Delirium can affect around 60%–80% of patients who are intubated in the intensive care unit (ICU). This state of disorientation and confusion can lead to longer hospital stays, increased mortality and higher rates of post-delirium cognitive decline.

Johns Hopkins researchers are hoping an AI-equipped headset can help lessen the effects of delirium. The headset tracks eye movements with an infrared camera and can send automated and recorded audio messages as well as play music. Clinicians will investigate if such a device can help reduce the duration of delirium, and they will also look at how decreased delirium affects long-term cognition by evaluating patients six months after discharge.

“This is a way to engage, orient and ground patients,” says study co-leader Haroon Burhanullah, a psychiatrist at Johns Hopkins Memory and Alzheimer’s Treatment Center and director of the Psychiatric Emergency Services Consultation Liaison Service in the emergency department at Johns Hopkins Bayview Medical Center.

“We hypothesize that the device will reduce the duration of delirium by making patients more aware and increasing circulation in the brain. It will give them more delirium-free days by improving communication, orientation and cognitive stimulation.”

Burhanullah is leading the pilot study, and co-investigators are Paul Rosenberg, co-director of the Memory and Alzheimer’s Treatment Center, and Wendy Ziai, co-director of the Johns Hopkins neurocritical care unit. It is supported by a $200,000 grant from the Johns Hopkins Artificial Intelligence and Technology Collaboratory for Aging Research (JH AITC), which is funded by a National Institute of Aging grant. It is among a number of JH AITC-funded studies that aim to determine whether AI can help neurological conditions and improve long-term health of older adults. Other research is looking at predicting cognitive fluctuations in patients with Alzheimer’s, enhancing deep sleep with acoustic stimulation through a smart headband to slow cognitive decline, and having virtual voice assistants perform cognitive tests.

Another Johns Hopkins researcher, intensivist Robert Stevens, published a study that successfully used AI algorithms to detect early warning signs of delirium and predict patients who are at risk for delirium in the ICU.

The headset Burhanullah and Rosenberg are studying, which was originally developed to help patients with ALS communicate, can play automated audio messages to orient patients and allow them to communicate. For example, it can ask patients if they are in pain and tell them to blink if so. It can tell patients the date and ask them if they want to hear the weather, or listen to music. (The headset can be programmed to play the patient’s favorite music.) Family members and friends can download an app and record audio messages to be played through the headset as well.

“Patients are going through different phases of awareness and consciousness during their recovery in the ICU, and this device can be adjusted according to their communication needs,” Burhanullah says. “As they open their eyes during the waxing and waning of alertness, it can send them messages and orient them on an hourly basis.”

The headset can also conduct the Confusion Assessment Method for the ICU (CAM-ICU) questionnaire, a tool used to assess delirium in patients in the ICU through a series of yes or no questions and commands. The research team will compare these results with those from manual CAM-ICU tests given by nurses. Clinicians can use these data and other data the headset collects, including material on patient alertness and activity. The headset can even help with physical therapy, instructing patients to move their hands and feet to a rhythm with audio guidance, Burhanullah says.

The study, which will be conducted in the Johns Hopkins Bayview ICU, will include about 30 patients who are age 50 or older, intubated and experiencing delirium. For half of them, the device will be activated, and for the other half, it will not.

Although there are already tests to gauge a patient’s level of delirium, this can be time-consuming for ICU staff members who may be monitoring multiple patients with the condition.

“This headset could be a game-changer,” Burhanullah says, “If it reduces duration of delirium, it will improve mortality. It would have secondary benefits of reducing hospital costs and nursing staff burnout.” Additionally, the device could help patients who have certain biomarkers that connect delirium with Alzheimer’s, he says.