Edwin S Wong1,2, Jean Yoon3,4, Rebecca I Piegari5, Ann-Marie M Rosland6,7, Stephan D Fihn5,8, Evelyn T Chang9,10. 1. Center of Innovation for Veteran-Centered and Value-Driven Care, VA Puget Sound Health Care System, 1660 S. Columbian Way, HSR&D MS S-152, Seattle, WA, 98108, USA. eswong@uw.edu. 2. Department of Health Services, University of Washington, Seattle, WA, USA. eswong@uw.edu. 3. Health Economics Resource Center, VA Palo Alto Healthcare System, Livermore, CA, USA. 4. Department of General Internal Medicine, UCSF School of Medicine, San Francisco, CA, USA. 5. Office of Clinical Systems Development and Evaluation, Veterans Health Administration, Seattle, WA, USA. 6. Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System, Pittsburgh, PA, USA. 7. Department of Internal Medicine, University of Pittsburgh, Pittsburgh, PA, USA. 8. Department of Medicine, University of Washington, Seattle, WA, USA. 9. Center for the Study of Healthcare Innovation, Implementation and Policy, VA Greater Los Angeles Health Care System, Los Angeles, CA, USA. 10. David Geffen School of Medicine, University of California, Los Angeles, CA, USA.
Abstract
OBJECTIVE: Many healthcare systems employ population-based risk scores to prospectively identify patients at high risk of poor outcomes, but it is unclear whether single point-in-time scores adequately represent future risk. We sought to identify and characterize latent subgroups of high-risk patients based on risk score trajectories. STUDY DESIGN: Observational study of 7289 patients discharged from Veterans Health Administration (VA) hospitals during a 1-week period in November 2012 and categorized in the top 5th percentile of risk for hospitalization. METHODS: Using VA administrative data, we calculated weekly risk scores using the validated Care Assessment Needs model, reflecting the predicted probability of hospitalization. We applied the non-parametric k-means algorithm to identify latent subgroups of patients based on the trajectory of patients' hospitalization probability over a 2-year period. We then compared baseline sociodemographic characteristics, comorbidities, health service use, and social instability markers between identified latent subgroups. RESULTS: The best-fitting model identified two subgroups: moderately high and persistently high risk. The moderately high subgroup included 65% of patients and was characterized by moderate subgroup-level hospitalization probability decreasing from 0.22 to 0.10 between weeks 1 and 66, then remaining constant through the study end. The persistently high subgroup, comprising the remaining 35% of patients, had a subgroup-level probability increasing from 0.38 to 0.41 between weeks 1 and 52, and declining to 0.30 at study end. Persistently high-risk patients were older, had higher prevalence of social instability and comorbidities, and used more health services. CONCLUSIONS: On average, one third of patients initially identified as high risk stayed at very high risk over a 2-year follow-up period, while risk for the other two thirds decreased to a moderately high level. This suggests that multiple approaches may be needed to address high-risk patient needs longitudinally or intermittently.
OBJECTIVE: Many healthcare systems employ population-based risk scores to prospectively identify patients at high risk of poor outcomes, but it is unclear whether single point-in-time scores adequately represent future risk. We sought to identify and characterize latent subgroups of high-risk patients based on risk score trajectories. STUDY DESIGN: Observational study of 7289 patients discharged from Veterans Health Administration (VA) hospitals during a 1-week period in November 2012 and categorized in the top 5th percentile of risk for hospitalization. METHODS: Using VA administrative data, we calculated weekly risk scores using the validated Care Assessment Needs model, reflecting the predicted probability of hospitalization. We applied the non-parametric k-means algorithm to identify latent subgroups of patients based on the trajectory of patients' hospitalization probability over a 2-year period. We then compared baseline sociodemographic characteristics, comorbidities, health service use, and social instability markers between identified latent subgroups. RESULTS: The best-fitting model identified two subgroups: moderately high and persistently high risk. The moderately high subgroup included 65% of patients and was characterized by moderate subgroup-level hospitalization probability decreasing from 0.22 to 0.10 between weeks 1 and 66, then remaining constant through the study end. The persistently high subgroup, comprising the remaining 35% of patients, had a subgroup-level probability increasing from 0.38 to 0.41 between weeks 1 and 52, and declining to 0.30 at study end. Persistently high-risk patients were older, had higher prevalence of social instability and comorbidities, and used more health services. CONCLUSIONS: On average, one third of patients initially identified as high risk stayed at very high risk over a 2-year follow-up period, while risk for the other two thirds decreased to a moderately high level. This suggests that multiple approaches may be needed to address high-risk patient needs longitudinally or intermittently.
