Multilevel models for survival analysis with random effects

A method for modeling survival data with multilevel clustering is described. The Cox partial likelihood is incorporated into the generalized linear mixed model (GLMM) methodology. Parameter estimation is achieved by maximizing a log likelihood analogous to the likelihood associated with the best linear unbiased prediction (BLUP) at the initial step of estimation and is extended to obtain residual maximum likelihood (REML) estimators of the variance component. Estimating equations for a three-level hierarchical survival model are developed in detail, and such a model is applied to analyze a set of chronic granulomatous disease (CGD) data on recurrent infections as an illustration with both hospital and patient effects being considered as random. Only the latter gives a significant contribution. A simulation study is carried out to evaluate the performance of the REML estimators. Further extension of the estimation procedure to models with an arbitrary number of levels is also discussed.     

On robustness and model flexibility in survival analysis: transformed hazard models and average effects

Yin and Ibrahim (2005a, Biometrics 61, 208-216) use a Box-Cox transformed hazard model to acknowledge uncertainty about how a linear predictor acts upon the hazard function of a failure-time response. Particularly, additive and proportional hazards models arise for particular values of the transformation parameter. As is often the case, however, this added model flexibility is obtained at the cost of lessened parameter interpretability. Particularly, the interpretation of the coefficients in the linear predictor is intertwined with the value of the transformation parameter. Moreover, some data sets contain very little information about this parameter. To shed light on the situation, we consider average effects based on averaging (over the joint distribution of the explanatory variables and the failure-time response) the partial derivatives of the hazard, or the log-hazard, with respect to the explanatory variables. First, we consider fitting models which do assume a particular form of covariate effects, for example, proportional hazards or additive hazards. In some such circumstances, average effects are seen to be inferential targets which are robust to the effect form being misspecified. Second, we consider average effects as targets of inference when using the transformed hazard model. We show that in addition to being more interpretable inferential targets, average effects can sometimes be estimated more efficiently than the corresponding regression coefficients.     

Relative impact of earlier diagnosis and improved treatment on survival for colorectal cancer: A US database study among elderly patients

PurposesTo estimate what proportion of improvement in relative survival was attributable to smaller stage/size due to early detection and what proportion was attributable to cancer chemotherapy in patients with colorectal cancer (CRC).MethodsWe studied 69,718 patients with CRC aged ≥66 years in 1992–2009 from Surveillance, Epidemiology, and End Results registries. Study periods were categorized into three periods according to the major changes or advances in screening and chemotherapy regimens: (1) Period-1 (1992–1995), during which there was no evidence-based recommendation for routine CRC screening and 5-fluorouracil was the mainstay for chemotherapy; (2) Period-2 (1996–2000), during which evidences and guidelines supported the use of fecal occult blood test (FOBT) and sigmoidoscopy for routine CRC screening; and (3) Period-3 (2001–2009), during which Medicare Program added the full coverage for colonoscopy screening to average-risk individuals, and several newly developed chemotherapy regimens were approved. Outcome variables included the likelihood of being diagnosed at an early stage or with a small tumor size, and improvement in relative survival.ResultsCompared to period-1, likelihood of being diagnosed with early stage CRC increased by 20% in period-2 (odds ratio = 1.2, 95%CI: 1.1–1.2) and 30% in period-3 (1.3, 1.2–1.4); and likelihood of being diagnosed with small-size CRC increased by 60% in period-2 and 110% in period-3. Similarly, 5-year overall relative survival increased from 51% in period-1 to 56% in period-2 and 60% in period-3. Increase in survival attributable to migration in stage/size was 9% in period-2 and 20% in period-3, while the remaining survival improvement during period-2 and period-3 were largely attributable to more effective chemotherapy regimens (≥71.6%) and other treatment factors (≤25%).ConclusionsImprovements in CRC screening resulted in a migration of CRC toward earlier tumor stage and smaller size, which contributed to ≤20% of survival increase. Survival improvement over the past 2 decades was largely explained by more effective chemotherapy regimens (≥71.6%).     

Bayesian influence measures for joint models for longitudinal and survival data

Summary This article develops a variety of influence measures for carrying out perturbation (or sensitivity) analysis to joint models of longitudinal and survival data (JMLS) in Bayesian analysis. A perturbation model is introduced to characterize individual and global perturbations to the three components of a Bayesian model, including the data points, the prior distribution, and the sampling distribution. Local influence measures are proposed to quantify the degree of these perturbations to the JMLS. The proposed methods allow the detection of outliers or influential observations and the assessment of the sensitivity of inferences to various unverifiable assumptions on the Bayesian analysis of JMLS. Simulation studies and a real data set are used to highlight the broad spectrum of applications for our Bayesian influence methods.     

