Today we are seeing a strong desire among healthcare organizations to incorporate Artificial Intelligence (AI) into their operating models as peers in the high tech and financial industries have already done. Over 50 percent of Accenture’s healthcare clients cite AI as one of their top three strategic initiatives within the next year1. In part this is due to the shift towards value-based care, which emphasizes improving patient outcomes and experience while reducing cost. AI holds tremendous promise in meeting these twin goals, particularly in the area of clinical efficiency. It has been touted as an enabler of improved clinical efficiency outcomes by supplementing, augmenting—and in some cases, supplanting—certain parts of clinicians’ roles in the delivery of care.
Our Accenture Health Strategy analysis finds that providers primarily want to use AI tools for productivity, data analysis, and cost savings. The clinical efficiency AI segment, if implemented correctly, has the potential to deliver over $10B of value to providers through cost avoidance via cognitive predictive analytics as well as cost savings via smart clinical decision and diagnostic support.2 The broader healthcare industry clearly believes that AI-powered clinical efficiency applications have the potential to deliver benefits. Large technology players are developing capabilities in the space and clinical efficiency start-ups are better funded than other players in the health AI space. However, we have yet to see the full value realized from this new generation of clinical efficiency AI.
Notwithstanding the general industry enthusiasm around AI, our analysis to date shows a lack of adoption or willingness to engage with clinical efficiency technologies specifically. We see that 68 percent of Healthcare providers have deployed AI within information technology, but have yet to leverage it at scale in the delivery of care.3 While providers highlight decision support as the primary application for AI, they cite limited use cases, discomfort with the technology, as well as data and workflow integration as key roadblocks to realizing its full potential in their organizations.4 Initial launches of clinical efficiency AI tools have yielded mixed results due to limited data interoperability and clinical workflow inefficiencies—heightening the healthcare industry’s resistance to adopting the technology as compared to other applications within the healthcare AI space.
Despite the lackluster debut of clinical efficiency AI at scale, there is significant opportunity in this space if providers are willing to embrace some of these technologies—particularly in areas where even early stage technologies can create efficiencies. To start realizing the value from clinical efficiency AI, providers must:
With increased collaboration between clinicians, software engineers, data scientists and user-experience experts, wide-spread adoption of clinical efficiency AI will follow.
One area within clinical efficiency that has begun to see success, due to high clinician engagement, is imaged-based diagnostics. Pathology image analysis can be subjective and requires the input of multiple physicians for complex cases. AI-enabled interpretation creates structure and standardization around diagnosis allowing clinicians to provide an answer more quickly and with more confidence. These tools have been implemented with limited disruption to clinical workflows through direct integration with existing imaging technology.
As AI-powered clinical efficiency applications and the broader health AI landscape continue to mature, it is critical that providers embrace AI to realize the technology’s full potential. Without buy-in from clinicians to expand use cases and ensure clinical utility, healthcare AI will continue to underdeliver on its promise of delivering clinical efficiency.
1 Accenture Strategy, Tech-led change for AI research, 2017
2 Accenture Strategy Health AI analysis, 2017
3 Accenture Strategy, Tech-led change for AI research, 2017