Lawrence Joseph

McGill University Health Centre, Montréal, Quebec, Canada

Are you Lawrence Joseph?

Claim your profile

Publications (366)1736.7 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Despite the increased risk of cardiovascular disease and type 2 diabetes associated with excess bodyweight, development of a clinically meaningful metric for health professionals remains a challenge. We estimated the years of life lost and the life-years lost from diabetes and cardiovascular disease associated with excess bodyweight. We developed a disease-simulation model to estimate the annual risk of diabetes, cardiovascular disease, and mortality for people with BMI of 25-<30 kg/m(2) (overweight), 30-<35 kg/m(2) (obese), or 35 kg/m(2) and higher (very obese), compared with an ideal BMI of 18·5-<25 kg/m(2). We used data from 3992 non-Hispanic white participants in the National Nutrition and Examination Survey (2003-10) for whom complete risk factor data and fasting glucose concentrations were available. After validation of the model projections, we estimated the years of life lost and healthy life-years lost associated with each bodyweight category. Excess bodyweight was positively associated with risk factors for cardiovascular disease and type 2 diabetes. The effect of excess weight on years of life lost was greatest for young individuals and decreased with increasing age. The years of life lost for obese men ranged from 0·8 years (95% CI 0·2-1·4) in those aged 60-79 years to 5·9 years (4·4-7·4) in those aged 20-39 years, and years lost for very obese men ranged from 0·9 (0-1·8) years in those aged 60-79 years to 8·4 (7·0-9·8) years in those aged 20-39 years, but losses were smaller and sometimes negligible for men who were only overweight. Similar results were noted for women (eg, 6·1 years [4·6-7·6] lost for very obese women aged 20-39 years; 0·9 years [0·1-1·7] lost for very obese women aged 60-79 years). Healthy life-years lost were two to four times higher than total years of life lost for all age groups and bodyweight categories. Our estimations for both healthy life-years and total years of life lost show the effect of excess bodyweight on cardiovascular disease and diabetes, and might provide a useful health measure for discussions between health professionals and their patients. Canadian Institutes of Health Research. Copyright © 2014 Elsevier Ltd. All rights reserved.
    The lancet. Diabetes & endocrinology. 12/2014;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Commuters may be exposed to increased levels of traffic-related air pollution owing to close proximity to traffic-emissions. We collected in-vehicle and roof-top air pollution measurements over 238 commutes in Montreal, Toronto, and Vancouver, Canada between 2010 and 2013. Voice recordings were used to collect real-time information on traffic density and the presence of diesel vehicles and multivariable linear regression models were used to estimate the impact of these factors on in-vehicle pollutant concentrations (and indoor/outdoor ratios) along with parameters for road type, land use, and meteorology. In-vehicle PM2.5 and NO2 concentrations consistently exceeded regional outdoor levels and each unit increase in the rate of encountering diesel vehicles (count/minute) was associated with substantial increases (>100%) in in-vehicle concentrations of ultrafine particles (UFPs), black carbon, and PM2.5 as well as strong increases (>15%) in indoor/outdoor ratios. A model based on meteorology and the length of highway roads within a 500 meter buffer explained 53% of the variation in in-vehicle UFPs; however, models for PM2.5 (R2=0.24) and black carbon (R2=0.30) did not perform as well. Our findings suggest that vehicle commuters experience increased exposure to air pollutants and that traffic characteristics, land use, road types, and meteorology are important determinants of these exposures.
