Gregory B Auffenberg1, Khurshid R Ghani2, Shreyas Ramani3, Etiowo Usoro3, Brian Denton4, Craig Rogers5, Benjamin Stockton6, David C Miller2, Karandeep Singh7. 1. Department of Urology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA. 2. Department of Urology, University of Michigan Medical School, Ann Arbor, MI, USA. 3. School of Information, University of Michigan, Ann Arbor, MI, USA. 4. Department of Urology, University of Michigan Medical School, Ann Arbor, MI, USA; Department of Industrial & Operations Engineering, University of Michigan College of Engineering, Ann Arbor, MI, USA. 5. Department of Urology, Henry Ford Health System, Detroit, MI, USA. 6. Lakeside Urology PC, St. Joseph, MI, USA. 7. School of Information, University of Michigan, Ann Arbor, MI, USA; Department of Learning Health Sciences, University of Michigan Medical School, Ann Arbor, MI, USA; Department of Internal Medicine, University of Michigan Medical School, Ann Arbor, MI, USA. Electronic address: kdpsingh@umich.edu.
Abstract
BACKGROUND: Clinical registries provide physicians with a means for making data-driven decisions but few opportunities exist for patients to interact with registry data to help make decisions. OBJECTIVE: We sought to develop a web-based system that uses a prostate cancer (CaP) registry to provide newly diagnosed men with a platform to view predicted treatment decisions based on patients with similar characteristics. DESIGN, SETTING, AND PARTICIPANTS: The Michigan Urological Surgery Improvement Collaborative (MUSIC) is a quality improvement consortium of urology practices that maintains a prospective registry of men with CaP. We used registry data from 45 MUSIC urology practices from 2015 to 2017 to develop and validate a random forest machine learning model. After fitting the random forest model to a derivation cohort consisting of a random two-thirds sample of patients after stratifying by practice location, we evaluated the model performance in a validation cohort consisting of the remaining one-third of patients using a multiclass area under the curve (AUC) measure and calibration plots. RESULTS AND LIMITATIONS: We identified 7543 men diagnosed with CaP, of whom 45% underwent radical prostatectomy, 30% surveillance, 17% radiation therapy, 5.6% androgen deprivation, and 1.8% watchful waiting. The personalized prediction for patients in the validation cohort was highly accurate (AUC 0.81). CONCLUSIONS: Using clinical registry data and machine learning methods, we created a web-based platform for patients that generates accurate predictions for most CaP treatments. PATIENT SUMMARY: We have developed and tested a tool to help men newly diagnosed with prostate cancer to view predicted treatment decisions based on similar patients from our registry. We have made this tool available online for patients to use.
BACKGROUND: Clinical registries provide physicians with a means for making data-driven decisions but few opportunities exist for patients to interact with registry data to help make decisions. OBJECTIVE: We sought to develop a web-based system that uses a prostate cancer (CaP) registry to provide newly diagnosed men with a platform to view predicted treatment decisions based on patients with similar characteristics. DESIGN, SETTING, AND PARTICIPANTS: The Michigan Urological Surgery Improvement Collaborative (MUSIC) is a quality improvement consortium of urology practices that maintains a prospective registry of men with CaP. We used registry data from 45 MUSIC urology practices from 2015 to 2017 to develop and validate a random forest machine learning model. After fitting the random forest model to a derivation cohort consisting of a random two-thirds sample of patients after stratifying by practice location, we evaluated the model performance in a validation cohort consisting of the remaining one-third of patients using a multiclass area under the curve (AUC) measure and calibration plots. RESULTS AND LIMITATIONS: We identified 7543 men diagnosed with CaP, of whom 45% underwent radical prostatectomy, 30% surveillance, 17% radiation therapy, 5.6% androgen deprivation, and 1.8% watchful waiting. The personalized prediction for patients in the validation cohort was highly accurate (AUC 0.81). CONCLUSIONS: Using clinical registry data and machine learning methods, we created a web-based platform for patients that generates accurate predictions for most CaP treatments. PATIENT SUMMARY: We have developed and tested a tool to help men newly diagnosed with prostate cancer to view predicted treatment decisions based on similar patients from our registry. We have made this tool available online for patients to use.
Authors: Paul R Womble; James E Montie; Zaojun Ye; Susan M Linsell; Brian R Lane; David C Miller Journal: Eur Urol Date: 2014-08-24 Impact factor: 20.096
Authors: David A Share; Darrell A Campbell; Nancy Birkmeyer; Richard L Prager; Hitinder S Gurm; Mauro Moscucci; Marianne Udow-Phillips; John D Birkmeyer Journal: Health Aff (Millwood) Date: 2011-04 Impact factor: 6.301
Authors: Nihar R Desai; Kevin F Kennedy; David J Cohen; Traci Connolly; Deborah B Diercks; Mauro Moscucci; Stephen Ramee; John Spertus; Tracy Y Wang; Robert L McNamara Journal: Am Heart J Date: 2017-08-12 Impact factor: 4.749
Authors: Matthew R Cooperberg; David J Pasta; Eric P Elkin; Mark S Litwin; David M Latini; Janeen Du Chane; Peter R Carroll Journal: J Urol Date: 2005-06 Impact factor: 7.450
Authors: Jinping Xu; James Janisse; Julie J Ruterbusch; Joel Ager; Joe Liu; Margaret Holmes-Rovner; Kendra L Schwartz Journal: Ann Fam Med Date: 2016-05 Impact factor: 5.166
Authors: Grace A Lin; David S Aaronson; Sara J Knight; Peter R Carroll; R Adams Dudley Journal: CA Cancer J Clin Date: 2009-10-19 Impact factor: 508.702
Authors: Michael L Cher; Apoorv Dhir; Gregory B Auffenberg; Susan Linsell; Yuqing Gao; Bradley Rosenberg; S Mohammad Jafri; Laurence Klotz; David C Miller; Khurshid R Ghani; Steven J Bernstein; James E Montie; Brian R Lane Journal: J Urol Date: 2016-07-14 Impact factor: 7.450
Authors: Paula Dhiman; Jie Ma; Constanza L Andaur Navarro; Benjamin Speich; Garrett Bullock; Johanna A A Damen; Lotty Hooft; Shona Kirtley; Richard D Riley; Ben Van Calster; Karel G M Moons; Gary S Collins Journal: BMC Med Res Methodol Date: 2022-04-08 Impact factor: 4.615