By Justin Wang, SBHA Board Member and Youth Advisory Council member
The following reflects Justin’s lived experiences, thoughts, and opinions.
Ever since I started my time at Stanford, I have been surrounded by constant reminders of artificial intelligence (AI) and its increasing role in our day-to-day lives. Stanford’s Human-Centered Artificial Intelligence (HAI) Lab and Artificial Intelligence Lab (SAIL) are two heavyweight research organizations on campus, and dozens of AI-powered startups are involved with the student body here. Curious about AI’s relevance to healthcare, I decided to research its applications in the healthcare domain. According to Deloitte’s 2025 US health care outlook, 65% of healthcare executives identified developing growth strategies to increase revenue as a “very important” priority, making it the top industry trend of the year. AI is emerging as a powerful way for healthcare practitioners to accomplish this, as AI-embedded technologies are enabling automated diagnostic applications, efficient treatment planning, and increased operational efficiency. AI’s increased exposure in industry is in part due to recent advances in computational power, big data analytics, and machine learning, which have all contributed to its ability to revolutionize healthcare workflows cost-effectively and scalably.
Diagnostics have been – and are continuing to be – transformed by AI. More specifically, medical imaging analysis has been automated in areas including radiology and early detection of diseases, allowing for both speed and accuracy improvements. AI’s ability to quickly recognize patterns in patient data has improved the identification of disease markers and predicting of potential patient health issues. Large amounts of imaging data can be analyzed leveraging computational power, allowing for abnormalities and patterns that may be undetectable to the bare human eye. In fact, the Food and Drug Administration has approved almost 400 AI models to-date in just the radiology field alone. Use cases for medical imaging analysis include: X-rays for bone fracture detection; CT scans for lung nodule identification; MRIs for brain tumor segmentation; and ultrasound for cardiac function assessments.
Another key application area is in that of treatment planning and personalized medicine. AI-powered treatment recommendations allow for analysis of patient history against similar cases as well as drug interaction predictions. Many factors may be included in the automated recommendation process, such as medical history, genetic information, and lifestyle factors. These factors can then be used to predict potential health issues and identify high-risk patients, enabling practitioners to determine cases in which early interventions and preventative measures are necessary. On a more granular level, AI tools have also been deployed so as to develop personalized medicine plans. These tools are able to receive genetic data as inputs and identify disease susceptibility and selected target therapies. Right now, leading models have been proven to predict patient responses to immunotherapies with a 70-80% success rate. Though adoption is not yet widespread, such AI-driven tools have shown promise, with tailored cancer treatments, for example, having shown improved response rates and reduced toxicity levels in patients.
A less expected area of healthcare that AI is transforming is that of administrative and operational efficiency. According to a study published by Google Cloud, practitioners dedicate more than 30% of their working time towards paperwork and related tasks, eating up time that could otherwise be used to treat patients. Now, however, new AI tools are able to dramatically reduce administrative burdens typically faced by healthcare professionals by automating low-level work such as appointment scheduling, billing, and claims processing. These tools are able to do so by searching patient documents, analyzing historical patient volumes, and accessing staff skills. The results? More time for healthcare professionals to focus on patient care, reduced turnover costs for facilities, and reducing potential for data errors.
Several challenges stand in the way of full implementation of AI in healthcare, most notably skepticism with regards to data privacy and security. AI tools are driven by data, and as such, must have access to large quantities of patient data to operate efficiently. This concern has drawn legal attention, with lawmakers putting increased emphasis on regulating data usage with regards to health care in specific. Additionally, health care’s complex regulatory structures impose barriers that easily lead to potential misuse of patient information by consumer apps and digital health companies. The digitization of health information also increases the risk of cyberattacks, raising security concerns. But to even transfer data from pre-existing systems to AI tools is a challenge in and of itself. Integrating AI with legacy systems carries with it data merging difficulties. This is because health care organizations typically store data in legacy systems that are dispersed across a wide range of disparate data sources, making data compatibility especially rare. Integration imposes costs on prospective users, and though advances are being made in interoperability and flexible infrastructures, widespread implementation is still proving difficult.
