RSRussell J. Steele

Russell J. Steele

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Russell J. Steele has a total of 68 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • The semiparametric accelerated trend-renewal process for recurrent event data

    by Chien-Lin Su, Russell J. Steele, Ian Shrier

    Recurrent event data arise in many biomedical longitudinal studies when health-related events can occur repeatedly for each subject during the follow-up time. In this article, we examine the gap times between recurrent events. We propose a new semiparametric accelerated gap time model based on the trend-renewal process which contains trend and renewal components that allow for the intensity function to vary between successive events. We use the Buckley–James imputation approach to deal with censored transformed gap times. The proposed estimators are shown to be consistent and asymptotically normal. Model diagnostic plots of residuals and a method for predicting number of recurrent events given specified covariates and follow-up time are also presented. Simulation studies are conducted to assess finite sample performance of the proposed method. The proposed technique is demonstrated through an application to two real data sets.


  • Doubly robust estimation and causal inference for recurrent event data

    by Chien‐Lin Su, Russell Steele, Ian Shrier

    Many longitudinal databases record the occurrence of recurrent events over time. In this article, we propose a new method to estimate the average causal effect of a binary treatment for recurrent event data in the presence of confounders. We propose a doubly robust semiparametric estimator based on a weighted version of the Nelson‐Aalen estimator and a conditional regression estimator under an assumed semiparametric multiplicative rate model for recurrent event data. We show that the proposed doubly robust estimator is consistent and asymptotically normal. In addition, a model diagnostic plot of residuals is presented to assess the adequacy of our proposed semiparametric model. We then evaluate the finite sample behavior of the proposed estimators under a number of simulation scenarios. Finally, we illustrate the proposed methodology via a database of circus artist injuries.


    • 27 September 2019
    • Published by HAL in HAL
    Open Access

    Bayesian mixtures of multiple scale distributions

    by Alexis Arnaud, Florence Forbes, Russel Steele, Benjamin Lemasson, Emmanuel Barbier

    Multiple scale distributions are multivariate distributions that exhibit a variety of shapes not necessarily elliptical while remaining analytical and tractable. In this work we consider mixtures of such distributions for their ability to handle non standard typically non-gaussian clustering tasks. We propose a Bayesian formulation of the mixtures and a tractable inference procedure based on variational approximation. The interest of such a Bayesian formulation is illustrated on an important mixture model selection task, which is the issue of selecting automatically the number of components. We derive promising procedures that can be carried out in a single-run, in contrast to the more costly comparison of information criteria. Preliminary results on simulated and real data show promising performance in terms of clustering and computation time.

  • Open Access

    A population-based study of chronic hepatitis C in immigrants and non-immigrants in Quebec, Canada

    by Christina Greenaway, Laurent Azoulay, Robert Allard, Joseph Cox, Viet Anh Tran, Claire Nour Abou Chakra, Russ Steele, Marina Klein

    Immigrants originating from intermediate and high HCV prevalence countries may be at increased risk of exposure to hepatitis C infection (HCV) in their countries of origin, however they are not routinely screened after arrival in most low HCV prevalence host countries. We aimed to describe the epidemiology of HCV in immigrants compared to the Canadian born population. Using the reportable infectious disease database linked to the landed immigration database and several provincial administrative databases, we assembled a cohort of all reported cases of HCV in Quebec, Canada (1998–2008). Underlying co-morbidities were identified in the health services databases. Stratum specific rates of reported cases/100,000, rate ratios (RRs) and trends over the study period were estimated. A total of 20,862 patients with HCV were identified, among whom 1922 (9.2%) were immigrants. Immigrants were older and diagnosed a mean of 9.8 ± 7 years after arrival. The Canadian born population was more likely to have behavior co-morbidities (problematic alcohol or drug use) and HIV co-infection. Immigrants from Sub-Saharan Africa, Asia and Eastern Europe had the highest HCV reported rates with RRs compared to non-immigrants ranging from 1.5 to 1.7. The age and sex adjusted rates decreased by 4.9% per year in non-immigrants but remained unchanged in immigrants. The proportion of HCV occurring in immigrants doubled over the study period from 5 to 11%. Immigrants from intermediate and high HCV prevalence countries are at increased risk for HCV and had a mean delay in diagnosis of almost 10 years after arrival suggesting that they may benefit from targeted HCV screening and earlier linkage to care.


