Institute for Clinical Research and Health Policy Studies

Predictive Analytics and Comparative Effectiveness (PACE) Center

Medical care is delivered one-patient-at-a-time. But the evidence for practicing is derived by aggregating many patients—typically thousands or tens of thousands of patients—into groups. This group-derived evidence would be highly informative for medical practice if all patients were identical. The dissimilarity of individual patients, however, potentially undermines clinical research as a scientific basis for the practice of medicine.

The Predictive Analytics and Comparative Effectiveness Center (PACE), led by Dr. David Kent, seeks to better understand and address the limitations of using group-derived evidence as the basis for decision making in individual patients. Our approach is based on the close integration of clinical and statistical reasoning. Our goal is to provide clinicians and patients with evidence better tailored to their particular circumstances; we have expertise in clinical medicine, risk modeling, individual patient meta-analysis, and observational comparative effectiveness studies.

The PACE Center seeks to provide clinicians and patients with evidence better tailored to their particular circumstances through risk modeling. Among other projects, Dr. Kent is the Principal Investigator of several federally-funded research grants related to these issues, including several methods grants from the Patient-Centered Outcomes Research Institute (PCORI) and grants from the National Institutes of Health (NIH)/National Institute of Neurological Disorders and Stroke (NINDS) focused on cerebrovascular disease. Dr. Kent works alongside 5 additional faculty investigators and statisticians.

 

PACE research group at their annual retreat

David Kent, MD, CM, MSc
Director and Professor of Medicine

June C. Baglione
Senior Research Administrator

Mabel Celestino
Research Project Coordinator

William Crown, PhD
Senior Adviser

Benjamin Koethe, MPH
Statistician

Keren Ladin, PhD, MSc
Associate Professor, Depts of Occupational Therapy and Community Health, Tufts University; Director, Research on Ethics, Aging, and Community Health (REACH Lab)

Lester Y. Leung, MD, MSc
Director, Comprehensive Stroke Center; Director, Stroke and Young Adults (SAYA) Program

Jennifer Lutz, MA
Research Assistant

C.H.M. (Carolien) Maas, MSc
Student Researcher 

Jinny Park, MPH
Research Project Coordinator

Jenica Upshaw, MD
Medical Director, Cardio-Oncology Program; Attending Physician, Advanced Heart Failure 

Ellen Vickery, MS
Clinical Studies Manager

Andy Y. Wang, AB
Medical Student 

Benjamin S. Wessler, MD
Staff Cardiologist and Assistant Professor of Medicine 

Greenwall Foundation "A Rubric for Predictive Analytics in Healthcare"

This multidisciplinary collaboration of experts will develop a rubric to provide guidance on algorithmic bias and fairness concerns within clinical prediction modeling. This project draws on the expertise of numerous professionals in fields such as machine learning, prediction modeling, health informatics, medical ethics, diversity and inclusion, and others, as well as the engagement of patient stakeholders to explore the issue of algorithmic fairness within healthcare settings.

PCORI “The Systematic, Collaborative, PFO closure Evaluation” (SCOPE)

The project aims to examine and provide definitive evidence as to the treatment effects of PFO closure versus medical therapy to support patient-centered decision making for patients with PFO and cryptogenic stroke. This project conducted a high-quality and rigorous IPD MA utilizing the 6 component databases of patients with PFO and CS.

W.L. Gore & Associates “How Well is PFO Closure Likely to Work in Patients Over 60 Years of Age? An Analysis of the Risk of Paradoxical Embolism Database”

The project aims to examine and provide definitive evidence as to the treatment effects of PFO closure versus medical therapy to support patient-centered decision making for patients with PFO and cryptogenic stroke who are over 60 years of age, using data from multiple databases.

 

Tufts Collaborates “Estimating Individual Treatment Effects from Randomized Clinical Trials using Machine Learning"

This project will provide a benchmark comparison of state of the art machine learning approaches for treatment effect prediction on 18 large clinical trials

NIH K23 “Palliative Care for High-Risk TAVR Patients: The Impact of Multimorbidity”

This proposal addresses knowledge gaps related to the significance of multiple chronic conditions (MCC), the nature of unmet care needs, and the use of early palliative care for high-risk patients treated with TAVR who are unlikely to do well.

NIH R01 “Enabling Comparative Effectiveness Research in Silent Brain Infarction Through Natural Language Processing and Big Data”

We aim to improve the evidence base for the prevention of stroke in patients with silent brain infarct (SBI). We will develop a natural language processing algorithm that can accurately identify cases of SBI, and estimate the comparative effectiveness of medications on the risk of future stroke and dementia.

