From symptom to treatment: Navigating the patient journey with AI

© Prostock-studio - stock.adobe.com

With health systems worldwide grappling with mounting pressures tied to workforce shortages and staff burnout, emerging technologies – specifically artificial intelligence (AI) integrated with virtual triage and care referral – are emerging as powerful tools to streamline patient care, improve outcomes, and reduce costs. As a medical advisor focused on AI applications in healthcare, I have been integrally involved in research and development that is transforming the patient journey, with a particular focus on care acuity alignment and early detection of critical conditions.

Recent research has broken new ground in two areas: care acuity alignment with respect to care de-escalation and escalation, and early detection of leading chronic diseases, including heart attack, stroke, pulmonary embolism, pneumonia, and asthma.

Aligning care with patient acuity levels

AI plays a crucial role in helping patients navigate the complex healthcare landscape, especially when they are uncertain about their care needs. AI has value for patients who are unclear about whether they need care in the first place, and if they do, what level of care is appropriate. Through virtual triage encounters, AI-powered systems can assess symptoms, generate potential diagnoses, and recommend appropriate levels of care acuity.

A recent study published in the International Journal of Healthcare evaluated the impact of AI-based virtual triage and care referral (VTCR) on patient care intent and seeking in an ambulatory setting. Researchers analyzed 8,088 online encounters to understand how VT influenced patient behavior in engaging various levels of care acuity. The technology is designed to evaluate patients’ care intentions and to help align patients with the appropriate care that is needed based on their clinical presentation. VTCR was found to reduce unnecessary in-person visits and promote virtual care among patients seeking care in a leading ambulatory care system.

Among the results:

• A 19.1% increase in patients opting for virtual care options such as e-visits and telephone consultations.
• 12.5% decrease in outpatient care seeking, including in-person and video consultations.
• 35% of all patients altered their care seeking behaviors following VT recommendations, and among patients whose care intent differed from VT, 50% altered their care seeking in alignment with the recommendation of VTCR.

VTCR technology reduces patient uncertainty by providing clear guidance on whether self-care at home is sufficient, if a visit to their regular physician is necessary and sufficient, or if urgent care and emergency department attention is required. The AI explains its recommendations, helping patients understand the rationale behind the suggested care path.

Early detection of critical conditions

One of the most promising applications of AI in healthcare is its ability to facilitate early detection of potentially life-threatening conditions. Recent research has focused on five high-morbidity, high-mortality conditions: heart attack, stroke, asthma, pneumonia, and pulmonary embolism.

Findings revealed a significant disconnect between patients’ self-perceived acuity levels and the actual urgency of their conditions as determined by AI-based virtual triage. Substantial numbers of individuals did not intend before triage to seek the level of urgent care that they needed clinically.

By identifying these discrepancies, AI can prompt patients to seek appropriate care sooner, potentially saving lives and reducing long-term health consequences. For conditions like stroke and heart attack (myocardial infarction), where time is critical, early intervention facilitated by AI VTCR could mean the difference between long-term disability and a full recovery.

Enhancing clinical decision-making

For healthcare providers, AI-powered virtual triage systems offer valuable insights that can expedite and improve clinical decision-making. By providing a ranked list of potential diagnoses based on the patient’s reported symptoms, AI gives clinicians a head start in their assessment.

It saves time and helps to organize information, offering a clinical workflow advantage by potentially expediting the ordering of therapeutics and delivery of care to the patient. This efficiency not only benefits the healthcare system but also ensures patients receive timely, appropriate care.

Challenges and considerations

While the potential of AI in healthcare is vast, its implementation comes with challenges. Healthcare providers and administrators may initially approach AI with skepticism, requiring education and transparency about the technology’s capabilities and limitations. It is crucial to prioritize patient safety in AI design. For instance, some systems are built to slightly “over-triage” to emergency departments, erring on the side of caution rather than risking under-diagnosis of serious conditions.

Looking ahead, several exciting developments in AI-powered healthcare are on the horizon:

• Integration of objective clinical data: As more patients use devices like pulse oximeters and blood pressure monitors at home, AI systems will incorporate this data to enhance diagnostic accuracy and care recommendations.
• Improved language models: Advancements in large language models (LLMs) will increase the fluidity and power of AI in healthcare applications as well as increasing patient comfort and satisfaction with VTCR.
• Expanded condition coverage: AI systems will become more adept at identifying and managing rarer conditions and mental health issues.
• Addressing socially stigmatized conditions: AI may provide a more comfortable platform for patients to discuss sensitive health issues like sexually transmitted diseases or substance abuse.
• Focus on chronic disease management: As the leading source of morbidity and mortality in most nations, an increasing focus on chronic diseases in AI-based VTCR will play a larger role in long-term patient care and monitoring.
• Tackling healthcare access and resource inequities: AI-powered virtual triage can improve healthcare access for underserved, high inequity populations, addressing long standing disparities in care.

We are at the beginning of the healthcare journey with AI-based VTCR. As AI continues to evolve, its ability to align care with patient needs, detect critical conditions earlier, refer patients for needed care, and support clinical decision-making will increase dramatically.

AI is poised to revolutionize the patient’s and the clinician’s journey, from initial symptom assessment to long-term care management. By leveraging VTCR technology to improve care acuity alignment and early detection of serious conditions, healthcare providers and plans can enhance patient outcomes, reduce costs, and create a more efficient, equitable healthcare system for all. For senior healthcare executives, embracing and integrating these AI technologies will be crucial in staying at the forefront of patient care and operational efficiency and organizational performance in the coming years.

George Gellert MD, MPH, MPA is a physician executive and epidemiologist focused on using health information technology to improve population and public health outcomes through transformative organizational strategy and services. With a proven track record of formulating and executing innovative and transformative organizational strategy and operations in industry, public, and non-profit sectors, Dr. Gellert has a history of results-oriented leadership within healthcare, health information technology, public/population health, and within domestic and international public health. Dr. Gellert serves as a peer reviewer for 31 biomedical, public health, and IT journals, and has over 180 journal and book publications.