Oral anticoagulants (OACs) are indicated in the first-line treatment of venous thromboembolism (VTE), which comprises of deep vein thrombosis (DVT) and pulmonary embolism (PE). While contemporary guidelines recommend extended-duration anticoagulation after the first diagnosis of unprovoked VTE, the benefits and harms associated with this approach remain unclear across age groups, especially among older adults. Crucially, contemporary estimates of VTE recurrence have not incorporated all-cause mortality as a competing event, the risk of which increases with age. Therefore, we evaluated and synthesized existing literature on of the risk of all-cause mortality by age following completion of limited-duration anticoagulation for a first episode of unprovoked VTE. In addition, we determined the risk of VTE recurrence after completion of limited-duration OAC therapy by age, with death as a competing outcome using data from a prospective cohort study.

No description available.

Indiana University-Purdue University Indianapolis (IUPUI) / Background: High blood pressure (HBP) is a common risk factor for dementia in elder population. Anti-hypertensive medications have been reported to associate with lower incidence rate of dementia in elder African Americans. The Apolipoprotein E (ApoE) epsilon 4 allele has been shown to be associated with both increased dementia and hypertension risk. However, previous studies had not examined the association between anti-hypertensive medications by ApoE status accounting for the competing risk from death. Methods: This is a prospective observational cohort study in 1236 community-dwelling hypertensive African Americans aged 65 years and older without dementia at baseline, with follow-up cognitive assessment and clinical evaluation for dementia diagnosis. Dementia-free mortality was considered as the competing risk. Of these, 707 participants were genotyped for ApoE status. Anti-hypertensive medication use was obtained from prescription records in the electronic medical records of the Indiana Network for Patient Care (INPC). Cox proportional cause-specific hazard (CSH) regression models were applied to assess the association between anti-hypertensive medication use and CSHs for dementia and death in ApoE epsilon 4 carriers and non-carriers separately. Key results: In ApoE epsilon 4 carriers, participants using anti-hypertensive medications had lower CSH of dementia compared to those not on anti-hypertensive medications before adjusting for blood pressure (BP) (hazard ratio (HR), 0.365; 95% CI, 0.170 – 0.785; p = 0.0099). The HR was no longer significant once BP control was adjusted (HR, 0.784; 95% CI, 0.197 – 3.123; p = 0.7303). Anti-hypertensive medications were not associated with dementia rate in non-carriers. In ApoE epsilon 4 non-carriers, participants on anti-hypertensive treatment showed significantly lower CSH of death compared to those not on mediations adjusting for covariates and BP control (HR, 0.237; 95% CI, 0.149 – 0.375; p < 0.0001). There was no significant association between anti-hypertensive medication use and death in ApoE epsilon 4 carriers. Conclusions: Anti-hypertensive medication was associated with lower dementia rate in ApoE epsilon 4 carriers and that rate was primarily mediated through BP control. In non-carriers, anti-hypertensive medication was significantly associated with lower mortality rate and this association appears to be independent of BP control.

Indiana University-Purdue University Indianapolis (IUPUI) / Objectives: The aims of this study are to estimate the cumulative incidence of lower extremity amputation (LEA) revision and reamputation adjusting for a competing risk of death, estimate the one-year event-free mortality rates for patients with peripheral vascular disease undergoing LEA, and develop predictive models for LEA revision and reamputation adjusting for a competing risk of death. Methods: This was a retrospective review of the prospectively collected Vascular Quality Initiative (VQI) registry between 2013 and 2018. Adults undergoing unilateral LEA were included. Demographics, comorbidities, medications, smoking status, history of vascular procedures and revascularization attempts, and procedure urgency were considered. Models to predict LEA revision and reamputation were developed using multivariable regression on the interval-censored competing risks data using semiparametric regression on the cumulative incidence function. Results: The cumulative incidences of LEA revision and revision-free mortality within one year of index amputation are 14.9% and 15.5% respectively. Patient BMI, smoking status, aspirin use, history of revascularization, and level of planned LEA are significantly associated with the odds of LEA revision. Age, amputation urgency, dialysis, and level of planned LEA are associated with the one-year odds of revision-free mortality. A patient receiving an index above knee amputation (AKA) has 61% lower odds of LEA revision (p < 0.0001) but 51% higher odds of revision-free mortality following LEA (p < 0.0001). Previous revascularization procedures increase the odds of revision by 23% (p < 0.0001). The cumulative incidences of reamputation and one-year reamputation-free mortality following LEA are 11.5% and 16.9% respectively. Urgency of the procedure, history of revascularization procedures, and level of planned LEA are statistically associated with the odds of reamputation when adjusting for the competing risk of death. Patients receiving index AKA have 62% lower odds of reamputation (p < 0.0001) compared to BKA. Dialysis is the strongest predictor of one-year mortality (OR 2.576, p < 0.0001). Conclusions: Patients with appropriately managed PVD, which still progresses to amputation have higher odds of LEA revision and reamputation. Revision risk can be predicted and compared on the basis of patient factors and the planned index amputation.

