Causal inference in survival analysis using pseudo-observations

Hazard Ratios and Alternative Effect Measures: An Applied Illustration

PURPOSE

Although the limitations of hazard ratios (HRs) for quantifying treatment effects in right-censored data have been widely discussed, HRs are still preferentially reported over other, more interpretable effect measures. This may stem from the fact that there are few applied examples that directly contrast the HR and its interpretation with alternative effect measures.

METHODS

We analyzed data from two randomized clinical trials comparing panitumumab plus standard-of-care chemotherapy (SOCC) with SOCC alone as first- and second-line treatment for metastatic colorectal cancer. We report the effect of treatment with panitumumab on progression-free survival (PFS) using a Cox proportional hazards model to estimate the HR and the Kaplan-Meier estimator of cumulative incidence (risk). Further analyses included examining the cumulative incidence curves; kernel-smoothed, non-parametric hazards curves; fitting the Cox model with a continuous time variable; and estimating restricted mean survival as well as median survival.

RESULTS

The HR was 0.82 (95% confidence interval [CI]: 0.71, 0.93), while the risk ratio (or relative risk [i.e., ratio of the cumulative incidence among the treated versus comparator]) was 0.99 (95% CI: 0.96, 1.02). These two measures suggest apparently different conclusions: either a treatment benefit or no effect. Through subsequent analyses, we demonstrated that, while the cumulative incidence of the outcome was similar by the end of follow-up regardless of treatment, the panitumumab treated group experienced longer PFS than those randomized to SOCC. Substantial nonproportional hazards were evident with panitumumab treatment reducing the hazard of progression/mortality during the first ~1.75 years but associated with an increased hazard of progress/mortality thereafter.

DISCUSSION

This example underscores the difficulties in interpreting HRs, particularly in the setting of qualitative violations of proportional hazards, and the value of quantifying treatment effects via multiple effect measures.

Prognostic score-based methods for estimating center effects based on survival probability: Application to post-kidney transplant survival

In evaluating the performance of different facilities or centers on survival outcomes, the standardized mortality ratio (SMR), which compares the observed to expected mortality has been widely used, particularly in the evaluation of kidney transplant centers. Despite its utility, the SMR may exaggerate center effects in settings where survival probability is relatively high. An example is one-year graft survival among U.S. kidney transplant recipients. We propose a novel approach to estimate center effects in terms of differences in survival probability (ie, each center versus a reference population). An essential component of the method is a prognostic score weighting technique, which permits accurately evaluating centers without necessarily specifying a correct survival model. Advantages of our approach over existing facility-profiling methods include a metric based on survival probability (greater clinical relevance than ratios of counts/rates); direct standardization (valid to compare between centers, unlike indirect standardization based methods, such as the SMR); and less reliance on correct model specification (since the assumed model is used to generate risk classes as opposed to fitted-value based 'expected' counts). We establish the asymptotic properties of the proposed weighted estimator and evaluate its finite-sample performance under a diverse set of simulation settings. The method is then applied to evaluate U.S. kidney transplant centers with respect to graft survival probability.

Incidence of type 2 diabetes, cardiovascular disease and chronic kidney disease in patients with multiple sclerosis initiating disease-modifying therapies: Retrospective cohort study using a frequentist model averaging statistical framework

Researchers are increasingly using insights derived from large-scale, electronic healthcare data to inform drug development and provide human validation of novel treatment pathways and aid in drug repurposing/repositioning. The objective of this study was to determine whether treatment of patients with multiple sclerosis with dimethyl fumarate, an activator of the nuclear factor erythroid 2-related factor 2 (Nrf2) pathway, results in a change in incidence of type 2 diabetes and its complications. This retrospective cohort study used administrative claims data to derive four cohorts of adults with multiple sclerosis initiating dimethyl fumarate, teriflunomide, glatiramer acetate or fingolimod between January 2013 and December 2018. A causal inference frequentist model averaging framework based on machine learning was used to compare the time to first occurrence of a composite endpoint of type 2 diabetes, cardiovascular disease or chronic kidney disease, as well as each individual outcome, across the four treatment cohorts. There was a statistically significantly lower risk of incidence for dimethyl fumarate versus teriflunomide for the composite endpoint (restricted hazard ratio [95% confidence interval] 0.70 [0.55, 0.90]) and type 2 diabetes (0.65 [0.49, 0.98]), myocardial infarction (0.59 [0.35, 0.97]) and chronic kidney disease (0.52 [0.28, 0.86]). No differences for other individual outcomes or for dimethyl fumarate versus the other two cohorts were observed. This study effectively demonstrated the use of an innovative statistical methodology to test a clinical hypothesis using real-world data to perform early target validation for drug discovery. Although there was a trend among patients treated with dimethyl fumarate towards a decreased incidence of type 2 diabetes, cardiovascular disease and chronic kidney disease relative to other disease-modifying therapies-which was statistically significant for the comparison with teriflunomide-this study did not definitively support the hypothesis that Nrf2 activation provided additional metabolic disease benefit in patients with multiple sclerosis.

