one hand the historical development of ICT over time and on the other hand its dynamics in the various subfields. So far other indicators like scientific publications or patent applications are used to describe the development especially of new technologies, e. g. biotechnology or nanotechnology. The analysis of publications of standard documents extends the former exercises by a new more market and diffusion related dimension. Second, the characteristics of standard documents are used to explain their life times. Here we borrow for the first time general approaches from bibliometrics and patent analysis in order to explain life times of standards as indicator for their value by documents’ characteristics. The remainder of the paper is structured as follows. First, we analyse the average lifetimes of standards in a quantitative manner, taking into account differences between countries. Since the simple approach of calculating the average lifetimes of historical standards does not allow us to include standards which are still alive, we have to apply a more sophisticated methodology, the so-called survival analysis, which was initially mainly applied in medical science. The application of this statistical approach produces average lifetimes of standards, taking into account the expected lifetime of standards which are still valid. This approach is crucial, especially for the analysis of ICT standards, because the number of valid standards relative to historical standards is rather high. Due to the very high relevance of international standards in the ICT sector and the high quality of this subsample, we concentrate the survival analysis espeically on the international standards including the standards released by the European standardisation bodies. The results of this analysis provide us with new insights about the expected lifetimes of standards differentiated by technology in the ICT area. The final step of our analysis tries to answer the question which causal factors influence the lifetimes of standards in the ICT sector. We present first insights by applying the so-called Cox regression, which allows us to identify whether some selected characteristics of a standard, like cross references or references to international standards, have a significant impact on its actual or expected lifetime. The approach to assess the importance of a technical document by analysing its references to other documents or being referenced in other documents has a long tradition in evaluating the value of patents by counting and analysing their citations. The paper concludes with a brief summary of the main results, but also with some general recommendations regarding standardisation processes and the maintenance of standards derived from the new insights.

scholarly journals Gastric cancer in East Azerbaijan, Iran: Five-year survival analysis of population-based cancer registry results

2020 ◽  
Vol 7 (11) ◽  
pp. 4114-4121
Author(s):  
Pooneh Jabbaripour ◽  
Mohammad Hossein Somi ◽  
Hossein Mashhadi Abdolahi ◽  
Roya Dolatkhah
Keyword(s):  
Gastric Cancer ◽  
Survival Analysis ◽  
Cancer Registry ◽  
Survival Data ◽  
Cox Regression ◽  
Survival Function ◽  
Population Based ◽  
Rank Test ◽  
Cancer Data ◽  
East Azerbaijan

Introduction: Gastric cancer is the most common cancer with significant increasing trends during the last decade in Iran. The aim of this study was to evaluate the epidemiologic profile of gastric cancer along with gastric cancer-specific survival analysis. Methods: This was an analytical cross-sectional study in which all gastric cancer data were analyzed using the database of the East Azerbaijan Population-Based Cancer Registry (EA-PBCR). The incidents of definitive gastric cancer diagnosis were between the period of March 20th, 2015 to March 19th, 2017 ( = 3 Iranian solar years). The survival analysis was performed using the Kaplan-Meier method and life tables for 1- to 5-year survival data. The Log-rank test and Cox regression were computed to test the equality of survival function and mortality hazard. Results: Overall, 2,631 newly diagnosed gastric cancer cases were registered for 3 years. Gastric cancer was 2.35 times more common in men than women. The most common age group was the 7th decade- with 531 (31.2%) gastric cancer cases. Most of the gastric cancer cases were non-cardia (n = 2,244, 85.29%) cancer, and the proportion of non-cardia to cardia gastric cancer was 5.8:1. Overall survival was 60.1%, and 1- to 5-year survival proportions were 91.61%, 64.21%, 58.53%, 30.14% and 24.77%, respectively. Cardia cancers had a worse survival rate than non-cardia cancers, and the hazard of mortality was 1.33 times higher in cardia than non-cardia cancers (hazard ratio or HR = 1.33; 95% CI: 1.05 - 1.68; P = 0.017). Conclusion: Non-cardia gastric cancer is still the most dominant subsite in East Azerbaijan, Iran. There was a higher 1- to 5- year survival proportion in East Azerbaijan, with lower overall mortality rates, compared to other regions of Iran.

scholarly journals Hospital outcomes of community-acquired COVID-19 versus influenza: Insights from the Swiss hospital-based surveillance of influenza and COVID-19

