Page 215 - GHR_NursingCMU2024_Final.indd

Risk Factor Analysis and Nomograph Construction of Models for Unplanned
Readmissions in Patients with Hematologic Malignancies

1
1
Y. Cui, M. Chen, and N. Jin 1
1
School of Nursing, Peking Union Medical College, Institute of Hematology & Blood Diseases Hospital,
Chinese Academy of Medical Sciences & Peking Union Medical College, China

Background: Patients with hematologic malignancies have a high mortality rate, with approximately
460,000 deaths from malignant hematological diseases worldwide. Hospitals in all regions of the world
incur significant healthcare costs each year due to unplanned readmissions, resulting in a waste of
healthcare resources. It is used internationally as a key indicator of quality of care evaluation.

Objective: This study aimed to develop a model for predicting the risk of unplanned readmissions in
hematologic malignancy patients and to validate its validity.

Methods: This study adopted a retrospective data collection method. A total of 3241 hematologic
malignancy patients in a tertiary hospital in Tianjin from 1 January 2017 to 31 December 2021 were enrolled. All
of the data has been divided into two groups at random (7:3): a modeling group consisting of 2269 cases
and a validation group consisting of 972 cases. According to whether unplanned readmission occurs, the
patients in the modeling group were divided into two groups: one group was unplanned readmission (329
cases) and the other group was planned readmission (1940 cases). Univariate analysis and multivariate
logistic regression analysis were applied to construct risk prediction models.

Results: The incidence of unplanned readmission in patients with hematologic malignancies was 14.50%.
Logistic regression analysis showed that the risk factors included in the model were disease diagnosis,
ECOG score, lung infection, central venous cannula, and comorbidities. The nomogram’s area under the
curve (AUC) for both the modeling and validation groups stood at 0.790 and 0.771, respectively, signifying
the effective discrimination capabilities of the risk prediction nomogram model. The real value roughly
Poster Presentation Abstracts
matched the predicted value, while the calibration curve almost perfectly matched the ideal curve.

Conclusion: Clinical caregivers can use the efficient risk prediction nomogram model developed in
this work as a guide to identifying patients with medium and high risks.

Keywords: hematologic malignancies, unplanned readmissions, risk factors, prediction model,
nursing care

_____________________________________________________________________________________________________
Correspondence: Yan Cui, School of Nursing, Peking Union Medical College, Institute of Hematology &
Blood Diseases Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, China
E-mail: sci888@qq.com

213