The AI healthcare revolution and what it means for medical communications
Blockbuster movies depict artificial intelligence (AI) as self-aware, sentient beings with a hunger for world domination. While this is still a far-fetched fantasy, AI is ever more present in our lives and looks set to transform healthcare too; Google Maps route predictions, Netflix movie suggestions and Amazon’s Alexa platform all rely on AI technology.
Evidence is mounting that medical research, payor decision-making, diagnostics and even the delivery of care could benefit from AI solutions. With an AI healthcare revolution on the horizon, it may be time for the medical communications industry to understand the implications of, and possibly even harness, AI technologies to better serve our clients.
AI in drug discovery and development
Though lagging behind other sectors, such as retail and social media, we are starting to see pharmaceutical companies implement AI strategies. For instance, machine learning (a form of AI) could accelerate drug discovery. Novartis has partnered with IBM’s machine learning platform ‘Watson’ to identify breast cancer drug targets using real-world data1. By increasing the efficiency of the drug discovery process, AI may be key in reversing the trend of rising research and development costs for new drugs2.
The drug development process could also benefit from machine learning technologies. The application of machine learning to computational modelling of chemical structures could reduce the economic and ethical costs of toxicity studies by alleviating the burden of, or even replacing, animal models3.
We are seeing an explosion of patient-facing AI solutions such as ‘Chatbot’ health assistants, with the potential to relieve the burden of medical triaging in primary care; and a good example of this is ‘GP at Hand,’ which was recently trialled by the NHS4. Chatbot technology mimics human conversation and is already being harnessed by medical communications agencies to modernise patient engagement. For example, Cohn & Wolfe developed ‘Tabatha’ – a Facebook-based Chatbot – as part of Boehringer Ingelheim’s ‘Think.Act.Breathe‘ (TAB) campaign, to encourage patients to act on their asthma symptoms5. A key benefit of this technology is that it can generate tangible metrics that illustrate how effective campaigns are at promoting behavioural change.
Another application of AI is in the automation of repetitive cognitive and decision-making tasks. The Hindsait6 platform, for example, applies predictive analytics and natural language processing (NLP) to patient data in order to provide scoring of clinical need, in accordance with clinical guidelines and regulatory requirements. This technology has the potential to improve the efficiency and reliability of payor decision-making.
Furthermore, in the secondary care setting, predictive analytics could improve patient outcomes by identifying inpatients who are most at risk of transfer to ICU7. We may also begin to see AI systems used in the automation of various pharmaceutical company processes, including compliance documentation and social media monitoring8.
Diagnostics and patient monitoring
AI is also showing potential in the areas of diagnostic techniques and patient monitoring. For instance, when supported by AI image-analysis technology, dermatologists have been shown to be more accurate at diagnosing melanomas9; similarly, the facial image-analysis framework ‘DeepGestalt’ has been shown to outperform clinicians in the identification of rare genetic disorders10.
Wearable patient-monitoring technologies such as the Apple Watch Series 4’s EKD (electrocardiogram)11 and Google’s blood glucose monitor ‘the Soli system’12 could lead to improved, more personalised patient care, as they enable the collection of considerably higher volumes of data than is possible by traditional medical consultation. Such quantities of patient data are likely to require interpretation by AI systems in order to be clinically useful.
Pharmaceutical companies could benefit from AI-based diagnostic and monitoring technologies through modernisation of treatment pathways and consequent optimisation of pharmaceutical intervention. As a result, there may be value in raising awareness of such technologies through medical education initiatives.
AI in med comms
From patient engagement to medical education, AI has the potential to impact the medical communications industry. However, the key will be in identifying which of our clients’ needs could be addressed by AI technology, whether it is financially viable to do so and how to implement these solutions. Armed with an understanding of how AI can be utilised, now may be the time for us to join, or possibly even lead, our clients into this exciting new paradigm of healthcare.
Sally Neath, HealthCare21 Communications, Macclesfield
Artificial intelligence (AI): a broad term, essentially encompassing all computer systems which can perform tasks that are traditionally seen as human behaviours. Some of the key areas of AI have been summarised in the diagram ‘Examples of AI in healthcare’
Machine learning: a form of AI in which computer systems/machines can autonomously learn and improve performance of tasks without being explicitly programmed
Deep learning: an advanced form of machine learning that has networks capable of learning unsupervised from data that is unstructured or unlabelled, based on learning data representations, as opposed to task-specific algorithms
NLP: Natural language processing
Chatbot: NLP developed to mimic human conversation
HCP: Healthcare professional
ICU: Intensive care unit
- Novartis announces ground-breaking collaboration with IBM Watson Health. https://www.novartis.com/news/media-releases/novartis-announces-ground-breaking-collaboration-ibm-watson-health-outcomes [Last accessed February 2019].
- Unlocking R&D productivity (2018). https://www2.deloitte.com/content/dam/Deloitte/uk/Documents/life-sciences-health-care/deloitte-uk-measuring-return-on-pharma-innovation-report-2018.pdf [Last accessed February 2019].
- Wu Y and Wang G. Machine Learning Based Toxicity Prediction: From Chemical Structural Description to Transcriptome Analysis. Int J Mol Sci. 2018;19(8):2358.
- GP at Hand: London launch. (2018). https://www.gpathand.nhs.uk/uploads/icons/GP-at-hand-London-launch.pdf [Last accessed February 2019].
- PMLive. http://www.pmlive.com/awards/communique/previous_winners/communique_awards_2018_results/healthcare_communications_awards/innovation_in_healthcare_communications [Last accessed February 2019].
- Hindsait. https://www.hindsait.com/ [Last accessed February 2019].
- Pearl R. Artificial Intelligence in Healthcare: Separating Reality from Hype. Forbes. https://www.forbes.com/sites/robertpearl/2018/03/13/artificial-intelligence-in-healthcare/#ea0be71d7506 [Last accessed February 2019].
- Shahid W and Woloszynski R. Regulatory Compliance in the Age of Artificial Intelligence. Anakura. https://ankura.com/insights/regulatory-compliance-in-the-age-of-artificial-intelligence/ [Last accessed February 2019].
- Agence France Presse. Computer learns to detect skin cancer more accurately than doctors. The Guardian. May 2018 https://www.theguardian.com/society/2018/may/29/skin-cancer-computer-learns-to-detect-skin-cancer-more-accurately-than-a-doctor [Last accessed February 2019].
- Gurovich Y, Hanani Y, Bar O, et al. Identifying facial phenotypes of genetic disorders using deep learning. Nature Medicine. 2019;25:60–64.
- Karston J and West D.M. Wearable Device Data and AI can reduce health care costs and paperwork. Brookings. October 2018. https://www.brookings.edu/blog/techtank/2018/10/18/wearable-device-data-and-ai-can-reduce-health-care-costs-and-paperwork/ [Last accessed February 2019].
- Shaker G, Smith K, Omer A.E, et al. Non-invasive Monitoring of Glucose Level Changes Utilizing a mm-Wave Radar System. Int J Mob Human Interaction Computer Interaction. 2018;10(3):1–20.