EUVIMEDEUVIMED entdecken
Crossref

Abstract 6311: Evaluating Patient Barriers and Phenotypes of the Cardiac Rehabilitation Participation Cascade: Decision Tree Analysis and Logistic Regression

Minkun Liu, Kunihiro Matsushita, Kerry J Stewart, Lena M Mathews

Circulation · 02.11.2026 · 154 (Suppl_1)

DOI: 10.1161/circ.154.suppl_1.6311 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Circulation, erschienen am 02.11.2026. Verfasst von Minkun Liu und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Traditional approaches to improving cardiac rehabilitation (CR) participation often evaluate risk factors in isolation, limiting their real-world clinical implications. Rather than isolated risk factors, multiple risk factors can interact and exist in the same patient. We aimed to identify patients' phenotypes across the cascade of CR enrollment and completion using a dual-analytical framework.

Methods

We analyzed 2,780 patients referred to CR within a large health system between 2018 and 2025. We performed a sequential analysis utilizing multivariable logistic regression and machine learning-based Classification and Regression Tree (CART) modeling to evaluate two outcomes: enrollment (attending ≥1 session) and completion (≥36 sessions). Models integrated clinical characteristics, social determinants, and geospatial real-world driving distances, with subgroup analyses for percutaneous coronary intervention (PCI) and medically managed myocardial infarction (MI).

Results

The cohort had median age of 68, (IQR 60 to 77), 38% women, 29% Black. Overall, 690 (25%) of referred patients enrolled, and only 115 (17%) of enrollees completed the program. Multivariable regression (Figure 1) revealed enrollment barriers including Medicaid, Medicare, or self-pay insurance, higher neighborhood deprivation, longer driving distance, and non-cardiac surgery discharging specialty. Conversely, program completion was driven by an older age. CART analysis unmasked distinct non-linear patient phenotypes: the cohort presenting the highest risk of non-enrollment comprised patients discharged from cardiac surgery who faced longer driving distances (Figure 2). For the medically managed MI subgroup (Figure 3), insurance type acted as the primary systemic gatekeeper, with Medicaid, Medicare, or self-pay defining the lowest enrollment phenotype.

Conclusions

CR participation involves distinct phenotypic transitions, where factors driving initial enrollment differ fundamentally from those dictating long-term adherence. While driving distance represents a highly modifiable logistical barrier to entry, patient age demonstrates a paradoxical effect, hindering initial enrollment while increasing the likelihood for long-term completion. Deploying this phenotype-driven framework provides actionable insights for clinical practice and population health, enabling targeted strategies for patients at the highest risk of attrition.

Zitieren

APA

Liu, M., Matsushita, K., Stewart, K. J., & Mathews, L. M. (2026). Abstract 6311: Evaluating Patient Barriers and Phenotypes of the Cardiac Rehabilitation Participation Cascade: Decision Tree Analysis and Logistic Regression. Circulation, 154(Suppl_1). https://doi.org/10.1161/circ.154.suppl_1.6311

Vancouver

Liu M, Matsushita K, Stewart KJ, Mathews LM. Abstract 6311: Evaluating Patient Barriers and Phenotypes of the Cardiac Rehabilitation Participation Cascade: Decision Tree Analysis and Logistic Regression. Circulation. 2026;154(Suppl_1). doi:10.1161/circ.154.suppl_1.6311

BibTeX
@article{Liu2026,
  author = {Liu, Minkun and Matsushita, Kunihiro and Stewart, Kerry J and Mathews, Lena M},
  title = {Abstract 6311: Evaluating Patient Barriers and Phenotypes of the Cardiac Rehabilitation Participation Cascade: Decision Tree Analysis and Logistic Regression},
  journal = {Circulation},
  year = {2026},
  volume = {154},
  number = {Suppl_1},
  doi = {10.1161/circ.154.suppl_1.6311}
}
Kostenlos starten →