Abstract
In the past, surgical (postoperative) outcome was reported as mortality and morbidity rates. While the mortality rate, especially when expressed for a specified time period, for example, 30 days, is non-controversial, the morbidity rate is less informative and subject to different interpretations. Moreover, it does not differentiate between mild and severe complications. The outcome for a patient with superficial thrombophlebitis is very different from a patient suffering from major pulmonary embolism. Besides, one complication may lead to a series of subsequent complications. When all complications are expressed in a table form, the total number of complications often exceeds the number of patients distorting the reader’s ability to evaluate the true morbidity. Finally, fatal complications are sometimes listed among complications, and sometimes left out from the list of complications (reported among survivors only).
It is also important to differentiate complications (that are potentially avoidable) from “failure to cure” and “sequelae.” While the former refers to the inability to achieve the goal of the operation (e.g. relieving intestinal obstruction in a patient with massive carcinomatosis), the latter is an inevitable result inherent to the procedure (e.g. diabetes after total pancreatectomy). Neither of these should be listed as a complication, whereas all other negative events should (1).
A major improvement in reporting complications was the publication of the Clavien–Dindo classification system where complications are graded by severity and required intervention from 1 to 5 with 5 being death (2, 3). Because the literature reports only the most severe complication, it does not take into account less severe complications and fails to represent the true overall “morbidity burden” of a procedure. A recently published “Comprehensive Complication Index, CCI” is designed to summarize the overall morbidity after surgery (4). It is calculated as the sum of all complications that are weighed for their severity and expressed in a continuous scale from 0 to 100 in a single patient. It can be computed with a formula available at http://www.assessurgery.com.
But not all patients are equal, and the risk of postoperative complications even for a same procedure is affected by patient and disease factors. Co-morbidities and the disease acuity, including the stage of preoperative organ dysfunction can have a profound effect on postoperative outcome. With the increasing trend of public reporting of hospital- or even surgeon-specific outcomes, it is important to include a proper risk adjustment system to account for the case-mix. Whether all this could fit into a single formula with external, inter-hospital validity requires further studies.
The ability to perform risk-adjusted outcome prediction could also be used to help the preoperative assessment and even selection of the type of procedure (if any) and required resources, such as the need for postoperative intensive care. A method to assess the probability of survival in trauma patients, the Trauma and Injury Severity Score (TRISS) methodology, takes into account the degree of physiological impairment on arrival, the anatomical severity and location of the injuries, and age (5). It does not, however, predict non-fatal complications or the “morbidity burden.” A similar predictive model could be useful for all patients undergoing surgical procedures, especially those requiring emergency surgery.
What would be the benefits of a risk-adjusted overall complication score? First, it would provide a tool for benchmarking surgical outcomes between different institutions worldwide. It could also be used to compare outcomes of individual surgeons performing a same, standard procedure, although it should be applied with great caution, because the outcomes obviously depend on many other factors than the surgeon’s skills, such as the type and quality of anesthesia, intensive care services, and the overall ability of the hospital to “rescue” patients after severe complications (6).
Finally, because postoperative complications are the strongest indicator of in-hospital costs, all efforts to lower the rate of postoperative complications have a significant savings capacity (7). Furthermore, if the risk-adjusted outcome index would be coupled with the amount of resources spent at the hospital level, this could be used as a measurement of efficiency, or at least could give a better idea of what the tax-payers get for their money.
Although surgeons have come a long way in honestly evaluating their results, there is still room for improvement. In the increasingly critical public scrutiny of surgical outcomes and costs, it is better that we create a solid, reproducible, and fair outcome assessment system. Otherwise, someone else will.