Impacting patient care through early disease detection with AI

• Traditional healthcare focuses on treatment after symptoms appear, but early intervention prevents many chronic conditions
• Artificial intelligence enables earlier diagnoses, impacts patient care, and optimizes resource allocation
• Healthcare leaders must integrate AI and digital tools into workflows responsibly to better support sustainable, value-based care

Early disease detection utilizing AI (artificial intelligence) is a departure from traditional healthcare, which often centers on treating illness after symptoms appear, rather than preventing disease.¹ In the United States, a significant portion of overall healthcare spending is allocated toward treating chronic conditions. Many of these, including heart disease, diabetes, and cancer, are diseases that are preventable and better managed with earlier intervention.²

At the recent Healthcare Information and Management Systems Society 2025 (HIMSS25) Global Conference, healthcare leaders, including Dr. Okan Ekinci of Roche, Robin Carver of Prenosis, Dr. Anwar Jebran of Oak Street Health, part of CVS Health, and Sunil Dadlani of Atlantic Health, discussed practical strategies for improving disease prevention and diagnostics by incorporating artificial intelligence (AI) and digital tools into clinical care. Their discussion focused on improving patient care, enhancing the clinician experience, and delivering long-term economic value for health systems.

The U.S and many countries in the EU are shifting their healthcare approaches to proactive, preventative strategies^.3,4 Investments are growing in early disease detection with AI and prevention programs, and the market for digital tools that support early screening and continuous disease monitoring is expanding, Dr. Ekinci noted: “There is spending in technology and innovation, and readiness to actually shift more upstream in the early detection of disease to prevent downstream cost.”

However, several challenges impede the widespread adoption of programs leveraging early disease detection with AI. These include shortages in the healthcare workforce, limited availability of physicians, long wait times for appointments, and low adherence in routine screening programs.^5,6

AI-driven tools generate vast amounts of data that can inform clinical decisions. This data must be translated into actionable insights that guide targeted, individualized care. “We need to better understand how we use the multidimensional data of the patients to really target the right drug, the right patient,” Dr. Ekinci commented.

For sepsis, early detection remains difficult due to vague and varied symptoms, inconsistent definitions among healthcare decision-makers, and the rapid deterioration that sepsis can cause, often without clear warning signs. “Sepsis presents differently across patient populations,

making it challenging for clinicians to recognize when someone is becoming critically ill,” said Carver.

According to Carver, predictive algorithms powered by AI and machine learning can play a key role in improving diagnosis and treatment. These tools help make sense of massive volumes of clinical data, enabling care teams to identify which patients are currently ill or are at risk of becoming critically ill.⁷ These technologies also support more accurate antibiotic use, ensuring the right patients receive treatment while avoiding unnecessary prescriptions.⁸ Finally, AI assists hospitals in meeting compliance standards while ensuring that sepsis cases are documented and coded accurately, so providers receive appropriate reimbursement.,^9,10

“By integrating AI models into clinical decision-making, we can better identify which patients are at highest risk and that need additional resources, while also recognizing those at lower risk who may be safely directed to alternative care pathways,” commented Carver.

Oak Street Health, a value-based primary care provider for Medicare patients and part of CVS Health, conducted a study to explore how machine learning can support clinical decision-making and improve patient care.

Published in the New England Journal of Medicine Catalyst, the study utilized a machine learning algorithm to predict hospital admissions, mortality, and healthcare costs.11 By analyzing historical data from electronic health records (EHRs), including past hospitalizations, documented diagnoses, and markers of health status, the model stratified patients into risk tiers. Compared to clinicians alone, the algorithm more accurately identified patients likely to be hospitalized or incur high medical costs. 

Dr. Jebran remarked, “This internally developed tool helped us accurately identify high-risk patients to enable Oak Street Health to appropriately allocate resources to those most in need and have an impact on their health outcomes, have an impact on their admission, and then impact on cost.” 

Another promising area discussed by Dr. Jebran is the use of AI to enhance clinical decision support (CDS). AI-powered CDS tools are being explored to extract meaningful data from external sources, which could significantly enhance population health efforts. However, these models require high-quality, standardized data. “70% of the patient chart is unstructured data, which is free text, and it’s so hard to retrieve this data,” said Dr. Jebran.

Data harmonization, making data from different sources consistent and interoperable, helps build effective AI-powered CDS tools, which are then used to generate actionable insights for physicians and care teams. “We push those insights within the appropriate workflows,” said Dr. Jebran, “this is the new frontier of Clinical Decision Support.”

Modern care is moving toward real-time, on-demand, and setting-agnostic delivery, whether in acute care, ambulatory settings, retail clinics, or at home.12 AI and digital tools are accelerating this shift.

Predictive analytics and risk stratification models help care teams to intervene sooner, thereby improving and optimizing resource utilization at the population level. AI-powered clinical decision support systems provide real-time insights during care, reducing cognitive load on physicians and promoting more consistent, evidence-based decisions. “These technologies will continue to become extremely smart, not only in terms of speed, in terms of accuracy and in terms of their predictions,” said Dadlani.

Digital tools also support more personalized care pathways. AI can tailor monitoring plans, send alerts for preventive screenings, and flag missed follow-ups, helping patients stay engaged in their care.13,14 In diagnostics, AI accelerates imaging analysis, reduces diagnostic errors, and supports point-of-care and at-home testing. By automating routine tasks such as documentation and triage, AI also frees up providers to focus on the complex or high-value patient needs.

According to Dadlani, to successfully deploy and scale AI, healthcare organizations must manage key enablers:

• Robust digital infrastructure and system interoperability
• Strong governance frameworks
• Effective management and clinician training
• Sustainable funding models that support long-term integration

Finally, security and compliance should be embedded into AI solutions from the start and not treated as an afterthought. This ensures safe, scalable, and responsible innovation as healthcare moves into its digital future. “Cybersecurity and data privacy are at the core, at the heart of these initiatives,” concluded Dadlani.

AI and digital tools are transforming healthcare by enabling earlier detection, more personalized care, and smarter resource allocation. Realizing this potential requires strong infrastructure, responsible implementation, and collaboration across the healthcare ecosystem. As experts emphasized, now is the time for healthcare organizations to start integrating early disease detection with AI into clinical workflows to improve care and reduce costs.