Cox regression with linked data.

Record linkage is increasingly used, especially in medical studies, to combine data from different databases that refer to the same entities. The linked data can bring analysts novel and valuable knowledge that is impossible to obtain from a single database. However, linkage errors are usually unavoidable, regardless of record linkage methods, and ignoring these errors may lead to biased estimates. While different methods have been developed to deal with the linkage errors in the generalized linear model, there is not much interest on Cox regression model, although this is one of the most important statistical models in clinical and epidemiological research. In this work, we propose an adjusted estimating equation for secondary Cox regression analysis, where linked data have been prepared by a third-party operator, and no information on matching variables is available to the analyst. Through a Monte Carlo simulation study, the proposed method is shown to lead to substantial bias reductions in the estimation of the parameters of the Cox model caused by false links. An asymptotically unbiased variance estimator for the adjusted estimators of Cox regression coefficients is also proposed. Finally, the proposed method is applied to a linked database from the Brest stroke registry in France.

Variance estimators for weighted and stratified linear dose-response function estimators using generalized propensity score.

Propensity score methods are widely used in observational studies for evaluating marginal treatment effects. The generalized propensity score (GPS) is an extension of the propensity score framework, historically developed in the case of binary exposures, for use with quantitative or continuous exposures. In this paper, we proposed variance estimators for treatment effect estimators on continuous outcomes. Dose-response functions (DRFs) were estimated through weighting on the inverse of the GPS, or using stratification. Variance estimators were evaluated using Monte Carlo simulations. Despite the use of stabilized weights, the variability of the weighted estimator of the DRF was particularly high, and none of the variance estimators (a bootstrap-based estimator, a closed-form estimator especially developed to take into account the estimation step of the GPS, and a sandwich estimator) were able to adequately capture this variability, resulting in coverages below the nominal value, particularly when the proportion of the variation in the quantitative exposure explained by the covariates was large. The stratified estimator was more stable, and variance estimators (a bootstrap-based estimator, a pooled linearized estimator, and a pooled model-based estimator) more efficient at capturing the empirical variability of the parameters of the DRF. The pooled variance estimators tended to overestimate the variance, whereas the bootstrap estimator, which intrinsically takes into account the estimation step of the GPS, resulted in correct variance estimations and coverage rates. These methods were applied to a real data set with the aim of assessing the effect of maternal body mass index on newborn birth weight.

Allostatic load and subsequent all-cause mortality: which biological markers drive the relationship? Findings from a UK birth cohort.

The concept of allostatic load (AL) refers to the idea of a global physiological 'wear and tear' resulting from the adaptation to the environment through the stress response systems over the life span. The link between socioeconomic position (SEP) and mortality has now been established, and there is evidence that AL may capture the link between SEP and mortality. In order to quantitatively assess the role of AL on mortality, we use data from the 1958 British birth cohort including eleven year mortality in 8,113 adults. Specifically, we interrogate the hypothesis of a cumulative biological risk (allostatic load) reflecting 4 physiological systems potentially predicting future risk of death (N = 132). AL was defined using 14 biomarkers assayed in blood from a biosample collected at 44 years of age. Cox proportional hazard regression analysis revealed that higher allostatic load at 44 years old was a significant predictor of mortality 11 years later [HR = 3.56 (2.3 to 5.53)]. We found that this relationship was not solely related to early-life SEP, adverse childhood experiences and young adulthood health status, behaviours and SEP [HR = 2.57 (1.59 to 4.15)]. Regarding the ability of each physiological system and biomarkers to predict future death, our results suggest that the cumulative measure was advantageous compared to evaluating each physiological system sub-score and biomarker separately. Our findings add some evidence of a biological embodiment in response to stress which ultimately affects mortality.

An omnibus test for several hazard alternatives in prevention randomized controlled clinical trials.