2022 ◽  
Vol 27 (1) ◽  
Author(s):  
Georg Marcus Fröhlich ◽  
Marlieke E. A. De Kraker ◽  
Mohamed Abbas ◽  
Olivia Keiser ◽  
Amaury Thiabaud ◽  
...  
Keyword(s):  
Influenza A ◽  
Seasonal Influenza ◽  
Cox Regression ◽  
Hospital Death ◽  
Rt Pcr ◽  
Hospital Outcomes ◽  
Subdistribution Hazard ◽  
Inverse Probability ◽  
Icu Admission ◽  
Competing Events

Background Since the onset of the COVID-19 pandemic, the disease has frequently been compared with seasonal influenza, but this comparison is based on little empirical data. Aim This study compares in-hospital outcomes for patients with community-acquired COVID-19 and patients with community-acquired influenza in Switzerland. Methods This retrospective multi-centre cohort study includes patients > 18 years admitted for COVID-19 or influenza A/B infection determined by RT-PCR. Primary and secondary outcomes were in-hospital mortality and intensive care unit (ICU) admission for patients with COVID-19 or influenza. We used Cox regression (cause-specific and Fine-Gray subdistribution hazard models) to account for time-dependency and competing events with inverse probability weighting to adjust for confounders. Results In 2020, 2,843 patients with COVID-19 from 14 centres were included. Between 2018 and 2020, 1,381 patients with influenza from seven centres were included; 1,722 (61%) of the patients with COVID-19 and 666 (48%) of the patients with influenza were male (p < 0.001). The patients with COVID-19 were younger (median 67 years; interquartile range (IQR): 54–78) than the patients with influenza (median 74 years; IQR: 61–84) (p < 0.001). A larger percentage of patients with COVID-19 (12.8%) than patients with influenza (4.4%) died in hospital (p < 0.001). The final adjusted subdistribution hazard ratio for mortality was 3.01 (95% CI: 2.22–4.09; p < 0.001) for COVID-19 compared with influenza and 2.44 (95% CI: 2.00–3.00, p < 0.001) for ICU admission. Conclusion Community-acquired COVID-19 was associated with worse outcomes compared with community-acquired influenza, as the hazards of ICU admission and in-hospital death were about two-fold to three-fold higher.

Survival Analysis and Data Mining

2011 ◽  
pp. 1077-1082
Author(s):  
Qiyang Chen ◽  
Alan Oppenheim ◽  
Dajin Wang
Keyword(s):  
Data Mining ◽  
Survival Analysis ◽  
Failure Analysis ◽  
Survival Data ◽  
Event History Analysis ◽  
Real Life ◽  
Event Analysis ◽  
History Analysis ◽  
Lifetime Data Analysis ◽  
The Times

Survival analysis (SA) consists of a variety of methods for analyzing the timing of events and/or the times of transition among several states or conditions. The event of interest can happen at most only once to any individual or subject. Alternate terms to identify this process include Failure Analysis (FA), Reliability Analysis (RA), Lifetime Data Analysis (LDA), Time to Event Analysis (TEA), Event History Analysis (EHA), and Time Failure Analysis (TFA), depending on the type of application for which the method is used (Elashoff, 1997). Survival Data Mining (SDM) is a new term that was coined recently (SAS, 2004). There are many models and variations of SA. This article discusses some of the more common methods of SA with real-life applications. The calculations for the various models of SA are very complex. Currently, multiple software packages are available to assist in performing the necessary analyses much more quickly.

scholarly journals The Pseudo-Observation Analysis of Time-To-Event Data. Example from the Danish Diet, Cancer and Health Cohort Illustrating Assumptions, Model Validation and Interpretation of Results

2018 ◽  
Vol 7 (1) ◽  
Author(s):  
Lotte Maxild Mortensen ◽  
Camilla Plambeck Hansen ◽  
Kim Overvad ◽  
Søren Lundbye-Christensen ◽  
Erik T. Parner
Keyword(s):  
Cox Regression ◽  
Point Of View ◽  
Right Censoring ◽  
Event Analysis ◽  
Event Data ◽  
Time To Event ◽  
Time To Event Data ◽  
Model Control ◽  
Additive Risk ◽  
Observation Method

