InterSystems Solution Architect at InterSystems considers how successful the Topol Review has been in meeting its objectives?
The Topol Review published and presented in February 2019 with grand publicity and attendance from Matt Hancock and others, is the culmination of a year’s work by Eric Topol and a large team, sponsored by Health Education England (HEE).
Its purpose was to identify what the requirements will be in the next 20 years for the digital skills that the NHS healthcare workforce will need to acquire and utilise. This is undoubtedly a good thing and needs to be in the public and NHS agenda. However, did it actually meet its own objectives?
My assessment is that there is not a fully coherent picture emerging from the review. There are several areas where there are gaps in scope, in thinking or in detail. These need to be addressed if the real value of the exercise, and achieving its objectives are to be actually realised and put into action.
Here are the gaps as I see them;
Gaps in which digital technology areas were considered
The review focused on 3 technology areas; genomics, digital health, and AI and Robotics. The immediate question is why these three? These are definitely strong candidates as areas where digital transformation of healthcare provision can develop, but are they the most applicable or achievable ones? Why were they the only 3 areas considered? Professor Topol lists these 3 areas as his areas of focus and expertise according to his LinkedIn profile. Is that sufficient reason to concentrate on them?
What about other areas that were not addressed; patient information security, clinical decision support, wider applications of analytics, how to address transformational change through the application of digital technologies. Each of these are relevant, particularly to determining what skills the workforce needs to acquire.
The genomics area is undoubtedly going to be important, but we don’t really know what that means in terms of outcomes, efficiencies or staff training needs. This gap is not addressed.
The “there’s an app for that” mentality represents a coherence gap. There is a surfeit of personal healthcare apps available, but unless they interoperate with underlying operational and informational systems, they serve little purpose but as marooned islands of health data. We must avoid the creation of an app gap!
Artificial Intelligence is a current buzzword, but the underlying themes are valid. The hype masks much of the reality; we need access to data and compute power harnessed with the appropriate data science skills. That convergence is now happening but is not well or widely understood. This is an education gap to close. Curiously, while the review report references AI and Robotics, there is no mention of Machine Learning (ML), while the review presentation concentrated on ML. A case of style over substance?
Gaps between current academic or science capabilities and putting them into practice
The reviews objectives were to identify the skills that were going to be required, based on the digital technology shifts and trends currently apparent. The technologies addressed were all considered from their current scientific, engineering or academic perspective. There is nothing wrong with this approach, but it raises the question of how to bridge the gap between the abstract capabilities developed and understood in the lab and putting them into practice in the field. How does the healthcare ‘industry’ and the associated professions go about industrialising, operationalising, or standardising the technologies available to it?
This is nothing new. Making this shift applies to all science and engineering innovations, but it is not a journey that is always made successfully. The ambition to make this happen is admirable. Anticipating what it will mean for those who will seek to make use of these capabilities in an operational setting is also both well intentioned and far-sighted. How we achieve this is, however, not addressed.
Gaps between technical focus points and the implications for training
To take this point one step further; how do the technologies on offer translate into the education and training needs? How can we understand this without understanding how the digital technologies are likely to be deployed? This is classic “people, process, technology”. The training needs are only one part of a broader transformative shift in how healthcare can be delivered. Not only is the translation into training requirements not made in the review, but it is not set in this wider context of the digital transformation to be undertaken.
Gaps in training needs
This is recognised in the review but not really addressed. Introducing a new set of digital skills, knowledge and experience to healthcare professionals is a quantum shift. Blithe statements about everyone needing to have a basic education in genomics is hard to quantify, qualify or comprehend. What does this really mean?
How does the NHS effect continual professional development? How do organisations set about creating the roles, skills and capabilities that will be required? Existing clinical roles will be required to acquire new skills. If this is an increasing skillset, how do we ensure we have the right calibre of individuals to fulfil the roles. How do we extend the skillset of those currently in compatible roles? This digital shift also generates an additional requirement for new roles or increased demand for capabilities; for informaticians, bio-informaticians, data scientists, health data scientists, technology evaluators, and health economists in an operational setting. This is not an exhaustive list.
How do we train clinicians (doctors and nurses) in their graduate/post-graduate studies effectively so that they are digital capable, digitally enabled, and digitally aligned? What fundamental changes need to be made to their current training syllabi and the way education is delivered to achieve this?
The review acknowledged that 50% of the workforce will turnover in 15 years. This presents both a challenge and an opportunity. For a start, this is actually a very low rate of attrition (~3% pa). What if this number is actually much higher? What if cost pressures mean either that we need to do more with less resources, or conversely that we need to recruit even more staff?
The next generation of healthcare professionals are already being trained. Mention was made of changes to their training, focusing more on specialisms and possibly less on generalisms. Providing healthcare professionals with the appropriate skills means that 5 to 15 years of training needs to start now. How do we effect radical change to course content and how do we do this in such a way as to accommodate digital technologies and their implications into education and training packages?
There are ~730k existing healthcare or clinical professionals working in the NHS (117k doctors in hospitals + 525k nurses, midwives, paramedics, therapists etc, 42k GPs + 41k primary care nurses & other clinical roles). These staff will need to be provided with additional training. If 50% of these roles are to be replaced in 15 years, then 365k new staff will require additional training from scratch. Assuming a linear rate of new staff introduction means training an additional 24k staff per year.
