National Congenital Heart Disease Audit (NCHDA) made two important changes since the last report.
1. Survival data. This relates to a revised statistical methodology, which has enabled us to capture and report an expanded breakdown of 30-day outcomes by age group for all procedures, as well as now reporting 83 Specific Procedures (previously 72 procedures). As a result, the analyses do not correspond to existing data displayed on the website, so we have placed all 2014-17 analyses on separate tab and the 2013-16 analyses is now available via the ‘Previous Analyses’ tab.
2. Antenatal diagnosis: This aspect of analysis has moved from procedure based counting to patient based counting. Previously antenatal diagnosis looked at the number of procedures in infancy where an antenatal diagnosis had been made. The Audit is now using the more appropriate metric of how many infants had had an antenatal diagnosis irrespective of how many procedures they may have had in the first year of life (excluding isolated procedures for a secundum atrial septal defect or persistent patent arterial duct, as these cannot be diagnosed before birth). This means that the same patient, who may have had more than one procedure in infancy, is no longer counted more than once.
It is important to also understand that these figures are probably an underestimate of the national antenatal detection rates as they do not take into account four scenarios: 1. fetal deaths that may occur during pregnancy (spontaneous or termination of pregnancy); 2. A perinatal death before a procedure was possible; 3. Less severe malformations that did not require a procedure in infancy; and 4. Where a decision is made not to intervene due to the complexity of the heart abnormality or associated comorbidities (compassionate care).
Historically, the Congenital Audit has looked at the overall detection rates of congenital heart disease in all infants requiring an intervention, and for the first time this year we have also examined the success of antenatal screening to detect two contrasting specific heart malformations (please see antenatal web pages for more detail :
· hypoplastic left heart syndrome (HLHS) - with a functionally single ventricle circulation and,
· transposition of the great arteries with an intact ventricular septum (TGA-IVS).
Our approach to survival data analyses:
It is absolutely essential to understand the way we analyse survival outcomes for patients undergoing procedures for congenital heart disease, in order to understand and interpret the analyses correctly.
In earlier years we used a standard statistical approach to survival analysis, in which follow up after a procedure stopped if the patient had a further procedure. In this case, the patient’s ‘life status’ was only related to the most recent procedure.
Because so many patients with congenital heart disease have multiple procedures, this standard approach of censorship resulted in a substantial proportion of our cases’ follow up appearing as “re-operated” rather than “alive, dead or unknown”. These re-operated cases were duly and properly ‘censored’ (left out of) our survival analyses with the result that the survival graphs contained fewer procedures than the total number of procedures actually carried out in each centre.
From feedback, NCHDA understood that what professionals and the public really want to know is how many patients survive after a given procedure, irrespective of whether they have further procedures, or not. In order to provide this information, we removed this censorship from our analyses. This results in complete outcome data for patients undergoing each procedure being included in the survival analyses and graphs, even if the patient has had multiple procedures (providing we know if they are alive or not).
This approach is much simpler and gives a more complete picture of outcome after a specific procedure. However, it is not standard statistical practice because it results in multiple procedures for a single patient being given the same life status. So, if a patient has one procedure followed within 30 days by a second procedure and survives, the outcome from both procedures will be “alive”. If the patient dies within 30 days of the original procedure, the outcome from both procedures will be “dead”.
If congenital heart patients or their families need more information about how to interpret NCHDA data, we suggest they contact Children’s Heart Federation email@example.com or 0300 561 0065.