Attribution: whose complication is it?

Abstract

To improve quality, programs such as accountable care organizations need to determine the part of the health care system most "responsible" for a complication. This is referred to as attribution. This provides a framework to compare physicians for patients and third-party payers. Traditionally, the attribution of complications has been to the admitting physician. This may misidentify the physician "responsible" for the complication. This is especially difficult in trauma patients who have multiple providers. We hypothesized that the current mechanism for attributing complications in trauma patients is inadequate and will need to be modernized. All trauma admissions during a 12-month period were reviewed. Patients with single-system trauma were excluded. We reviewed our trauma database for mechanism of injury, complications, and readmissions. The trauma director and the medical director of our accountable care organizations reviewed all complications and attributed them to the appropriate health care provider. These were compared with the hospital decisions using the traditional definition. The trauma service had 1,526 admissions. After exclusions, 1,019 patients were reviewed. One hundred twenty-five complications occurred in 73 patients. Using the traditional definition, the acute care surgery service was assigned all 125 complications. Using the trauma director and medical director method, the neurosurgical attending accounted for 36% (45 of 125) of complications. The acute care surgery attending was responsible for 34% (43 of 125) of complications, and orthopedic surgery was identified as the causative factor in 22% (27 of 125). The remaining 8% (10 of 125) were attributed to various other services. Seven patients had unexpected readmissions. Most (6 of 7) of these were related to orthopedics. Hospital complications are now being assigned to individual surgeons. Which physician is responsible for each complication will be a controversial matter. Without a critical review process with physician input, up to two thirds of complications could be attributed incorrectly. The attribution process needs to be refined. Epidemiologic study, level IV.