A Clinical Prediction Rule to Identify High‐Risk Patients During Interhospital Transfer

1University of Maryland, Baltimore, MD
2University of Maryland, Baltimore, MD
3University of Maryland, Baltimore, MD
4University of Maryland, Baltimore, MD
5University of Maryland, Baltimore, MD
6University of Maryland, Baltimore, MD
7University of Maryland School of Medicine, Baltimore, MD

Meeting: Hospital Medicine 2010, April 8-11, Washington, D.C.

Abstract number: 36

Background:

Interhospital transfer (IHT) is a common event whereby patients are transferred from community hospitals to tertiary‐care hospitals for specialized care or further workup in the setting of a diagnostic dilemma. Hospitalists are frequently the principal mediators for IHT, both as sending and receiving physicians. Data from an initial chart review at our tertiary‐care hospital suggested that patients admitted to internal medicine floor or telemetry beds via IHT had a 5‐fold higher incidence of death during the index hospitalizalion than patients admitted to the same level of care via the emergency department. Our objective was to identify specific high‐risk features among those patients who experienced an adverse outcome and to develop a clinical prediction rule that could be used prospectively to improve patient safety among patients undergoing IHT.

Methods:

We conducted a retrospective cohort study of patients admitted to the University of Maryland Medical Center via IHT over a 3‐year period (December 2005‐November 2008) to internal medicine floor or telemetry beds, defined as nursing units with vital sign monitoring once per 4 hours without continuous pulse oximetry. Data were collected via chart review and using the hospital's electronic clinical data repository, We used descriptive statistics to identify potential components of our prediction rule based on established definitions of abnormal vital signs, clinical features, and laboratory data. These variables were entered into a logistic regression model to identify independent risk factors associated with our outcome. Boolean logic terms were then used to create several potential prediction rules. Sensitivity and specificity were calculated to assess each rule's predictive ability.

Results:

Among 1179 patients admitted via IHT over the study period, 104 (8.8%) had the primary outcome of ICU admission within 48 hours or death during the index hospitalization. Variables most predictive of this outcome were pulse ≥ 110 beats/min, mean arterial pressure < 60 mm Hg, leukocyte count ≥ 18,000 cells/μL or hemoglobin < 7 g/dL A clinical prediction rule including 1 or more of these criteria would have idenlified 15% of the total study population as high risk and would have 43% sensitivity and 88% specificity,

Conclusions:

Patients admitted via IHT to internal medicine floor or telemetry beds were at highest risk for transfer to an ICU within 48 hours of arrival or death in the presence of extremes of tachycardia, hypotension, leukocytosis, or anemia. Our clinical prediction rule can be implemented at no cost, has a positive predictive value of 26% and a negative predictive value of 94% in our population, and may be a practical tool to reduce adverse patient outcomes by prompting over‐the‐phone triage to a higher level of care prior to transfer.

Author Disclosure:

M. Cina, none; M. Gulati, none; K. Silva, none; R. Habicht, none; R. Anderson, none; E. Lamos, none; J. Furuno, none.

To cite this abstract:

Cina M, Gulati M, Silva K, Habicht R, Anderson R, Lamos E, Furuno J. A Clinical Prediction Rule to Identify High‐Risk Patients During Interhospital Transfer. Abstract published at Hospital Medicine 2010, April 8-11, Washington, D.C. Abstract 36. Journal of Hospital Medicine. 2010; 5 (suppl 1). https://www.shmabstracts.com/abstract/a-clinical-prediction-rule-to-identify-highrisk-patients-during-interhospital-transfer/. Accessed September 15, 2019.

« Back to Hospital Medicine 2010, April 8-11, Washington, D.C.