Clinical outcomes are broadly agreed, measurable changes in health or quality of life that result from our care. Constant review of our clinical outcomes establishes standards against which to continuously improve all aspects of our practice.
About the cardiology and cardiac surgery services
The cardiology and cardiac surgery services at Great Ormond Street Hospital (GOSH) provide medical and surgical treatment to children with congenital and acquired heart disease. The cardiology service sees children that are treated from birth until adolescence, when they are transferred to the Grown Up Congenital Heart Unit (GUCH). The team provides a tertiary referral service both nationally and internationally and performs the largest number of paediatric cardiothoracic operations in the United Kingdom. The cardiology and cardiac surgery services also manage children within a group of highly specialised services, namely extracorporeal life support (ECLS), heart and lung transplantation, bridge to transplantation, tracheal surgery and the treatment of pulmonary hypertension.
Clinical outcome measures
1. 30-day survival for paediatric cardiac surgery (overall, expected and prediction interval)
The 30-day survival rate for paediatric cardiac surgery is a nationally accepted benchmark that is used to judge outcomes. It has to be born in mind, however, that the outcomes should be considered in the context of case mix severity. Furthermore, that 30-day outcome is a relatively limited measure of outcome, with longer term survival and other measures of morbidity being important to consider. The GOSH cardiac team is completely committed to developing other means to monitor outcomes in children with heart disease.
In the three years 2014 to 2017, there were over 2000 cardiothoracic operations performed in our unit, of which 99.2% of patients survived to 30 days. When these outcomes are benchmarked using the Partial Risk Adjustment in Surgery (PRAiS2) model, the results are better than expected based on the confidence limits selected by the National Congenital Heart Audit (NCHDA).
The data are shown in more detail below. For those readers interested in the results for individual specific operations, these can be found at: the National Institute for Cardiovascular Outcomes Research Congenital Heart Disease website.
In 2016 a new website, developed together with parents, was launched that explains what the survival rates for hospitals are and what they mean.
|Cardiorespiratory and Intensive Care Unit - Cardiac Surgery||April 2014 to March 2017||April 2013 to March 2016|
|Actual 30-day survival rate||99.20%||99.20%|
|Expected survival rate using PRAiS||98.00%||98.00%|
|95% prediction interval for observed survival rate||(97.4%, 98.6%)||(97.3%, 98.6%)|
|Ratio of observed survival rate to expected survival rate||1.011||1.012|
|95% prediction interval for ratio of observed survival to expected survival rate||(0.993, 1.006)||(0.993, 1.006)|
2. Our annual Variable Life-Adjusted Display (VLAD) plot for paediatric cardiac surgery outcomes
The following VLAD plot shows the trend in 30-day outcome of all cardiac surgery patients under 16 years old during 2015-16, benchmarked against expected based on the Partial Risk Adjustment in Surgery (PRAiS2) model. The number of procedures carried out and the number of deaths within the year are written at the top of the plot. Using the national risk adjustment method for paediatric cardiac surgery, the VLAD plot displays how many fewer (or more) deaths there are over time compared to “what would be expected”. As some readers may be less familiar with VLAD plots, which are now used in all children’s cardiac programs in the UK for quality assurance, we have added some information to guide you through interpretation below.
What would be expected?
We use a risk model (Rogers et al, The Annals of Thoracic Surgery, 2017) to estimate the risk of death, m, for each patient, taking into account risk factors such as procedure, diagnosis, age and weight.
Interpreting the VLAD chart
Each point on the VLAD chart represents an episode of care (the first surgical procedure for a child in a 30-day care period). If the 30-day outcome is a survival then the VLAD plot goes up by m and if it is a death the VLAD plot goes down by (1-m). The vertical axis is the total number of (expected – actual) deaths: when this is positive (negative) there have been fewer (more) than expected deaths.
- A run of survivors will cause the VLAD plot to go up and a run of deaths will cause it to go down.
- Over time, if outcomes are as expected by the risk model, the end of the VLAD plot will tend to be close to zero. Ending close to zero is not a sign that things are not going well! The risk model essentially benchmarks the unit’s outcomes against recent national outcomes in paediatric heart surgery. Despite this being one of the most complex areas of surgery and lifesaving for the children involved, the UK programme has excellent outcomes with very low mortality rates. So typically, m, the estimated risk of death for a patient is small (e.g. about 85 per cent of GOSH patients have estimated risks of 0.1 per cent to 5 per cent, and the highest risk is about 20 per cent for the most complex procedures such as some Norwoods). This means that the VLAD will rise much more slowly for a run of survivors than it will fall for a run of deaths (but of course there are many more survivors than deaths).
What is a VLAD most useful for?
- Spotting trends in outcomes (whether positive or negative) that might prompt discussion.
- A visual aid to gain an overall perspective on how things are going. The VLAD plot is not intended to judge outcomes, nor does it provide statistical control limits. Any risk model can only partially adjust for risks associated with any individual child.