Gaps in training funding
How should this best be delivered? Where does the funding come from? If we treat this initial investment and ongoing training as a capital + revenue model, then we need to provide training for 24k new staff per year. If we assume this takes at least 5 years then an ongoing cohort of 120k being trained each year full time, of which digital training is, say 10%. That means ~12k FTE in training each year specifically on the application of digital technology in healthcare settings. With average costs of ~£30k per FTE that means £360m per year on digital training investment.
Ongoing digital training of 730k staff at, say, 1 week a year means ~18k staff FTE on digital training. At £30k per FTE that means £550m revenue training cost. It is therefore not unreasonable to consider a total annual cost of digital enablement of staff to be ~£1b, at current rates and assuming a linear delivery over time. That is a significant proportion of the total healthcare annual budget (~1%).
Gaps in overall staff funding
This is the fundamental question. Where is the cost of a digitally enabled workforce calculated or budgeted for? This is a combination of new roles and enhanced existing roles, as described above. What cost take-outs can the NHS consider to counter-balance this? What roles can be removed and replaced with the digital specialists identified above?
The review panel did suggest, in response to a question, that seeking additional sources of revenue for the NHS through the “monetisation of patient data” was one route to closing the affordability gap. Given that the review had looked carefully at the ethical issues associated with the introduction of digital technologies, it was something of a surprise to hear this so willingly endorsed. The boundary between public services funding and private enterprise is necessarily blurred, but taking further steps in this direction without general public support seems cavalier.
Noting the potential training cost above, the wider staffing cost is significant. It will require a revolutionary approach to healthcare funding in general.
Gaps in patient/carer access?
The review correctly identified that parity of access to or appreciation of digital technology enablement for the public whatever their role, for citizens, patients and carers, is crucial. We cannot afford to either establish or widen any real or perceived gap between those who can and those who cannot access the necessary care because of digital technology constraints. This is a broader concern since it does not directly relate to workforce education (although there must be a strong relationship), and it raises the wider issues of validating and encouraging patients’ direct active involvement in their own care.
However, little focus seemed to have been made on the practicalities of preventing this gap. This itself is a credibility or conceptual gap. The ONS reports that 90% of UK households had internet access in 2017, and 73% of adults have smartphone or similar access. Other studies suggest that these percentage may be higher still. However, a significant patient group is the elderly. There are >3m people in the UK over 75 years old. Within this group Age UK report that 60% do not have internet access, and that those with disabilities and/or effected by socio-economic factors are even more likely to not use internet-based digital technologies. These are a key patient group.
The suggestion from the review panel that these patient cohorts should access digital healthcare provision via kiosks in public libraries because GP surgeries are too busy is preposterous. There are ~3k public libraries in England, while there are ~11k GP surgeries and other community-based healthcare locations. Even if they were open, patients cannot access them. The cost benefit of providing free or subsidised appropriate and applicable digital devices might be more advantageous and realistic.
Gaps in digital technology impact
Various case studies are cited in the review. They represent practical academic studies or pilot projects aimed at utilising data or technology more effectively to support care provision. These case studies have then been used to indicate the potential financial or staffing benefits if the use cases were scaled out nationally. This is always a dubious exercise and is therefore suitably caveated. Understandably, no advantages to do with patient health outcomes are claimed.
Savings in healthcare professional time, Topol’s “the gift of time” are suggested. Here again there is both a credibility and impact gap. Most of the case studies do not suggest either large improvements in HCP time saved. Many of the published calculations suggest an overall impact of ~1% of the relevant workforce. No account is made of the transformative cost of both delivering these initiatives or of managing them i.e. what is the net effect of introducing potentially 100’s or 1000’s of small digital initiatives, each of which has value, but the combination of which introduces significant technical and organisational change? It is worth noting that the Digital dictation use case has the largest impact of those cited, the technology has been available for many years but has not been widely adopted. This is more of an organisational change management issue than a technical one. It indicates the scale of the challenges to be overcome.
Gaps in contributors
Reviewing the review body, there were 62 experts named as being actively involved in the review, excluding the Health Education England (HEE) secretariat. 22 of these work for HEE or are patient representatives. Of the HEE team only 3 have a clinical background and 2 have a health technology background.
Of the remainder, 30 are UK university academics across a range of subjects. 16 of these are in clinical fields. The final 10 include 3 NHS clinicians and 6 employees of NHS England or NHS Digital. Babylon Health, the golden child of GP care provision models, was the only private sector input. With the exception of Professor Topol, no-one from outside the UK is involved. This seems a myopic failing.
This is an opportunity missed in many senses. The only health technologists actively involved in the operational delivery of digital technology to support care provision are CIOs Rachel Dunscombe and Will Smart. Both of whom are in NHSE/D roles. No-one else from the NHS health informatics or health technology profession was involved in the review. A practical view reflecting the experience on the ground should have provided an injection of reality and an opinion on the challenges currently faced. These would certainly inform the anticipated future challenges.
Furthermore, no-one from the healthcare technology industry was involved in or is acknowledged as having provided input to the review. Again, an opportunity missed. The experience both inside and outside the NHS in the health tech market is crucial to understanding how new technologies can be introduced in support of digital transformation programmes. The risk is that without this input the review, like the examples investigated and cited, becomes an academic exercise.
The Topol review is a valid and well-intentioned exercise. Possibly a unique attempt to address the implications of a digital future of healthcare for the NHS. However, its recommendations are relatively weak and insubstantial. There are no SMART objectives leading from these recommendations or concrete plans or actions to be taken forward.
There are significant gaps in both the method undertaken and the scope of what has been addressed which undermine the value of the exercise. Fundamentally, the review misses the opportunity to address how digital transformation is to be effectively implemented with the future NHS workforce.