Indiana University-Purdue University Indianapolis (IUPUI) / Context. Prediction models for dementia are crucial for informing clinical decision making in older adults. Previous models have used genotype and age to obtain risk scores to determine risk of Alzheimer’s Disease, one of the most common forms of dementia (Desikan et al., 2017). However, previous prediction models do not account for the fact that the time to dementia onset is unknown, lying between the last negative and the first positive dementia diagnosis time (interval censoring). Instead, these models use time to diagnosis, which is greater than or equal to the true dementia onset time. Furthermore, these models do not account for the competing risk of death which is quite frequent among elder adults. Objectives. To develop a prediction model for dementia that accounts for interval censoring and the competing risk of death. To compare the predictions from this model with the predictions from a naïve analysis that ignores interval censoring and the competing risk of death. Methods. We apply the semiparametric sieve maximum likelihood (SML) approach to simultaneously model the cumulative incidence function (CIF) of dementia and death while accounting for interval censoring (Bakoyannis, Yu, & Yiannoutsos, 2017). The SML is implemented using the R package intccr. The CIF curves of dementia are compared for the SML and the naïve approach using a dataset from the Indianapolis Ibadan Dementia Project. Results. The CIF from the SML and the naïve approach illustrated that for healthier individuals at baseline, the naïve approach underestimated the incidence of dementia compared to the SML, as a result of interval censoring. Individuals with a poorer health condition at baseline have a CIF that appears to be overestimated in the naïve approach. This is due to older individuals with poor health conditions having an elevated risk of death. Conclusions. The SML method that accounts for the competing risk of death along with interval censoring should be used for fitting prediction/prognostic models of dementia to inform clinical decision making in older adults. Without controlling for the competing risk of death and interval censoring, the current models can provide invalid predictions of the CIF of dementia.

Master of Science / Department of Clinical Sciences / Bob L. Larson / The 2007 NAHMS (National Animal Health Monitoring System) survey indicated that early lactation health issues are major factors influencing reproduction and culling on U.S. dairy herds. The objective of this study was to evaluate fresh-cow health during the first 30 days in milk, and its association with days to pregnancy in the concurrent lactation. Data were collected on cattle that calved over a two month period (July and August 2009) on a dairy farm located in the Upper Midwest region of the U.S. Health and production data were collected daily for each cow from the beginning of lactation until the majority of the study population was confirmed pregnant. Both a competing risk analysis and a semi-parametric Cox regression model were used to test the association between specific health-related events and days to pregnancy and the outcomes of the two models were compared. These analyses showed metritis and dystocia in the first 30 days of lactation were associated with greater days to pregnancy. The only difference noted between parities was that lactation-five and greater cows were significantly associated with greater days to pregnancy. The two analyses showed conflicting significance of association between retained placenta, ketosis, twinning, lameness, and other non-specific illnesses with days to pregnancy. This study found that a competing risk analysis and a semi-parametric regression model were appropriate methods to analyze time sensitive data such as reproductive efficiency. This study supports the evidence that parity, metritis, retained placenta, ketosis, dystocia, twinning, lameness, and other non-specific illnesses can have an impact on reproductive efficiency.