Analysis of survival data with nonproportional hazards: A comparison of propensity-score-weighted methods

One of the most common ways researchers compare cancer survival outcomes across treatments from observational data is using Cox regression. This model depends on its underlying assumption of proportional hazards, but in some real-world cases, such as when comparing different classes of cancer therapies, substantial violations may occur. In this situation, researchers have several alternative methods to choose from, including Cox models with time-varying hazard ratios; parametric accelerated failure time models; Kaplan-Meier curves; and pseudo-observations. It is unclear which of these models are likely to perform best in practice. To fill this gap in the literature, we perform a neutral comparison study of candidate approaches. We examine clinically meaningful outcome measures that can be computed and directly compared across each method, namely, survival probability at time T, median survival, and restricted mean survival. To adjust for differences between treatment groups, we use inverse probability of treatment weighting based on the propensity score. We conduct simulation studies under a range of scenarios, and determine the biases, coverages, and standard errors of the average treatment effects for each method. We then demonstrate the use of these approaches using two published observational studies of survival after cancer treatment. The first examines chemotherapy in sarcoma, which has a late treatment effect (i.e., similar survival initially, but after 2 years the chemotherapy group shows a benefit). The other study is a comparison of surgical techniques for kidney cancer, where survival differences are attenuated over time.

Estimation of time-specific intervention effects on continuously distributed time-to-event outcomes by targeted maximum likelihood estimation

This work considers targeted maximum likelihood estimation (TMLE) of treatment effects on absolute risk and survival probabilities in classical time-to-event settings characterized by right-censoring and competing risks. TMLE is a general methodology combining flexible ensemble learning and semiparametric efficiency theory in a two-step procedure for substitution estimation of causal parameters. We specialize and extend the continuous-time TMLE methods for competing risks settings, proposing a targeting algorithm that iteratively updates cause-specific hazards to solve the efficient influence curve equation for the target parameter. As part of the work, we further detail and implement the recently proposed highly adaptive lasso estimator for continuous-time conditional hazards with L1 -penalized Poisson regression. The resulting estimation procedure benefits from relying solely on very mild nonparametric restrictions on the statistical model, thus providing a novel tool for machine-learning-based semiparametric causal inference for continuous-time time-to-event data. We apply the methods to a publicly available dataset on follicular cell lymphoma where subjects are followed over time until disease relapse or death without relapse. The data display important time-varying effects that can be captured by the highly adaptive lasso. In our simulations that are designed to imitate the data, we compare our methods to a similar approach based on random survival forests and to the discrete-time TMLE.

Causal inference for oncology: past developments and current challenges

In this paper, we review some important early developments on causal inference in medical statistics and epidemiology that were inspired by questions in oncology. We examine two classical examples from the literature and point to a current area of ongoing methodological development, namely the estimation of optimal adaptive treatment strategies. While causal approaches to analysis have become more routine in oncology research, many exciting challenges and open problems remain, particularly in the context of censored outcomes.

Multiply robust estimator for the difference in survival functions using pseudo-observations

BACKGROUND

When estimating the causal effect on survival outcomes in observational studies, it is necessary to adjust confounding factors due to unbalanced covariates between treatment and control groups. There is no study on multiple robust method for estimating the difference in survival functions. In this study, we propose a multiply robust (MR) estimator, allowing multiple propensity score models and outcome regression models, to provide multiple protection.

METHOD

Based on the previous MR estimator (Han 2014) and pseudo-observation approach, we proposed a new MR estimator for estimating the difference in survival functions. The proposed MR estimator based on the pseudo-observation approach has several advantages. First, the proposed estimator has a small bias when any PS and OR models were correctly specified. Second, the proposed estimator considers the advantage pf the pseudo-observation approach, which avoids proportional hazards assumption. A Monte Carlo simulation study was performed to evaluate the performance of the proposed estimator. And the proposed estimator was used to estimate the effect of chemotherapy on triple-negative breast cancer (TNBC) in real data.

RESULTS

The simulation studies showed that the bias of the proposed estimator was small, and the coverage rate was close to 95% when any model for propensity score or outcome regression is correctly specified regardless of whether the proportional hazard assumption holds, finite sample size and censoring rate. And the simulation results also showed that even though the propensity score models are misspecified, the bias of the proposed estimator was still small when there is a correct model in candidate outcome regression models. And we applied the proposed estimator in real data, finding that chemotherapy could improve the prognosis of TNBC.

CONCLUSIONS

The proposed estimator, allowing multiple propensity score and outcome regression models, provides multiple protection for estimating the difference in survival functions. The proposed estimator provided a new choice when researchers have a "difficult time" choosing only one model for their studies.

