Donna Tjandra1, Raymond Q Migrino2,3, Bruno Giordani4, Jenna Wiens1. 1. Computer Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA. 2. Phoenix Veterans Affairs Healthcare System, Phoenix, Arizona, USA. 3. University of Arizona, College of Medicine-Phoenix, Phoenix, Arizona, USA. 4. Neuropsychology Program, Department of Psychiatry, and Michigan Alzheimer's Disease Research Center, University of Michigan, Ann Arbor, Michigan, USA.
Abstract
INTRODUCTION: Studies investigating the relationship between blood pressure (BP) measurements from electronic health records (EHRs) and Alzheimer's disease (AD) rely on summary statistics, like BP variability, and have only been validated at a single institution. We hypothesize that leveraging BP trajectories can accurately estimate AD risk across different populations. METHODS: In a retrospective cohort study, EHR data from Veterans Affairs (VA) patients were used to train and internally validate a machine learning model to predict AD onset within 5 years. External validation was conducted on patients from Michigan Medicine (MM). RESULTS: The VA and MM cohorts included 6860 and 1201 patients, respectively. Model performance using BP trajectories was modest but comparable (area under the receiver operating characteristic curve [AUROC] = 0.64 [95% confidence interval (CI) = 0.54-0.73] for VA vs. AUROC = 0.66 [95% CI = 0.55-0.76] for MM). CONCLUSION: Approaches that directly leverage BP trajectories from EHR data could aid in AD risk stratification across institutions.
INTRODUCTION: Studies investigating the relationship between blood pressure (BP) measurements from electronic health records (EHRs) and Alzheimer's disease (AD) rely on summary statistics, like BP variability, and have only been validated at a single institution. We hypothesize that leveraging BP trajectories can accurately estimate AD risk across different populations. METHODS: In a retrospective cohort study, EHR data from Veterans Affairs (VA) patients were used to train and internally validate a machine learning model to predict AD onset within 5 years. External validation was conducted on patients from Michigan Medicine (MM). RESULTS: The VA and MM cohorts included 6860 and 1201 patients, respectively. Model performance using BP trajectories was modest but comparable (area under the receiver operating characteristic curve [AUROC] = 0.64 [95% confidence interval (CI) = 0.54-0.73] for VA vs. AUROC = 0.66 [95% CI = 0.55-0.76] for MM). CONCLUSION: Approaches that directly leverage BP trajectories from EHR data could aid in AD risk stratification across institutions.
Authors: M Kivipelto; E L Helkala; M P Laakso; T Hänninen; M Hallikainen; K Alhainen; H Soininen; J Tuomilehto; A Nissinen Journal: BMJ Date: 2001-06-16
Authors: Chike C Nwabuo; Yuichiro Yano; Henrique T Moreira; Duke Appiah; Henrique D Vasconcellos; Queen N Aghaji; Anthony Viera; Jamal S Rana; Ravi V Shah; Venkatesh L Murthy; Norrina B Allen; Pamela J Schreiner; Donald M Lloyd-Jones; João A C Lima Journal: JAMA Cardiol Date: 2020-07-01 Impact factor: 14.676
Authors: Rianne A A de Heus; Marcel G M Olde Rikkert; Phillip J Tully; Brian A Lawlor; Jurgen A H R Claassen Journal: Hypertension Date: 2019-09-23 Impact factor: 10.190