    Environmental science & technology. 12/2014;
  • [Show abstract] [Hide abstract]
    ABSTRACT: To estimate the population-based prevalence and healthcare use for osteoarthritis (OA) by First Nations (FN) and non-First Nations (non-FN) in Alberta, Canada. A cohort of adults with OA (≥ 2 physician claims in 2 yrs or 1 hospitalization with ICD-9-Clinical Modification code 715x or ICD-10-Canadian Adaptation code M15-19, 1993-2010) was defined with FN determination by premium payer status. Prevalence rates (2007/8) were estimated from the cohort and the population registered with the Alberta Health Care Insurance Plan. Rates of outpatient primary care and specialist visits (orthopedics, rheumatology, internal medicine), arthroplasty (hip and knee), and all-cause hospitalization were estimated. OA prevalence in FN was twice that of the non-FN population [16.1 vs 7.8 cases/100 population, standardized rate ratio (SRR) adjusted for age and sex 2.06, 95% CI 2.00-2.12]. The SRR (adjusted for age, sex, and location of residence) for primary care visits for OA was nearly double in FN compared with non-FN (SRR 1.88, 95% CI 1.87-1.89), and internal medicine visits were increased (SRR 1.25, 95% CI 1.25-1.26). Visit rates with an orthopedic surgeon (SRR 0.49, 95% CI 0.48-0.50) or rheumatologist (SRR 0.62, 95% CI 0.62-0.63) were substantially lower in FN with OA. Hip and knee arthroplasties were performed less frequently in FN with OA (SRR 0.48, 95% CI 0.47-0.49), but all-cause hospitalization rates were higher (SRR 1.59, 95% CI 1.58-1.60). We estimate a 2-fold higher prevalence of OA in the FN population with differential healthcare use. Reasons for higher use of primary care and lower use of specialty services and arthroplasty compared with the general population are not yet understood.
    The Journal of rheumatology. 11/2014;
  • [Show abstract] [Hide abstract]
    ABSTRACT: To estimate and compare sex-specific screening polypectomy rates to quality benchmarks of 40% in men and 30% in women. A prospective cohort study was undertaken of patients aged 50-75, scheduled for colonoscopy, and covered by the Québec universal health insurance plan. Endoscopist and patient questionnaires were used to obtain screening and non-screening colonoscopy indications. Patient self-report was used to obtain history of gastrointestinal conditions/symptoms and prior colonoscopy. Sex-specific polypectomy rates (PRs) and 95%CI were calculated using Bayesian hierarchical logistic regression. In total, 45 endoscopists and 2134 (mean age = 61, 50% female) of their patients participated. According to patients, screening PRs in males and females were 32.4% (95%CI: 23.8-41.8) and 19.4% (95%CI: 13.1-25.4), respectively. According to endoscopists, screening PRs in males and females were 30.2% (95%CI: 27.0-41.9) and 16.6% (95%CI: 16.3-28.6), respectively. Sex-specific PRs did not meet quality benchmarks at all ages except for: males aged 65-69 (patient screening indication), and males aged 70-74 (endoscopist screening indication). For all patients aged 50-54, none of the CI included the quality benchmarks. Most sex-specific screening PRs in Québec were below quality benchmarks; PRs were especially low for all 50-54 year olds.
    World journal of gastroenterology : WJG. 11/2014; 20(43):16300-5.
  • [Show abstract] [Hide abstract]
    ABSTRACT: We conducted a systematic review to examine the efficacy of the Atkins, South Beach, Weight Watchers (WW), and Zone diets, with a particular focus on sustained weight loss at ≥12 months.
    Circulation Cardiovascular Quality and Outcomes 11/2014; · 5.66 Impact Factor
  • Source
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The role that animals play in the transmission of Schistosoma japonicum to humans in the Philippines remains uncertain and prior studies have not included several species, adjustment for misclassification error and clustering, or used a cohort design. A cohort study of 2468 people providing stool samples at 12 months following praziquantel treatment in 50 villages of Western Samar, the Philippines, was conducted. Stool samples from dogs, cats, rats, and water buffaloes were collected at baseline (2003–2004) and follow-up (2005). Latent-class hierarchical Bayesian log-binomial models adjusting for misclassification errors in diagnostic tests were used. The village-level baseline and follow-up prevalences of cat, dog, and rat S. japonicum infection were associated with the 12-month cumulative incidence of human S. japonicum infection, with similar magnitude and precision of effect, but correlation between infection levels made it difficult to divide their respective effects. The cumulative incidence ratios associated with a 1% increase in the prevalence of infection in dogs at baseline and in rats at follow-up were 1·04 [95% Bayesian credible interval (BCI) 1·02–1·07] and 1·02 (95% BCI 1·01–1·04), respectively, when both species were entered in the model. Dogs appear to play a role in human schistosomiasis infection while rats could be used as schistosomiasis sentinels.