Propensity score weighting methods for causal subgroup analysis with time-to-event outcomes

Evaluating causal effects of an intervention in pre-specified subgroups is a standard goal in comparative effectiveness research. Despite recent advancements in causal subgroup analysis, research on time-to-event outcomes has been lacking. This article investigates the propensity score weighting method for causal subgroup survival analysis. We introduce two causal estimands, the subgroup marginal hazard ratio and subgroup restricted average causal effect, and provide corresponding propensity score weighting estimators. We analytically established that the bias of subgroup-restricted average causal effect is determined by subgroup covariate balance. Using extensive simulations, we compare the performance of various combinations of propensity score models (logistic regression, random forests, least absolute shrinkage and selection operator, and generalized boosted models) and weighting schemes (inverse probability weighting, and overlap weighting) for estimating the causal estimands. We find that the logistic model with subgroup-covariate interactions selected by least absolute shrinkage and selection operator consistently outperforms other propensity score models. Also, overlap weighting generally outperforms inverse probability weighting in terms of balance, bias and variance, and the advantage is particularly pronounced in small subgroups and/or in the presence of poor overlap. We applied the methods to the observational Comparing Options for Management: PAtient-centered REsults for Uterine Fibroids study to evaluate the causal effects of myomectomy versus hysterectomy on the time to disease recurrence in a number of pre-specified subgroups of patients with uterine fibroids.

An inverse probability weighted regression method that accounts for right-censoring for causal inference with multiple treatments and a binary outcome

Comparative effectiveness research often involves evaluating the differences in the risks of an event of interest between two or more treatments using observational data. Often, the post-treatment outcome of interest is whether the event happens within a pre-specified time window, which leads to a binary outcome. One source of bias for estimating the causal treatment effect is the presence of confounders, which are usually controlled using propensity score-based methods. An additional source of bias is right-censoring, which occurs when the information on the outcome of interest is not completely available due to dropout, study termination, or treatment switch before the event of interest. We propose an inverse probability weighted regression-based estimator that can simultaneously handle both confounding and right-censoring, calling the method CIPWR, with the letter C highlighting the censoring component. CIPWR estimates the average treatment effects by averaging the predicted outcomes obtained from a logistic regression model that is fitted using a weighted score function. The CIPWR estimator has a double robustness property such that estimation consistency can be achieved when either the model for the outcome or the models for both treatment and censoring are correctly specified. We establish the asymptotic properties of the CIPWR estimator for conducting inference, and compare its finite sample performance with that of several alternatives through simulation studies. The methods under comparison are applied to a cohort of prostate cancer patients from an insurance claims database for comparing the adverse effects of four candidate drugs for advanced stage prostate cancer.

Causal inference in survival analysis under deterministic missingness of confounders in register data

Long-term register data offer unique opportunities to explore causal effects of treatments on time-to-event outcomes, in well-characterized populations with minimum loss of follow-up. However, the structure of the data may pose methodological challenges. Motivated by the Swedish Renal Registry and estimation of survival differences for renal replacement therapies, we focus on the particular case when an important confounder is not recorded in the early period of the register, so that the entry date to the register deterministically predicts confounder missingness. In addition, an evolving composition of the treatment arms populations, and suspected improved survival outcomes in later periods lead to informative administrative censoring, unless the entry date is appropriately accounted for. We investigate different consequences of these issues on causal effect estimation following multiple imputation of the missing covariate data. We analyse the performance of different combinations of imputation models and estimation methods for the population average survival. We further evaluate the sensitivity of our results to the nature of censoring and misspecification of fitted models. We find that an imputation model including the cumulative baseline hazard, event indicator, covariates and interactions between the cumulative baseline hazard and covariates, followed by regression standardization, leads to the best estimation results overall, in simulations. Standardization has two advantages over inverse probability of treatment weighting here: it can directly account for the informative censoring by including the entry date as a covariate in the outcome model, and allows for straightforward variance computation using readily available software.

A comparison of different methods to adjust survival curves for confounders

Treatment specific survival curves are an important tool to illustrate the treatment effect in studies with time-to-event outcomes. In non-randomized studies, unadjusted estimates can lead to biased depictions due to confounding. Multiple methods to adjust survival curves for confounders exist. However, it is currently unclear which method is the most appropriate in which situation. Our goal is to compare forms of inverse probability of treatment weighting, the G-Formula, propensity score matching, empirical likelihood estimation and augmented estimators as well as their pseudo-values based counterparts in different scenarios with a focus on their bias and goodness-of-fit. We provide a short review of all methods and illustrate their usage by contrasting the survival of smokers and non-smokers, using data from the German Epidemiological Trial on Ankle-Brachial-Index. Subsequently, we compare the methods using a Monte-Carlo simulation. We consider scenarios in which correctly or incorrectly specified models for describing the treatment assignment and the time-to-event outcome are used with varying sample sizes. The bias and goodness-of-fit is determined by taking the entire survival curve into account. When used properly, all methods showed no systematic bias in medium to large samples. Cox regression based methods, however, showed systematic bias in small samples. The goodness-of-fit varied greatly between different methods and scenarios. Methods utilizing an outcome model were more efficient than other techniques, while augmented estimators using an additional treatment assignment model were unbiased when either model was correct with a goodness-of-fit comparable to other methods. These "doubly-robust" methods have important advantages in every considered scenario.