  • Open Access

    Simple Estimation of Patient-Oriented Effects From Randomized Trials: An Open and Shut CACE

    by Russell J. Steele, Ian Shrier, Jay S. Kaufman, Robert W. Platt

    In randomized controlled trials, the intention-to-treat estimator provides an unbiased estimate of the causal effect of treatment assignment on the outcome. However, patients often want to know what the effect would be if they were to take the treatment as prescribed (the patient-oriented effect), and several researchers have suggested that the more relevant causal effect for this question is the complier average causal effect (CACE), also referred to as the local average treatment effect. Sophisticated approaches to estimating the CACE include Bayesian and frequentist methods for principal stratification, inverse-probability-of-treatment–weighted estimators, and instrumental-variable (IV) analysis. All of these approaches exploit information about adherence to assigned treatment to improve upon the intention-to-treat estimator, but they are rarely used in practice, probably because of their complexity. The IV principal stratification estimator is simple to implement but has had limited use in practice, possibly due to lack of familiarity. Here, we show that the IV principal stratification estimator is a modified per-protocol estimator that should be obtainable from any randomized controlled trial, and we provide a closed form for its robust variance (and its uncertainty). Finally, we illustrate sensitivity analyses we conducted to assess inference in light of potential violations of the exclusion restriction assumption.


  • Open Access

    The diagnostic accuracy of the Patient Health Questionnaire-2 (PHQ-2), Patient Health Questionnaire-8 (PHQ-8), and Patient Health Questionnaire-9 (PHQ-9) for detecting major depression: protocol for a systematic review and individual patient data meta-analyses

    by Brett D Thombs, Andrea Benedetti, Lorie A Kloda, Brooke Levis, Ioana Nicolau, Pim Cuijpers, Simon Gilbody, John P A Ioannidis, Dean McMillan, Scott B Patten, et al.

    Major depressive disorder (MDD) may be present in 10%–20% of patients in medical settings. Routine depression screening is sometimes recommended to improve depression management. However, studies of the diagnostic accuracy of depression screening tools have typically used data-driven, exploratory methods to select optimal cutoffs. Often, these studies report results from a small range of cutoff points around whatever cutoff score is most accurate in that given study. When published data are combined in meta-analyses, estimates of accuracy for different cutoff points may be based on data from different studies, rather than data from all studies for each possible cutoff point. As a result, traditional meta-analyses may generate exaggerated estimates of accuracy. Individual patient data (IPD) meta-analyses can address this problem by synthesizing data from all studies for each cutoff score to obtain diagnostic accuracy estimates. The nine-item Patient Health Questionnaire-9 (PHQ-9) and the shorter PHQ-2 and PHQ-8 are commonly recommended for depression screening. Thus, the primary objectives of our IPD meta-analyses are to determine the diagnostic accuracy of the PHQ-9, PHQ-8, and PHQ-2 to detect MDD among adults across all potentially relevant cutoff scores. Secondary analyses involve assessing accuracy accounting for patient factors that may influence accuracy (age, sex, medical comorbidity). Data sources will include MEDLINE, MEDLINE In-Process & Other Non-Indexed Citations, PsycINFO, and Web of Science. We will include studies that included a Diagnostic and Statistical Manual or International Classification of Diseases diagnosis of MDD based on a validated structured or semi-structured clinical interview administered within 2 weeks of the administration of the PHQ. Two reviewers will independently screen titles and abstracts, perform full article review, and extract study data. Disagreements will be resolved by consensus. Risk of bias will be assessed with the Quality Assessment of Diagnostic Accuracy Studies-2 tool. Bivariate random-effects meta-analysis will be conducted for the full range of plausible cutoff values. The proposed IPD meta-analyses will allow us to obtain estimates of the diagnostic accuracy of the PHQ-9, PHQ-8, and PHQ-2. PROSPERO CRD42014010673


  • Open Access

    Projection scenarios of body mass index (2013–2030) for Public Health Planning in Quebec

    by Ernest Lo, Denis Hamel, Yun Jen, Patricia Lamontagne, Sylvie Martel, Colin Steensma, Chantal Blouin, Russell Steele

