AI-Powered Healthcare Tools Address Diagnosis, Drug Discovery, and Data Privacy
Healthcare organizations are deploying artificial intelligence to tackle rising costs, limited access, fragmented data systems, and security threats. But these tools only work if clinicians and patients trust them-a challenge that remains unsolved.
At the Edison Awards panel Thursday, healthcare leaders discussed the practical obstacles to scaling AI innovations: proving clinical value, protecting patient data, and building confidence in new systems.
Using Voice to Diagnose Heart Disease
Mayo Clinic teams built an AI classifier that detects aortic stenosis-a narrowing of the heart's aortic valve-by analyzing a patient's voice. The system identifies disease signatures in speech patterns that would otherwise require elaborate testing.
Dr. Matthew Callstrom of Mayo Clinic said the approach matters because early detection allows patients to receive valve replacement and live a normal lifespan. Undiagnosed cases lead to heart failure.
"Now we're making a diagnosis for aortic stenosis with a phone rather than having to do more elaborate testing," Callstrom said.
The voice data itself presents a security problem. Voice is unique to each person and cannot be anonymized like other medical information.
Privacy Concerns Grow With AI Use
Even when images are stripped of names and identifiers, AI systems can infer a patient's race, age, and location from medical data alone, according to Merage Ghane of the Coalition for Health AI.
The problem intensifies when working with rare diseases or small patient populations, where fewer data points exist to hide behind.
Ghane recommended that organizations invest in cybersecurity training, hire privacy experts, and obtain relevant certifications to protect patient information.
Callstrom acknowledged Mayo Clinic's responsibility to secure voice data used in diagnosis. "It is characteristically you, and we know that it's super important to protect," he said.
AI Accelerates Drug Discovery-If Data Is Representative
Algen Biotechnologies uses AI and CRISPR gene-modulation technology to identify disease pathways and potential treatments. The company's product, AlgenBrain, won an Edison Award for its approach to drug discovery.
The traditional drug development process fails 90% of the time and takes years to complete. AI systems can compress timelines, but only if the underlying data reflects real patient diversity.
Co-founder Chun-Hao Huang said drugs developed on data from a narrow patient population will not work effectively for everyone. "If we're able to capture at an earlier stage the diversity of the patients, then in the end, that can ensure the drug that evolved is effective for every single patient," he said.
No FDA-approved drugs using this technology exist yet, but the pace of development suggests approval is coming. The bottleneck now is ensuring that data sets used to train these systems represent the full range of patients who will eventually use the drugs.
Learn more about AI for Healthcare and AI Data Analysis to understand how these tools work in practice.
Your membership also unlocks:
