3 proposals to empower AI startups in healthcare
The use of AI technologies in healthcare is accelerating exponentially. The global market size is expected to grow from $4.9 billion in 2020 to $45.2 billion in 2026. AI startups are developing solutions for diagnostics, clinical decision support, precision medicine, drug discovery and patient monitoring. Recently, we have witnessed startups pivoting to use their AI-based technology to help tackle the pandemic. Examples include COVID-19 assessment chatbots, digital monitoring of COVID-19 patients and tools for real-time mapping of the spread of infection.
Cutting edge AI startups are unlocking the potential to improve health outcomes in a variety of ways, but multiple challenges remain. In this blog post, we share 3 ways policy makers can empower AI startups that are contributing to the digital transformation of healthcare.
- Develop understandable and implementable rules – Provide a clear opportunity for AI startups to scale up
Startups seek to be compliant with regulation from Day 1. However, a growing number of regulations require additional resources and contribute to uncertainty.
Earlier this year, the European Commission released the White paper on AI, outlining its AI strategy. According to the White Paper, a risk-based approach could be introduced, with a distinction between high risk and low risk AI application. In the healthcare sector, AI would be considered ‘high-risk’, and thus subject to prior conformity assessments – procedures for testing, inspection or certification.
We believe that the ambition of the AI regulatory framework should be that AI startups that operate in the high-risk area should have a clear perspective to scale-up too. The framework should be developed in dialogue with innovation ecosystems and take into account the interconnected nature of rules currently in place, such as the GDPR and Medical Device Regulation. We also advocate for regulatory sandboxes at the EU level as a tool for reducing barriers to entry.
- Address data fragmentation and data collection issues – Ensure interoperability of Electronic Health Records and improve access to secondary data
Digital health entrepreneurs cite the need for widespread adoption of electronic health records and enhanced access and sharing of health data. This is especially true for AI startups in healthcare: according to a report from EIT Health and McKinsey, lack of interoperability and systems for data sharing was cited by startup executives and investors as the number one barrier for scaling AI in healthcare.
As the first step, governments should support health care providers in the adoption of electronic health records systems (EHRs). Lack of EHR interoperability should also be addressed – many independent systems used by the healthcare organisations are not interoperable between each other. Thus they cannot interact between different providers and systems. Last year, the European Commission adopted a recommendation on a European Electronic Health Record exchange format. This common framework aims to facilitate the interoperability of health data and its cross-border exchange.
Secondly, governments should facilitate the use of secondary health data for research and innovation. One of the best practise examples is legislation in Finland that provides a GDPR-compliant legal basis for health data processing and access, including for research and development in the health sector. At the European level, the development of a European Health Data Space holds potential for harmonising health data processing frameworks across the EU. (You can find our post on European Health Data Space here.)
- Foster public-private partnerships – Incentivise healthcare providers to adopt digital solutions
Partnerships and pilots with hospitals and care organisations are key to success for AI startups. It not only enables startups to validate their products but also allows healthcare professionals to influence their development and gain valuable skills.
However, one of the barriers for startups is difficulties in connecting with and partnering healthcare providers. It is a sector which has not had many incentives to embrace innovation or try out new solutions. Moreover, hospitals and clinics rarely have resources or dedicated personnel to deal with the potential partnerships and adaptation of new technologies.
Another problem is the lack of funding and skills for adopting digital solutions in healthcare organisations. According to a report from EIT Health and McKinsey, lack of skills in data analytics and shortage of funds was cited by healthcare professionals as the primary barriers for scaling AI in their organisations.
Governments should make efforts to facilitate partnerships and enable healthcare providers to be better prepared to adopt digital solutions, including AI applications. At European level, the proposed EU4Health programme aims to support the digital transformation and innovation of healthcare systems. If an agreement is reached, the programme could be a part of the solution to this issue.
Overall, artificial intelligence technologies are unlocking numerous opportunities for startups to innovate and for health care systems to improve health outcomes. We believe these 3 recommendations to empower AI startups have the potential to greatly support the digital transformation of healthcare while enabling startups to focus on what they do best – innovation.