Hot topics – Funding an innovative future with AI in healthcare
Key themes and emerging challenges: Enabling investment into – and reimbursement of – AI
Artificial intelligence (AI) is set to transform multiple sectors, from finance to education as well as healthcare.1 However, in healthcare settings there are some unique financial hurdles – firstly to obtain funding to develop AI solutions, and secondly to enable its deployment by securing reimbursement once the technology has been developed. How can these hurdles be cleared to make sure Europe can keep up with the global pace of change?
Investing in innovation
AI’s potential to transform healthcare means it is attracting the attention of various parties looking to invest in the technology, from the EU and national governments to the private sector.1 However, while this diversity of funding sources is encouraging, it can also be a challenge itself – participants at the Round Table Meetings highlighted that innovators may often be unaware of the investment options available to them.1 To combat this, participants suggested that EU and national public funding streams should be consolidated to make it easier for innovators to understand what funding there is and how to obtain it. This could be supported by defining specific priorities for AI in European healthcare so that funding streams can be aligned with these.1 Participants at the Round Table Meetings advised that it can be a huge administration task to get access to public money, which is why small and medium-sized enterprises (SMEs) and entrepreneurs could benefit from connections with expert consultants, or ‘innovation clusters’, who understand what is needed to secure EU funding.
Increasing private investment in the form of venture capital (VC) funding for AI in healthcare will also be key to scaling the technology within European healthcare. VC funding in this arena has risen steadily in recent years, but Europe remains behind other regions such as the US and Asia.1 Further growth could be supported by public-private partnerships, for example between governments and pharmaceutical companies. Such partnerships can support the creation of dynamic innovation-driven ecosystems and Israel, in particular, has done this successfully.1
The Israeli government has long invested in and encouraged partnerships between start-ups, research institutes, hospitals and other stakeholders. As a result, investment in Israeli digital health companies has grown, with the country’s share of global investment for digital health doubling between 2014 and 2019 (from 1.5% to 4.5%).2 This is significant for a country representing just 0.1% of the global population and shows Israel as an emerging player in the race to lead AI innovation in healthcare.2 Indeed, of the top 50 best-funded AI companies in healthcare since 2010, four are Israeli (36 are US based, six are Chinese, two are based in the UK, one in Ireland and one in Switzerland).1
Whilst the Israeli figures are impressive, pace in Europe is also gathering speed – between 2015 and 2019, VC funding for AI companies in Europe increased 22 fold. Moreover, the EU’s Horizon Europe programme, as an example, offers funding on research and development (R&D) and innovation on health and ageing, which includes opportunities for AI.1
Part of the difficulty in securing investment for innovative AI solutions in the healthcare sector is in demonstrating their value and possible return on investment.3 Participants across the Round Table Meetings concurred that AI tools that improve operations or business functions are easier to fund, while those aimed at treating patients tend to receive less investment.3 This may partly relate to the black box nature of AI (i.e., being unable to see or fully understand its inner workings), which can make investors uneasy and increase perceived risk.1 Evidence for the value of AI will be supported by use cases as AI is deployed in health systems across the globe. These real-world examples will be important in scaling AI solutions and ensuring they deliver against clinical and operational endpoints, and can show how the technology is creating return on investment.1
It’s important to work out what the value an AI solution brings – what is the return? Does it save costs, does it make revenue? The biggest challenge in moving AI projects from proof of concept to pilot and through to being fully operational is establishing the project’s value, and the associated investment required to launch it.
Of course, investing in AI for healthcare goes beyond financing the development of AI tools for use in healthcare settings. Funding is also needed to create an ecosystem of innovation in which the advantages of AI can be best exploited. For example, expert participants at our Round Table Meetings agreed that there must be investment in infrastructure that can support the digitisation of healthcare systems, as well as in education and training to enable healthcare professionals (HCPs) to use AI in their practice.4,5
Key insights – The Netherlands
Access to funding for complex and specific projects in AI is challenging in the Netherlands, with funding likely to go to large institutions, advised participants at the Round Table Meeting in the Netherlands. More agile investment opportunities are needed, otherwise it can take several years for smaller instutitions to get a project off the ground.
The participants also noted that there can be an impact on working practices within an organisation when AI applications are adopted and clinical practice must be changed to accommodate them. This can incur initial costs due to workforce changes (e.g., new roles that must be recruited) or training and education requirements. While some hospitals and primary care providers have specific transformation budgets to cover these costs, it can be barrier in organisations that do not.
Participants considered that existing reimbursement policies in the Netherlands are compatible with the implementation of AI. However, it’s crucial that the AI can demonstrate value to payors.