Analysis of survival data with cure fraction and variable selection: A pseudo-observations approach

In biomedical studies, survival data with a cure fraction (the proportion of subjects cured of disease) are commonly encountered. The mixture cure and bounded cumulative hazard models are two main types of cure fraction models when analyzing survival data with long-term survivors. In this article, in the framework of the Cox proportional hazards mixture cure model and bounded cumulative hazard model, we propose several estimators utilizing pseudo-observations to assess the effects of covariates on the cure rate and the risk of having the event of interest for survival data with a cure fraction. A variable selection procedure is also presented based on the pseudo-observations using penalized generalized estimating equations for proportional hazards mixture cure and bounded cumulative hazard models. Extensive simulation studies are conducted to examine the proposed methods. The proposed technique is demonstrated through applications to a melanoma study and a dental data set with high-dimensional covariates.

A Review of Causal Inference for External Comparator Arm Studies

Randomized controlled trials (RCTs) are the gold standard design to establish the efficacy of new drugs and to support regulatory decision making. However, a marked increase in the submission of single-arm trials (SATs) has been observed in recent years, especially in the field of oncology due to the trend towards precision medicine contributing to the rise of new therapeutic interventions for rare diseases. SATs lack results for control patients, and information from external sources can be compiled to provide context for better interpretability of study results. External comparator arm (ECA) studies are defined as a clinical trial (most commonly a SAT) and an ECA of a comparable cohort of patients-commonly derived from real-world settings including registries, natural history studies, or medical records of routine care. This publication aims to provide a methodological overview, to sketch emergent best practice recommendations and to identify future methodological research topics. Specifically, existing scientific and regulatory guidance for ECA studies is reviewed and appropriate causal inference methods are discussed. Further topics include sample size considerations, use of estimands, handling of different data sources regarding differential baseline covariate definitions, differential endpoint measurements and timings. In addition, unique features of ECA studies are highlighted, specifically the opportunity to address bias caused by unmeasured ECA covariates, which are available in the SAT.

Regression modeling of restricted mean survival time for left-truncated right-censored data

The restricted mean survival time (RMST) is a clinically meaningful summary measure in studies with survival outcomes. Statistical methods have been developed for regression analysis of RMST to investigate impacts of covariates on RMST, which is a useful alternative to the Cox regression analysis. However, existing methods for regression modeling of RMST are not applicable to left-truncated right-censored data that arise frequently in prevalent cohort studies, for which the sampling bias due to left truncation and informative censoring induced by the prevalent sampling scheme must be properly addressed. The pseudo-observation (PO) approach has been used in regression modeling of RMST for right-censored data and competing-risks data. For left-truncated right-censored data, we propose to directly model RMST as a function of baseline covariates based on POs under general censoring mechanisms. We adjust for the potential covariate-dependent censoring or dependent censoring by the inverse probability of censoring weighting method. We establish large sample properties of the proposed estimators and assess their finite sample performances by simulation studies under various scenarios. We apply the proposed methods to a prevalent cohort of women diagnosed with stage IV breast cancer identified from surveillance, epidemiology, and end results-medicare linked database.

Timing of adjuvant chemotherapy initiation and mortality among colon cancer patients at a safety-net health system

Background

Prior studies reported survival benefits from early initiation of adjuvant chemotherapy for stage III colon cancer, but this evidence was derived from studies that may be sensitive to time-related biases. Therefore, we aimed to estimate the effect of initiating adjuvant chemotherapy ≤8 or ≤ 12 weeks on overall and disease-free survival among stage III colon cancer patients using a study design that helps address time-related biases.

Methods

We used institutional registry data from JPS Oncology and Infusion Center, a Comprehensive Community Cancer Program. Eligible patients were adults aged < 80 years, diagnosed with first primary stage III colon cancer between 2011 and 2017, and received surgical resection with curative intent. We emulated a target trial with sequential eligibility. We subsequently pooled the trials and estimated risk ratios (RRs) along with 95% confidence limits (CL) for all-cause mortality and recurrence or death at 5-years between initiators and non-initiators of adjuvant chemotherapy ≤8 or ≤ 12 weeks using pseudo-observations and a marginal structural model with stabilized inverse probability of treatment weights.

Results

Our study population comprised 222 (for assessing initiation ≤8 weeks) and 310 (for assessing initiation ≤12 weeks) observations, of whom the majority were racial/ethnic minorities (64–65%), or uninsured with or without enrollment in our hospital-based medical assistance program (68–71%). Initiation of adjuvant chemotherapy ≤8 weeks of surgical resection did not improve overall survival (RR for all-cause mortality = 1.04, 95% CL: 0.57, 1.92) or disease-free survival (RR for recurrence or death = 1.07, 95% CL: 0.61, 1.88). The results were similar for initiation of adjuvant chemotherapy ≤12 weeks of surgical resection.

Conclusions

Our results suggest that the overall and disease-free survival benefits of initiating adjuvant chemotherapy ≤8 or ≤ 12 weeks of surgical resection may be overestimated in prior studies, which may be attributable to time-related biases. Nevertheless, our estimates were imprecise and differences in population characteristics are an alternate explanation. Additional studies that address time-related biases are needed to clarify our findings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-022-09688-w.