Background: Dilated Cardiomyopathy (DCM) is the most common functional type of cardiomyopathy in children with significant morbidity and the leading indication for cardiac transplant over 5 years of age. Identification of baseline risk factors for failing medical management by etiologic grouping remain to be elucidated in a large populationbased study. The competing risk for heart death between all-cause mortality and heart transplantation is often overestimated in the literature and may obscure additional novel risk factors associated with poor clinical outcomes. Methods: The National Heart Lung and Blood Institute Pediatric Cardiomyopathy Registry collected longitudinal data from 1731 children with DCM in North America from 1990 to 2007. Composite endpoint (CEP) was the earlier occurrence of death or heart transplant. Univariate and multivariate predictors were identified from demographic and echocardiographic data (expressed as z-scores) collected within 30 days of diagnosis. A competing risk analysis was performed calculating cumulative incidence and identifying novel prognostic factors. All analyses were performed by etiologic group. Results: Multivariate Cox regression identified the highest mortality risk among children with idiopathic disease (N=1192, CEP: 41%) when diagnosed over age 6 years, and with congestive heart failure (CHF) and decreased left ventricular fractional shortening (FS). Risk factors for those with myocarditis (N=272, CEP: 26%) were older age, CHF, and increased left ventricular (LV) end-diastolic dimension (EDD); while for neuromuscular disease (N=139, CEP: 40%), it was a decreased FS and increased EDD. Only univariate predictors were identified for children with familial isolated cardiomyopathy (N=79, CEP: 44%) including: CHF, increased EDD, end-systolic dimension, or LV mass, and decreased FS or ejection fraction), while for children with inborn errors of metabolism (N=43, CEP: 33%) risk factors included: a positive family history of cardiomyopathy or genetic syndromes. The group of children with malformation syndromes (N=6, CEP: 50%) was not large enough to model. Comparison of cause-specific event rates between Kaplan-Meier and cumulative incidence demonstrated an overestimation with the former method. Competing risk multivariate regression showed similar models to those for CEP, with the following exceptions: for neuromuscular disease, an increased EDD had a larger hazard ratio for transplant than for death; for idiopathic disease, an increased EDD was associated with transplant, but not with death, and growth retardation (height-for-age zscore) was associated with death but not transplant. Conclusions: Within etiologic grouping, demographics and echocardiographic values at diagnosis have varying predictive value. Generally, the presence of symptomatic disease in the form of CHF, echocardiographic evidence of more severe DCM, and increased age were indicative of worse outcomes. These results help to validate those from conflicting studies; however, they suggest that etiology modifies the importance of particular factors. Analysis of competing risk provides an alternate interpretation of studies with composite endpoints and assists in the transfer of clinically relevant information. For children with idiopathic and neuromuscular disease, the degree of dilation had a differential effect that has gone unrecognized. The novel finding of reduced stature and its effect on mortality suggests a potential for treatment and mitigation of poor outcomes in idiopathic DCM. Both increased dilation and reduced stature could be used to improve the triage process and refer children to cardiac transplantation who otherwise might die prematurely and unnecessarily. Subsequent studies on the utility of these factors and their effect on improving survival are warranted.

There has been a wealth of research conducted on the high school dropouts spanning several decades. It is estimated that compared with those who complete high school, the average high school dropout costs the economy approximately $250,000 more over his or her lifetime in terms of lower tax contributions, higher reliance on Medicaid and Medicare, higher rates of criminal activity, and higher reliance on welfare (Levin & Belfield, 2007). The nation suffers not only because of the loss in revenue but also as a result of the education level of the population. Individuals who choose to drop out of high school are less likely to be in the labor force than adults who earned a high school credential, and they fare worse in many aspects of life. In many studies on high school dropouts, an important challenge is how to determine an appropriate structural form for a statistical model to be used in making inferences and predictions. Many useful statistical modeling for survival analysis have been developed to study the competing risks frame of probability of dropping out and the probability of graduating; however, few methods exist for establishing the actual competing risks structural form of a model when the data contains two educational milestones – drop out and graduation. In this dissertation, we first utilized the data collected from the National Education Longitudinal Study (NELS: 88/2000) and proposed a discrete time competing risks hazard model and the corresponding model selection process to study the contributions of student’s academic ability, family background, school characteristics and vocational education to the probabilities of students graduating from or dropping out of high school. This model finds a way to overcome the shortcomings of the traditional models existing in the previous research. Within educational research, missing data is very common occurrence and can easily complicate the model selection problem. Handling missing data inappropriately can lead to bias and inaccurate inferences. This dissertation applies four missing data techniques to the key attributes including listwise deletion, dummy variable adjustment, mean imputation, and multiple imputation. Recommendations were offered for future endeavors and research in finding solutions to handle missing data in educational research. Finally, we outline the implementation of the proposed methodology. This research has the potential for both theoretical merit and implications for affecting educational policy. My dissertation adds to the limited body of literature of quantitative studies of the high school dropouts. A discrete time competing risks hazard model for predicting the probability of dropping out could become part of a powerful tool to identify students at risk of dropping out.

