Joel Dubin

Associate Professor

Joel DubinContact Information:
Joel Dubin

Research interests

My primary research interest is in the area of methodological development in longitudinal data analysis, including for multivariate longitudinal data, where more than one outcome, (e.g., systolic and diastolic blood pressure) are each followed for individuals over time. Methods pursued for this type of data include the correlation of different longitudinal outcomes over time using curve-based methods, and incorporating lags and derivatives of the curves. I am also interested in change point and latent response models for longitudinal data, as well as prediction models.

An additional area of interest is developing graphical methods for censored survival data, which simultaneously display individual-level as well as cohort-level survival information. This research has resulted in two separate graphical methods, the event chart and the event history graph; the latter displays time-dependent covariate information embedded within the Kaplan-Meier survival curve.

I work in a variety of application areas, including nephrology, cancer, smoking cessation, aging, environmental issues, and the analysis of adverse events from clinical trials.

Along with Professor Joon Lee in the University of Waterloo School of Public Health & Health Systems, I am the co-leader of the Health Data Science Lab.

Education/biography

After completing my masters degree in Applied Statistics at Villanova University in 1993, I worked at Veteran Affairs Health Services and Research in Houston, Texas; and at the University of Texas M.D. Anderson Cancer Center, also in Houston.

I then went on to receive my PhD in Statistics from the University of California at Davis in 2000, after which I worked as an assistant professor at the Yale University Department of Epidemiology and Public Health in the Division of Biostatistics, forging several collaborations with researchers in public health and medicine at Yale, some of which continue today.

I arrived as an associate professor at the University of Waterloo in 2005, with a joint appointment in the Department of Statistics and Actuarial Science, and the Department of Health Studies and Gerontology, the latter which is now the School of Public Health and Health Systems.

Selected publications

  • Raffa JD, Dubin JA.  Multivariate longitudinal data analysis with mixed effects hidden Markov models. Biometrics, 71(3): 821-831, 2015.
  • Lee J, Maslove DM, Dubin JA.  Personalized mortality prediction driven by electronic medical data and a patient similarity metric. PLOS ONE, 10(5): e0127428, 2015.
  • Leatherdale ST, Brown KS, Carson V, Childs RA, Dubin JA, Elliot SJ, Faulkner G, Hammond D, Manske S, Sabiston CM, Laxer RE, Bredin C, Thompson-Haile A.  The COMPASS study: a longitudinal hierarchical research platform for evaluating natural experiments related to changes in school-level programs, policies and built environment resources. BMC Public Health, 14: 331, 2014.
  • Khan, S.A., Chiu, G.S., Dubin, J.A. Therapeutic hypothermia: quantification of the transition of core body temperature using the flexible mixture bent-cable model for longitudinal data. Australian & New Zealand Journal of Statistics, 55: 369-385, 2013 (published in 2014).
  • Foebel A.D., Heckman G.A., Ji K., Dubin J.A., Turpie I.D., Hussack P., McElvie R.S. Heart failure-related mortality and hospitalization in the year following admission to a long-term care facility: The Geriatric Outcomes and Longitudinal Decline in Heart Failure (GOLD-HF) Study. Journal of Cardiac Failure, 19(7): 468-477, 2013.
  • Khan, S.A., Rana, M., Li, L., Dubin, J.A. A statistical investigation to monitor and understand atmospheric CFC decline with the spatial-longitudinal bent-cable model. International Journal of Statistics & Probability, 1(2): 56-68, 2012.
  • Hobin, E., Leatherdale, S., Manske, S., Dubin, J.A., Elliot, S., Veugelers, P.  A multilevel examination of gender differences in the association between features of the school environment and time spent in physical activity among a sample of grades 9 to 12 students in Ontario, Canada. Accepted, BMC Public Health, 12: article 74, DOI: 10.1186/1471-2458-12-74, 2012.
  • Xiong, X, Dubin, J.A.  A binning method for analyzing mixed longitudinal data measured at distinct time points. Statistics in Medicine, 29: 1919-1931, 2010.
  • Dubin J.A., O'Malley S.S.  Event charts for the analysis of adverse events in longitudinal studies: an example from a smoking cessation pharmacotherapy trial.  The Open Epidemiology Journal, 3: 34-41, 2010.
  • Khan, S.A., Chiu, G., Dubin, J.A.  Atmospheric concentration of chlorofluorocarbons: addressing the global concern with the longitudinal bent-cable model.  CHANCE, 22: 8-17, 2009.
  • Hall P.A., Dubin J.A., Crossley M., Holmqvist M., D'Arcy, C.   Does executive function explain the IQ-mortality association? Evidence from the Canadian Study on Health and Aging.  Psychosomatic Medicine, 71: 196-204, 2009.
  • Wainwright, P.E., Leatherdale, S.E., Dubin, J.A.  Advantages of mixed effects models over traditional ANOVA models in developmental studies: A worked example in a mouse model of fetal alcohol syndrome.  Developmental Psychobiology, 49: 664-674, 2007.
  • Dubin J.A., Han, L., Fried, T.R.  Triggered sampling could help improve longitudinal studies of persons with elevated mortality risk.  Journal of Clinical Epidemiology, 60: 288-293, 2007.
  • Dubin J.A., Müller H.G.  Dynamical correlation for multivariate longitudinal data.  Journal of the American Statistical Association, 100: 872-881, 2005.
  • Hardy, S.E., Dubin, J.A., Holford, T.R., Gill, T.M.  Transitions between states of disability and independence among older persons.  American Journal of Epidemiology, 161: 575-584, 2005.
  • Dubin J.A., Müller H.G., Wang J.L.  Event history graphs for censored survival data.  Statistics in Medicine, 20: 2951-2964, 2001.
  • Lee JJ, Hess KH, Dubin JA.  Extensions and applications of event charts.  American Statistician, 54(1): 63-70, 2000.

Software development

  • Dubin JA, Qiao, D, Müller HG. R package dynCorr to perform dynamical correlation analysis on multivariate longitudinal data, as described in Dubin and Müller (2005). On CRAN. Completed (version 0.1-1) in February, 2009. Updated (version 0.1-2) in October, 2012.
  • Dubin JA, Müller HG, Wang, JL. R (and S-Plus) function event.history, which is contained in Frank Harrell’s Hmisc package. The function is based on the method from Dubin, Müller, and Wang (2001). Completed in 2001.
  • Lee JJ, Hess KH, Dubin JA.  R (and S-Plus) function event.chart, which is contained in Frank Harrell’s Hmisc package. The function is based on the method from Lee, Hess, and Dubin (2000), and Dubin, Lee, and Hess (1997). Completed in 1997, with updates in 2000 and 2008.
Affiliation: 
University of Waterloo
Contact information: