About 90% of the U.S.’s annual $4.1 trillion health care expenditure goes to treatment and management of chronic diseases.1 These diseases present a significant burden for healthcare providers, accounting for around 74% of all deaths globally.2 A global shortage of 10 million healthcare workers predicted by 2030 compounds this burden.3
To mitigate these challenges, screening and early detection can play a crucial role in reducing care costs for these diseases and improving patient outcomes. However, current programs have low adherence; fewer than 70% of individuals aged 50-75 get colorectal cancer screening in the U.S.4 And age-related guidelines don’t always correspond to risk; according to Aamir Ali, MD, gastroenterologist and Co-Director of Clinical Research at Capital Digestive Care (CDC), colon cancer rates have risen sharply in individuals younger than 45. “We don’t have a good playbook on how or whether to screen these individuals,” he added.
How AI enhances at-risk patient identification and improves care quality
Artificial intelligence (AI)-based algorithms can help provide that playbook and augment clinical decision-making. These algorithms analyze data to find at-risk patients even outside recommended guidelines. Ali said that once CDC’s AI-powered, non-invasive risk stratification tool identifies high-risk patients, their screening procedures are expedited.
AI-based solutions can also compensate for varying levels of quality in colonoscopy procedures. CDC uses an AI tool to detect possible adenomas (pre-cancerous polyps), highlighting these areas to assist endoscopists in examining them further. Results from this program indicate that adenoma detection rates increased and variation among endoscopists decreased.
Anwar Jebran, MD, primary care physician and Medical Director of Population Health Informatics at Oak Street Health, said his organization uses AI-based algorithms to detect chronic conditions such as congestive heart failure, COPD and diabetes using health signals from medical records.* This effort factors in patients’ social determinants of health to find patients’ needs that community health workers can help fill. Such assistance includes guiding patients through scheduling screenings and even arranging transportation. “Machine learning helps us shift our resources when appropriate,” he said.
Best practices for building leadership and clinician support of AI tools
Matt Prime, MD, International Business Leader for Clinical Decision Support and Algorithms at Roche, recommended that organizations looking to implement AI-driven disease screening programs frame their efforts in terms of the quadruple aim: better outcomes, lower cost and improved experiences for patients and clinicians. “Roche has consistently demonstrated expertise at integrating new technologies into clinical practice and effectively developing products that gain regulatory approval,” he noted. To support this mission, Roche is dedicated to creating innovative digital health tools and fostering trust in their use among clinicians.
According to Ali, using an organization’s own data can communicate the value of AI screening tools more clearly to clinicians. “It’s one thing to hear that, nationwide, [more than] 30% of people are not adhering to colon cancer screening,” he added. “It’s entirely another to hear that 25% of [one] physician’s patients are not adhering. It’s more impactful when the data is local.”
Jebran said that leadership support is vital for introducing disease screening algorithms, as well as clearly communicating algorithms’ trustworthiness and how they are used to clinicians and patients. “We are in an inflection point in terms of innovation,” he concluded. “To us, [leveraging AI tools] is the future. Investing in them is great, and it will only get better.”
*Views and opinions expressed by Dr. Jebran are his own and do not represent Oak Street Health or CVS Health.
References
- U.S. Centers for Disease Control and Prevention. July 12, 2024. Fast facts: Health and economic costs of chronic conditions. https://www.cdc.gov/chronic-disease/data-research/facts-stats/.
- World Health Organization. December 23, 2024. Noncommunicable diseases: Mortality. https://www.who.int/data/gho/data/themes/topics/topic-details/GHO/ncd-mortality.
- World Health Organization. Health workforce. https://www.who.int/health-topics/health-workforce#tab=tab_1.
- National Colorectal Cancer Roundtable. Colorectal cancer is a major public health problem. https://nccrt.org/our-impact/data-and-progress/.