🔒Gaylord Healthcare tests AI for scheduling, care insights

Gaylord Specialty Healthcare is stepping into new territory, exploring machine learning, artificial intelligence and automation software in the hopes of finding new efficiencies and even identifying challenges (and solutions) it didn't know existed.

Wallingford-based Gaylord, one of two long-term, acute-care hospitals in the state, recently hired Groton-based Decision Options, which will put its software to work analyzing the hospital's reams of data on medical care and outcomes, as well as patient appointment scheduling.

The dual aim for the 137-bed hospital, which also has outpatient facilities, is to produce insights that lead to better business and care decisions.

Trying to predict the future is part of the goal, such as identifying patients who are more likely to miss an appointment or need to be transferred to a regular hospital, said Peter Grevelding, Gaylord's vice president of clinical operations.

Traditional tools like spreadsheets can be helpful in analyzing data and trends, but Grevelding said Decision Options' software platform works on another level, with the potential to analyze a much more complicated universe of data from multiple sources and systems.

The next few months will be spent importing Gaylord data and making other tweaks before going live with the platform. The partners plan to start small, with one specific project, but Gaylord already has plenty of ideas about what to investigate.

One focus area is analyzing ways to reduce the number of times Gaylord inpatients — who often have long, slow recoveries ahead — are transported back to a traditional emergency hospital for more intensive care.

Decision Options' prediction models could lead to customized-care interventions before a potentially avoidable problem arises.

Another focus is developing more efficient scheduling. Certain patients may take longer with their doctors than others, or cancel more often, and scheduling practices may not account for it. That also affects staffing levels and needs.

“Ideally, we want our staff busy and providing care and we don't want big openings in the schedule,” Grevelding said.

First hospital client

Gill Eapen, who during his time as a Pfizer executive in the late 1990s helped decide the best balance for the pharmaceutical maker's R&D investments, founded Decision Options in 2001 as a boutique advisory firm that helped companies make better decisions.

That experience led to the development of its software, which the company markets to a broad range of industries, including health care, manufacturing and professional-service firms.

Though Decision Options counts some Michigan behavioral clinics as customers, Gaylord represents its first foray into a hospital. The company has helped behavioral-health clinics predict patient volumes and make better patient-therapist matches, said Eapen, who is CEO.

What if the platform spits out strange insights about patient care at Gaylord? Could it be dangerous?

Decision Options could certainly make some wrong calls, but Eapen said those will decline over time because the software will learn from prior experiences and outcomes, and incorporate them into future predictions.

Patients and their families need not worry, Grevelding said, as physicians will be scrutinizing the insights and whether or not to act on them in some new way.

The humans are very much still in charge.

“It's a decision-making guide, more than anything,” he said.