Doubly-robust estimator of the difference in restricted mean times lost with competing risks data

In the context of competing risks data, the subdistribution hazard ratio has limited clinical interpretability to measure treatment effects. An alternative is the difference in restricted mean times lost (RMTL), which gives the mean time lost to a specific cause of failure between treatment groups. In non-randomized studies, the average causal effect is conventionally used for decision-making about treatment and public health policies. We show how the difference in RMTL can be estimated by contrasting the integrated cumulative incidence functions from a Fine-Gray model. We also show how the difference in RMTL can be estimated by using inverse probability of treatment weighting and contrasts between weighted non-parametric estimators of the area below the cumulative incidence. We use pseudo-observation approaches to estimate both component models and we integrate them into a doubly-robust estimator. We demonstrate that this estimator is consistent when either component is correctly specified. We conduct simulation studies to assess its finite-sample performance and demonstrate its inherited consistency property from its component models. We also examine the performance of this estimator under varying degrees of covariate overlap and under a model misspecification of nonlinearity. We apply the proposed method to assess biomarker-treatment interaction in subpopulations of the POPLAR and OAK randomized controlled trials of second-line therapy for advanced non-small-cell lung cancer.

Doubly-robust methods for differences in restricted mean lifetimes using pseudo-observations

In clinical studies or trials comparing survival times between two treatment groups, the restricted mean lifetime (RML), defined as the expectation of the survival from time 0 to a prespecified time-point, is often the quantity of interest that is readily interpretable to clinicians without any modeling restrictions. It is well known that if the treatments are not randomized (as in observational studies), covariate adjustment is necessary to account for treatment imbalances due to confounding factors. In this article, we propose a simple doubly-robust pseudo-value approach to effectively estimate the difference in the RML between two groups (akin to a metric for estimating average causal effects), while accounting for confounders. The proposed method combines two general approaches: (a) group-specific regression models for the time-to-event and covariate information, and (b) inverse probability of treatment assignment weights, where the RMLs are replaced by the corresponding pseudo-observations for survival outcomes, thereby mitigating the estimation complexities in presence of censoring. The proposed estimator is double-robust, in the sense that it is consistent if at least one of the two working models remains correct. In addition, we explore the potential of available machine learning algorithms in causal inference to reduce possible bias of the causal estimates in presence of a complex association between the survival outcome and covariates. We conduct extensive simulation studies to assess the finite-sample performance of the pseudo-value causal effect estimators. Furthermore, we illustrate our methodology via application to a dataset from a breast cancer cohort study. The proposed method is implementable using the R package drRML, available in GitHub.

Impact of prognostic factor distributions on mortality disparities for socioeconomically disadvantaged cancer patients

PURPOSE

We aimed to assess whether differences in the distributions of prognostic factors explain reported mortality disparities between urban safety-net and Surveillance, Epidemiology, and End Results (SEER) cancer populations.

METHODS

We used data from SEER and a safety-net cancer center in Texas. Eligible patients were adults aged ≤64 years and diagnosed with first primary female breast, colorectal, or lung cancer between 2008 and 2016. We estimated crude and adjusted risk differences (RD) in 3- and 5-year all-cause mortality (1- and 3-year for lung cancer), where adjustment was based on entropy balancing weights that standardized the distribution of sociodemographic and tumor characteristics between the two populations.

RESULTS

Our study populations comprised 1914 safety-net patients and 389,709 SEER patients. For breast cancer, the crude 3- and 5-year mortality RDs between safety-net and SEER populations were 7.7% (95% confidence limits [CL]: 4.3%, 11%) and 11% (95% CL: 6.7%, 16%). Adjustment for measured prognostic factors reduced the mortality RDs (3-year adjusted RD = 0.049%, 95% CL: -2.6%, 2.6%; 5-year adjusted RD = 5.6%, 95% CL: -0.83%, 12%). We observed similar patterns for colorectal and lung cancer albeit less magnitude.

CONCLUSIONS

Sociodemographic and tumor characteristics may largely explain early mortality disparities between safety-net and SEER populations but not late mortality disparities.

Replacing Red Meat with Other Nonmeat Food Sources of Protein is Associated with a Reduced Risk of Type 2 Diabetes in a Danish Cohort of Middle-Aged Adults

BACKGROUND

Few cohort studies have modelled replacements of red meat with other sources of protein on subsequent risk of type 2 diabetes using dietary changes.

OBJECTIVES

To determine whether replacing red meat with other food sources of protein is associated with a lower risk of type 2 diabetes.

METHODS

We used data from the Danish Diet, Cancer, and Health cohort (n = 39,437) of middle-aged (55-72 years old) men and women who underwent 2 dietary assessments roughly 5 years apart to investigate dietary changes. The pseudo-observation method was used to model the average exposure effect of decreasing the intake of red meat while increasing the intake of either poultry, fish, eggs, milk, yogurt, cheese, whole grains, or refined grains on the subsequent 10-year risk of developing type 2 diabetes, compared with no changes in the intakes of these foods.