    Epidemiology and Infection 10/2014; · 2.87 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: To estimate systemic autoimmune rheumatic disease (SARD) prevalence using administrative data for pediatric populations in four Canadian provinces. Physician billing claims and inpatient hospitalizations from Alberta, Manitoba, Quebec, and Saskatchewan were used to define cases aged ≤18 years with a SARD diagnosis code in: one or more hospitalization, two or more physician visits within 2 years and at least 2 months apart, or one or more physician visit to a rheumatologist. Estimates ranged from 15.9/100,000 in Quebec [95 % confidence interval (95 % CI) 14.1, 18.0] to 23.0/100,000 in Manitoba (95 % CI 17.9, 29.2). SARDs were more common in females than in males across all provinces. There was a slightly higher prevalence among those living in urban compared to rural areas of Alberta (rate difference 14.4, 95 % CI 8.6, 20.1) and Saskatchewan (rate difference 13.8, 95 % CI 1.0, 26.6). Our results provide population-based prevalence estimates of pediatric SARDs in four Canadian provinces.
    Rheumatology international. 09/2014;
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background Studies suggest that siblings of children with peanut allergy (PNA) have a higher prevalence of PNA than the general population.Objectives The Canadian Peanut Allergy Registry was used to assess the percentage of siblings of registered index PNA children who were 1) never exposed to peanut or 2) reportedly diagnosed with PNA by a physician without either a history of allergic reaction or confirmatory testing. Sociodemographic and clinical factors that may be associated with either outcome were evaluated.Methods Parents completed a questionnaire on siblings’ sociodemographic characteristics, exposure and reaction to peanut, confirmatory tests performed and whether PNA had been diagnosed.ResultsOf 932 Registry families, 748 families responded, representing 922 siblings. 13.6% of siblings had never been exposed to peanut, 70.4% (n=88) of which were born after the index child. Almost 9% of siblings (80) were reported as PNA, but almost half of this group had no history of an allergic reaction to peanut, including 5 children who also had no testing to confirm PNA. Of these 5, 4 were born after PNA diagnosis in the index child. In a multivariate regression analysis for siblings at least 3 years old, those born after PNA diagnosis in the index child were more likely to have never been exposed to peanut. In a univariate analysis, siblings born after the diagnosis of PNA in the index child were more likely to be diagnosed with PNA without supportive history or confirmatory testing.Conclusions and clinical relevanceThese data estimate that more than 10% of siblings of PNA patients will avoid peanut and that siblings born after the diagnosis of PNA in an index child are more likely to have never been exposed. Educational programs and guidelines that caution against unnecessary avoidance are required.This article is protected by copyright. All rights reserved.
    Clinical & Experimental Allergy 09/2014; · 4.79 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background: The Cross-Canada Anaphylaxis Registry (C-CARE) assesses the triggers and management of anaphylaxis and identifies predictors of the development of severe allergic reactions and of epinephrine use. Here, we present data from an urban adult tertiary care emergency department (ED) in Montreal, Canada. Methods: Potential anaphylaxis cases were identified using ICD-10 codes related to anaphylaxis or allergic reactions. Putative cases underwent chart review to ensure they met anaphylaxis diagnostic criteria. Demographic, clinical and management data were collected. Multivariate logistic regressions were conducted to assess the effect of demographic characteristics, triggers, and comorbidities on severity and management of reactions. Results: Among 37,730 ED visits, 0.26% (95% CI 0.21, 0.32) fulfilled the definition of anaphylaxis. Food was the suspected trigger in almost 60% of cases. Epinephrine was not administered in almost half of moderate-to-severe cases, and similar numbers of individuals with moderate-to-severe reactions were not prescribed an epinephrine autoinjector. Reaction to shellfish was associated with more severe reactions (OR 13.9; 95% CI 2.2, 89.4). Older individuals and those not receiving steroids were more likely managed without epinephrine (OR 1.04; 95% CI 1.01, 1.07 and OR 2.97; 95% CI 1.05, 8.39, respectively). Conclusions: Anaphylaxis accounted for a substantial number of ED visits in adults, and the most common trigger was food. There is non-adherence to guidelines recommending epinephrine use for all cases of anaphylaxis. We postulate that this may be related to concerns regarding the side effects of epinephrine in adults. © 2014 S. Karger AG, Basel.
