Patients with acute myeloid leukemia (AML) frequently experienced unplanned hospital readmission or death within 30 days after discharge following induction chemotherapy, according to findings reported by Chae et al in JCO Clinical Cancer Informatics. The investigators found that incorporating comorbidities and symptom information extracted from clinical notes into electronic health record (EHR)-based models improved the ability to identify patients at risk for these adverse outcomes.
Study Details
The retrospective observational study included 305 adults with AML treated at a single Midwestern academic hospital between August 2006 and December 2021. Patients had received induction chemotherapy and were hospitalized during or after the treatment; both patients with newly diagnosed and relapsed/refractory AML requiring reinduction were included. The analysis ultimately encompassed 1,111 inpatient encounters.
The primary outcome of the study was an adverse event defined as an unplanned readmission or all-cause death within 30 days after discharge from the initial induction hospitalization. Clinical variables included demographics, treatment era, chemotherapy intensity, comorbidities, laboratory results, and symptoms documented during the induction admission. Symptoms were categorized into 10 domains (pain, gastrointestinal, myelosuppression, cardiopulmonary, skin changes, fatigue, anxiety/anger, central nervous system, depression, and sleep disturbance), while multiple logistic regression was used to identify predictors. A base model incorporating age, sex, treatment era, chemotherapy intensity, and an interaction between chemotherapy intensity and age was subsequently expanded with comorbidities, laboratory variables, and symptom data.
Key Results
The mean patient age was 64.6 years, and the mean induction hospitalization lasted 25.8 days. Overall, 121 patients (39.7%) experienced an adverse event within 30 days after discharge: 67 (22%) had an unplanned readmission, 67 (22%) died, and 13 (4.3%) experienced both outcomes. Among the 67 patients with unplanned readmissions, investigators documented 79 readmission encounters and 99 reasons for readmission. Fever or infection accounted for 53.5% of documented reasons, followed by metabolic, gastrointestinal, or renal complications (16.2%) and pain or discomfort (14.1%).
In the final model, a history of solid neoplasm was associated with more than twice the odds of an adverse event (odds ratio [OR] = 2.34, 95% confidence interval [CI] = 1.02–5.38, P = .046). Greater daily frequency of cardiopulmonary symptoms documented in clinical notes was also associated with increased risk (OR = 1.16, 95% CI = 1.07–1.26, P < .001). Chemotherapy intensity interacted significantly with age (P ≤ .001): for every 5-year increase in age, high-intensity chemotherapy was associated with a 22% increase in the odds of an adverse event (OR = 1.22, 95% CI = 1.03–1.45), whereas low-intensity chemotherapy was associated with a 27% decrease in the odds (OR = 0.73, 95% CI = 0.58–0.93).
The authors concluded: “Unplanned readmissions or early mortality occurred in 39.7% of patients with AML within 30 days postdischarge. Integrated risk stratification tools leveraging structured and unstructured EHR data may inform timely interventions to improve survival and reduce avoidable hospitalizations.”
Sena Chae PhD, RN, of the College of Nursing, University of Iowa, Iowa City, Iowa, is the corresponding author for the Journal of Clinical Oncology Clinical Cancer Informatics article.
DISCLOSURE: For full disclosures of the study authors, visit ascopubs.org.
