Invited Commentary

Abstract

Reoperation is an important outcome to track and understand and is increasingly being used as a quality metric. For the present study, the authors used data from the United Kingdom National Congenital Heart Disease Audit Database to conduct a retrospective cohort study of reoperations occurring within 30 days of index operations [1Fiorentino F. Stickley J. Dorobantu D. et al.Early reoperations in a 5-year national cohort of pediatric patients with congenital heart disease.Ann Thorac Surg. 2016; 101: 1522-1529Abstract Full Text Full Text PDF Scopus (10) Google Scholar]. They found that, like another common quality metric, mortality, the incidence of reoperation was low, with only 3.5% of patients being affected. They also found that the patterns of reoperation were dictated by the index operations. Very young patients undergoing palliative procedures were most likely to undergo reoperations, and these patients experienced a higher mortality rate than patients who did not undergo reoperation. The low overall incidence of reoperation and its distribution across the spectrum of congenital heart operations influences its utility as a quality measure. A reduction in the incidence of reoperation would affect a minority of patients. Also, the clustering of reoperations in a small number of index operations limits one's ability to use reoperation broadly as a measure of quality. Thirty-one percent of reoperations occurred after just seven index operations. Thirty-seven percent of reoperations were redo arterial shunt procedures, and an additional 9% were arterial shunt procedures after Norwood procedures. This may reflect the incidence of shunt thrombosis and be an area for directed improvement. Perhaps reoperation would be best used as a quality metric for selected groups, whereas outcomes such as functional and neurologic status that are experienced by all patients are used to measure success in the broader population. In addition to calculating the incidence of reoperation, the authors hoped to determine whether operations subsequent to the index operations were planned or unplanned. However, they decided that this was not possible because the data did not allow for a retrospective understanding of the clinicians’ thought processes that resulted in the operation sequences. This issue is common to large databases, both administrative and clinical. Instead, reoperations were designated as “related” or “unrelated.” Related was defined so as to include operations technically similar to the index operation. Unrelated was defined as both technically dissimilar and not a direct consequence of a surgical step of the index operation. The authors appropriately noted the difficulty in determining whether a reoperation was planned or unplanned from a retrospective viewpoint; however, how useful is the concept of related/unrelated for measuring quality? Successful surgical outcomes rely on both planning and execution. Not only must A be sewn to B adequately, but the decision to sew A to B must be appropriate. The related/unrelated concept may capture most errors related to the former, but it may miss those related to the latter. As defined by the authors, the index procedure of “Norwood operation” and the reoperation of “isolated coarctation repair” are unrelated. However, this reoperation could be the result of a technical error that caused a residual coarctation at the distal aortic anastomosis. The index operations “aortic valve repair” and “interrupted aortic arch repair“ were both considered unrelated to the reoperation of “Norwood procedure”; yet, these reoperations could both result from the failed choice of a two-ventricle repair in a marginal patient. In the second example, there may have been no technical errors, but the reoperations were necessary as a result of unsuccessful decision making. These examples highlight the difficulty of using the related/unrelated concept to determine whether or not a reoperation is the consequence of an index procedure. What we want to know is whether a reoperation was planned or unplanned. Rather than spending too much time creating an imperfect construct, effort should be directed toward modifying databases to include the information necessary to answer the question of interest. Early Reoperations in a 5-Year National Cohort of Pediatric Patients With Congenital Heart DiseaseThe Annals of Thoracic SurgeryVol. 101Issue 4PreviewThe exact types, frequency, and consequences of early congenital cardiac reoperations are not well known. We aim to describe them and evaluate the potential of early reoperations as a metric for quality of care. Full-Text PDF