Difference between revisions of "Healthcare AI Use Cases"

Revision as of 10:25, 10 July 2020

ClosedLoop AI Use Cases

Hospitals

• Appointment No-Shows - Predict patients most likely to miss appointments
• Predict patients most likely to acquire infections like sepsis
• Predict patients most likely to benefit from periodic check ups
• Leakage - Network Integrity - Out of Network

Health Insurers

• Risk Adjustment - Suspect Diagnosis
• Preventable Hospitalizations
• Readmissions - Predict patient readmissions
• Chronic Disease Onset and Progression - Predict onset or progression of chronic diseases.
• Patients with highest overall risk in the next 6 months or year

Providers & ACOs

• Total Risk: Who will be my most expensive patients this year
• ED Over-Utilization: Which of my patients would most benefit from establishing a relationship with a primary care provider?
• Readmissions: Which patients are most likely to be readmitted to the hospital?
• Transportation: Which patients are most likely to need help with transportation in order to make it to their appointments
• Rising Risk: Which patients are likely to see large increases in their overall health risk over the next 30/60 days?
• Chronic Disease: Which patients are most likely to develop or progress in the severity of a chronic disease
• Medication Adherence: Which patients are most likely to be noncompliant with medications?

Payers and Health Plans

• Trend: Which members are likely to see large increases in their overall health risk over the next three to six months?
• Readmissions: Which members are most likely to be readmitted to the hospital?
• Medication Adherence: Which members are most likely to be noncompliant with medication plans?
• Palliative Care: Which members are most likely to benefit from discussions with a palliative care specialist?
• ED Over-Utilization: Which members are most likely to benefit from establishing a relationship with a primary care provider?
• Disenrollment: Which members will most likely to disenroll?
• Fraud: How do new data sources allow health plans to spot fraud sooner?
• Payment Integrity: How can AI help ensure proper reimbursement?

Pharma & Life Science

• Biomarkers: Which patients will have an increased success rate based on biological factors?
• Drug-Combinations: Which drug combinations are most likely to be successful?
• Segmentation: Which groups of patients respond differently to treatment?
• Strategy: Which subpopulations should be included/excluded based off of predicted success rates?
• Responsive: Which patients are responding to treatment?
• Events: Which patients are most likely to experience adverse reactions?
• Effectiveness: How will clinical trial results translate into real world effectiveness?
• Value: What improvement in outcomes will a new treatment generate over existing therapies?
• Switching: Which factors are most relevant in understanding which patients switch drugs?
• Marketing: Which physicians can we market to?

Other

• Identify patients' with collapsed lungs - artificial intelligence-enabled mobile X-ray system flags images of patients who need immediate intercession, ensuring that patients with collapsed lungs receive more timely care. (7/8/20)
• Predicts patients' hospitalization risk and triages them to appropriate care - uses AI tech, EHR data and a questionnaire to perform clinical intake of patients visiting ER, Urgent Care or Home (6/30/20)
• Diagnosed breast cancer at a higher rate than 11 pathologists. AI model using algorithms and deep learning (12/12/17)
• Reducing cost of Rx at a system level without sacrificing quality

• Emergency Department Demand Prediction
• Emergency Department Utilization Prediction
• Length of Stay Prediction