By Kumar Chebrolu, principal, Deloitte Consulting LLP

Global health care systems are facing increasing challenges due to rising demand, unsustainable staff shortages and funding and resource constraints. Advanced digital technologies, such as AI, can provide innovative, cost-effective solutions to these challenges. This week’s blog by Kumar Chebrolu, a principal in our US Consulting practice, appeared first as a US Center for Health Solutions Health Forward blog. It explores how AI technologies are increasingly being seen as a critical tool to optimise the use of data to improve outcomes, spot potential health risks and generate predictive clinical insights. The use cases that US health care systems are beginning to scale, present useful insights for UK and European health care systems to consider.

Introduction

Remember fax machines? At least 70 per cent of health care providers still rely on fax machines to exchange medical information!¹ Knowing that, it’s not too surprising that the health care sector has been slow to embrace artificial intelligence (AI). That situation, however, appears to be changing.

According to our recent report, Health care's quest for an enterprise-wide AI strategy, 220 hospital, health system, and health plan executives consider AI to be ‘important’ (30 per cent) or ‘very important’ (64 per cent) for remaining competitive. Moreover, 85 per cent of respondents expect their AI investments will increase during the 2022-2023 fiscal year. The survey responses were a subset of Deloitte’s recent State of AI in the Enterprise survey of nearly 3,000 global technology executives across all industries. Our 2022 survey follows our 2020 report, Smart use of AI in health care.

Many of our large health care clients have introduced AI and are now working to standardise data so that the technology can be scaled and expanded enterprise wide. These clients are also focusing on ethics/trustworthy AI, and ensuring that the technology doesn’t inadvertently automate biases (see Trustworthy AI: Bridging the ethics gap surrounding AI). Biases that wind up in algorithms could result in inaccurate clinical decisions, missed diagnoses, worsened clinical outcomes, and substandard patient experiences (see Could advanced analytics automate racism in health care?). It’s also important that the data and algorithms are kept current. An algorithm built on year-old data likely won’t be effective. In addition, some clinicians still need to be convinced that this technology will help them spend less time on paperwork and other tasks and more time providing direct care to patients.

Did COVID-19 help accelerate AI use?

The COVID-19 pandemic exposed the dangers of outdated technology as many local and state health departments struggled to stay on top of a rapidly spreading virus. It also helped to illustrate the potential of AI in helping respond rapidly to a health emergency. A growing number of health care organisations are using data and AI/machine learning to improve outcomes, to spot potential health risks in patients, and to generate predictive clinical insight. Since the pandemic emerged two years ago, some health care stakeholders have made significant investments in the technology. According to our interviewees, these investments turned some AI pilots into full-scale implementation initiatives.

Mount Sinai in New York City, for example, might have been the first US health system to use AI (along with imaging and clinical data) to identify COVID-19 patients. Last year, the Icahn School of Medicine at Mount Sinai launched a department of AI and human health.² Here are a few other examples of how AI is being used by health care organisations:

• The Mayo Clinic has a dedicated Department of Artificial Intelligence and Informatics. The health system has spent the last several years consolidating and standardising all its data. AI is now being used ‘to program computers to process and respond to data quickly and consistently for better treatment outcomes’, according to the health system.³ In cardiology, for example, AI is being used in screening for left ventricular dysfunction in people who do not have noticeable symptoms. The AI-assisted screening tool identified people at risk of left ventricular dysfunction 93 per cent of the time, according to a press release. The health system also is using AI to predict breast cancer in patients. The model uses 84 input variables, of which EHR data is only a small portion.⁴
• New Jersey-based Hackensack University Medical Center recently received funding to implement AI-assisted colonoscopy technology for underserved communities. The hospital intends to use the tool to increase screening and early detection of colorectal cancer.⁵
• Elevance Health (formerly Anthem Inc.) uses AI algorithms to search for fraud and abuse in insurance claims. The company is working to develop a synthetic data platform that will expand on its use of AI while reducing privacy issues surrounding personal medical information.⁶
• A new AI-empowered system at John’s Hopkins University scours medical records and clinical notes to identify patients who might be at risk of infection before symptoms develop. A study conducted by the health system found that patients are 20 per cent less likely to die of sepsis as a result of the new tool.⁷

Establishing an enterprise-wide AI strategy

Our survey findings showed that just one in three executives strongly agree that their organisation has an enterprise-wide AI strategy. In interviews with health care executives, we heard that AI sometimes loses out to other organisational priorities and might not be featured in the organisation’s enterprise-wide strategy. “There are a growing number of AI-powered applications across the enterprise, but most are utility-based, and there isn’t a unified organisation-wide strategy on AI,” one executive told us.

At this point, the use of AI is typically siloed or focused on specific projects, but many health care organisations are working towards more of a platform solution. Some organisations might start small and expand an AI application after demonstrating some success. According to our survey, some health care organisations are considering ways to implement an enterprise-wide strategy for AI. Some health plans are collaborating with health systems to develop AI models to improve early intervention among high-risk patients.

While the health care sector has long lagged other industries in the adoption of technology, the implementation of AI-based tools appears to be picking up steam. Five years from now, I expect mature AI applications will be integrated into a wide range of health care enterprise applications. I also expect the costs of implementation will be substantially lower.

Acknowledgements: Hemnabh Sandip Varia and Maulesh Jagdish Shukla

Kumar Chebrolu – Principal, Deloitte Consulting LLP

Kumar is a Principal in Deloitte Consulting LLP’s Life Sciences and Health Care practice. He leads Deloitte’s Applied AI and Digital Analytics practice for the health care sector, and his work ranges from identifying growth opportunities, to implementing transformational programs, building innovation capabilities, and creating disruptive products and services by leveraging artificial intelligence and machine learning, expert decision support systems, digital analytics, and cloud computing.

Email | LinkedIn

___________________________________________________________________________

¹ Health care clings to faxes as US pushes electronic records, Bloomberg Law, November 4, 2021; Office of the National Coordinator for Health Information Technology

² Mount Sinai first in US to use artificial intelligence to analyze coronavirus patients, Mount Sinai press release, May 19, 2020

³ Overview of Mayo Clinic’s Department of Artificial Intelligence and Informatics, Research Departments and Divisions, Mayo Clinic

⁴ Artificial intelligence in cardiovascular medicine, Mayo Clinic

⁵ NJ hospital to implement colonoscopy AI tech in underserved communities, Health IT analytics, August 5, 2022

⁶ Six trends driving digital transformation in healthcare, Elevance Health press release, June 28, 2022

⁷ AI speeds sepsis detection to prevent hundreds of deaths, How payers are using predictive analytics and virtual care to improve care delivery, MedCity News, August 8, 2022, July 21, 2022