存活分析主要在研究事件的發生時間；傳統的存活分析並不考慮治癒者(或免疫者)的存在。若以失敗為事件，且造成失敗的可能原因不止一種，但它們不會同時發生，則這些失敗原因就是失敗事件的競爭風險。競爭風險可分為有參數的競爭風險與無母數的競爭風險。本文同時考慮了有治癒與有參數的混合廣義Gamma分配，並將預估計的位置參數與失敗機率有關的參數與解釋變數結合，代入Choi及Zhou(2002)提出的最大概似估計量的大樣本性質。並考慮在治癒情況下，利用電腦模擬來估計在型一設限及無訊息(non-informative)的隨機設限(random censoring)下之一個失敗原因與兩個失敗原因下的參數平均數與標準差。 / The purpose of survival analysis is aiming to analyze the timeline of events. The typically method of survival analysis don’t take account of the curer (or the immune). If the event is related to failure and there are more than one possible reason causing the failure but are not happening at the same time, we called the possible reasons a competing risk for failed occurrence. competing risk can be categorized as parameter and non-parameter. This research has considered the generalized gamma distribution over both cure and parameter aspects. In addition, it combines anticipated parameter with covariate which affected to the possibilities of failure. Follow by the previous data, it is then substituted by the large-sample property of the maximum likelihood estimator which is presented by Choi and Zhou in 2002. With considering the possibilities of cure, it uses computer modeling to investigate that under the condition of type-1 censoring and non-informatively random censoring, we will find out the parameter mean and standard error that is resulted by one and two reason causes failure.

