I’ve been really interested to see the level of AI-related bullishness expressed by European Healthcare executives in the recent Accenture AI in Public Service survey. We polled business executives in seven public service sectors¹ across five countries² to test how governments and public sector leaders can: 

  • Better scale with AI investments. 
  • Measure their AI spending trends. 
  • Understand the key success factors and barriers in the scaling of AI projects.

The reason I was so interested was that their bullishness is matched by neither their AI investment, nor the extent of actual or expected benefit realisation in the next year. With 60 respondents from each country, and 50 from Healthcare across the five countries, Healthcare was second only to Postal in terms of optimism with respect to AI, both leading the pack by a significant margin. That’s really bullish and, given the result, you’d expect enthusiasm for investment to be high, and current and expected future scaling of AI projects in health to rank strongly against their counterparts in other sectors. Not so? 

<<< Start >>>

<<< End >>>

 We’re keen on AI in healthcare, it seems, and optimistic. Yet, we’re not putting our money where our mouths are, and we haven’t derived that many benefits yet, compared to other sectors. In terms of investment, Healthcare is doing moderately well when compared to laggards like Defense, but still occupies the lower middle of the pack when comparing its spending profile to the Education, Postal and Revenue sectors.  

That doesn’t line up with the optimism expressed above. The glaring question is: why wouldn’t you invest more enthusiastically in something you believe so strongly in? 

<<< Start >>>

<<< End >>>

When I look at anticipated AI scaling by use case by segment, Healthcare once again seems largely in line with other public service sectors and doesn’t reflect the massive additional optimism about the future of AI in Healthcare as compared to other sectors. 

<<< Start >>>

<<< End >>>

So, on the one hand we have healthcare executives telling us it’s looking easy compared to other sectors and saying there’s lots of opportunity. On the other, they’re not investing and (apparently) not realising current or future benefits to any significantly greater degree than any other sector. 
 
Are healthcare executives being realistic? 

Is it really that easy to increase the AI success quotient and transform healthcare without spending much money? Our research suggests it’s well past the time to wonder about whether to use AI – that’s more or less common cause. It’s about when and how – what parameters will ensure that everyone gets the benefits they’re hoping for and are entitled to. 

That’s my view too – it’s time to turn the page to the next chapter: the “how” chapter, and the “how” centres on issues of scaling up AI initiatives, and scaling up means dealing with dependencies on data, ethics and security.  

Time for grown up conversations 

Are we talking enough about what’s next, and addressing the difficult issues in a grown-up way? The first way to close the “reality gap” between enthusiasm about AI, and the extent of its actual success to date (or likely success in the short term) is to remove the rose-tinted glasses and see what real challenges lie ahead. 

AI might enjoy too much trust (or too little) from healthcare practitioners and patients. Here are some reality checks³: 

  • AI does not, by definition, replace human contact/company. 
  • Only complete population datasets can eliminate AI’s inherent bias and incompleteness  
  • Scalability of AI applications is biased by the rate at which consumer products are adopted, and the consumer-led approach is fast and messy because AI products reach the market long before testing and NHS approval. 
  • Organisational challenges in the NHS can hamper scalability of any new technology because of fragmentation across 150 acute trusts, 220 clinical commissioning groups and 40 Mental Health Trusts. 

I’d suggest that, until we tackle some of these challenges in an open and adult way, and follow up those discussions with judicious investment, the gap between AI enthusiasm and reality will remain and patients are unlikely to start enjoying the full benefits of appropriately tested technology as fast as they ideally should. 

1 Healthcare, Revenue, Social Services, Postal, Policing, Defense, Education (Higher)
2 Germany, France, United Kingdom, Finland and Norway
3 AI & Healthcare report, Windsor Castle and College of St George, 30th September – 1st October 2019

Niamh McKenna

Managing Director – Accenture Health, UK

Subscription Center
Subscribe to Accenture's Insight Driven Health Blog Subscribe to Accenture's Insight Driven Health Blog