    International archives of allergy and immunology. 08/2014; 164(3):246-252.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background Studies suggest that individuals of low education and/or income, new Canadians (immigrated <10 years ago), and individuals of Aboriginal identity may have fewer food allergies than the general population. However, given the difficulty in recruiting such populations (hereafter referred to as vulnerable populations), by using conventional survey methodologies, the prevalence of food allergy among these populations in Canada has not been estimated. Objectives To estimate the prevalence of food allergy among vulnerable populations in Canada, to compare with the nonvulnerable populations and to identify demographic characteristics predictive of food allergy. Methods By using 2006 Canadian Census data, postal codes with high proportions of vulnerable populations were identified and households were randomly selected to participate in a telephone survey. Information on food allergies and demographics was collected. Prevalence estimates were weighted by using Census data to account for the targeted sampling. Multivariable logistic regression was used to identify predictors of food allergy. Results Of 12,762 eligible households contacted, 5734 households completed the questionnaire (45% response rate). Food allergy was less common among adults without postsecondary education versus those with postsecondary education (6.4% [95% CI, 5.5%-7.3%] vs 8.9% [95% CI, 7.7%-10%]) and new Canadians versus those born in Canada (3.2% [95% CI, 2.2%-4.3%] vs 8.2% [95% CI, 7.4%-9.1%]). There was no difference in prevalence between those of low and of high income or those with and without Aboriginal identity. Conclusion Analysis of our data suggests that individuals of low education and new Canadians self-report fewer allergies, which may be due to genetics, environment, lack of appropriate health care, or lack of awareness of allergies, which reduces self-report.
    The Journal of Allergy and Clinical Immunology: In Practice. 08/2014;
  • [Show abstract] [Hide abstract]
    ABSTRACT: There is a paucity of published population-based estimates of the prevalence of chronic inflammatory arthritis in the pediatric population. We used administrative health data to estimate the prevalence of chronic inflammatory arthritis in individuals ≤18 years in three Canadian provinces: Quebec, Manitoba, and Saskatchewan. Cases aged ≤18 years were identified by meeting any one of the following criteria: (a) ≥1 hospital discharge abstract with an ICD-9 code of 714 or ICD-10-CA codes of M05, M06 or M08, or (b) ≥2 ICD-9 714 billing codes ≥8 weeks apart, but within 2 years, or (c) ≥1 ICD-9 714 billing code by a rheumatologist. Crude prevalence estimates per 10,000 population were estimated with 95 % confidence intervals (CIs). Prevalence estimates were 11.7 per 10,000 individuals ≤18 years of age in Manitoba, 9.8 per 10,000 in Saskatchewan, and 8.0 per 10,000 in Quebec. In pairwise comparisons of rate differences, Manitoba and Saskatchewan had higher estimates than Quebec. Prevalence estimates were higher for females than males, with a difference of 5.9 cases per 10,000 residents (95 % CI 5.1, 6.7). Saskatchewan was the only province with a higher estimate in urban compared to rural residents (5.2, 95 % CI 2.5, 8.0). Variations in provincial estimates may be due to differences in underlying population characteristics. Although these estimates have face validity and are in keeping with the range of previously published pediatric prevalence estimates, studies to establish the empiric validity of case-finding algorithms are needed to advance research in pediatric chronic disease epidemiology.
    Rheumatology International 07/2014; · 2.21 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: There is evidence that land use mix based on the Shannon (1948) entropy formula may be misspecified in some studies. The aim of this study was to quantify the bias arising from this misspecification. Spatial coordinates were obtained from Statistics Canada for 9348 unique point locations. Five hundred-metre polygon-based network buffers were drawn around each coordinate (ArcGIS 10.1). Land use mix was calculated for each buffer using the true and misspecified land use mix formulas. Linear regression models were used to estimate the associations between a simulated dataset of daily steps and the true and misspecified measures. Misspecification of the land use mix formula resulted in a systematic underestimation of the true association by 26.4% (95% CI 25.8-27.0%). To minimize measurement bias in future studies, researchers are encouraged to use a constant definition of N in the denominator of the Shannon entropy formula.
    Health & Place 07/2014; 29C:79-83. · 2.42 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Women with gestational diabetes history are at increased risk for type 2 diabetes. They face specific challenges for behavioural changes, including childcare responsibilities. The aim of this study is to test a tailored type 2 diabetes prevention intervention in women within 5 years of a pregnancy with gestational diabetes, in terms of effects on weight and cardiometabolic risk factors.
