The path towards integrating digital technologies

In addition to the challenges posed by privacy, data biases, data sharing infrastructure and the digital skills gap, there is one key component for the successful clinical translation of AI-based technologies: practical integration into the healthcare system. The wrong approach would risk introducing unwanted biases or inequalities, and challenge the longterm sustainability of the system, both socially and financially. The rapid pace at which healthcare AI is developing means that decisions are being taken quickly and sometimes without direct precedents. We have investigated the main challenges involved in the integration process by studying a recent, very closely related example of a disruptive digital healthcare technology: virtual primary care services (VPCs). We have found that VPCs tend to service specific populations, particularly those who are younger and healthier. This, especially if combined with policies that require that patients de-register from their in-person GP to access virtual services, could lead to in-person GPs being left with increasingly complex patient cohorts, therefore threatening their operability. The precedent of VPCs suggests that the establishment of an effective parallel virtual /physical system may introduce risk segmentation, risk selection and cost-shifting. In order to improve the health of those who need it most, policies directed towards AI technologies may need to move towards unified systems to achieve balance and equity in terms of access, services and financial sustainability