The logrank test is optimal for testing the equality of survival distributions against a proportional hazards alternative. Under a late effects alternative, it is no longer appropriate, and one may turn to Fleming-Harrington's class of weighted logrank tests instead. In some settings, such as in preventive clinical trials where the statistical analysis has to be designed before the trial begins, it can be difficult to choose a priori between the logrank and Fleming-Harrington tests. A solution to this issue is provided. A decision rule is constructed for the problem of testing the equality of two survival distributions when the expected alternative may be one of the proportional hazards and late effects. A formula for computing the necessary sample size is obtained for this decision rule. A comprehensive simulation study is conducted to assess finite sample properties of the proposed test statistic. The proposed test improves both the logrank test and Fleming-Harrington's test for late effects. Finally, the methodology is illustrated on a data set in the field of prevention of Alzheimer's disease.

Influence of exposure assessment methods on associations between long-term exposure to outdoor fine particulate matter and risk of cancer in the French cohort Gazel.

BACKGROUND: Many studies investigated the relationship between outdoor fine particulate matter (PM2.5) and cancer. While they generally indicated positive associations, results have not been fully consistent, possibly because of the diversity of methods used to assess exposure. OBJECTIVES: To investigate how using different PM2.5 exposure assessment methods influences risk estimates in the large French general population-based Gazel cohort (20,625 participants at enrollment) with a 26-year follow-up with complete residential histories. METHODS: We focused on two cancer incidence outcomes: all-sites combined and lung. We used two distinct exposure assessment methods: a western European land use regression (LUR), and a chemistry-dispersion model (Gazel-Air) for France, each with a time series ≥20-years annual concentrations. Spearman correlation coefficient between the two estimates of PM2.5 was 0.71 across all person-years; the LUR tended to provide higher exposures. We used extended Cox models with attained age as time-scale and time-dependent cumulative exposures, adjusting for a set of confounders including sex and smoking, to derive hazard ratios (HRs) and their 95% confidence interval, implementing a 10-year lag between exposure and incidence/censoring. RESULTS: We obtained similar two-piece linear associations for all-sites cancer (3711 cases), with a first slope of HRs of 1.53 (1.24-1.88) and 1.43 (1.19-1.73) for one IQR increase of cumulative PM2.5 exposure for the LUR and the Gazel-Air models respectively, followed by a plateau at around 1.5 for both exposure assessments. For lung cancer (349 cases), the HRs from the two exposure models were less similar, with largely overlapping confidence limits. CONCLUSION: Our findings using long-term exposure estimates from two distinct exposure assessment methods corroborate the association between air pollution and cancer risk.

Contribution of Long-Term Exposure to Outdoor Black Carbon to the Carcinogenicity of Air Pollution: Evidence regarding Risk of Cancer in the Gazel Cohort.

BACKGROUND: Black carbon (BC), a component of fine particulate matter [particles with an aerodynamic diameter ≤2.5 µm (PM2.5)], may contribute to carcinogenic effects of air pollution. Until recently however, there has been little evidence to evaluate this hypothesis. OBJECTIVE: This study aimed to estimate the associations between long-term exposure to BC and risk of cancer. This study was conducted within the French Gazel cohort of 20,625 subjects. METHODS: We assessed exposure to BC by linking subjects' histories of residential addresses to a map of European black carbon levels in 2010 with back- and forward-extrapolation between 1989 and 2015. We used extended Cox models, with attained age as time-scale and time-varying cumulative exposure to BC, adjusted for relevant sociodemographic and lifestyle variables. To consider latency between exposure and cancer diagnosis, we implemented a 10-y lag, and as a sensitivity analysis, a lag of 2 y. To isolate the effect of BC from that of total PM2.5, we regressed BC on PM2.5 and used the residuals as the exposure variable. RESULTS: During the 26-y follow-up period, there were 3,711 incident cancer cases (all sites combined) and 349 incident lung cancers. Median baseline exposure in 1989 was 2.65 10-5/m [interquartile range (IQR): 2.23-3.33], which generally slightly decreased over time. Using 10 y as a lag-time in our models, the adjusted hazard ratio per each IQR increase of the natural log-transformed cumulative BC was 1.17 (95% confidence interval: 1.06, 1.29) for all-sites cancer combined and 1.31 (0.93, 1.83) for lung cancer. Associations with BC residuals were also positive for both outcomes. Using 2 y as a lag-time, the results were similar. DISCUSSION: Our findings for a cohort of French adults suggest that BC may partly explain the association between PM2.5 and lung cancer. Additional studies are needed to confirm our results and further disentangle the effects of BC, total PM2.5, and other constituents. https://doi.org/10.1289/EHP8719.

