In addition to the broad use cases above, business and technology leaders cite four additional use cases as high priority for deeper exploration. While they do not currently meet the strict criteria for immediate acceleration (impact potential, global applicability and ripeness for public-private collaboration), they each have a strong value proposition and potential to positively impact global health and healthcare. Many of these use cases strongly leverage emerging generative AI methodologies. As AI capabilities rapidly advance, these use cases should be closely monitored, and strategies for collaborative scale-up must be defined. A key opportunity for lessening healthcare workers’ burden is automating triage of patients for follow-up or acute care. AI-powered chatbots and call centres can educate patients and help assess whether an evaluation is needed, especially in low- and middle-income countries with a shortage of frontline healthcare resources. For example, healthtech organizations such as MyndYou21 have made progress using AI to automatically assess patient symptoms, decide if a patient needs to see a healthcare provider and furnish follow-up education. However, surveyed leaders agreed these tools need to better integrate with electronic medical records systems, care management platforms and remote patient monitoring systems before scaling is possible. Continued refinement based on dialect and culture presents another roadblock that must be overcome. New applications of generative AI may solve some of these issues, and enable this use case to further scale. Major technology firms are showing strides here, for example Google’s Med-PaLM tool, which has demonstrated clinical acumen similar to doctors in some dimensions. The COVID-19 pandemic exacerbated an already dire shortage of frontline healthcare workers, leading to patient backlogs and provider burnout. In the US, administrative tasks account for roughly a third of healthcare costs and up to half of a provider’s time for some specialities.22 Experts interviewed agreed that administrative AI could automate many repetitive tasks, such as recording patient notes and managing patient coding and billing. Employing AI to reduce administrative burden and health system costs has a clear value proposition, especially in high-income countries like the US and those within the EU, by helping ease healthcare’s labour shortage and allowing providers to practice “at the top of their license”. However, leaders cite poor data management of medical records as one of the primary barriers to automation, often making the problem to solve less of an AI issue and more of a foundational systems-and-processes one. Algorithms need data to flourish. Microsoft is making headway using conversational AI during patient visits to transcribe notes and fill in billing and coding information.23 Automating this process will lessen the heavy provider burden for notetaking, billing, coding and documentation, which can require significant time outside patient care. Microsoft likely will continue making advances in administrative AI with its push into generative AI through OpenAI and ChatGPT; alongside other key technology firms. Identifying redundancy and waste is the most ‘boring’ application of AI and possibly more impactful than we would give it credit for. Between 40% and 60% of every physician’s hour is spent on admin tasks… Easing the administrative burdens on doctors reduces burnout and improves patient care, two really important goals in healthcare. Genya Dana, Head, Global Health Policy, Avellino Labs Employing AI to reduce administrative burden and health system costs has a clear value proposition, especially in highincome countries like the US and those within the EU. AI-powered chatbots and call centres can educate patients and help assess whether an evaluation is needed, especially in low- and middleincome countries. 2.1 Patient triage AI 2.2 Administrative AI Scaling Smart Solutions with AI in Health: Unlocking Impact on High-Potential Use Cases 22
RkJQdWJsaXNoZXIy NzQwMjQ=