    Projection analyses can provide estimates of the future health burden of increasing BMI and represent a relevant and useful tool for public health planning. Our study presents long-term (2013–2030) projections of the prevalence and numbers of individuals by BMI category for adult men and women in Quebec. Three applications of projections to estimate outcomes more directly pertinent to public health planning, as well as an in-depth discussion of limits, are provided with the aim of encouraging greater use of projection analyses by public health officers. The weighted compositional regression method is applied to prevalence time series derived from sixteen cross-sectional survey cycles, for scenarios of linear change and deceleration. Estimation of the component of projected change potentially amenable to intervention, future health targets and the projected impact on type 2 diabetes, were done. Obesity prevalence in Quebec is projected to rise steadily from 2013 to 2030 in both men (from 18.0-19.4% to 22.2-30.4%) and women (from 15.5-16.3% to 18.2-22.4%). Corresponding projected numbers of obese individuals are (579,000-625,000 to 790,000-1,084,000) in men and (514,000-543,000 to 661,000-816,000) in women. These projected increases are found to be primarily an ‘epidemiologic’ rather than ‘demographic’ phenomenon and thus potentially amenable to public health intervention. Assessment of obesity targets for 2020 illustrates the necessity of using projected rather than current prevalence; for example a targeted 2% drop in obesity prevalence relative to 2013 translates into a 3.6-5.4% drop relative to 2020 projected levels. Type 2 diabetes is projected to increase from 6.9% to 9.2-10.1% in men and from 5.7% to 7.1-7.5% in women, from 2011–2012 to 2030. A substantial proportion of this change (25-44% for men, and 27-43% for women) is attributable to the changing BMI distribution. Obesity in Quebec is projected to increase and should therefore continue to be a public health priority. Application of projections to estimate the proportion of change potentially amenable to intervention, feasible health targets, and future chronic disease prevalence are demonstrated. Projection analyses have limitations, but represent a pertinent tool for public health planning.


  • Open Access

    Procedures for assessing psychological predictors of injuries in circus artists: a pilot prospective study

    by Ian Shrier, John S Raglin, Emily B Levitan, Murray A Mittleman, Russell J Steele, Janette Powell

    Research on psychological risk factors for injury has focused on stable traits. Our objective was to test the feasibility of a prospective longitudinal study designed to examine labile psychological states as risk factors of injury. We measured psychological traits at baseline (mood, ways of coping and anxiety), and psychological states every day (1-item questions on anxiety, sleep, fatigue, soreness, self-confidence) before performances in Cirque du Soleil artists of the show “O”. Additional questions were added once per week to better assess anxiety (20-item) and mood. Questionnaires were provided in English, French, Russian and Japanese. Injury and exposure data were extracted from electronic records that are kept as part of routine business practices. The 43.9% (36/82) recruitment rate was more than expected. Most artists completed the baseline questionnaires in 15 min, a weekly questionnaire in <2 min and a daily questionnaire in <1 min. We improved the formatting of some questions during the study, and adapted the wording of other questions to improve clarity. There were no dropouts during the entire study, suggesting the questionnaires were appropriate in content and length. Results for sample size calculations depend on the number of artists followed and the minimal important difference in injury rates, but in general, preclude a purely prospective study with daily data collection because of the long follow-up required. However, a prospective nested case-crossover design with data collection bi-weekly and at the time of injury appears feasible. A prospective study collecting psychological state data from subjects who train and work regularly together is feasible, but sample size calculations suggest that the optimal study design would use prospective nested case-crossover methodology.


  • Association Between Anesthesiologist Age and Litigation

    by Michael J. Tessler, Ian Shrier, Russell J. Steele


  • Open Access

    Clinical significance of antibodies to Ro52/TRIM21 in systemic sclerosis

    by Marie Hudson, Janet Pope, Michael Mahler, Solène Tatibouet, Russell Steele, Murray Baron, Canadian Scleroderma Research Group (CSRG), Marvin J Fritzler

    Introduction: Autoantibodies to Ro52 recently identified as TRIM21 are among the most common autoantibodies in systemic autoimmune rheumatic diseases, but their clinical association remains poorly understood. We undertook this study to determine the clinical and serologic associations of anti-Ro52/TRIM21 antibodies in patients with systemic sclerosis (SSc). Methods: Detailed clinical data and sera from 963 patients with SSc enrolled in a multicenter cohort study were collected and entered into a central database. Antibodies to Ro52/TRIM21 and other autoantibodies were detected with an addressable laser-bead immunoassay and different enzyme-linked immunosorbent assay (ELISA) systems. Associations between anti-Ro52/TRIM21 antibodies and clinical and other serologic manifestations of SSc were investigated. Results: Anti-Ro52/TRIM21 antibodies were present in 20% of SSc patients and overlapped with other main SSc-related antibodies, including anti-centromere (by immunofluorescence and centromere protein (CENP)-A and CENP-B ELISA), anti-topoisomerase I, anti-RNA polymerase III, and anti-Pm/Scl antibodies. Anti-Ro52/TRIM21 antibodies were strongly associated with interstitial lung disease (odds ratio (OR), 1.53; 95% confidence interval (CI), 1.11 to 2.12; P = 0.0091) and overlap syndrome (OR, 2.06; 95% CI, 1.01 to 4.19; P = 0.0059). Conclusions: Anti-Ro52/TRIM21 antibodies were the second most common autoantibodies in this SSc cohort. In SSc, anti-Ro52/TRIM21 antibodies may be a marker of interstitial lung disease and overlap syndrome.