An Automated Electronic Health Record Score to Estimate Length of Stay and Readmission in Patients Undergoing Radical Cystectomy for Bladder Cancer.

Abstract

Patients treated with radical cystectomy experience a high rate of postoperative complications and frequent hospital readmissions. We sought to explore the utility of the Care Assessment Need (CAN) score, derived from electronic health data, to estimate the risk of these adverse clinical outcomes, thereby aiding patient counseling and informed treatment decision-making. We retrospectively examined data from 982 patients with bladder cancer who underwent radical cystectomy between 2013 and 2018 within the national Veterans Health Administration system. We tested for associations between the preoperative CAN score and length of stay, discharge location, and readmission rates. We observed a correlation between higher CAN scores and longer hospital stays (adjusted relative risk = 1.03 [95% CI: 1.02-1.05]). An increased CAN score was also linked to greater odds of discharge to a skilled nursing facility or death (adjusted odds ratio = 1.16 [95% CI: 1.06-1.26]). Furthermore, the score was associated with hospital readmission at both 30 and 90 days postdischarge (adjusted HR = 1.03 [95% CI: 1.00-1.07] and 1.04 [95% CI: 1.00-1.07], respectively). The CAN score is associated with length of hospital stay, discharge to a skilled nursing facility, and readmission within 30 and 90 days after radical cystectomy. These findings highlight the potential of health care systems leveraging electronic health records for automatically calculating multidimensional tools, such as the CAN score, to identify patients at risk of adverse clinical outcomes after radical cystectomy.