The (McKinsey) Procurement (part 2)

When McKinsey was advising on the structure of the then “new” NHS England in 2013, McKinsey was simultaneously advising other clients how to take advantage of the structure they were recommending.

This year McKinsey won a £1m contract to advise on the structure of the (2023) “new NHS England” following the takeover of NHS Digital (and Health Education England). Presumably it continues to advise other clients how best to take advantage of those new structures, and past practice suggests McKinsey will be paid more money by others to subvert the model they proposed.

McKinsey doesn’t talk about their clients, but sometimes they are forced to by courts. One such client was IMS Health, which was set up to be the “information intermediary” between doctors and the makers of oxycontin, the drug whose sales practices were partially responsible for the opioid epidemic in the US, and which still operates in the UK (under the current brand of “IQVIA”) doing much the same thing as they have done before.

Does NHS England know who McKinsey’s other clients are? Does NHS England know whether they’ll benefit from knowing McKinsey’s advice to NHS England? Does NHS England know whether McKinsey advice was written in a way which might help those other clients? 

The ongoing trade in NHS information

NHS England is both a consumer of data via their analytics, and a producer of data for themselves and others. McKinsey’s report should have recognised this conflict of interest, and potentially managed it in better than the usual way (either of McKinsey or NHS England). The functions of the data safe haven, which should be to hold data, be accountable for what data is used and how, and offer multiple environments in which it can be analysed, should be transparently separated from the functions of the analysts who consume data they need to do their work.

Realisation will creep across NHS England that the data they hold is now almost all identifiable patient data, as they have the Personal Demographics Service, identifiable copies of HES, and the ability to match across different datasets on fields which they take no steps to protect. 

Indeed, Palantir is very proud of the fact that it offers exactly that functionality to clients, and Palantir never ceases to point out that whether any functionality is used is purely a choice of their client – it’s up to NHS England and the government of the day. Of course, not everyone at NHS England is racist and incompetent, but there are informed individuals with legitimate fears that someone elsewhere in the organisation is doing something stupid with the identifiable patient data that NHS England now hold; and they’re probably right.

McKinsey and Palantir aside, there’s a different contract with our old friends at PA Consulting for implementing the recommended changes, PA Consulting being the company who agreed in contract not to upload a lot of data to google’s cloud, and then did so anyway.

NHS England is not a data literate organisation

The new NHS England is not (yet) a data literate organisation – you only need to look at the difference between NHS Digital’s board papers, full of numbers, RAG ratings and trajectories of change over time, and the NHS England’s board papers, of essays which contain the minimal numbers. The old NHS Digital showed what it really was, whereas NHS England describes what it thinks something will be, with enough people commenting on drafts that anything interesting will be taken out.

Insight into flows of data between NHS Digital and NHS England disappeared when NHS Digital got abolished. We were expecting NHS England to restore that transparency by publishing their “internal data flow records” this week; they didn’t.

If the new model goes as expected, McKinsey may advertise a case study of the leadership of Tim Ferris, epitomised by his monologue to the first post-takeover NHS England Board meeting. “Taking the paper as read”, he then talks through it, (probably correctly) knowing that even this superficial detail was below the attention of the board. The integrity of his examples is clear from his anecdote about the value of the NHS App, delegated access, and his kids’ records.

It is possible that the papers of the digital subcommittee of NHS England’s board (which takes over the oversight role that used to be managed in public by NHS Digital’s board) will have such information, but none of it will be public.

After all, the structure of the “new NHS England” data functions will be reflective of the late-but-still-forthcoming statutory guidance for data functions in NHS England, which should have been in place before the merger happened. They weren’t, and still aren’t.

Is McKinsey’s “rightsizing” recommendation to get rid of experts who know something?

Professor Mazzucato’s recent book on consultants and consultancies explains how the choices and outsourcing of key work results in a hollowing out of Government, and a brain drain that makes them ever more dependent on ever more consultants. 

The opening chapter of the McKinsey book covers how those with the most experience are let go as McKinsey helps “rightsize” organisations, and the deaths that resulted from those choices. As McKinsey give the same advice over and over again, did they do something new this time?

Large consultancies only offer solutions which involve some future role for large consultancies. Approaches like Reproducible Analytical Pipelines, which are cheaper and more effective for all kinds of analysis, get deprioritised by the consultancy world as there’s little consulting money from that approach. 

Consultants everywhere, so how long until the NHS spend around Palantir costs more than the NPfIT? The currently published £480m tender only includes NHS England’s role, and NHS England is increasingly saying that Trusts, ICSs, GPs, and others will be expected to shoulder their own burdens for interacting with the system, and the way to minimise those costs is to pay Palantir more money, because interaction between Palantir and other systems is still manual (and will be unless a Trust cedes decision making to NHS England, importing the US model with NHS England acting as the insurer and decision maker rationing care).