    Cardiovascular Diabetology 06/2014; 13(1):104. · 4.21 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To prospectively assess changes in health-related quality of life (HRQOL) over 10 years, by age and sex, and to compare measured within-person change to estimates of change based on cross-sectional data.
    Quality of life research : an international journal of quality of life aspects of treatment, care and rehabilitation. 06/2014;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The Diabetes Prevention Program (DPP) is highly effective in promoting weight loss in overweight and obese individuals. However, one-on-one DPP sessions are costly. As a cost-saving alternative, a group version of the DPP, called Group Lifestyle Balance program (GLB), has been developed but has been shown to be less effective. The aim of this two-arm parallel randomized controlled trial is to increase the effectiveness of the GLB by integrating habit formation techniques, namely if-then plans and their mental practice, into the program.Methods/design: A total of 154 participants will be randomized to a standard or enriched GLB program. For the enriched GLB program, if-then plans and their mental practice will be integrated into the standard GLB program. Participants will be overweight or obese men and women (BMI of 28 to 45 kg/m2, waist circumference >= 88 for women, >= 102 for men, 18 to 75 years of age) who do less than 200 minutes of self-reported moderate or vigorous exercise per week. Measures will be completed at baseline, 3 months, post-intervention (12 months), and 12 months post-intervention (24 months). The primary outcome measure is weight loss at 3, 12, and 24 months. Secondary outcomes include percent reaching weight loss goal, physical activity at 3, 12, and 24 months, and weight-related risk factors (waist circumference, hemoglobin A1c, systolic/diastolic blood pressure, total cholesterol/HDL ratio). Standardized training of the life-style coaches, use of standardized manuals, and audio taping and reviewing of the sessions will ensure intervention fidelity.
    BMC Public Health 05/2014; 14(1):470. · 2.08 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Increasing evidence suggests that ultrafine particles (UFPs) may contribute to cardiorespiratory morbidity. We examined the relationship between near road UFPs and several traffic and built environment factors to identify predictors that may be used to estimate exposures in population-based studies. Black carbon (BC) was also examined. Methods Data were collected on up to 6 occasions at 73 sites in Montreal, Canada over 6-week period during summer, 2012. After excluding highly correlated variables, road width, truck ratio (trucks/total traffic), building height, land zoning parameters, and meteorological factors were evaluated. Random-effect models were used to estimate percent changes in UFP and BC concentrations with interquartile changes in each candidate predictor adjusted for meteorological factors. Results Mean pollutant concentrations varied substantially across sites (UFP range: 1977–94, 798 particles/cm3; BC range: 29–9460 ng/m3). After adjusting for meteorology, interquartile increases in road width (14%, 95% CI: 0, 30), building height (13%, 95% CI: 5, 22), and truck ratio (13%, 95% CI: 3, 23) were the most important predictors of mean UFP concentrations. Road width (28%, 95% CI: 9, 51) and industrial zoning (18%, 95% CI: 2, 37) were the strongest predictors of maximum UFP concentrations. Industrial zoning (35%, 95% CI: 9, 67) was the strongest predictor of BC. Conclusions A number of traffic and built environmental factors were identified as important predictors of near road UFP and BC concentrations. Exposure models incorporating these factors may be useful in evaluating the health effects of traffic related air pollution.
    Environmental Research 05/2014; 132:305-310. · 3.24 Impact Factor
  • International journal of cardiology 03/2014; · 6.18 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Current guidelines recommend smoking cessation and weight management for secondary prevention in post-myocardial infarction patients. However, little is known about the effects of smoking cessation on weight change post-myocardial infarction. We examined patterns of weight change and its effects on blood pressure and glycemic control using data from a randomized trial investigating the effect of bupropion on smoking cessation in post-myocardial infarction patients. Weight change was compared among 3 groups of patients: those who were completely abstinent (n=92), those who smoked intermittently (n=49), and those who smoked persistently (n=38) during the 12-month follow-up. Analyses were restricted to patients who attended all follow-up visits. The median weight at baseline was 77.1 kg (interquartile range [IQR] 66.0, 87.5), and 64.3% of patients were overweight/obese (Body Mass Index ≥ 25.0 kg/m(2)). Median weight gain at 12 months was 4.0 kg (IQR 0, 7.0), with more than a third gaining > 5 kg. The proportion of patients that were overweight/obese increased by ∼ 10%, and 23.2% of patients moved up a body mass index category. Abstainers gained a median of 4.8 kg (IQR 1.0, 8.6), intermittent smokers 2.0 kg (IQR -2.0, 5.0), and persistent smokers 3.0 kg (IQR -0.8, 6.0). Weight gain was associated with an increase in blood pressure and requirements for hypoglycemic medications at 12 months. The majority of patients attempting to quit smoking gain weight 12 months post-myocardial infarction, with abstainers gaining more weight than those who return to smoking. Weight gain was associated with an increased prevalence of hypertension and diabetes.