Contexte : Dans les analyses de survie, un risque concurrent est un événement qui empêche l'observation de l'événement d'intérêt (le décès le plus souvent). Si la probabilité de survenue d'un risque concurrent dépend de la probabilité de l'événement d'intérêt, alors il ne peut pas être traité comme une censure. Les patients ayant une insuffisance rénale chronique terminale peuvent être traités par hémodialyse, dialyse péritonéale et greffe rénale. Ces traitements sont complémentaire et les patients peuvent passer d'une modalité de traitement à une autre au cours de leur prise en charge. La dépendance entre les changements de traitement et la probabilité de décès n'a pas été étudiée et ces changements sont traités comme des censures dans les analyses de survie.Objectifs : Analyser la dépendance entre les probabilités de décès en dialyse et de greffe rénale, et entre les probabilités de décès en dialyse péritonéale et de transfert en hémodialyse. Nous démontrerons les conséquences néfastes de la non-prise en compte de cette dépendance dans les analyses de survie en dialyse. Méthodes : (1) Nous avons comparé les estimations de probabilité d'événement obtenues par la méthode de Kaplan-Meier et la méthode de Kalbfleisch et Prentice sur 383 patient indicent consécutifs traités par dialyse péritonéale à Lille. (2) Nous avons analysé les données de 7318 patients incidents traités par hémodialyse en France grâce au registre national REIN. Nous avons utilisé un modèle multi-états pour analyse l'influence de l'inscription sur liste d'attente de greffe sur la probabilité de décès en dialyse. (3) Sur une cohorte de 2790 patients âgés de plus de 65 ans et traités par dialyse péritonéale issus du Registre de Dialyse Péritonéale de Langue Française (RDPLF), nous avons analysé les facteurs de contre-indication au transfert en HD en prenant en compte le décès comme risque concurrent à l’aide du modèle de Fine et Gray. Cette analyse a été complétée par un questionnaire réalisé auprès 55 des néphrologues pratiquant la dialyse péritonéale en France. Résultats : (1) La méthode de Kaplan Meier surestimait systématiquement la probabilité de décès du fait de la violation de l'hypothèse d'indépendance entre le décès et les risques concurrents. Cette méthode n'apparaît donc pas valide dans les analyses de survie en dialyse. La méthode de Kalbfleisch et Prentice était valide mais l'interprétation des incidences cumulées doit prendre en compte tous les risques concurrents. (2) La greffe rénale est un risque concurrent dépendant de la probabilité de décès des patients. Les patients inscrits sur liste d'attente de greffe avaient un risque de décès significativement plus bas que les autres patients, après ajustement sur l'âge et la présence de comorbidités. (3) Le transfert en hémodialyse est un risque concurrent qui semble dépendre de la probabilité de décès des patients. En effet, l'âge et la présence de comorbidités étaient à la fois des facteurs de risque de décès et des facteurs de contre-indications au transfert en hémodialyse. De plus, la plupart des néphrologues ayant répondu à notre enquête ont déclaré qu'une espérance de vie limitée pouvait constituer une contre-indication au transfert. Conclusion : Dans les études de cohorte de patients en insuffisance rénale chronique terminale, les analyses de survie devraient prendre en compte les changements de traitement car ce sont des risques concurrents dépendants de la probabilité de décès. Notre travail a montré que les modèles multi-états sont des outils statistiques flexibles qui permettent de bien représenter l'inter-dépendance entre les différentes modalités de traitement entre dialyse péritonéale, hémodialyse, greffe rénale et décès. / In survival analysis, a competing risk is an event that hinders the observation of the event of interest (usually death). If the probability of a competing risk depends on the probability of the event of interest, then it can not be treated as censoring. Patients with ESRD can be treated with hemodialysis, peritoneal dialysis and renal transplantation. These treatments are complementary and patients can move from one treatment modality to another. The dependence between changes in treatment modality and the probability of death has not been studied and these changes are censored in survival analysis.Objectives: To analyze the dependence between the probability of death in dialysis and kidney transplant, and the probability of death on peritoneal dialysis and transfer to hemodialysis. We demonstrate the negative consequences if this dependence is not taken into account in the survival analysis. Methods: (1) We compared estimates of event probability obtained by the Kaplan-Meier method and Kalbfleisch and Prentice on 383 consecutive indicent patients treated by peritoneal dialysis in Lille. (2) We analyzed data on 7318 incident patients undergoing hemodialysis in France from the national registry REIN. We used a multistate model to analyze the influence of inclusion on the transplant waiting list on the probability of death on dialysis. (3) In a cohort of 2790 patients aged over 65 and treated with peritoneal dialysis from the registry of the French Language Peritoneal Dialysis (RDPLF), we analyzed the factors against transfer-indication in HD taking into account death as competing risk using the Fine and Gray model. This analysis was complemented by a survey conducted among 55 nephrologists practicing Peritoneal dialysis in France.Results: (1) The Kaplan-Meier method systematically overestimated the probability of death due to violation of the assumption of independence between death and competing risks. This method does not appear valid in the analyzes of survival on dialysis. The method of Kalbfleisch and Prentice was valid but the interpretation of cumulative impacts must take into account all the competing risks. (2) Kidney transplantation is a competing risk depending on the probability of dying patients. Patients on the transplant waiting list had a risk of death significantly lower than other patients, after adjustment for age and comorbidity. (3) The transfer is a risk in hemodialysis competitor who seems to depend on the probability of dying patients. Indeed, age and comorbidities were both risk factors and death factors against transfer-indications for hemodialysis. Moreover, most nephrologists who responded to our survey reported that limited life expectancy could be an indication to the transfer-cons.Conclusion: In cohort studies of patients with ESRD, the survival analyzes should take into account changes in treatment because they are competing risks dependent on the probability of death. Our work has shown that multi-state models are statistical tools that enable flexible to adequately represent the interdependence between the different modalities of treatment for peritoneal dialysis, hemodialysis, kidney transplantation and death.