Value-based healthcare (VBHC) models* should be considered when validating and deploying AI driven innovations, ensuring that high costs to implement the technology are balanced with the possibility of improved long-term outcomes.6
Participants quoted a statement from Eindhoven University, which notes that the improvement in care quality that AI brings means that reimbursement should be based on outcomes.**
* VBHC is a model in which healthcare providers are rewarded for improvement in patient outcomes, which differs from traditional fee-for-service models where providers are simply paid for the amount of healthcare delivered. The value is therefore not based purely on how much it costs to deliver an intervention but how these costs are balanced against measured patient outcomes.7
** Currently unpublished. Discussed by participants of the Think Tank Round Table Series in The Netherlands.
The need for new reimbursement models
Another fiscal obstacle for innovators looking to get their AI solutions into healthcare systems is reimbursement. AI solutions are likely to lead to efficiencies that can reduce spending on care delivery, which is why they are key to a more sustainable European healthcare system.1,6 However, AI is not routinely funded because the technology often does not fit current reimbursement frameworks or standards, which are complicated and inconsistent across Europe.1,6 Member states, and even individual regions or hospitals, make their own reimbursement decisions and therefore have different rules or criteria. 1,6
Participants at the Round Table Meetings advised that the EU has a role to play in providing clear criteria for reimbursement models that can support the adoption of AI solutions at scale to enable some form of centralisation similar to the role that the EMA plays in regulation.1 Key to this will be defining standards for evidence to support reimbursement.6 This will ensure innovators can invest in, and generate, meaningful evidence to demonstrate the specific value of their AI solutions, that healthcare providers need to see. There will also be a need for new models to incentivise solutions that can demonstrate specific and real-time value such as reducing the number of patients requiring care. Therefore, it is likely that we will see a shift away from fee-for-service models for digital solutions such as AI, with a much more pronounced shift to value-based health care.1
Key insights – Denmark
Round Table Meeting participants in Denmark advised that standard systems and processes, or ‘business models’ for the implementation of AI solutions in healthcare should be developed to accommodate all phases – from creation to prototype testing, clinical validation, pilot implementation and normal operation.
These should feature specific key performance indicators (KPIs) assessing the value of the solution and what improvements it makes to a process or patient.
Denmark’s newly established Treatment Council (which assesses whether the price of treatments and health technology measure up to the benefits for the patients) should play a role in AI adoption.
The Council is well-placed to approve solutions and so should be helped to understand how AI creates value for patients, and what approval of AI will entail.
Current funding and reimbursement models for AI in healthcare are lagging behind the technology, yet these are crucial to enable the scaling-up of AI in healthcare settings. It is vital that innovators capitalise on the increasing interest in AI solutions for healthcare to benefit from both private and public funding. But it is becoming clear that, once funding has been secured and AI technology developed, the financial challenges don’t stop. A pan-EU framework for reimbursement, that acknowledges the unique characteristics of AI and offers guidelines as to what evidence is required to secure reimbursement for AI solutions would enable successful solutions to be implemented in clinical practice efficiently and the ensuing benefits swiftly recognised.
Let us know your thoughts on how AI can be funded and reimbursed in healthcare by tagging @EITHealth on Twitter or Facebook, and use the hashtag #EITHealthAI.
References
1 EIT Health and McKinsey & Company. Transforming healthcare with AI: The impact on the workforce and organisations. 2020. Available from: https://eithealth.eu/wp-content/uploads/2020/03/EIT-Health-and-McKinsey_Transforming-Healthcare-with-AI.pdf (accessed January 2021).
2 Ministry of Economy and Industry, State of Israel. Digital Health: The Israeli Promise. 2020. Available from: https://investinisrael.gov.il/HowWeHelp/downloads/Digital%20Health%20-%20The%20Israeli%20Promise.pdf (accessed February 2021).
3 Halminen O, Tenhunen H, Heliste A, Seppälä T. Factors affecting venture funding of healthcare AI companies. Stud Health Technol Inform 2019; 262: 268–271.
4 Deloitte. The socio-economic impact of AI in healthcare. 2020. Available from: https://www.medtecheurope.org/wp-content/uploads/2020/10/mte-ai_impact-inhealthcare_oct2020_report.pdf (accessed January 2021).
5 Medtech Europe. Artificial Intelligence in MedTech: Delivering on the Promise of Better Healthcare in Europe. 2019. Available from: https://www.medtecheurope.org/wp-content/uploads/2019/11/MTE_Nov19_AI-in-MedTech-Delivering-on-the-Promise-of-Better-Healthcare-in-Europe.pdf (accessed January 2021).
6 European Commission. Proposed Guiding Principles for Reimbursement of Digital Health Products and Solutions. 2019. Available from: https://www.medtecheurope.org/wp-content/uploads/2019/04/30042019_eHSGSubGroupReimbursement.pdf (accessed January 2021).
7 Gray M. Value based healthcare. BMJ 2017; 356: j437.