New AI laboratory for the NHS
BMJ 2019; 366 doi: https://doi.org/10.1136/bmj.l5434 (Published 13 September 2019) Cite this as: BMJ 2019;366:l5434
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
I read the editorial article by Steventon et al with great hope and enthusiasm(1). I strongly agree with what was mentioned in the article. I would like to make some further contributions as an AI interested Radiologist. I find the NHSx initiative is very exciting but it needs to be activated and should be beyond waffle. As an active Radiologist working for the NHS, I would like to give a snapshot of the situation in the current NHS in terms of issues including staffing and existing and future AI issues.
I would like to draw attention to the importance of team work without any discrimination or people feeling discriminated against. We, as doctors, should be more open minded and learn from our mistakes and learn from each other. We don’t discriminate against our patients, do we? We can easily adapt similar skills for ourselves.
* Transparency and fairness amongst doctors and healthcare staff are needed in the current and future AI enhanced NHS.
* There is a shortage of Radiologists in the UK according to the RCR 2017 census report (2) and the shortage may increase in the future and be deepened after Brexit.
* Current IT and PACS have shortcomings and almost every week we face IT failures. Our Voice recognition (VR) has failures in terms of recognition of the words and frequent collapses; never ever fixed properly. I have also used different VR for tele-radiology private company with no hassle for 10 years. Remedy: new NHS should be —at least partly— privatised. This is, however, debatable and would remain controversial.
* AI interested staff should be encouraged to develop their expertise. My personal experience has not been great so far. It could have been better.
Solution: AI interested staff should be supported by all means. I personally published 3 opinion articles about AI in radiology and medicine in 2019, all out of my own pocket. Funding is limited in terms of supporting research projects. It would be a nice dream for me to be working with an AI algorithm in day to day practice one day in the near future, but who knows? Perhaps NHSx will make my dream happen.
* I feel the current NHS response to AI innovations is slowly but I hope it will become more robust.
When it comes to future projections:
* To me AI is a friend not a foe (3). However, it needs to be a firm, robust Q@A (Quality and Assurance)
* AI algorithms need to be ratified as quickly as possible by Royal College of Radiologists (RCR).
FDA (The Food and Drug Administration, USA) has already endorsed several AI algorithms so far and RCR needs to catch up the similar pace ASAP.
* AI will not replace healthcare professionals but will be a helpful adjunct indispensable tool for diagnosis and triage.
I feel the current clumsy NHS would transform into an agile and more productive one soon with the aid of AI.
References:
1. Steventon A, Deeny SR, Keith J and Wolters AT. New AI laboratory for the NHS. BMJ 2019;366:l5434.
2. Royal College of Radiologists (RCR). UK workforce census 2017 report in Clinical Radiology. London: RCR, 2017. Available at: https://www.rcr.ac.uk/system/files/publication/field_publication_files/b... (accessed 14.09.2019).
3. Pakdemirli E. Artificial intelligence in Radiology: friend or foe? Where are we know and where are we heading. Acta Radiologica Open 2019; 8(2):1-5.
Competing interests: No competing interests
Re: New AI laboratory for the NHS
EEGs (my specialty) generate vast amounts of data, which have attracted numerous forms of analysis. Though analytical challenges are very substantial, they are matched by the challenge of dealing with data protection in any project that involves data sharing. Secure transmission and protection of the data and associated clinical information are particularly convoluted. For the new AI laboratory to be maximally useful, there is a strong case for devising and implementing protocols which overcome these necessary but extremely tiresome hurdles in providing its diet of data.
Competing interests: No competing interests