AI Model Predicts Hospitalization Outcomes Among Dementia Patients

Researchers from Houston Methodist created an artificial intelligence (AI) model that can successfully predict hospitalization outcomes among dementia patients, leading to improved timeliness of care, care coordination, and resource allocation.

The use of AI to enhance treatment for various conditions is growing, and researchers continue to determine how to further improve its application in healthcare.

According to the Centers for Disease Control and Prevention (CDC), about 5 million adults over 65 battled dementia in 2014, and this number is on track to reach 14 million in 2060. 

To bolster care for dementia patients, Houston Methodist created an AI model that has proved successful in predicting hospitalization outcomes among geriatric adults with dementia.

Using data from 15,678 encounters representing 8,407 patients over 10 years, researchers defined risk factors linked to demographics, hospital complications, and pre-admission and post-admission data. After reviewing the risk factors associated with each type of dementia, they found that the most common ones were encephalopathy, falls, and pressure ulcers. This information assisted researchers in creating the AI model.

Researchers found that the model not only achieved an accuracy level of 95.6 percent but also performed better than other risk assessment methods for dementia.

Further, researchers noted that the model showed the potential to help reduce hospital stays.

“The study showed that if we can identify geriatric patients with dementia as soon as they are hospitalized and recognize the significant risk factors, then we can implement some suitable interventions right away,” said Eugene C. Lai, MD, PhD, the Robert W. Hervey Distinguished Endowed Chair for Parkinson’s Research and Treatment in the Stanley H. Appel Department of Neurology at Houston Methodist, in a press release. “By mitigating and correcting the modifiable risk factors for undesirable outcomes immediately, we are able to improve outcomes and shorten their hospital stays.”

Lai has been interested in determining subjective Parkinson’s disease patient responses to hospitalization for years. This led him to collaborate with Stephen T.C. Wong, PhD, who provided him access to a large data warehouse of Houston Methodist patients.

In the future, Lai and the researchers intend to use mitigation measures to assist clinical interventions designed to reduce adverse outcomes among these patients.

The use of AI to detect, diagnose, and treat various conditions has grown in recent years.

A blood test system created by researchers at Tulane University in September showed that the pediatric tuberculosis diagnosis process could be improved using AI and nanotechnology. Using these technologies, the system provided a view into lipoarabinomannan and LprG, which make up tuberculosis-causing bacteria.

An AI-based strategy created in September showed the ability of AI to detect atrial fibrillation. Researchers concluded that AI-guided electrocardiograms provide many benefits, such as precise measurements.