The future outlook for AI’s applications in healthcare is promising, as it can serve as a powerful vehicle that enables practitioners to dedicate a larger portion of their time toward treating patients and expertise toward tackling the nuanced challenges that present themselves in niche patient cases. As the underlying technology powering these tools continues to develop, we are sure to see more and more powerful application areas with health. School-based health centers (SBHCs) are positioned to greatly benefit from these developments: administrative efficiency improvements could lift bottlenecks on resource-constrained SBHCs; AI-powered diagnostic tools could provide nurses and care providers with a means to validate their concerns; and treatment planning could enable continued care solutions for students. As discussed in the previous paragraph, however, there are several challenges that must be addressed, and responsible AI implementation methods must be established. Clear guidelines and regulations must be developed, allowing for safe, effective usage in healthcare contexts; measures must be taken to protect patient data, preventing the leakage of vital information to malicious hackers; and transparency must be incorporated into tools such that explainability does not become a crux to AI-practitioner workflows. Taking these actions is vital in creating a future where AI’s powerful capabilities contribute to improved care delivery, patient outcomes, and practitioner efficiency.
Sources
- 2025 US Health Care Outlook. Deloitte Insights. December 13, 2024. Accessed January 27, 2025. https://www2.deloitte.com/us/en/insights/industry/health-care/life-sciences-and-health-care-industry-outlooks/2025-us-health-care-executive-outlook.html.
- How ai is improving diagnostics, decision-making and care: AHA. American Hospital Association. Accessed January 27, 2025. https://www.aha.org/aha-center-health-innovation-market-scan/2023-05-09-how-ai-improving-diagnostics-decision-making-and-care.
- Artificial Intelligence in medical imaging: AI in Imaging. SpectralAI. August 14, 2024. Accessed January 27, 2025. https://www.spectral-ai.com/blog/artificial-intelligence-in-medical-imaging/.
- Medscience O, About The Author Open Medscience Open MedScience is a dedicated platform designed to facilitate discussions on a broad spectrum of medical imaging technologies. Our focus encompasses proton therapy. Artificial Intelligence in Healthcare: Revolutionising diagnosis and treatment. Open MedScience. December 2, 2024. Accessed January 27, 2025. https://openmedscience.com/artificial-intelligence-in-healthcare-revolutionising-diagnosis-and-treatment/.
- wplabru2020. The role of Artificial Intelligence in personalized medicine. Laboratorios Rubió. February 19, 2024. Accessed January 27, 2025. https://www.laboratoriosrubio.com/en/ai-personalized-medicine/.
- Butler N. AI-powered personalised medicine. HGF. September 19, 2023. Accessed January 27, 2025. https://www.hgf.com/blogs/ai-powered-personalised-medicine/.
- Staff R. AI models may predict patient response to immunotherapies. Respiratory Therapy. January 30, 2024. Accessed January 27, 2025. https://respiratory-therapy.com/products-treatment/industry-regulatory-news/business-news/ai-models-may-predict-patient-response-to-immunotherapies/.
- Shoaib Karim, Jack Shaw, et al. How AI can lead to Personalized Medicine. How AI Can Lead to Personalized Medicine | Insurance Thought Leadership. June 6, 2024. Accessed January 27, 2025. https://www.insurancethoughtleadership.com/life-health/how-ai-can-lead-personalized-medicine.
- Gupta A. How gen ai can help doctors and nurses ease their administrative workloads. Google. October 17, 2024. Accessed January 27, 2025. https://blog.google/products/google-cloud/generative-ai-healthcare-administration/.
- NetSuite.com. How AI is Transforming Patient Care & Health. Oracle NetSuite. Accessed January 27, 2025. https://www.netsuite.com/portal/resource/articles/erp/ai-healthcare.shtml.
- Olsen E. Lawmakers stress data privacy in Health Ai Oversight. Healthcare Dive. November 30, 2023. Accessed January 27, 2025. https://www.healthcaredive.com/news/healthcare-ai-data-privacy-house-energy-commerce-subcommittee/701126/.
- Risk management tools & resources. Artificial Intelligence Risks: Data Privacy and Security | MedPro Group. Accessed January 27, 2025. https://www.medpro.com/artificial-intelligence-risks-privacysecurity.