FROM

Leslie Hamilton

Cardiac (adult and transplant) surgeon Freeman Hospital, Newcastle

Past President SCTS

Vice chair Safe+Sustainable Steering Group

 

Subject: “Safe and Sustainable”: the Jarman data
Date: Mon, 5 Nov 2012 20:40:58 +0000
From: Leslie.Hamilton

Dear Phil

 

We met at the very first national “Stakeholders”meeting when the review started. I have always enjoyed your MD column in Private Eye and have very much appreciated your support for the principles of the review. Your column in the current issue (Eye 1326) is no exception.

Nonetheless, I have significant reservations about Professor Jarman’s analysis of the outcome (mortality) data. I know he was a member of the panel for the Bristol Inquiry and am aware of his work through the Dr Foster organisation.

 

It would seem obvious to use outcome data to assess the quality of care when deciding which units to designate as the surgical units for the future. However the clinical Steering Group (who advised the JCPCT) was very clear that the data should not be used in this way.

 

In adult cardiac surgery we have a well established, internationally accepted risk stratification system (EuroSCORE) which was developed using a robust statistical analysis of a database containing tens of thousands of patients undergoing a small range of operations – this enables us to allow for patient and operation factors to give a predicted mortality so that outcomes can be compared in a fair manner. Indeed we have shown that mortality has fallen steadily over the years so that the score has had to be recalibrated – EuroSCORE 2 was presented at the annual meeting of the European Association for Cardiothoracic Surgery last week in Barcelona (http://www.euroscore.org/calc.html).

In paediatric cardiac surgery, we have long recognised that we needed a similar process to enable outcomes to be compared. However we have the opposite situation – a small number of children undergoing a large range (149 on the database but only the 50 commonest shown on CCAD) of procedures (with sometimes several procedures combined in an individual child). Also some of the higher risk procedures currently performed do not have a specific code in HES data.

 

Several international groups (including the one at UCL / NICOR) have been working to develop a scoring system which would allow for both case complexity and patient factors in paediatric cardiac surgery:

 

• · The Aristotle Complexity Score (http://www.aristotleinstitute.org/aboutScore.asp). I was involved with this group – it allows for most of the factors influencing outcomes but surgeons felt it was too complex! (http://ejcts.oxfordjournals.org/content/25/6/911.abstractand http://ats.ctsnetjournals.org/cgi/content/full/84/6/2027

 

• · The RACHS score (http://www.ncbi.nlm.nih.gov/pubmed/18377539) which divides the operations into 6 categories of technical difficulty and is similar to the Aristotle Basic Score (http://jtcs.ctsnetjournals.org/cgi/content/full/133/4/865)

 

• · More recently there has been further work between the American (STS) and European (EACTS) societies to produce a “STAT”score (http://www.chop.edu/service/cardiac-center/cardiac-outcomes/volumes-mortality-rates-complexity.html) but this is not too far away from RACHS.

 

Ultimately though these are empirical scores based on professional opinion. It was for this reason that the Steering Group advised the JCPCT not to use outcome data to assess units. We are currently seeing challenges to the decision making process used by the JCPCT – if outcome data had been used, dissatisfied units would have had endless ammunition with which to challenge the decision! It was felt better to ask Sir Ian Kennedy to assess the units against their ability to meet the standards proposed for the future (which he and his team did).

 

We know that there is significant variation in the complexity of cases undertaken by the current units. We know that neonates have the more complex operations and have the highest mortality – so using age up to 5 years (as Prof Jarman has done) does not make sense. We also know that primary repair at a young age is usually better but an initial palliative operation (a shunt?) will have a lower mortality risk initially – so initial outcome may look better. We also know that the hazard risk for surgery probably goes out to 90 days so using 30 day mortality can be misleading.

 

One of the strong arguments for having bigger centres with a higher volume of cases is that they will do a similar range of cases and the statistical analysis will be more robust. And we can look at outcomes other than mortality (parents are especially interested in neurological injury). Work needs to continue on producing a score which is helpful to parents and fair to surgical teams.

 

My real concern about the Jarman analysis and your conclusion is that parents will be mislead – if my grandchild needed heart surgery I would have no hesitation in going to Birmingham or Guys/St Thomas’s. So I think the JCPCT got it right. I can say this with no vested interest – I stopped doing paediatric cardiac surgery 5 years ago when the physical, emotional and mental pressures of a 1 in 2 rota got to me. Which is why I feel so strongly about seeing this through.

 

Yours sincerely

 

Leslie Hamilton

 

Cardiac (adult and transplant) surgeon Freeman Hospital, Newcastle

Past President SCTS

Vice chair S+S Steering Group
MY RESPONSE

Thanks Leslie

 

Ever since the Bristol Inquiry, UK experts have told me that it is too difficult to collect properly risk adjusted mortality in paediatric cardiac patients, that the heterogeneity of the population make it not worthwhile, that the numbers are too small, that there is no validated model and so on and so forth.

Odd then that New York state have been collecting and publishing this pooled data in 3 yearly batches since 1997, with the latest report published in October 2011 (see http://www.health.ny.gov/statistics/diseases/cardiovascular/index.htm, and scroll down to ‘Pediatric Congenital Cardiac Surgery in New York State’near the bottom (where paediatrics always is)).

 

I have already corresponded with David Cunningham on this, and put the correspondence up on my website where I shall put your response too. My own view is that first we need to get the data, models and analysis as good as they possibly can be – and that means sharing CCAD data with Brian Jarman, Paul Aylin and their team to allow them to contribute to this process. I think it is appalling that their request for the data has been denied – had it not been, there may well have been a more complete analysis, and Guys and Birmingham may have had a lower mortality ratio. But they can only work with the data and codings that they have. Professor Jarman has sent his data to Bruce Keogh and the DH twice, with no response. Time to converse and share, please.

 

Although the final analysis of pooled data may not be perfect and there may be mitigating factors, I believe it should be in the public domain – as per New York – over an agreed period of time, say three years. How else will you spot another Bristol? Or Oxford? And be sure to act on it? In Bristol and Oxford, the data simply was not acted on – it took whistleblowers to expose the scandals. You could argue that publication of data alone is not enough – the Mid Staffs scandal tells us that. You may have good arguments for Birmingham and Guys being high – and you should share them with Professor Jarman. I believe parents have a right to see this data and have it explained to them. But if CCAD won’t share the data with DFI, you shouldn’t be surprised if the analysis isn’t as complete as it could be.

 

I understand why the Steering Group advised the JCPCT not to use outcome data to assess units for the Safe and Sustainable review, but I won’t agree with it until I’ve seen the best statistical analysis of the pooled data. Do we have any idea how many excess deaths there have been in the 11 years since the Bristol inquiry? Have there been any significant statistical outliers over that time? HMSRs aren’t perfect either, but I feel a lot safer knowing they’re out in the public domain.

 

Yours Sincerely

 

Phil Hammond