Missing the Signs: How to Trust Data, Dashboards and People. Part One
The data to spot failures in care has existed for years. The question is whether anyone looked at it, believed it, and had the courage to act on it. In this session from the Connected Health & Care Summit 2026, Alex Kafetz, who works part-time for RLDatix and is a non-executive director at the Care Quality Commission, walks through a series of NHS scandals and asks a direct question for each: was there data that could have identified the problem earlier, and why was it ignored?
Watch Alex Kafetz draw on Mid Staffs, Shipman, Ian Paterson and his own experience on public inquiries to show how data could have spotted failures earlier, and why board culture is the deciding factor.
What the session covers and the key lessons from each case
This session walks through three major NHS scandals and examines the data that existed at the time. The key lessons were:
- Mid Staffordshire was flagged as a mortality outlier every single year for an extended period. The board commissioned a paper to challenge the methodology rather than investigate the underlying concerns about care.
- Harold Shipman could have been detected earlier through cumulative sum analysis. A time-of-death chart showed a significant spike at 3pm, coinciding with his home visits. Professor David Spiegelhalter’s analysis makes it visually obvious
- Ian Paterson harmed over 1,000 people. Financial data showed mastectomy patients being sent for radiotherapy against NICE guidelines. Nurses knew. The first chief executive dismissed the data. The second suspended Paterson immediately with the same information
- In every case, the data existed. The session argues that what was missing was the board culture and curiosity to act on it.
What the data showed and why it was ignored
Alex Kafetz traces the thread that connects each failure. At Mid Staffordshire, the hospital flagged on standardised mortality data year after year. Patients were meeting in a local cafe to share their experiences. The board chose to commission an academic paper challenging the methodology rather than investigate the care. Eventually, a public inquiry chaired by Robert Francis concluded there were horrific failures. Alex Kafetz argues that most of the inquiry’s recommendations remain unimplemented more than a decade later.
At the Paterson inquiry, Alex Kafetz was asked to examine whether data could have spotted the surgeon earlier. The answer was yes, both clinically and financially. But the first chief executive decided the data was inconclusive. The second suspended Paterson the next day with no new information, just a different lens. Alex Kafetz draws a direct parallel with the Ronan Farrow account of Harvey Weinstein: “Stories of abuse by powerful people are also a story of a failure of board culture.”
What this means for boards today and who it is relevant to
Alex Kafetz is direct about what boards should be doing. He argues there are three pieces of information that every board member and non-executive director should know: the risk-adjusted mortality rate, the friends and family test results, and the staff survey question asking whether staff would want their loved ones treated at the hospital where they work. That final measure was reported in the session as averaging approximately 65%, meaning more than three in 10 NHS staff would say no.
He maps every hospital in England across four mortality measures and shows substantial unwarranted variation. Very few hospitals perform well across all four, and the pattern is not consistent, meaning boards cannot assume that good performance on one measure means safety across the board.
On the Paterson inquiry, Alex Kafetz notes that he believes key recommendations have not been implemented. There is still no database showing how many times a surgeon has performed an operation. Patients cannot verify a surgeon’s experience before consenting. He describes this as a failure of informed consent. Recent investigations at Salford Royal, Great Ormond Street and elsewhere show that rogue operators continue to exist.
This session is relevant to board members, non-executive directors, medical directors, patient safety leads, governance teams, regulators, quality improvement leaders and anyone responsible for data-driven oversight of care quality.
FAQs
The session is presented by Alex Kafetz, who works part-time for RLDatix and is a non-executive director at the Care Quality Commission. He previously worked at Dr Foster, where he was involved in producing the Good Hospital Guide and the mortality data that helped surface the Mid Staffordshire scandal. He was asked to join the Ian Paterson inquiry and gave evidence to the Mid Staffordshire public inquiry.
Dr Foster’s data showed Stafford Hospital flagging as a mortality outlier every single year for an extended period. It was the only hospital to flag consistently on the Hospital Standardised Mortality Ratio. At the same time, patients were gathering in a local cafe to share stories of failures in their care. The board’s response was to commission the University of Birmingham to challenge the methodology rather than investigate the clinical concerns.
Alex Kafetz presents cumulative sum analysis showing that Shipman would have tripped a statistical threshold and theoretically been investigated. Professor David Spiegelhalter’s time-of-death analysis shows a significant spike at approximately 3pm, corresponding to Shipman’s home visits when he was murdering patients. Other outliers in the same dataset were explained by case mix factors. Alex Kafetz concludes the data would have flagged him, but investigation and curiosity were still needed to confirm the signal.
The inquiry found that Paterson harmed over 1,000 people, not the 17 he was convicted of. Both clinical and financial data pointed to problems. Mastectomy patients were being sent for radiotherapy against NICE guidelines because nurses knew the operations were incomplete. The first chief executive dismissed the data. The second suspended Paterson immediately with the same information. The inquiry classified witnesses as “should have known, could have known, must have known.”
Alex Kafetz argues that many recommendations from the Mid Staffordshire inquiry remain unimplemented more than a decade later, including non-disclosure agreements with NHS staff and the absence of a single regulator for quality and finance. From the Paterson inquiry, Alex Kafetz states that there is still no database showing how many operations a surgeon has performed across the NHS and private sector, and patients cannot verify surgeon experience before consenting to treatment.
Alex Kafetz identifies three pieces of information. First, the risk-adjusted mortality rate at the hospital. Second, the friends and family test results, where most trusts score well. Third, the staff survey question asking whether staff would want their loved ones treated at the hospital where they work. That figure was reported in the session as averaging approximately 65%, with significant variation, meaning more than three in 10 staff would say no.
Alex Kafetz draws a direct parallel between the Paterson case and the Harvey Weinstein scandal, quoting: “Stories of abuse by powerful people are also a story of a failure of board culture.” Paterson was making money for his hospital, was well thought of and was hard to challenge. People knew and did not act. Alex Kafetz argues this pattern is not necessarily unique to one case and can persist where organisational culture discourages challenge.
Yes. Alex Kafetz cites recent investigations at Salford Royal, Great Ormond Street and a South London hospital where a surgeon was linked to 15 unnecessary deaths. He states that concerns about individual practitioners continue to emerge and that data may be available to help identify potential warning signs. The question remains whether boards have the culture and curiosity to look at the data and act on what it tells them.