Comparison of procedures to assess non-linear and time-varying effects in multivariable models for survival data

The focus of many medical applications is to model the impact of several factors on time to an event. A standard approach for such analyses is the Cox proportional hazards model. It assumes that the factors act linearly on the log hazard function (linearity assumption) and that their effects are constant over time (proportional hazards (PH) assumption). Variable selection is often required to specify a more parsimonious model aiming to include only variables with an influence on the outcome. As follow-up increases the effect of a variable often gets weaker, which means that it varies in time. However, spurious time-varying effects may also be introduced by mismodelling other parts of the multivariable model, such as omission of an important covariate or an incorrect functional form of a continuous covariate. These issues interact. To check whether the effect of a variable varies in time several tests for non-PH have been proposed. However, they are not sufficient to derive a model, as appropriate modelling of the shape of time-varying effects is required. In three examples we will compare five recently published strategies to assess whether and how the effects of covariates from a multivariable model vary in time. For practical use we will give some recommendations.     

Determinants of survival in children with congenital abnormalities: a long-term population-based cohort study

BACKGROUND: Today more children with birth defects survive early childhood because of improved medical care; however, little information is available about patterns of long-term mortality and survival in this population. In particular, it is not clear whether other birth characteristics, apart from birth defects, have any role in their mortality. METHODS: Two large cohorts of children with and without birth defects were followed for up to 17 years. More than 45,000 children with birth defects, and 45,000 matched children without birth defects born in Ontario between 1979 and 1986 were followed. Throughout the study period long-term survival rates and the risk of death were compared between the 2 cohorts. Birth characteristics were also examined to determine their effect on the risk of death. RESULTS: During the study the deaths of 3620 and 301 children with and without birth defects, respectively, were recorded, indicating that those with birth defects had a 13 times higher rate of mortality (relative risk [RR], 12.9, 95% confidence interval [CI], 12.1-13.7). Mortality rates in the birth-defects cohort remained higher even after 10-15 years. In both groups children of low gestational age and low birth weight had a higher risk of death. There was a strong dose-response relationship between the number of defects and the risk of death. CONCLUSIONS: Children born with abnormalities face many challenges throughout their lifetimes. If they survive the high mortality risk of the first year of life, they still have to face the considerably higher risk of death in the years to come. In addition to birth defects, other birth characteristics play an independent role in their mortality. These indicators could be used to identify high-risk children.     

Estimation of the causal effects on survival of two-stage nonrandomized treatment sequences for recurrent diseases

In the treatment of cancer, patients commonly receive a variety of sequential treatments. The initial treatments administered following diagnosis can vary, as well as subsequent salvage regimens given after disease recurrence. This article considers the situation where neither initial treatments nor salvage treatments are randomized. Assuming there are no unmeasured confounders, we estimate the joint causal effects on survival of initial and salvage treatments, that is, the effects of two-stage treatment sequences. For each individual treatment sequence, we estimate the survival distribution function and the mean restricted survival time. Different treatment sequences are then compared using these estimates and their corresponding covariances. Simulation studies were conducted to evaluate the performance of the methods, including their sensitivity to the violation of the assumption of no unmeasured confounders. The methods are illustrated by a retrospective study of patients with soft tissue sarcoma, which motivated this research.     

MicroRNAs as possible biomarkers for screening of aortic aneurysms: a systematic review and validation study

Context: There is an urgent need to identify non-invasive biomarkers for the early detection of aortic aneurysms, preceding a fatal event. The potential role for MicroRNAs (miRNAs) as diagnostic markers for aortic aneurysms was investigated through the present systematic review.

Objective: To perform a comprehensive review on published studies examining the association of miRNAs with aortic aneurysms and further validate these results with plasma samples collected from thoracic aortic aneurysm (TAA) patients.

Methods: The literature search was performed via numerous databases and articles were only included if they fulfilled the predefined eligibility criteria. The miRNAs reported three times or more with expression consistency were validated using plasma samples from TAA patients collected before and following surgery.

Results: Twenty-four articles were selected from the literature search and 11 miRNAs were chosen for validation using our samples. The miRNAs which were further validated were found to follow the trend in the regulation pattern as with the majority of the published data. MiRNA hsa-miR-193a-5p was found to be significantly down-regulated in the plasma samples collected before the aneurysmal removal when compared with postsurgical serum samples.

Conclusions: Numerous miRNAs have been associated with aortic aneurysms, and specifically hsa-miR-193a-5p and hsa-miR-30b-5p; therefore they warrant further investigation as potential biomarkers.