PCORI Dissemination & Implementation “Improving Diabetes Prevention with Benefit-Based Tailored Treatment

Disseminating Patient-Centered Estimates of Benefit": The goals of this project are to disseminate a clear and clinically-actionable understanding of the variation in patient-centered estimates of benefits of diabetes prevention interventions based on our PCORI-funded prediction model derived directly from the Diabetes Prevention Program (DPP) trial. The Tufts PACE DPP Risk model was developed for use in different electronic health record systems to provide information on a patient's individualized risk for developing diabetes and how well preventive treatment—an intensive lifestyle program or taking Metformin—was likely to reduce risk for the patient.

Publicly Accessible Project Materials

PCORI Methods Covid Enhancement “Generalizable Prognostic Models for Patient-Centered Decisions in COVID-19”

This project seeks to develop and validate clinical prediction models in 2 large hospital systems (including approximately 20,000 hospitalized COVID-19 patients) to predict the need for mechanical ventilation and mortality in patients hospitalized with COVID-19.  These tools will be piloted in at least one setting.

Featured

Gulati G, Upshaw J, Wessler BS, Brazil RJ, Nelson J, van Klaveren D, Lundquist CM, Park JG, McGinnes H, Steyerberg EW, Van Calster B, Kent DM. Generalizability of Cardiovascular Disease Clinical Prediction Models: 158 Independent External Validations of 104 Unique Models. Circ Cardiovasc Qual Outcomes. 2022 Apr;15(4):e008487. doi: 10.1161/CIRCOUTCOMES.121.008487. Epub 2022 Mar 31.

 

Leung LY, Zhou Y, Fu S, Zheng C, Luetmer PH, Kallmes DF, Liu H, Chen W, Kent DM. Risk Factors for Silent Brain Infarcts and White Matter Disease in a Real-World Cohort Identified by Natural Language Processing. Mayo Clin Proc. 2022 Jun;97(6):1114-1122. doi: 10.1016/j.mayocp.2021.11.038. Epub 2022 Apr 27.

 

Kent DM, Saver JL, Kasner SE, Nelson J, Carroll JD, Chatellier G, Derumeaux G, Furlan AJ, Herrmann HC, Jüni P, Kim JS, Koethe B, Lee PH, Lefebvre B, Mattle HP, Meier B, Reisman M, Smalling RW, Soendergaard L, Song JK, Mas JL, Thaler DE. Heterogeneity of Treatment Effects in an Analysis of Pooled Individual Patient Data From Randomized Trials of Device Closure of Patent Foramen Ovale After Stroke. JAMA. 2021 Dec 14;326(22):2277-2286.

 

Kent DM, Nelson J, Pittas A, Colangelo F, Koenig C, van Klaveren D, Ciemins E, Cuddeback J. An Electronic Health Record-Compatible Model to Predict Personalized Treatment Effects From the Diabetes Prevention Program: A Cross-Evidence Synthesis Approach Using Clinical Trial and Real-World Data. Mayo Clin Proc. 2021 Nov 12:S0025-6196(21)00708-4.

 

Wessler BS, Nelson J, Park JG, McGinnes H, Gulati G, Brazil R, Van Calster B, van Klaveren D, Venema E, Steyerberg E, Paulus JK, Kent DM. External Validations of Cardiovascular Clinical Prediction Models: A Large-Scale Review of the Literature. Circ Cardiovasc Qual Outcomes. 2021 Aug;14(8):e007858.

 

Kent DM, Leung LY, Zhou Y, Luetmer PH, Kallmes DF, Nelson J, Fu S, Zheng C, Liu H, Chen W. Association of Silent Cerebrovascular Disease Identified Using Natural Language Processing and Future Ischemic Stroke. Neurology. 2021 Sep 28;97(13):e1313-e1321.

 

Paulus JK, Kent DM. Predictably Unequal: Understanding and Addressing Concerns that Algorithmic Clinical Prediction May Increase Health Disparities. Nature Digital Medicine. 2020 Jul 30;3(99).

 

Kent DM, Paulus JK, van Klaveren D, D'Agostino R, Goodman S, Hayward R, Ioannidis JPA, Patrick-Lake B, Morton S, Pencina M, Raman G, Ross JS, Selker HP, Varadhan R, Vickers A, Wong JB, Steyerberg EW. The Predictive Approaches to Treatment effect Heterogeneity (PATH) Statement. Ann Intern Med. 2020 Jan 7;172:35-45.

