Jarrod E Dalton,  PhD

Jarrod E Dalton, PhD

Assistant Staff

Lerner Research Institute, 9500 Euclid Avenue, Cleveland, Ohio 44195
Location: JJN3-426
Email: daltonj@ccf.org
Phone: (216) 444-9924


The overarching objective in my research is to conduct relevant, problem-oriented research through continual establishment of collaborations with experts from many disciplines. I have steadily sought to develop and promote systems science as a platform for transdisciplinary research.

I currently classify systems science into a) the specification and analysis of mechanistic systems (e.g., through structural equation models, Bayesian networks and influence diagrams), b) the analysis of network (or connected) data structures, c) the modeling of dynamic systems (i.e., “stock-and-flow” or differential equation models), and d) the study of interactive systems (e.g., agent-based models). To date, I have established particular expertise with modeling Bayesian networks and influence diagrams. In my current role as a KL2 Scholar in the Clinical and Translational Science Collaborative of Cleveland, I led a large statistical software development project which yielded the ‘HydeNet’ R package. The HydeNet package allows for streamlined analysis of Bayesian network and influence diagram model structures through an interface to JAGS.

Recently, on behalf of my research team, I submitted a NIH systems science R01 proposal that seeks to model the complex environmental and behavioral mechanisms involved with atherosclerotic risk. In this proposed research, we are applying forecasting methodology from the atmospheric science community to combine two distinct systems-based prediction models. Essentially, we are assuming that prediction models for atherosclerotic disease-related variables like LDL cholesterol and blood pressure are fundamentally different (i.e., different predictors, different relationships, etc.) depending on the time horizon. The first model will be designed to predict nearer-term outcomes of primary preventative care (forecast time horizon of approximately 1-2 years), while the second will be designed to capture overall trends for key clinical variables over the life course among different subpopulations (forecast time horizon greater than approximately 1-2 years). These two models will be combined to form more personalized assessments of cardiovascular risk via a convolution function over time (i.e., more weight is given to the primary practice model for nearer-term forecast horizons, and more weight is given to the climatological model for longer-term forecast horizons). The planned use of Bayesian network models, at least with respect to the climatological component, will allow for a variety of applications ranging from population health management (e.g., targeted implementation of preventative care services and health education programs) to dynamically-evolving risk assessments for specific patients as clinical data are obtained.

I am currently involved with other modeling efforts to promote population health, improve healthy eating environments, and other efforts to reduce health disparities. These projects are in various stages – with some already attracting federal funding – and are being done in collaboration with faculty from across the Clinical and Translational Science Collaborative of Cleveland.

In other words ...

My research expertise is in modeling complex systems by combining data from a variety of sources (for instance, electronic medical records, and neighborhood characteristics from agencies such as the U.S. Census and the U.S. Centers for Disease Control and Prevention). In particular, I am currently leading an effort to understand the extent to which various neighborhood and environmental characteristics can be helpful in identifying who is likely to have a heart attack or stroke.

Nikolas Ivan Krieger  MA, MS Nikolas Ivan Krieger, MA, MS
Data Scientist
Phone:(216) 445-4470

  1. Dalton JE, Zidar DA, Udeh BL, Patel MR, Schold JD, and Dawson NV. Practice Variation among Hospitals in Revascularization Therapy and Its Association with Procedure-Related Mortality. Medical Care 54(6): 623-31. 2016 [PubMed]

  2. Dalton JE, Dawson NV, Sessler DI, Schold JD, Love TE, and Kattan MW.  Empirical Treatment Effectiveness Models for Binary Outcomes. Medical Decision Making 36(1):101-14. 2016 [PubMed]

  3. Dalton JE and Nutter B. HydeNet: Hybrid Bayesian Networks Using R and JAGS. R package version 0.9.0. URL: http://CRAN.R-project.org/package=HydeNet. 2015

  4. Dalton JE. Flexible Recalibration of Binary Clinical Prediction Models. Statistics in Medicine. 32(2): 282-289. 2013 [PubMed]

  5. Dalton JE, Glance LG, Mascha EJ, Ehrlinger J, Chamoun N, Sessler DI. Impact of present-on-admission indicators on risk-adjusted hospital mortality measurement. Anesthesiology 118(6): 1298-1306. 2013 [PubMed] [software link for POARisk model]

  6. Dalton JE, Kurz A, Turan A, Mascha EJ, Sessler DI, Saager L. Development and Validation of a Risk Quantification Index for 30-Day Postoperative Mortality and Morbidity in Noncardiac Surgical Patients. Anesthesiology 114(6):1336-1344. 2011 [PubMed] [software link for RQI model]

  7. Dalton JE, Kattan MW. Recent advances in evaluating the prognostic value of a marker. Scand J Clin Lab Invest Suppl.242:59-62. 2010 [PubMed]

10/03/2017 |  

Cardiovascular Risk Models Exclude Socioeconomic Factors; New Grant to Develop Enhanced Model

A team of Lerner researchers led by Jarrod Dalton, PhD, assistant staff in the Department of Quantitative Health Sciences, have shown that current cardiovascular risk prediction models systematically underestimate risk in individuals from low income and disadvantaged neighborhoods. With a newly awarded four-year, $2.2 million grant from the National Institute on Aging, part of the National Institutes of Health, the team will develop a more sophisticated model—one that incorporates both clinical and socioeconomic factors—to better predict risk for major cardiovascular disease-related health events such as stroke and heart attack. In addition to helping clinicians provide more targeted prevention and treatment options for patients, the team hopes that better understanding the underlying neighborhood-level risks associated with these events might ultimately inform community development initiatives to create healthier environments.