4.1 Strategy and culture

Confidence that the right strategic decisions are being made about artificial intelligence (AI) technologies in a culture that supports innovation and collaboration.

The successful introduction and ongoing implementation of artificial intelligence (AI) technologies in healthcare settings will depend, amongst other factors, on developing strong related business cases, maintaining effective relationships with industry innovators, and establishing organisational cultures conducive to innovation, collaboration, and public engagement.

These factors, which demand specific knowledge and skills, can influence how receptive workers will be to AI technologies and contribute to developing confidence in AI as a strategic and organisational asset.

The business case for adopting AI technologies

Interviewees for this research cautioned that some AI technologies are potentially being developed (including in proof of concept and trial stages) without a clear understanding of how they might affect healthcare provision. This underlines the importance of a strategic approach to the deployment of AI in healthcare settings, including through the development of value propositions and business cases.

For the interviewees, understanding the value, benefits, and risks of AI technologies (including in relation to patient outcomes, financial and human resourcing considerations, and alignment with related local and national strategies and priorities) are key to establishing strong business cases for deploying AI. 

Despite available guidelines (including in the National Institute for Health and Care Excellence’s evidence standards framework), the development of these business cases can be complicated by the minimal clinical evidence for most AI technologies (as discussed in section 3.2) and by workforce perceptions founded on limited knowledge and experiences of AI technologies.

Many interviewees for this research shared their concerns about the potential impact of AI technologies in their settings; for example, technologies that may lead to a higher number of patients being recalled and resulting in further costs for assessments and in unnecessary stress for their patients. Others wished that they were more cautious when they introduced AI in their settings by taking more time to debate and understand the related challenges.

On the flipside, interviewees voiced their hopes for using AI to address the major challenges in healthcare provision, including the increasing needs of an ageing population and the current backlog and waiting times in secondary care.

Relationships with industry innovators

Once AI technologies are procured, many are successfully embedded by establishing collaborative and sustainable relationships with industry innovators. These relationships can involve optimising and evaluating AI technologies.

Interviewees for this research noted that effective relationships between healthcare settings and industry innovators are built on shared values, and that they require a significant commitment of time and resources from both parties. Adept management of these relationships demands appropriate skills and resources, and shared understanding of the different perspectives within the AI ecosystem.

Interviewees noted that effective relationships can ensure buy-in and development of confidence in AI technologies among healthcare workers. However, they cautioned that the general lack of knowledge and experiences with AI technologies in most health and care settings limits their ability to critically appraise information provided by industry innovators, including during due diligence. These limitations in knowledge and experience may also lead some health and care settings to adopt industry-set parameters around the transparency and bias of AI technologies.

Interviewees noted that while some industry innovators support sites to establish a strategic approach to adopting AI technologies, others are not as proactive in their engagement or transparent about their products. Some interviewees spoke of ‘hidden’ costs involving AI technologies (including the costs of monitoring performance and reporting errors) that are not explicitly disclosed by some developers.

The importance of culture and leadership

Interviewees for this research highlighted that organisational cultures and leadership are key to the successful introduction and deployment of AI technologies. This insight is not surprising, and not exclusive to AI, as organisational cultures and leaders who support innovation and collaboration are crucial to the broader digital transformation of health and care settings.⁹

A focus on the digitalisation of health and care services is an important prerequisite to the adoption of AI technologies, as supported by NHS’s 'What Good Looks Like' framework (see also section 3.3).

Interviewees suggested a few key cultural and leadership features that can support these broader efforts, and increase confidence in AI, including:

• developing AI technologies from the ‘ground up‘ by involving multi-disciplinary teams (including clinicians, information technology and governance specialists, clinical domain experts and data scientists) and internal decision-makers early in discussions about their needs and implementation challenges. Interviewees for this research noted that multi-disciplinary teams tend to be the most successful structure for implementing AI
• establishing senior leadership and clinical lead buy-in, and identifying and supporting internal champions for change
• conducting early and ongoing engagement of patients and the public to inform AI development or co-design of technologies
• focusing on ongoing learning and development of staff

Interviewees for this research suggested that peer and expert endorsement and support can enhance confidence in AI technologies. This highlights the importance of developing and resourcing mechanisms to establish and encourage connections to share AI-related knowledge and experiences amongst peers and sites adopting AI technologies.

An example is the NHS AI Virtual Hub, an online platform for discussions, shared resources, and collaboration about AI technologies in health and care.⁶⁵ Other potential approaches can include developing and distributing case studies to highlight challenges and success stories (similar to existing NHS digital playbooks).⁶⁶

References

⁹ Sinha S, Al Huraimel K. Transforming Healthcare with AI. In: Reimagining Businesses with AI; 2020:33-54. doi:10.1002/9781119709183.ch3

⁶⁵ NHS AI Virtual Hub - NHS Transformation Directorate. https://www.nhsx.nhs.uk/ai-lab/ai-lab-virtual-hub/. Accessed March 8, 2022.

⁶⁶ Dermatology digital playbook - Digital playbooks - NHS Transformation Directorate. https://www.nhsx.nhs.uk/key-tools-and-info/digital-playbooks/dermatology-digital-playbook/. Accessed March 7, 2022.