Entities:
Keywords:
high risk; latent subgroups; machine learning; patient-centered medical home; risk stratification; trajectory
Authors: S T Normand; M B Landrum; E Guadagnoli; J Z Ayanian; T J Ryan; P D Cleary; B J McNeil Journal: J Clin Epidemiol Date: 2001-04 Impact factor: 6.437
Authors: Julie C Lauffenburger; Jessica M Franklin; Alexis A Krumme; William H Shrank; Troyen A Brennan; Olga S Matlin; Claire M Spettell; Gregory Brill; Niteesh K Choudhry Journal: Med Care Date: 2017-01 Impact factor: 2.983
Authors: Seppo T Rinne; Anashua R Elwy; Lori A Bastian; Edwin S Wong; Renda S Wiener; Chuan-Fen Liu Journal: Med Care Date: 2017-07 Impact factor: 2.983
Authors: Chuan-Fen Liu; Willard G Manning; James F Burgess; Paul L Hebert; Chris L Bryson; John Fortney; Mark Perkins; Nancy D Sharp; Matthew L Maciejewski Journal: Med Care Date: 2011-10 Impact factor: 2.983
Authors: Laura A Petersen; Margaret M Byrne; Christina N Daw; Jennifer Hasche; Brian Reis; Kenneth Pietz Journal: Health Serv Res Date: 2010-04-06 Impact factor: 3.402
Authors: Denise M Hynes; Kristin Koelling; Kevin Stroupe; Noreen Arnold; Katherine Mallin; Min-Woong Sohn; Frances M Weaver; Larry Manheim; Linda Kok Journal: Med Care Date: 2007-03 Impact factor: 2.983
Authors: Donna M Zulman; Christine Pal Chee; Stephen C Ezeji-Okoye; Jonathan G Shaw; Tyson H Holmes; James S Kahn; Steven M Asch Journal: JAMA Intern Med Date: 2017-02-01 Impact factor: 21.873
Authors: Cari Levy; Mary Ersek; Winifred Scott; Joan G Carpenter; Jennifer Kononowech; Ciaran Phibbs; Jill Lowry; Jennifer Cohen; Marybeth Foglia Journal: J Gen Intern Med Date: 2020-02-24 Impact factor: 5.128
Authors: Ravi B Parikh; Kristin A Linn; Jiali Yan; Matthew L Maciejewski; Ann-Marie Rosland; Kevin G Volpp; Peter W Groeneveld; Amol S Navathe Journal: PLoS One Date: 2021-02-19 Impact factor: 3.240
Authors: James Wick; David J T Campbell; Finlay A McAlister; Braden J Manns; Marcello Tonelli; Reed F Beall; Brenda R Hemmelgarn; Andrew Stewart; Paul E Ronksley Journal: CMAJ Open Date: 2022-04-19
Authors: Evelyn T Chang; Jean Yoon; Aryan Esmaeili; Donna M Zulman; Michael K Ong; Susan E Stockdale; Elvira E Jimenez; Karen Chu; David Atkins; Angela Denietolis; Steven M Asch Journal: Health Serv Res Date: 2021-06-18 Impact factor: 3.734