Registration: The protocol of the review was registered in Prospero Databases (ID: CRD42016039953)     

Thrombosis during pregnancy: Risks,prevention, and treatment for mother and fetus—harvesting the power of omic technology,biomarkers and in vitro or in vivo models to facilitate the treatment of thrombosis

Maternal thromboembolism and a spectrum of placenta‐mediated complications including the pre‐eclampsia syndromes, fetal growth restriction, fetal loss, and abruption manifest a shared etiopathogenesis and predisposing risk factors. Furthermore, these maternal and fetal complications are often linked to subsequent maternal health consequences that comprise the metabolic syndrome, namely, thromboembolism, chronic hypertension, and type II diabetes. Traditionally, several lines of evidence have linked vasoconstriction, excessive thrombosis and inflammation, and impaired trophoblast invasion at the uteroplacental interface as hallmark features of the placental complications. “Omic” technologies and biomarker development have been largely based upon advances in vascular biology, improved understanding of the molecular basis and biochemical pathways responsible for the clinically relevant diseases, and increasingly robust large cohort and/or registry based studies. Advances in understanding of innate and adaptive immunity appear to play an important role in several pregnancy complications. Strategies aimed at improving prediction of these pregnancy complications are often incorporating hemodynamic blood flow data using non‐invasive imaging technologies of the utero‐placental and maternal circulations early in pregnancy. Some evidence suggests that a multiple marker approach will yield the best performing prediction tools, which may then in turn offer the possibility of early intervention to prevent or ameliorate these pregnancy complications. Prediction of maternal cardiovascular and non‐cardiovascular consequences following pregnancy represents an important area of future research, which may have significant public health consequences not only for cardiovascular disease, but also for a variety of other disorders, such as autoimmune and neurodegenerative diseases. Birth Defects Research (Part C) 105:209–225, 2015. © 2015 Wiley Periodicals, Inc.     

Proteomics approaches in cervical cancer: focus on the discovery of biomarkers for diagnosis and drug treatment monitoring

Introduction: The HPV virus accounts for the majority of cervical cancer cases. Although a diagnostic tool (Pap Test) is widely available, cervical cancer incidence still remains high worldwide, and especially in developing countries, attributed to a large extent to suboptimal sensitivities of the Pap test and unavailability of the test in developing countries.

Areas covered: Proteomics approaches have been used in order to understand the HPV virus correlation to cervical cancer pathology, as well as to discover putative biomarkers for early cervical cancer diagnosis and drug mode of action.

Expert commentary: The present review summarizes the latest in vitro and in vivo proteomic studies for the discovery of putative cervical cancer biomarkers and the evaluation of available drugs and treatments.     

A joint frailty model for survival and gap times between recurrent events

Therapy for patients with a recurrent disease focuses on delaying disease recurrence and prolonging survival. A common analysis approach for such data is to estimate the distribution of disease-free survival, that is, the time to the first disease recurrence or death, whichever happens first. However, treating death similarly as disease recurrence may give misleading results. Also considering only the first recurrence and ignoring subsequent ones can result in loss of statistical power. We use a joint frailty model to simultaneously analyze disease recurrences and survival. Separate parameters for disease recurrence and survival are used in the joint model to distinguish treatment effects on these two types of events. The correlation between disease recurrences and survival is taken into account by a shared frailty. The effect of disease recurrence on survival can also be estimated by this model. The EM algorithm is used to fit the model, with Markov chain Monte Carlo simulations in the E-steps. The method is evaluated by simulation studies and illustrated through a study of patients with heart failure. Sensitivity analysis for the parametric assumption of the frailty distribution is assessed by simulations.     

Regression survival analysis with an assumed copula for dependent censoring: a sensitivity analysis approach

SUMMARY: In clinical studies, when censoring is caused by competing risks or patient withdrawal, there is always a concern about the validity of treatment effect estimates that are obtained under the assumption of independent censoring. Because dependent censoring is nonidentifiable without additional information, the best we can do is a sensitivity analysis to assess the changes of parameter estimates under different assumptions about the association between failure and censoring. This analysis is especially useful when knowledge about such association is available through literature review or expert opinions. In a regression analysis setting, the consequences of falsely assuming independent censoring on parameter estimates are not clear. Neither the direction nor the magnitude of the potential bias can be easily predicted. We provide an approach to do sensitivity analysis for the widely used Cox proportional hazards models. The joint distribution of the failure and censoring times is assumed to be a function of their marginal distributions. This function is called a copula. Under this assumption, we propose an iteration algorithm to estimate the regression parameters and marginal survival functions. Simulation studies show that this algorithm works well. We apply the proposed sensitivity analysis approach to the data from an AIDS clinical trial in which 27% of the patients withdrew due to toxicity or at the request of the patient or investigator.     

Bacterial pili,with emphasis on Mycobacterium tuberculosis curli pili: potential biomarkers for point-of care tests and therapeutics

Context: Novel biomarkers are essential for developing rapid diagnostics and therapeutic interventions

Objective: This review aimed to highlight biomarker characterisation and assessment of unique bacterial pili.

Methods: A PubMed search for bacterial pili, diagnostics, vaccine and therapeutics was performed, with emphasis on the well characterised pili.

Results: In total, 46 papers were identified and reviewed.

Conclusion: Extensive analyses of pili enabled by advanced nanotechnology and whole genome sequencing provide evidence that they are strong biomarker candidates. Mycobacterium tuberculosis curli pili are emphasised as important epitopes for the development of much needed point-of-care diagnostics and therapeutics.