AI can be instrumental in helping hospitals to boost decision-making by using quantifiable metrics to assess the quality of healthcare. As with medical practices, robotic process automation plays a key role in streamlining hospital operations. We mentioned Olive earlier in the context of billing and coding for medical practices. Their use of robotic process automation, computer vision and machine learning to automate healthcare administrative workflows like checking claim statuses and managing accounts receivable adds significant value to hospital operations.
AI optimization can be critical for managing healthcare assets like hospital rooms, operating rooms and radiology scans, as well as vaccine administration, bed allocation and sequencing activities (room cleaning, discharges, etc.). Companies like LeanTaas are using AI to help hospitals to create capacity. To make this a reality, we need electronic data sources and optimal access to different hospital databases, as well as the ability to aggregate and analyze. This may mean redesigning data systems and processes and introducing technology for optimization and forecasting.
There are a million hospital beds in the United States. Room turnover in hospitals is far less predictable than it is in hotels, and so it’s harder to manage capacity and operations. Prior to being discharged from a hospital, a patient might need to wait for labs, post-surgery observations and even a discharging physician who’s dealing with an emergency. Using sophisticated AI models that are far better than traditional forecasting tools, you can better forecast and redesign processes to unlock significant capacity.
Given that AI solutions can make sense of data in an ongoing manner and feed those insights to a medical facility, they can improve both clinical and operational decisions, which are very much intertwined. If sepsis is a big issue in a hospital, patients stay longer and that affects the availability of beds, along with other hospital economics. If hospital capacity is better managed using AI’s predictive capabilities, more beds will be available, which means patients will get admitted and receive treatment much faster, leading to better outcomes. As such, many of the AI solutions that improve hospital operations are addressing ongoing clinical issues. Examples of those include:
- Optimizing scheduling and staffing tasks
- Improving clinical workflows by scanning documents and providing recommendations
- Automating tasks like reading scans or interpreting results to expedite decision-making
- Using algorithms to reduce scanning time and radiation to the patients
- Using algorithms to expedite image reconstruction
- Using segmentation to improve radiation therapy
- Predicting hospital-acquired infections such as sepsis and C. Diff
- Using infrared, computer vision and sensors to reduce nosocomial infections and track handwashing
- Tapping into reinforcement learning to wean patients off ventilators
- Providing decision support for the use of IVF, pressors and ventilations
- Powering ICU surveillance of patient vitals and other metrics, as well as ICU discharge
- Monitoring conditions in the operating room
- Preventing patient falls
- Using remote monitoring and algorithms to pick up on brewing issues such as sepsis or heart failure
Much of this revolves around better outcome prediction, which in turn leads to better resource allocation. The information required for this type of predictive modeling comes from EHR data and hospital bed data, using metrics like hospitalization length, the number of procedures and the number of patients with sepsis and other nosocomial infections.