On the Use of Optimal Transportation Theory to Recode Variables and Application to Database Merging.

Merging databases is a strategy of paramount interest especially in medical research. A common problem in this context comes from a variable which is not coded on the same scale in both databases we aim to merge. This paper considers the problem of finding a relevant way to recode the variable in order to merge these two databases. To address this issue, an algorithm, based on optimal transportation theory, is proposed. Optimal transportation theory gives us an application to map the measure associated with the variable in database A to the measure associated with the same variable in database B. To do so, a cost function has to be introduced and an allocation rule has to be defined. Such a function and such a rule is proposed involving the information contained in the covariates. In this paper, the method is compared to multiple imputation by chained equations and a statistical learning method and has demonstrated a better average accuracy in many situations. Applications on both simulated and real datasets show that the efficiency of the proposed merging algorithm depends on how the covariates are linked with the variable of interest.

Suicidal Ideation and Traumatic Exposure Should Not Be Neglected in Epileptic Patients: A Multidimensional Comparison of the Psychiatric Profile of Patients Suffering From Epilepsy and Patients Suffering From Psychogenic Nonepileptic Seizures.

Introduction: Patients with psychogenic nonepileptic seizures (PNESs) have often been exposed to traumatic events, which is a risk factor for suicidal behavior. This would suggest that the severity of suicidal ideation is greater in PNES than in patients suffering only from epileptic seizures (ESs). However, these psychiatric symptoms may be underestimated in the ES population. The specific features or similarities between the psychiatric clinical profiles of these two groups should be elaborated to improve therapeutic management. Our study is the first to compare suicidal ideation, suicide risk, posttraumatic stress disorder (PTSD), and depression disorder simultaneously in both groups, in a tertiary care epilepsy center. Material and methods: We prospectively enrolled patients hospitalized for video-electroencephalography (EEG) monitoring to assess repeated seizures before an ES or a PNES diagnosis was made. During the psychiatric consultation that accompanied the video EEG, we rated the severity of suicidal ideation and depressive symptoms, suicidal risk, traumatic exposure history, and PTSD symptoms. Results: Eighteen subjects were enrolled and diagnosed with PNES, and 42, with ES. The PNES group reported more exposures to traumatic events and more intense PTSD symptoms (median: 17 vs. 27; p = 0.001). The severity of suicidal ideation did not differ significantly between the two groups. Conclusion: It is the severity of PTSD symptoms in PNES patients that differentiates them from ES patients, although exposure to traumatic events is also frequent in ES patients. We demonstrated that suicidal ideation and suicide risk are equally high in the ES and PNES groups. Therefore, both groups require extreme vigilance in terms of suicidal risk.

Data maturity and follow-up in time-to-event analyses.

We propose methods to determine the minimum number of subjects remaining at risk after which Kaplan-Meier survival plots for time-to-event outcomes should be curtailed, as, once the number remaining at risk drops below this minimum, the survival estimates are no longer meaningful in the context of the investigation. The size of the decrease of the Kaplan-Meier survival estimate S(t) at time t if one extra event should occur is considered in two ways. In the first approach, the investigator sets a maximum acceptable absolute decrease in S(t) should one extra event occur. In the second, a minimum acceptable number of subjects still at risk is calculated by comparing the size of the decrease in S(t) if an extra event should occur with the variability of the survival estimate had all subjects been followed to that time (confidence interval approach). We recommend calculating both limits for the number still at risk and then making an informed choice in the context of the particular investigation. We explore further how the amount of information actually available can assist in considering issues of data maturity for studies whose outcome of interest is a survival percentage at a particular time point. We illustrate the approaches with a number of published studies having differing sample sizes and censoring issues. In particular, one study was the subject of some controversy regarding how far in time the Kaplan-Meier plot should be extended. The proposed methods allow for limits to be calculated simply using the output provided by most statistical packages.

Homologous Recombination DNA Repair Pathway Disruption and Retinoblastoma Protein Loss Are Associated with Exceptional Survival in High-Grade Serous Ovarian Cancer.