RESULTS

Replacing 1 serving/day (100 g/day) of red meat with 1 serving/day of eggs [risk difference (RD), -2.7%; 95% CI: -4.0 to -1.1%; serving size: 50 g/day], milk (RD, -1.2%; 95% CI: -2.1 to -0.4%; 200 g/day), yogurt (RD, -1.5%; 95% CI: -2.4 to -0.7%; 70 g/day), whole grains (RD, -1.7%; 95% CI: -2.5 to -0.9%; 30 g/day), or refined grains (RD, -1.2%; 95% CI: -2.0 to -0.3%; 30 g/day) was associated with a reduced risk of type 2 diabetes. Analyses of replacements with poultry or cheese, but not fish, also suggested a lower risk, but with wide CIs. After further adjustment for potential mediators (BMI, waist circumference, and history of hypertension or hypercholesterolemia), only the replacement with eggs was associated with a reduced risk (RD, -1.7%; 95% CI: -3.0 to -0.5%; 50 g/day).

CONCLUSIONS

Replacing red meat with eggs in middle-aged adults may reduce the risk of type 2 diabetes. In models not adjusted for potential mediators, replacing red meat with milk, yogurt, whole grains, or refined grains was also associated with a reduced risk of type 2 diabetes.

The Effect of Body Mass Index and Weight-Adjusted Fluid Dosing on Mortality in Sepsis

PURPOSE

The Surviving Sepsis Campaign guidelines recommend 30 mL/kg of fluids within 3 hours (30by3) of sepsis-induced hypoperfusion, but a national mandate released an allowance for dosing based on ideal instead of actual body weight (IBW/ABW) for obese patients. This study aims to determine the dose-effect of 30by3 for patients with severe sepsis or septic shock (SS/SS) with respect to body mass index (BMI) categories and secondarily, examine the clinical impact of IBW vs. ABW-based dosing.

METHODS

Retrospective cohort study of adults (≥18 years; n = 1,032) with SS/SS presenting to an urban, tertiary-care emergency department. Models include MEDS score, antibiotic timing, lactate, renal and heart failure, among others.

RESULTS

The cohort was 10.2% underweight and 28.7% obese. Overall mortality was 17.1% with 20.4% shock mortality. An exponential increase in mortality was observed for each 5 mL/kg under 30by3 for underweight (p = 0.02), but not obese patients. ABW vs IBW-30by3 dosing was reached in 80.0 vs 52.4% (underweight), 56.4 vs 56.9% (normal/overweight), and 23.3 vs 46.0% (obese). Across all BMI categories, there was increased mortality for not reaching ABW-based 30by3 dosing (OR 1.78, 95% CI 1.18-2.69) with no significant impact for IBW (OR 1.28, 95% CI 0.87 -1.91). The increased mortality for failing to reach ABW-dosed 30by3 remained for underweight patients ABW (OR 5.82, 95% CI 1.32-25.57) but not obese patients. Longer ICU stays were observed for not reaching 30by3 based on ABW (β = 2.40, 95% CI 0.84-3.95) and IBW dosing (β = 1.58, 95% CI 0.07-3.08) overall. This effect remained for obese and underweight (except IBW dosing) patients.

CONCLUSIONS

An exponential, dose-effect increase in mortality was observed for underweight patients not receiving 30by3. Therefore, the mortality impact of under-dosing may be amplified using ABW for underweight patients. Fluid dosing did not impact mortality for obese patients, but we caution against deviation from guidelines without further studies.

G-estimation of structural nested mean models for competing risks data using pseudo-observations

This article provides methods of causal inference for competing risks data. The methods are formulated as structural nested mean models of causal effects directly related to the cumulative incidence function or subdistribution hazard, which reflect the survival experience of a subject in the presence of competing risks. The effect measures include causal risk differences, causal risk ratios, causal subdistribution hazard ratios, and causal effects of time-varying exposures. Inference is implemented by g-estimation using pseudo-observations, a technique to handle censoring. The finite-sample performance of the proposed estimators in simulated datasets and application to time-varying exposures in a cohort study of type 2 diabetes are also presented.

Non-parametric individual treatment effect estimation for survival data with random forests

MOTIVATION

Personalized medicine often relies on accurate estimation of a treatment effect for specific subjects. This estimation can be based on the subject's baseline covariates but additional complications arise for a time-to-event response subject to censoring. In this paper, the treatment effect is measured as the difference between the mean survival time of a treated subject and the mean survival time of a control subject. We propose a new random forest method for estimating the individual treatment effect with survival data. The random forest is formed by individual trees built with a splitting rule specifically designed to partition the data according to the individual treatment effect. For a new subject, the forest provides a set of similar subjects from the training dataset that can be used to compute an estimation of the individual treatment effect with any adequate method.

RESULTS

The merits of the proposed method are investigated with a simulation study where it is compared to numerous competitors, including recent state-of-the-art methods. The results indicate that the proposed method has a very good and stable performance to estimate the individual treatment effects. Two examples of application with a colon cancer data and breast cancer data show that the proposed method can detect a treatment effect in a sub-population even when the overall effect is small or nonexistent.

AVAILABILITY AND IMPLEMENTATION

The authors are working on an R package implementing the proposed method and it will be available soon. In the meantime, the code can be obtained from the first author at sami.tabib@hec.ca.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Smoking cessation and survival among people diagnosed with non-metastatic cancer

Background

We aimed to estimate the effects of smoking cessation on survival among people diagnosed with cancer.

