This content is part of the Essential Guide: How healthcare AI technology can improve patient outcomes

Patient readmission rates may fall with more AI use in healthcare

AI has a number of uses in healthcare, including analyzing patient data to determine which patients have a high risk of being readmitted after discharge.

AI offers healthcare significant opportunities to improve patient care, costs and outcomes. Since AI adoption varies from hospital to hospital, CIOs that haven't implemented it yet are looking to introduce AI to their health systems by looking at where opportunities exist for its use. This is a critical first step to justifying the investment and its adoption.

Hospitals today sit on troves of data related to billing, clinical, scheduling, administrative and rich media. Advancements in AI make it a useful tool for patient data analysis that allows hospitals to mine that information and put it to good use. When applied to health data, predictive and analytical algorithms can provide tremendous value that can impact not only daily operations but patient care as well. 

For healthcare organizations just beginning their AI journey, the following examples show how hospitals can successfully use AI for patient data analysis to deliver meaningful results, including lowering patient readmission rates.

Resource and patient planning

Hospitals wrestle daily with staff and resource planning for the incoming flow of admitted patients into their facilities. Since it can be difficult to predict the number of patients coming through and how long their stay will be, hospitals struggle to maintain a balance of having enough staff to care for patients. With the help of AI and its predictive models, hospitals can now feed those models historical and new data to predict the expected length of stay. Hospital administrators and staff can use this patient data analysis to better prepare for the workloads they will face and ensure patients are getting the right level of care. These predictions are generally algorithms written in Python or R programming language and can be used by any hospital. They are generally available in AI platforms from Google, Amazon and Microsoft.

High-risk patient detection

The next area that AI has been able to assist with is the detection of high-risk patients. Since different patients have different risk factors based on their age, sex, medical history and vitals, it can be difficult for an EHR to accurately detect which patients are considered high risk. An EHR looks for specific criteria that allow it to flag a patient as high risk, but the system does not offer predictions based on probability and historical data. This is where AI comes into play. AI can perform patient data analysis within an EHR and compare it to patterns it recognized from previously analyzed data, and then can accurately detect the presence of risks for some patients. Several algorithms are already in use to detect patients with heart disease.

Patient readmission prediction models

Another area that has been a concern for healthcare is patient readmission rates. Since patient readmission rates impact patient outcomes and incur penalties from CMS, hospitals recognize that it is critical for them to keep those rates as low as possible. One method has been to monitor discharged patients that are considered high risk for readmission. This can be achieved by prioritizing the list of patients using an AI model that can perform patient data analysis to score and predict which patients are considered high risk. Not only does the algorithm help improve the patient's outcomes by getting to them before their condition worsens after they leave the hospital, but the healthcare organization avoids increasing their admission rates.

The use of AI continues to expand in healthcare. As more hospitals evaluate and engage in advanced analytics and other disciplines of AI, it can support financial performance goals, resource planning and productivity, as well as patient care and outcomes.

Dig Deeper on Healthcare IT systems and applications

Cloud Computing
Mobile Computing