Purpose: Women with epithelial ovarian cancer generally have a poor prognosis; however, a subset of patients has an unexpected dramatic and durable response to treatment. We sought to identify clinical, pathological, and molecular determinants of exceptional survival in women with high-grade serous cancer (HGSC), a disease associated with the majority of ovarian cancer deaths.Experimental Design: We evaluated the histories of 2,283 ovarian cancer patients and, after applying stringent clinical and pathological selection criteria, identified 96 with HGSC that represented significant outliers in terms of treatment response and overall survival. Patient samples were characterized immunohistochemically and by genome sequencing.Results: Different patterns of clinical response were seen: long progression-free survival (Long-PFS), multiple objective responses to chemotherapy (Multiple Responder), and/or greater than 10-year overall survival (Long-Term Survivors). Pathogenic germline and somatic mutations in genes involved in homologous recombination (HR) repair were enriched in all three groups relative to a population-based series. However, 29% of 10-year survivors lacked an identifiable HR pathway alteration, and tumors from these patients had increased Ki-67 staining. CD8+ tumor-infiltrating lymphocytes were more commonly present in Long-Term Survivors. RB1 loss was associated with long progression-free and overall survival. HR deficiency and RB1 loss were correlated, and co-occurrence was significantly associated with prolonged survival.Conclusions: There was diversity in the clinical trajectory of exceptional survivors associated with multiple molecular determinants of exceptional outcome in HGSC patients. Concurrent HR deficiency and RB1 loss were associated with favorable outcomes, suggesting that co-occurrence of specific mutations might mediate durable responses in such patients. Clin Cancer Res; 24(3); 569-80. ©2017 AACRSee related commentary by Peng and Mills, p. 508.

A Longitudinal Study of Transitions Between Informal and Formal Care in Alzheimer Disease Using Multistate Models in the European ICTUS Cohort.

BACKGROUND: We aimed to describe longitudinal patterns of care in community-dwelling European patients with Alzheimer disease (AD), and determine patient-, caregiver-, and country-related predictors of transitions across different care levels. METHODS: Two-year follow-up data from ICTUS cohort (1375 patients with AD, 12 countries) were analyzed using multistate Markov models to describe transitions across states of care and identify their predictors. RESULTS: Of the patients, 61.3% stayed in the same state during follow-up, and only 9.5% experienced ≥2 changes between states. Six-month transition probabilities were 11% for informal to formal care and 13% for formal to informal care (in the community). Older age, male gender, poorer cognitive and behavioral scores, and country of residence were associated with transitioning from informal to formal care, but only country of residence was associated with the reverse transition. DISCUSSION: Changes between different types of care were rare during follow-up, and country factors in particular influenced these transitions.

Living Alone with Alzheimer's Disease and the Risk of Adverse Outcomes: Results from the Plan de Soin et d'Aide dans la maladie d'Alzheimer Study.

OBJECTIVES: To examine whether living alone predicted hospitalizations, nursing home admission, weight loss, and death in individuals with mild to moderate Alzheimer's disease (AD) over 2 years of follow-up. DESIGN: Data are from the Plan de Soin et d'Aide dans la maladie d'Alzheimer study, a 24-month trial with a cluster randomization of memory clinics in two arms: a multidomain intervention and usual care. SETTING: Memory clinics in France (N = 50). PARTICIPANTS: Community-dwelling individuals with AD with a Mini-Mental State Examination score between 12 and 26 and an identified caregiver. MEASUREMENTS: A neurogeriatric evaluation was conducted twice a year in the intervention group and annually in the control group. Hospitalizations, nursing home admission, weight loss, and death occurring during the past year were recorded. Information on sociodemographic characteristics, clinical conditions, therapy, and physical and cognitive status was recorded. RESULTS: At inclusion, 348 (30.8%) of the 1,131 participants lived alone. Living alone did not increase the risk of mortality or weight loss in individuals with mild to moderate AD, but significant associations with risk of hospitalization (hazard ratio (HR) = 1.33, 95% confidence interval (CI) = 1.01-1.74) and institutionalization (HR = 2.53, 95% CI = 1.84-3.47) were reported. A protective effect of physical function on institutionalization and mortality was found. CONCLUSION: These results might support clinicians in making decisions about institutionalization of individuals with AD living alone or improving home health care, such as increasing screening and managing functional impairment in this complex population.