Potentially avoidable hospital admissions in older patients with acute myeloid leukaemia in the USA: a retrospective analysis.

BACKGROUND: Older adults (≥60 years of age) with acute myeloid leukaemia spend a substantial proportion of their life in hospital after diagnosis. We examined reasons for their hospital admissions and identified potentially avoidable hospital admissions (PAH) in this age group in the USA.
METHODS: In this retrospective analysis, we examined the reasons for hospital admissions in older patients diagnosed with and treated for acute myeloid leukaemia at two tertiary care hospitals in the USA. We included patients receiving intensive induction chemotherapy or non-intensive treatment. We excluded those with acute promyelocytic leukaemia, those seen only for a one-time consultation who received primary treatment elsewhere, and those who received supportive care alone. We identified the eligible cohort using the Dana-Farber Cancer Institute and Massachusetts General Hospital Leukemia Clinical Research Information Systems database. Practising oncologists used a consensus-driven medical record review process to identify the primary reason for each hospital admission and categorise it as potentially avoidable or not avoidable on the basis of an adaptation of Graham's criteria for PAH. We used multivariable logistic regression analyses to identify predictors of PAH.
FINDINGS: Between May 1, 2005, and Dec 23, 2011, we assessed 1040 hospital admissions (excluding initial admission for diagnosis) in 329 consecutively admitted patients. The most common primary reasons for hospital admissions were: fever or infection (396 [38%]), planned admission for chemotherapy or transplantation (391 [38%]), and uncontrolled symptoms (102 [10%]). We identified 172 (27%) of 649 unplanned hospital admissions as potentially avoidable; among these admissions, 82 (48%) were readmissions because of previous premature hospital discharge, 32 (19%) because of problems that could have been managed in the outpatient setting, and 26 (15%) because of failure of timely outpatient follow-up. In a mixed logistic regression model, higher education (odds ratio 1·43 [95% CI 1·01-2·00]; p=0·04) and receipt of non-intensive induction chemotherapy (1·97 [1·25-3·10]; p=0·003) were predictors of PAH.
INTERPRETATION: Although many hospital admissions in older patients with acute myeloid leukaemia are unavoidable and driven by the illness course and its treatment, a substantial proportion are potentially avoidable. Future interventions to reduce PAH in this population are clearly warranted. FUNDING: National Cancer Institute, National Palliative Care Research Center, and Cambia Health Foundation.