As the Junior Doctor’s contract dispute continues with no end in sight, I revisit the saga of certain claims made by the Health Secretary last year – and find, as always, nothing is quite as it seems.
There have been numerous debates surrounding the now-infamous Jeremy Hunt speech in July last year, when he claimed that 6,000 additional deaths were attributable to NHS staffing levels at the weekends and that mortality increased 15% on a Sunday compared to a Wednesday – statistics which he used in the context of arguing for a change in the Junior Doctor’s contract, to enable a “7-day NHS”.
It has always been speculated that the statistics he used must have been taken from an article (authored by Sir Bruce Keogh, Nick Freemantle, Domenico Pagano and several others) for the British Medical Journal into the association between the day of hospital admission and mortality, prior to its release in September 2015.
In a report for the BMJ journalist Abi Rimmer, after speaking to the authors of the report, numerous people involved at the BMJ, the Department of Health and NHS England, tentatively concluded that the “6000 figure was derived from the 2012 paper, both by NHS England and by Deloitte during its work for NHS England” and that NHS England had “shared this figure with the Department of Health (DoH). It said that the figure had been discussed widely within the organisation, and it had been calculated both internally within NHS England and by Deloitte as part of its work for NHS England on seven day services.”
So far? Interesting. This aspect of her investigation partially stacks-up; all parties involved claim that their 2015 report was not shared with anyone prior to publication, and the 2012 paper in question was a collaboration between Keogh and the University Hospital Birmingham (UHB), which used “Hospital Episode Statistics” (HES) to conclude that there may be a link between increased mortality and the weekend. But this does ignore that fact that originally it was claimed that the information came from the 2015 BMJ article, pre-publication.
But what of the “15% increase in mortality on a Sunday” statistic? This is where it gets complicated – and it’s also where I believe she’s missed several vital points.
The “story” goes that Keogh claimed that the 15% figure was also derived from the 2012 paper – although originally it was 16%, until the UHB Research Group (the same people who authored the BMJ report) were asked to update the findings using 2013/14 HES-ONS data on mortality in February 2015 (and this same data was used in the report published in September).
After much flip-flopping NHS England told the BBC that “top level findings from the article, including the finding that mortality was 15% higher on a Sunday, were shared by Keogh with policy colleagues at NHS England, the Department of Health, health unions, Royal Colleges, the Health Select Committee, and Deloitte before publication.” They also said “Keogh also shared the top level findings with the Review Body on Doctors’ and Dentists’ Remuneration (DDRB) in March 2015” (all this after originally denying that he had shared any information).
Rimmer draws no real conclusions from this information, except that it is seemingly approved practice to share interim findings from review work before its publication, and that Keogh and Freemantle (who was the lead author of both the 2012/2015 reports) claimed much the same thing.
But there are several worrying points that Rimmer and the BMJ’s investigation into this fail to address.
The HES-ONS linked data to ascertain the 15% mortality figure was only released to the UHB team between January and May 2015 by the Health and Social Care Information Centre (HSCIC), where the UHB specifically said that: “HES-ONS linked data specifically will be used within this work to look at outcomes analysis and form analytical overviews relating to post discharge mortality”, stating “Any analysis produced using ONS data will not be made available to non-NHS organisations” – this is the terms of its licence to have this data.
This correlates with a response from Pagano on the BMJ website, who commented that “The analysis for mortality by days of the week was conducted under the governance umbrella provided by the (HSCIC) licenses. For the avoidance of doubt, our work published in the BMJ was not conceived, executed, or published in collaboration with the Secretary of State for Health.”
But an email between Deborah Williams of NHS England and Pagano, Freemantle and Keogh, however, clearly states the Deloitte had been “sighted” on the provisional paper for their “modelling work”; The DoH told Full Fact the 15% figure came from Keogh; NHS England told the BBC it came from HES statistics from 2013/14. The UK Statistics authority questioned the Head of Analytical Service of NHS England on the fact that statistics had been released without the source being revealed (albeit in an advisory way), however no response to this can be found.
There is also the matter of a Guardian expose from February this year. The article, based on a leaked Department of Health report, shows that a “7-day NHS” may not have any impact on weekend mortality rates. The DoH report “Seven-day NHS – update on progress and plans”, from January, says that it will cost an additional £900m to fund Hunt’s plans after taking out the “benefits such as reduced length of stay and reduction admissions”. But crucially to all this, is that the findings and specifically £900m figure came from an NHS England-commissioned study in the cost of a “7-day NHS” from 2015 – by Deloitte.
This study appears to have been sighted in a paper, “NHSPRB & DDRB 2015/16 Special Remit Evidence from NHS England”, released in January 2015 where evidence is given to the pay review bodies over the possible staffing costs of a “7-day NHS”. The study is referred to as “Deloitte interim findings”, and cites that evidence used was drawn from (amongst others) the HSCIC and Department of Health. The Deloitte study is not publicly available. This NHS England paper also refers to the previously contested “16% increase in Sunday mortality rates” figure from the 2012 Keogh and UHB report.
Worryingly, I can find no record on any HSCIC release documents of Deloitte ever being granted access to HES-ONS linked data.
The links between Deloitte and the UHB are also interesting. Deloitte not only audit the hospitals accounts, but they also provide “Counter-Fraud Services” and are permitted to analyse the hospitals Quality Account in three areas – one of which being 28-day readmission figures, which gives them access to usually confidential data.
Furthermore, “Cure Leukaemia” is Deloitte Birmingham’s nominated charity; a charity which has as its Trustee Richard Turnball, a Deloitte partner who advises David Cameron and a charity which operates out of the UHB newly-founded “Institute of Transitional Medicine” (ITM), via the co-founder of the former and the Transitional Director of the latter, Professor Charlie Craddock CBE – who appears rather friendly with Deloitte (who also conduct the ITM’s external reviews).
Little is mentioned, either, of Daniel Ray – another of the 2015 reports’ authors. Ray is the Chair of the “Health CIO Network” (a professional body for those in the data/informatics industry), and was also Director of Informatics at UHB until January of this year – cited as being involved in the development of the UHB’s ITM.
Interestingly, Deloitte released an NHS briefing in June 2015 on “Data Governance” – essentially a pitch to NHS Trusts on the services it provides surrounding analytics.
Are the links surrounding Deloitte and the UHB merely background noise? Possibly, but they do display the “six degrees of separation” between private companies and decision-makers within the NHS – which in the instance of the “leak” of data in question, seem rather pertinent.
Regarding some claims made by Rimmer’s report in the BMJ, the timescales appear bizarre, at best – if UHB were asked on the 5th February to redo their HES-ONS research, it seems implausible that they could have requested the data from HSCIC, assessed the information and had findings ready for Keogh to present to the DDRB in March.
Also, the first peer review was submitted at the end of June, and the statistical analysis was not returned until 23rd of July (analysis which questioned the report methodology behind the Wednesday and Saturday/Sunday comparisons, that is, why were figures fifths?) – so why was the top-line data still taken as accurate?
Overall, there remain unanswered questions and issues.
Why did David Cameron state in PMQs that Jeremy Hunt received the statistics he quoted in July 2015 directly from Keogh? Furthermore, if he did is this not a breach of the latter’s remit as a senior member of NHS England – supposedly at “arm’s length of Government”?
The BMJ say’s it is “not unusual” for research from papers to be shared before publication; fine – but what is unusual is how an external auditing company (Deloitte) appeared to have access to confidential HSCIC data, when their have categorically not been authorised to use it – and why did Deloitte deny that they had seen this information, when NHS England state on numerous occasions they did?
Furthermore, BMJ guidance makes it very clear that any data that is submitted to any government review/report pre-publication should be declared to the journal at acceptance – why wasn’t this done? They also seem to ignore the fact that Hunt used statistics which hadn’t (on the 16th July) been peer-reviewed. Is this normal BMJ procedure?
Why did the UHB allow mortality rates analysis (using HES-ONS data) to be “allegedly” widely shared, when it appears to be in clear breach of their licence to have such data?
Why were the initial confidential responses clear that the data Jeremy Hunt used had been provided by Deloitte, and then subsequently this statement thoroughly blurred?
Why have NHS England been colluding with the DoH in this whole debacle? NHS England is supposed to be wholly independent from the Department, as its raison d’etre clearly states – so the level of communication, and apparent attempts to “cover-up” Jeremy Hunt’s faux pas (as shown in several emails which specifically talk of “avoid(ing) undue criticism of both the DoH and NHS England) is staggering, if not highly worrying.
Why has Jeremy Hunt still not been questioned over using statistics which (at the time) had no verifiable source for the public to view, nor had been properly peer reviewed? This is a clear breach of the Code of Practice for Official Statistics, and a possible breach of the Ministerial code.
The BMJ article appears, from an outsider’s perspective, to be somewhat attempting to put this matter to bed. From where I’m standing, it does little but further the speculation as to whether there was systemic collusion, mismanagement of licenced data and ultimately scapegoating to cover-up and protect what appears to be another direct attempt to undermine the Junior Doctors in the eyes of the public.
My conclusion? Deloitte had access to data which they categorically were not supposed to have (which is probably a fairly standard occurrence) and used this data to strategically advise individuals in the Department of Health to further leverage the Governments straw man argument that a change in the Junior Doctor’s contract is vital for a “7-day NHS” in the eyes of the public.
Individuals in NHS England and the Department of Health were fully aware of this, however were caught “on the hop” when questions were asked over Jeremy Hunt’s 6,000 deaths and 15% mortality increase claims – hence the back-peddling, claim then counter-claim, and seeming inability to “get their stories straight”.
This whole issue also raises serious concerns about transparency within the NHS, as so many documents from third-parties (like Deloitte) are seemingly unavailable for external scrutiny.
More questions need to be asked, and further external investigations conducted surrounding this whole saga, as it serves to add another layer to the debacle of incessant Government propaganda that is the Junior Doctor’s Contract – which, with possible joint industrial action with teaching unions on the horizon, is a situation that is not going away.