Methods

We used data from a Comprehensive Community Cancer Program that is part of a large urban safety-net hospital system. Eligible patients were diagnosed with primary invasive solid tumors between 2013 and 2015, and were current smokers at time of diagnosis. Our exposure of interest was initiation of smoking cessation within 6 months of cancer diagnosis. We estimated inverse probability weighted restricted mean survival time (RMST) differences and risk ratio (RR) for all cause 3-year mortality.

Results

Our study population comprised 369 patients, of whom 42% were aged < 55 years, 59% were male, 44% were racial/ethnic minorities, and 59% were uninsured. The 3-year RMST was 1.8 (95% CL: − 1.5, 5.1) months longer for individuals who initiated smoking cessation within 6 months of cancer diagnosis. The point estimate for risk of 3-year mortality was lower for initiation of smoking cessation within 6 months of diagnosis compared with no initiation within 6 months (RR = 0.72, 95% CL: 0.37, 1.4).

Conclusions

Our point estimates suggest longer 3-year survival, but the results are compatible with 1.5 month shorter or 5.1 longer 3-year overall survival after smoking cessation within 6 months of cancer diagnosis. Future studies with larger sample sizes that test the comparative effectiveness of different smoking cessation strategies are needed for more detailed evidence to inform decision-making about the effect of smoking cessation on survival among cancer patients.

Implications for Cancer survivors

The benefits of smoking cessation after cancer diagnosis may include longer survival, but the magnitude of benefit is unclear.

Regression models using parametric pseudo-observations

Pseudo-observations based on the nonparametric Kaplan-Meier estimator of the survival function have been proposed as an alternative to the widely used Cox model for analyzing censored time-to-event data. Using a spline-based estimator of the survival has some potential benefits over the nonparametric approach in terms of less variability. We propose to define pseudo-observations based on a flexible parametric estimator and use these for analysis in regression models to estimate parameters related to the cumulative risk. We report the results of a simulation study that compares the empirical standard errors of estimates based on parametric and nonparametric pseudo-observations in various settings. Our simulations show that in some situations there is a substantial gain in terms of reduced variability using the proposed parametric pseudo-observations compared with the nonparametric pseudo-observations. The gain can be measured as a reduction of the empirical standard error by up to about one third; corresponding to an additional 125% larger sample size. We illustrate the use of the proposed method in a brief data example.

Evaluation and Predictors of Fluid Resuscitation in Patients With Severe Sepsis and Septic Shock

OBJECTIVES

Rapid fluid resuscitation has become standard in sepsis care, despite "low-quality" evidence and absence of guidelines for populations "at risk" for volume overload. Our objectives include as follows: 1) identify predictors of reaching a 30 mL/kg crystalloid bolus within 3 hours of sepsis onset (30by3); 2) assess the impact of 30by3 and fluid dosing on clinical outcomes; 3) examine differences in perceived "at-risk" volume-sensitive populations, including end-stage renal disease, heart failure, obesity, advanced age, or with documentation of volume "overload" by bedside examination.

DESIGN

Retrospective cohort study. All outcome analyses controlled for sex, end-stage renal disease, heart failure, sepsis severity (severe sepsis vs septic shock), obesity, Mortality in Emergency Department Sepsis score, and time to antibiotics.

SETTING

Urban, tertiary care center between January 1, 2014, and May 31, 2017.

PATIENTS

Emergency Department treated adults (age ≥18 yr; n = 1,032) with severe sepsis or septic shock.

INTERVENTIONS

Administration of IV fluids by bolus.

MEASUREMENTS AND MAIN RESULTS

In total, 509 patients received 30by3 (49.3%). Overall mortality was 17.1% (n = 176), with 20.4% mortality in the shock group. Patients who were elderly (odds ratio, 0.62; 95% CI, 0.46-0.83), male (odds ratio, 0.66; CI, 0.49-0.87), obese (odds ratio, 0.18; CI, 0.13-0.25), or with end-stage renal disease (odds ratio, 0.23; CI, 0.13-0.40), heart failure (odds ratio, 0.42; CI, 0.29-0.60), or documented volume "overload" (odds ratio, 0.30; CI, 0.20-0.45) were less likely to achieve 30by3. Failure to meet 30by3 had increased odds of mortality (odds ratio, 1.52; CI, 1.03-2.24), delayed hypotension (odds ratio, 1.42; CI, 1.02-1.99), and increased ICU stay (~2 d) (β = 2.0; CI, 0.5-3.6), without differential effects for "at-risk" groups. Higher fluid volumes administered by 3 hours correlated with decreased mortality, with a plateau effect between 35 and 45 mL/kg (p < 0.05).

CONCLUSIONS

Failure to reach 30by3 was associated with increased odds of in-hospital mortality, irrespective of comorbidities. Predictors of inadequate resuscitation can be identified, potentially leading to interventions to improve survival. These findings are retrospective and require future validation.