Abstract Regression analyses for time-to-event data are commonly performed by Cox regression. Recently, an alternative method, the pseudo-observation method, has been introduced. This method offers new possibilities of analyzing data exploring cumulative risks on both a multiplicative and an additive risk scale, in contrast to the multiplicative Cox regression model for hazard rates. Hence, the pseudo-observation method enables assessment of interaction on an additive scale. However, the pseudo-observation method implies more strict model assumptions regarding entry and censoring but avoids the assumption of proportional hazards (except from combined analyses of several time intervals where assumptions of constant hazard ratios, risk differences and relative risks may be imposed). Only few descriptions of the use of the method are accessible for epidemiologists. In this paper, we present the pseudo-observation method from a user-oriented point of view aiming at facilitating the use of this relatively new analytical tool. Using data from the Diet, Cancer and Health Cohort we give a detailed example of the application of the pseudo-observation method on time-to-event data with delayed entry and right censoring. We discuss model control and suggest analytic strategies when assumptions are not met. The introductory model control in the data example showed that data did not fulfill the assumptions of the pseudo-observation method. This was caused by selection of healthier participants at older baseline ages and a change in the distribution of study participants according to outcome risk during the inclusion period. Both selection effects need to be addressed in any time-to-event analysis and we show how these effects are accounted for in the pseudo-observation analysis. The pseudo-observation method provides us with a statistical tool which makes it possible to analyse cohort data on both multiplicative and additive risk scales including assessment of biological interaction on the risk difference scale. Thus, it might be a relevant choice of method – especially if the focus is to investigate interaction from a public health point of view.

scholarly journals COVID-19: more than “a little flu”? Insights from the Swiss hospital-based surveillance of Influenza and COVID-19

2020 ◽  
Author(s):  
Georg Marcus Fröhlich ◽  
Marlieke E. A. De Kraker ◽  
Mohammed Abbas ◽  
Olivia Keiser ◽  
Amaury Thiabaud ◽  
...  
Keyword(s):  
Influenza A ◽  
Seasonal Influenza ◽  
Cox Regression ◽  
Hospital Death ◽  
Time Dependency ◽  
Prospective Surveillance ◽  
Subdistribution Hazard ◽  
Inverse Probability ◽  
Icu Admission ◽  
Competing Events

AbstractBackgroundCoronavirus disease 19 (COVID-19) has frequently been colloquially compared to the seasonal influenza, but comparisons based on empirical data are scarce.AimsTo compare in-hospital outcomes for patients admitted with community-acquired COVID-19 to patients with community-acquired influenza in Switzerland.MethodsPatients >18 years, who were admitted with PCR proven COVID-19 or influenza A/B infection to 14 participating Swiss hospitals were included in a prospective surveillance. Primary and secondary outcomes were the in-hospital mortality and intensive care unit (ICU) admission between influenza and COVID-19 patients. We used Cox regression (cause-specific models, and Fine & Gray subdistribution) to account for time-dependency and competing events with inverse probability weighting to account for confounders.ResultsIn 2020, 2843 patients with COVID-19 were included from 14 centers and in years 2018 to 2020, 1361 patients with influenza were recruited in 7 centers. Patients with COVID-19 were predominantly male (n=1722, 61% vs. 666 influenza patients, 48%, p<0.001) and were younger than influenza patients (median 67 years IQR 54-78 vs. median 74 years IQR 61-84, p<0.001). 363 patients (12.8%) died in-hospital with COVID-19 versus 61 (4.4%) patients with influenza (p<0.001). The final, adjusted subdistribution Hazard Ratio for mortality was 3.01 (95% CI 2.22-4.09, p<0.001) for COVID-19 compared to influenza, and 2.44 (95% CI, 2.00-3.00, p<0.001) for ICU admission.ConclusionEven in a national healthcare system with sufficient human and financial resources, community-acquired COVID-19 was associated with worse outcomes compared to community-acquired influenza, as the hazards of in-hospital death and ICU admission were ∼3-fold higher.

scholarly journals Testing the proportional hazards assumption in cox regression and dealing with possible non-proportionality in total joint arthroplasty research: methodological perspectives and review

2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Ilari Kuitunen ◽  
Ville T. Ponkilainen ◽  
Mikko M. Uimonen ◽  
Antti Eskelinen ◽  
Aleksi Reito
Keyword(s):  
Survival Analysis ◽  
Regression Model ◽  
Total Joint Arthroplasty ◽  
Full Text ◽  
Proportional Hazards ◽  
Cox Regression ◽  
Joint Arthroplasty ◽  
Central Research ◽  
Cox Regression Model ◽  
Pubmed Database