    The American journal of medicine 03/2014; · 5.30 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To estimate systemic autoimmune rheumatic disease (SARD) prevalence across 7 Canadian provinces using population-based administrative data evaluating both regional variations and the effects of age and sex. Using provincial physician billing and hospitalization data, cases of SARD (systemic lupus erythematosus, scleroderma, primary Sjögren syndrome, polymyositis/dermatomyositis) were ascertained. Three case definitions (rheumatology billing, 2-code physician billing, and hospital diagnosis) were combined to derive a SARD prevalence estimate for each province, categorized by age, sex, and rural/urban status. A hierarchical Bayesian latent class regression model was fit to account for the imperfect sensitivity and specificity of each case definition. The model also provided sensitivity estimates of different case definition approaches. Prevalence estimates for overall SARD ranged between 2 and 5 cases per 1000 residents across provinces. Similar demographic trends were evident across provinces, with greater prevalence in women and in persons over 45 years old. SARD prevalence in women over 45 was close to 1%. Overall sensitivity was poor, but estimates for each of the 3 case definitions improved within older populations and were slightly higher for men compared to women. Our results are consistent with previous estimates and other North American findings, and provide results from coast to coast, as well as useful information about the degree of regional and demographic variations that can be seen within a single country. Our work demonstrates the usefulness of using multiple data sources, adjusting for the error in each, and providing estimates of the sensitivity of different case definition approaches.
    The Journal of Rheumatology 03/2014; · 3.26 Impact Factor

Publication Stats

9k Citations
1,736.70 Total Impact Points


  • 2000–2014
    • McGill University Health Centre
      • Epidemiology Clinic
      Montréal, Quebec, Canada
  • 1994–2014
    • McGill University
      • • Department of Epidemiology, Biostatistics and Occupational Health
      • • Department of Medicine
      Montréal, Quebec, Canada
  • 2012–2013
    • Laval University
      Québec, Quebec, Canada
    • The University of Calgary
      • Department of Medicine
      Calgary, Alberta, Canada
  • 2008–2009
    • Institut Bergonié
      Burdeos, Aquitaine, France
    • The University of Hong Kong
      Hong Kong, Hong Kong
  • 2000–2009
    • Centre hospitalier de l'Université de Montréal (CHUM)
      Montréal, Quebec, Canada
  • 2004–2007
    • Queen's University
      • Department of Community Health and Epidemiology
      Kingston, Ontario, Canada
    • Université de Montréal
      • Faculty of Pharmacy
      Montréal, Quebec, Canada
    • Procter & Gamble
      Cincinnati, Ohio, United States
    • Imperial College London
      • Department of Infectious Disease Epidemiology
      London, ENG, United Kingdom
  • 2006
    • University of Oklahoma Health Sciences Center
      • Department of Biostatistics and Epidemiology
      Oklahoma City, OK, United States
  • 2005
    • Hôpital La Pitié Salpêtrière (Groupe Hospitalier "La Pitié Salpêtrière - Charles Foix")
      • Service de Rhumatologie
      Lutetia Parisorum, Île-de-France, France
    • Université du Québec à Montréal
      Montréal, Quebec, Canada
    • Imperial Valley College
      Imperial, California, United States
    • Hôpital Maisonneuve-Rosemont
      Montréal, Quebec, Canada
  • 2002
    • Jewish General Hospital
      Montréal, Quebec, Canada
  • 1995
    • Centre Hospitalier de Verdun
      Verdun-sur-Meuse, Lorraine, France
  • 1992
    • Dalhousie University
      Halifax, Nova Scotia, Canada