Back in the summer I wrote about an unusual Commons statement made by the then Health Secretary, Jeremy Hunt, that some 450,000 women over a near ten year period had not been properly invited for their final breast screening and this error had potentially led to hundreds having “had their lives shortened” as a result. I was, in case you did not read the earlier post, somewhat incredulous that such a major and shocking process failure had happened at all and gone undetected for so long in what is the largest and longest-running health screening programme in the world, particularly because Hunt stated that the error was due to a “computer algorithm failure”. Naturally, the impact of his announcement attracted much Press comment & criticism and the impact on the affected women and their families, many learning of this bureaucratic failure in the Press or via a standard letter well after the event, might have been life-changing for many families, particularly those who may since have contracted or even died of the disease. This was a major incident and Mr Hunt apologised repeatedly in his statement on behalf of all the DoH agencies involved.
The independent breast screening review Hunt commissioned to look into the incident has duly completed and its report makes for interesting, if somewhat depressing, reading. Basically, it found there was no ‘incident’. There was no major failure of the screening programme, there was nothing wrong with the algorithm and the IT systems were generally doing what they were designed to do. The reasons behind Mr Hunt’s statement were basically due to a huge muddle by those charged with administering the programme or, in the words used in the report, ‘…the Health Secretary’s announcement was made following advice based on an incomplete understanding of what had happened.’ So primarily, it was a spectacular failure of management and governance.
The report’s Executive Summary states:
‘The Secretary of Health was advised that women had not been invited to their final breast screenings between the ages of 70 and 71 because of a problem with a computer algorithm. In fact, the reason that the women were not invited between those ages is that the way the breast screening programme had been run since the late 1980s meant they had already received their final screening three years earlier. The misunderstanding arose because of a Specification document written in November 2013 which stated that women should be invited for screening “within 36 months of their previous screening, until they reach the age of 71”. We believe this document [written by the DoH and NHS England] was based on a misunderstanding of how the programme was being delivered in practice. The vast majority of women who were told earlier this year that they may have missed an invitation to screening were only affected under the definition of this document, and of them some (those eligible for final screening before the Specification was written) should not have had the document applied to them at all.
For many years, from the beginning of the breast screening programme in the late 1980s, the specific age range of women to be invited was not set out in sufficient detail, and there was variability across breast screening units. In the hand-over of responsibilities to the newly-formed Public Health England and NHS England in 2013 we have found no evidence that there was a shared understanding of how the screening programme was being delivered. A new Service Specification was written to provide specific instructions against which the programme could be commissioned and quality assured, but it included a level of specificity which did not align with the IT system then in use, and was not consistently implemented by the breast screening units. We believe this to have been written in error. It appears that no-one in the Department of Health and Social Care, Public Health England or NHS England realised that this was not consistent with past policy or understood that this change then caused a misalignment between policy, delivery, the IT system and the AgeX trial algorithm. This lack of understanding of the root cause of the confusion continued even as the incident was being investigated and announced.’
The report further states: ‘The Review team is surprised that an incident was declared before the full extent of the problem had been understood or uncovered. The impact of this appears to have changed the focus of Public Health England to addressing the problem they thought they had discovered. So, despite the lack of understanding of the issue, work shifted in this period to how to respond operationally to the perceived problem.’
Comments relating to the programme’s failed governance include:
‘There is no strategic oversight to ensure the separate strands of the breast screening programme are functioning as a whole – there is no senior responsible owner. The quarterly section 7A agreement meetings are too large in scope to address individual issues in a robust way.
The November 2013 Service Specification (and its yearly updates) is central to the incident, but none of the tripartite organisations have been able to confirm how it was signed off. What is clear is that it was not implemented properly. This was a failure of governance.
NHS England failed to hold breast screening units to account for delivering against the contracts, which were based on the Service Specifications.’