Causal inference for recurrent event data using pseudo-observations

Recurrent event data are commonly encountered in observational studies where each subject may experience a particular event repeatedly over time. In this article, we aim to compare cumulative rate functions (CRFs) of two groups when treatment assignment may depend on the unbalanced distribution of confounders. Several estimators based on pseudo-observations are proposed to adjust for the confounding effects, namely inverse probability of treatment weighting estimator, regression model-based estimators, and doubly robust estimators. The proposed marginal regression estimator and doubly robust estimators based on pseudo-observations are shown to be consistent and asymptotically normal. A bootstrap approach is proposed for the variance estimation of the proposed estimators. Model diagnostic plots of residuals are presented to assess the goodness-of-fit for the proposed regression models. A family of adjusted two-sample pseudo-score tests is proposed to compare two CRFs. Simulation studies are conducted to assess finite sample performance of the proposed method. The proposed technique is demonstrated through an application to a hospital readmission data set.

On estimands and the analysis of adverse events in the presence of varying follow-up times within the benefit assessment of therapies

The analysis of adverse events (AEs) is a key component in the assessment of a drug's safety profile. Inappropriate analysis methods may result in misleading conclusions about a therapy's safety and consequently its benefit-risk ratio. The statistical analysis of AEs is complicated by the fact that the follow-up times can vary between the patients included in a clinical trial. This paper takes as its focus the analysis of AE data in the presence of varying follow-up times within the benefit assessment of therapeutic interventions. Instead of approaching this issue directly and solely from an analysis point of view, we first discuss what should be estimated in the context of safety data, leading to the concept of estimands. Although the current discussion on estimands is mainly related to efficacy evaluation, the concept is applicable to safety endpoints as well. Within the framework of estimands, we present statistical methods for analysing AEs with the focus being on the time to the occurrence of the first AE of a specific type. We give recommendations which estimators should be used for the estimands described. Furthermore, we state practical implications of the analysis of AEs in clinical trials and give an overview of examples across different indications. We also provide a review of current practices of health technology assessment (HTA) agencies with respect to the evaluation of safety data. Finally, we describe problems with meta-analyses of AE data and sketch possible solutions.

Time-to-event analysis for sports injury research part 1: time-varying exposures

BACKGROUND

'How much change in training load is too much before injury is sustained, among different athletes?' is a key question in sports medicine and sports science. To address this question the investigator/practitioner must analyse exposure variables that change over time, such as change in training load. Very few studies have included time-varying exposures (eg, training load) and time-varying effect-measure modifiers (eg, previous injury, biomechanics, sleep/stress) when studying sports injury aetiology.

AIM

To discuss advanced statistical methods suitable for the complex analysis of time-varying exposures such as changes in training load and injury-related outcomes.

CONTENT

Time-varying exposures and time-varying effect-measure modifiers can be used in time-to-event models to investigate sport injury aetiology. We address four key-questions (i) Does time-to-event modelling allow change in training load to be included as a time-varying exposure for sport injury development? (ii) Why is time-to-event analysis superior to other analytical concepts when analysing training-load related data that changes status over time? (iii) How can researchers include change in training load in a time-to-event analysis? and, (iv) Are researchers able to include other time-varying variables into time-to-event analyses? We emphasise that cleaning datasets, setting up the data, performing analyses with time-varying variables and interpreting the results is time-consuming, and requires dedication. It may need you to ask for assistance from methodological peers as the analytical approaches presented this paper require specialist knowledge and well-honed statistical skills.

CONCLUSION

To increase knowledge about the association between changes in training load and injury, we encourage sports injury researchers to collaborate with statisticians and/or methodological epidemiologists to carefully consider applying time-to-event models to prospective sports injury data. This will ensure appropriate interpretation of time-to-event data.

Getting the most out of intensive longitudinal data: a methodological review of workload-injury studies

OBJECTIVES

To systematically identify and qualitatively review the statistical approaches used in prospective cohort studies of team sports that reported intensive longitudinal data (ILD) (>20 observations per athlete) and examined the relationship between athletic workloads and injuries. Since longitudinal research can be improved by aligning the (1) theoretical model, (2) temporal design and (3) statistical approach, we reviewed the statistical approaches used in these studies to evaluate how closely they aligned these three components.

DESIGN

Methodological review.

METHODS

After finding 6 systematic reviews and 1 consensus statement in our systematic search, we extracted 34 original prospective cohort studies of team sports that reported ILD (>20 observations per athlete) and examined the relationship between athletic workloads and injuries. Using Professor Linda Collins' three-part framework of aligning the theoretical model, temporal design and statistical approach, we qualitatively assessed how well the statistical approaches aligned with the intensive longitudinal nature of the data, and with the underlying theoretical model. Finally, we discussed the implications of each statistical approach and provide recommendations for future research.

RESULTS

Statistical methods such as correlations, t-tests and simple linear/logistic regression were commonly used. However, these methods did not adequately address the (1) themes of theoretical models underlying workloads and injury, nor the (2) temporal design challenges (ILD). Although time-to-event analyses (eg, Cox proportional hazards and frailty models) and multilevel modelling are better-suited for ILD, these were used in fewer than a 10% of the studies (n=3).

CONCLUSIONS

Rapidly accelerating availability of ILD is the norm in many fields of healthcare delivery and thus health research. These data present an opportunity to better address research questions, especially when appropriate statistical analyses are chosen.