OBJECTIVE: Proposed value-based care models, such as bundled payments, will place providers at greater risk for financial outcomes. It is important for providers to identify which patients are at high-risk for increased spending so that they can be more aggressively managed during the global period. Our goal is to use surgical episodes of care to identify high-cost head and neck surgical patients and determine drivers of cost and outcomes in this group.
METHODS: We identified 3,459 adult patients who were admitted post-operatively after head and neck cancer surgery in the Optum claims database from 2003 to 2016. We measured episode of care costs from the index hospitalization through 30 days after discharge. We used Kaplan-Meier analysis, multivariate Cox proportional hazards regression, and multivariable linear regression to determine factors associated with high costs.
RESULTS: Our patients had a mean age of 61.7 years (interquartile range [IQR], 54-71 years) and a mean non-malignancy Charlson comorbidity index (CCI) of 1 (IQR, 0-2). The median cost of a 30-day episode of care after head and neck surgery was $37,682 (IQR, $22,995-$59,109). Patients in the top cost decile were responsible for 31.7% of the aggregate total cost and their surgical episodes cost nearly 15-fold more than patients in the lowest cost decile and over 3-fold more than the overall median cost. The cost variability was driven by the index hospitalization costs more than post-acute care costs. The highest cost patients were more likely to be younger patients and have multiple comorbidities than the lower cost patients (24.6% vs 11.9% were ≤60 years and CCI≥1, P< .001). We then created groups of patients based on age and CCI: 947 young healthy patients (≤60 years, CCI<1), 667 young multimorbid patients (≤60, CCI≥1), 525 older healthy patients (>60, CCI<1), and 1320 older multimorbid patients (>60, CCI≥1). Compared with young healthy patients, older multimorbid patients had similar costs (odds ratio [OR], 0.91; 95% confidence interval [CI] 0.61-1.35), but young multimorbid patients had twice the odds of being in the highest cost decile (OR, 1.94; 95% CI, 1.18-3.21). Young multimorbid patients generated $10,588 more in total episode costs than their healthy counterparts (P< .001). In terms of quality outcomes, the young multimorbid patients had decreased 5-year overall survival relative to their healthy counterparts (61.6% vs 74.6%, P< .001). These patients also had the highest readmissions rate (12.9%) out of all the groups and was nearly double that of young healthy patients (7.2%).
CONCLUSIONS: In this privately insured cohort, younger multimorbid patients are the most likely patients to be in the highest decile of costs and are responsible for nearly a third of total costs, and have worse survival and quality outcomes. This suggests that targeting these high-cost young multimorbid patients for cost-saving strategies would be beneficial in our head and neck cancer surgical cohort.