As to how the ‘incident’ was dealt with, it states:
‘The Review has found that the initial handling of the incident was not sufficiently gripped within Public Health England, which resulted in delays in fully understanding the nature of the incident and ultimately led to a response that did not reflect the precise failings in the service. In particular, we would have expected the most senior management to have been involved in the initial handling to provide overall direction to the response and to ensure those responding to the incident had the necessary resources available to complete a thorough review at pace. We did not find evidence that such an approach was adopted. Public Health England has an excellent track record of responding to public health incidents but the approach adopted for this incident meant that some three months elapsed between the initial realisation that an incident had occurred and the Secretary of State being advised to make a statement to Parliament. Not only is this arguably too long a period of investigation but, worse, the failure to fully understand what had occurred led to the Secretary of State being advised to make a statement that we now know to have lacked the complete evidence and analysis. This was due to assumptions being made that the November 2013 Service Specification amounted to policy, and a lack of understanding of what was being delivered by the breast screening programme in practice.
The Review has also found that there were failings in the handling of the data which resulted in the severity of the incident being exaggerated publicly and in thousands of women being warned unnecessarily that they may have been affected by a failure in the IT system.
The Review believes that the way in which the public was alerted to the incident and the lack of communication with screening units and the devolved administrations placed an avoidable strain on resources. ‘
“Mr Hunt told Parliament in May “a number of linked causes” had been identified “including issues with the system’s IT and how age parameters are programmed into it” and regional “variations in how local services send out invitations”. He also pointed to a major trial known as “AgeX”, which was set up in 2009 to test extending screening to age 73, indicating that the trial was central to the problem.
The review by management consultants PwC for Public Health England looked into the causes of the problem, which in May saw Jeremy Hunt tell Parliament that some women were likely to have died as a result of missed scans – potentially up to 270.
The PwC report states that a small but significant mistake was made in 2013 in a new specification of who should be screened and when under the programme, and has gone uncorrected for five years. The programme itself continued unchanged.
The PwC paper, completed in June, suggests there has also been ongoing confusion during this year and that the number of women who missed a screening invitation may have been substantially overestimated.
Robust governance
The report says it is unclear exactly how the error was introduced because of “a lack of clear documentation and comprehensive audit trail as to why and how decisions were made”. It was also unclear whether the new, apparently mistaken, specification was approved by the “tripartite” which took over national running of the screening programme under the 2013 Lansley reforms: PHE, NHS England and the DHSC.
Although PHE was put in charge of drawing up the new specifications, PwC said: “There does not appear to have been a robust governance process, across the tripartite, surrounding the approval of the specifications.”
It added: “The priority must be to review the current [screening programme specification] and update it to align with the practical realities of how the system operates.”
The report, entitled Project Crystal Briefing Paper, is part of the agency’s ongoing internal inquiry into how its national screening programme caused a national incident.
The PwC review is expected to contribute to a wider internal review by PHE which is ongoing. The DHSC also commissioned an independent inquiry into it. Both are due to be published in November.”
Quite why three reviews of this incident were necessary, PwC, PHE and the independent review commissioned by DHSC is unclear to me.
PwC’s report itself indicates the AgeX system was not the primary cause, instead focusing on the misunderstanding surrounding the difference between the national specification, and how the screening programme operates and has long operated.
The PwC report says the screening programme lost “a significant amount of corporate memory” in 2015 when a “longstanding lead for the cancer screening programme left PHE resulting in a lack of detailed understanding of how the [breast screening programme] operates at a granular level”.
“There is a lack of corporate knowledge across the screening team as to how and why historical decisions were made,” the paper added.’
Executive apology:
Apart from Jeremy Hunt’s apology in his original Parliamentary Statement I can find no apology or explanation for the incident from either PHE or NHS England online. Considering the specific report criticisms one might have expected the health agency heads, including PHE chief executive, Duncan Selbie, to make some public comment, given three separate investigations were undertaken. Selbie writes a weekly blog, presumably for the benefit of PHE staff, but all I could find in them related to this incident is the following in his year end message: