SAN DIEGO–Uniquely derived mortality scores, based exclusively on adult congenital data, can be used to create risk models that more accurately and reliably compare mortality outcomes across institutions with differing case mixes than do existing empirically based tools, according to researchers.
“Several established tools exist – including the STAT [Society of Thoracic Surgery–European Association for Cardiothoracic Surgery Congenital Heart Surgery Mortality] score, and the Aristotle Basic Complexity [ABC] score; however, none of these are age-specific,” Dr. Stephanie M. Fuller reported at the annual meeting of the Society of Thoracic Surgeons.
“Prior analyses have demonstrated varying degrees of discrimination using existing tools as supplied to adults; however, no empirically based system of assessing risk specifically in adults exists,” said Dr. Fuller of the University of Pennsylvania, Philadelphia.
Dr. Fuller and her colleagues conducted an exploratory analysis with a twofold purpose:
1. To compare procedure-specific risk of in-hospital mortality for adults, compared with an aggregate of all age groups and compared with pediatric patients using the Society of Thoracic Surgeons Congenital Heart Surgery Database (STS-CHSD).
2. To develop an empirically based mortality score unique to adults undergoing congenital heart surgery that can be easily applied to assess case-mix across hospitals as well as procedural risk.
Dr. Fuller and her coinvestigators estimated mortality risks based on more than 200,000 indexed operations from 120 centers found in the STS-CHSD between January 2000 and June 2013, eventually whittling the cases down to 52 distinct types and groups of operative procedures and 12,842 operations on adults aged 18 years and older for inclusion in the study.
All 52 procedural groups were ranked in ascending order according to their unadjusted mortality rates, at which point Bayesian modeling was used to determine the best estimates of procedural mortality risk to adjust for the relatively small numbers in some of these groups. Subsequently, each procedure or group of procedures was assigned a numerical score based on the estimated risk of mortality from 0.1 to 3.0 – this score was called the STS Adult Congenital Heart Surgery Mortality Score.
The researchers’ validation sample was done “by looking at those performed during the most recent year, July 2013 through June 2014, and all cases from this year were separated into the same 52 procedural groups,” explained Dr. Fuller. “These were analyzed comparing the adult congenital heart surgery score to both the STAT score and the ABC complexity level; [however] not all procedural groups were common to both STAT and ABC level analysis, so validation was performed only for those [that] were common in both groups.”
Dr. Fuller and her colleagues found that the number of adult congenital heart procedures per year increased drastically from 2000 to 2012 – from 85 to 1,961, respectively – and the rate of survival over this period was 1.8% without adjustment to the Bayesian model. After adjustment, model-based estimates for mortality in each of the main procedure types were determined: 7.3% for Fontan revisions, 7.2% for heart transplantations, 6.6% for lung transplantations, 6.0% for Fontan procedures, 5.5% for coronary artery intervention, 0.4% for partial anomalous pulmonary venous return, and 0.2% for repair of atrial septal defect operations.
“Future directions [for this study] include, primarily, augmenting the data sample, which can be done by incorporating data from either the European Association for Cardiothoracic Surgery database, or by incorporating data from the STS Adult Cardiac Surgery database,” said Dr. Fuller, adding that she and her colleagues would also have liked to include clinical characteristics and risk factors of patients, such as presence of genetic syndromes and medical comorbidities in the adult population.
Dr. Fuller did not report any financial disclosures.