 

Kent DM, van Klaveren D, Paulus JK, D'Agostino R, Goodman S, Hayward R, Ioannidis JPA, Patrick-Lake B, Morton S, Pencina M, Raman G, Ross JS, Selker HP, Varadhan R, Vickers A, Wong JB, Steyerberg EW. The Predictive Approaches to Treatment effect Heterogeneity (PATH) Statement: Explanation and Elaboration. Ann Intern Med. 2020 Jan 7;172:W1-W25.

 

Individual Investigators

View Dr. Kent’s recent publications on PubMED>

View Mr. Nelson’s recent publications on PubMED>

View Dr. Paulus’ recent publications on PubMED>

View Ms. Ruthazer’s recent publications on PubMED>

View Dr. van Klaveren’s recent publications on PubMED>

View Dr. Wessler’s recent publications on PubMED>

“Anatomic Features Key for Selecting Patients Best Suited to PFO Closure.” Yael L. Maxwell. TCTMD, produced by the Cardiovascular Research Foundation (CRF). Oct 14, 2022.

"New Formula Predicts with High Accuracy Which Patients with Recurring Stroke Would Benefit from PFO Closure" Press Release, Tufts Medical Center, Boston, MA December 14, 2021

“Heterogeneity of treatment effect and risk-stratified approach for reporting/ implementing clinical trial results.” 16th Global Cardiovascular Clinical Trialists (CVCT) Forum. Washington, DC. December 7, 2019.

“Prediabetes Predictive Model to Personalize Diabetes Risk.” OptumLabs Research and Translation Forum. Boston, MA. November 20, 2019.

“Improving Diabetes Prevention Based on Predicted Benefits of Treatment.” What’s Right For Me? Practical Approaches to Personalized Medicine. PCORI Annual Meeting. Washington, DC. September 18, 2019. 

“Personalized evidence based medicine: predictive approaches to heterogeneous treatment effects in randomized trials.” Society for Clinical Trials Annual Meeting. New Orleans, LA. May 19, 2019.

“Point-Hemoglobin A1c Goals: Is Lower Better? ” Tufts Medical Center Point-Counterpoint Grand Rounds. Boston, MA. October 12, 2018. 

“Evidence and the Individual Patient: Understanding Heterogeneous Treatment Effects for Patient-Centered Care.” National Academy of Medicine. Washington, DC. May 31, 2018.

"Selecting patients for lung cancer screening by personalized risk offers limited long term gains" Tufts Medical Center. Boston, MA. February 1, 2018 

“Personalized Risk Information in Cost Effectiveness Studies (PRICES).” Chapter 14. Health Economics Common Fund Research Symposium. Bethesda, MD. September 25, 2017. 

“Moving Beyond Averages” Patient-Centered Outcomes Research Institute. January 2017

“PACE Symposium: Using Group Data to Treat Individuals” Tufts Medical Center, Boston, MA. June 4, 2015.

“Getting it Right the First Time: Can We Predict Who is Likely to Respond?” The Myth of Average: Why Individual Patient Differences Matter Conference. Omni Shoreham Hotel, Washington DC, Nov 30, 2012.  

"Risk Modeling for Targeting Therapies to those who can Benefit: The Example of PFO Closure in Cryptogenic Stroke.” Center for Clinical and Translational Sciences (CCaTS) Grand Rounds. Mayo Clinic, Rochester, MN, September 27, 2013.

“An index to identify stroke-related versus incidental patent foramen ovale in cryptogenic stroke.”  Neurology Podcast. August 13 2013 Issue.

To better understand the extent of Clinical Prediction Model (CPM) development and to help researchers and clinicians, we have created the Tufts PACE CPM Registry, a field synopsis of over 1,000 CPMs that predict clinical outcomes for patients with and at risk for cardiovascular disease.

Predictive Analytics and Comparative Effectiveness (PACE) Center
The Institute for Clinical Research and Health Policy Studies
Tufts Medical Center
800 Washington St., Box 63 
Boston, MA 02111

 

Direct inquiries to:
Becca Maunder and/or Jinny Park
Research Project Coordinator, Predictive Analytics and Comparative Effectiveness (PACE) Center
pacecenter@tuftsmedicalcenter.org
Fax: 617-636-0022
https://twitter.com/Tufts_PACE