Is Deep Reinforcement Learning Ready for Practical Applications in Healthcare? A Sensitivity Analysis of Duel-DDQN for Hemodynamic Management in Sepsis Patients.
Mingyu LuZachary ShahnDaby SowFinale Doshi-VelezLi-Wei H. LehmanPublished in: AMIA (2020)
Keyphrases
- sensitivity analysis
- reinforcement learning
- clinical trials
- patient monitoring
- chronic disease
- home care
- managerial insights
- hospital information systems
- heart disease
- health care
- information systems
- initial stage
- cardiac surgery
- medical care
- post operative
- influence diagrams
- healthcare professionals
- health services
- intensive care
- clinical practice
- variational inequalities
- patient care
- decision making
- decision support
- health care services
- function approximation
- electronic patient record
- clinical decision support systems
- decision support system
- medical practice
- event related
- high risk
- clinical decision support
- dynamic programming
- hospital discharge
- medical data
- medical devices
- intensive care unit
- continuous monitoring
- state space
- computational fluid dynamics
- medical practitioners
- learning algorithm
- health information systems
- electronic health records
- cardiovascular disease