Thirty-day all-cause readmission rate at a cancer center as a quality of care measure

• Published in: Journal of clinical oncology Vol. 30; no. 34_suppl; p. 247
• Main Authors: Walters, Ronald Stewart, Kidin, Lisa M., Roquemore, Joyce, Jordan, Victoria S., Browning, Douglas
• Format: Journal Article
• Language: English
• Published: 01.12.2012
• ISSN: 0732-183X; 1527-7755
• DOI: 10.1200/jco.2012.30.34_suppl.247

Abstract only 247 Background: Recently, a 30-day all-cause readmission rate has been proposed as a measure of quality of care. Readmissions are assumed to reflect failure by the discharging physician, hospital, or post acute care. These rates are generally easily calculated from available administrative data, and classifiable as "related to the previous discharge diagnosis" or not. Present on admission modifiers may enhance classification and assignment to "preventable" or "non-preventable," "expected or non-expected." This methodology is not generally applicable to the oncology population. The experience with one major cancer center is presented as an example of the limitations of such an approach. Methods: We analyzed 52,097 oncology admissions in an all-payer population that occurred between January 2010 and January 2012. Results: A mean of 32.5% (n=16,918) were readmitted within 30 days, compared to a "peer" group in the database of the University Health Consortium, median of 15%. The attached graph demonstrates the stability of this proportion. Leukemia, lymphoma, stem cell patients (46%), all patients with intense medical needs and frequent readmissions, n=7,635, were the largest subgroup. 42% (n=7,099) were readmissions for chemotherapy or immunotherapy, both planned and expected, and 11% (n=1,803) due to neutropenic fever, pneumonia, or sepsis, all common in this population and neither unexpected nor usually preventable. The most preventable, unexpected, and unplanned readmissions were for postoperative infection, dehydration, and urinary tract infection, accounting for 3.6% (n=609). Thus, the majority of readmissions were planned, expected, or not preventable. Conclusions: In a cancer population at an academic cancer hospital, the majority of readmission are not only planned, but also expected for this population of patients and should not be construed as representative of a quality of care issue. Proper stratification and classification of readmissions is essential to the interpretation of such a measure.