Abstract Background Survival analysis and effect of covariates on survival time is a central research interest. Cox proportional hazards regression remains as a gold standard in the survival analysis. The Cox model relies on the assumption of proportional hazards (PH) across different covariates. PH assumptions should be assessed and handled if violated. Our aim was to investigate the reporting of the Cox regression model details and testing of the PH assumption in survival analysis in total joint arthroplasty (TJA) studies. Methods We conducted a review in the PubMed database on 28th August 2019. A total of 1154 studies were identified. The abstracts of these studies were screened for words “cox and “hazard*” and if either was found the abstract was read. The abstract had to fulfill the following criteria to be included in the full-text phase: topic was knee or hip TJA surgery; survival analysis was used, and hazard ratio reported. If all the presented criteria were met, the full-text version of the article was then read. The full-text was included if Cox method was used to analyze TJA survival. After accessing the full-texts 318 articles were included in final analysis. Results The PH assumption was mentioned in 114 of the included studies (36%). KM analysis was used in 281 (88%) studies and the KM curves were presented graphically in 243 of these (87%). In 110 (45%) studies, the KM survival curves crossed in at least one of the presented figures. The most common way to test the PH assumption was to inspect the log-minus-log plots (n = 59). The time-axis division method was the most used corrected model (n = 30) in cox analysis. Of the 318 included studies only 63 (20%) met the following criteria: PH assumption mentioned, PH assumption tested, testing method of the PH assumption named, the result of the testing mentioned, and the Cox regression model corrected, if required. Conclusions Reporting and testing of the PH assumption and dealing with non-proportionality in hip and knee TJA studies was limited. More awareness and education regarding the assumptions behind the used statistical models among researchers, reviewers and editors are needed to improve the quality of TJA research. This could be achieved by better collaboration with methodologists and statisticians and introducing more specific reporting guidelines for TJA studies. Neglecting obvious non-proportionality undermines the overall research efforts since causes of non-proportionality, such as possible underlying pathomechanisms, are not considered and discussed.

scholarly journals IMPDH2 and HPRT expression and a prognostic significance in preoperative and postoperative patients with osteosarcoma

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Parunya Chaiyawat ◽  
Areerak Phanphaisarn ◽  
Nutnicha Sirikaew ◽  
Jeerawan Klangjorhor ◽  
Viraporn Thepbundit ◽  
...  
Keyword(s):  
Overall Survival ◽  
Cox Regression ◽  
Expression Patterns ◽  
Prognostic Significance ◽  
Drug Repurposing ◽  
Survival Times ◽  
Metabolizing Enzymes ◽  
Hypoxanthine Guanine Phosphoribosyltransferase ◽  
High Grade Osteosarcoma ◽  
Enneking Stage

AbstractOsteosarcoma is one of the most aggressive bone tumors in children and adolescents. Development of effective therapeutic options is still lacking due to the complexity of the genomic background. In previous work, we applied a proteomics-guided drug repurposing to explore potential treatments for osteosarcoma. Our follow-up study revealed an FDA-approved immunosuppressant drug, mycophenolate mofetil (MMF) targeting inosine-5′-phosphate dehydrogenase (IMPDH) enzymes, has an anti-tumor effect that appeared promising for further investigation and clinical trials. Profiling of IMPDH2 and hypoxanthine–guanine phosphoribosyltransferase (HPRT), key purine-metabolizing enzymes, could deepen understanding of the importance of purine metabolism in osteosarcoma and provide evidence for expanded use of MMF in the clinic. In the present study, we investigated levels of IMPDH2, and HPRT in biopsy of 127 cases and post-chemotherapy tissues in 20 cases of high-grade osteosarcoma patients using immunohistochemical (IHC) analysis. Cox regression analyses were performed to determine prognostic significance of all enzymes. The results indicated that low levels of HPRT were significantly associated with a high Enneking stage (P = 0.023) and metastatic status (P = 0.024). Univariate and multivariate analyses revealed that patients with low HPRT expression have shorter overall survival times [HR 1.70 (1.01–2.84), P = 0.044]. Furthermore, high IMPDH2/HPRT ratios were similarly associated with shorter overall survival times [HR 1.67 (1.02–2.72), P = 0.039]. Levels of the enzymes were also examined in post-chemotherapy tissues. The results showed that high IMPDH2 expression was associated with shorter metastasis-free survival [HR 7.42 (1.22–45.06), P = 0.030]. These results suggest a prognostic value of expression patterns of purine-metabolizing enzymes for the pre- and post-chemotherapy period of osteosarcoma treatment.

Sign in / Sign up

Export Citation Format

Share Document