Griffin Health is anchored by Griffin Hospital in Derby, Connecticut, a 160-bed acute care community hospital serving more than 130,000 residents of the Lower Naugatuck Valley Region.
THE CHALLENGE
Like many health systems, Griffin Health struggled to ensure patients received timely follow-up when an imaging study uncovered a finding that required additional imaging to diagnose an issue or provide a clear answer for an ordering provider.
"Explicit and incidental radiology findings often signal the earliest stages of disease, but across the industry, about 50% of additional imaging that's recommended does not get performed," said Todd Liu, executive vice president and COO at Griffin Health.
"The reasons are varied, but include inconsistent handoffs, lack of centralized tracking, poor communication, and gaps between radiology and primary or specialty care teams.
"We had manual systems in place to notify ordering providers, but we lacked an efficient way to manage and organize follow-ups, coordinate with our patients and monitor what happened next," he continued. "Was the appropriate follow-up imaging study scheduled? Did the patient have the study done? Was there a diagnosis or clear answer to the ordering provider's reason for ordering the imaging study?"
Griffin Health's care navigators worked hard to ensure patients had any follow-up imaging studies ordered and completed, but there was no way to manage the workload manually and staff couldn't always be certain a patient made it to their appointment. That was a safety issue the health system no longer could afford to tolerate.
"We needed a solution to this challenge that would help close this loop, not just at the time of the initial report, but all the way through to resolution," Liu said.
PROPOSAL
Inflo Health, a vendor of follow-up care IT, approached Griffin Health with a proposal: If the health system could use artificial intelligence to identify radiology reports with a recommended follow-up imaging study, automate their escalation into care coordination workflows and orchestrate care, Griffin Health could significantly improve follow-up compliance and patient safety without overwhelming staff.
"The platform would help identify open recommendations for additional imaging studies in real time, put each patient into an appropriate, predefined care pathway for follow-up, and nudge our patients and providers into action," Liu explained.
"What made the approach particularly compelling was its flexibility," he continued. "The system was designed to plug into our existing processes and increase our internal capacity to track and to act on findings – not entirely replace our existing workflows."
The goal wasn't just to flag at-risk patients, but to automate outreach and deliver insights into every step of the process, from identification to closure.
"This helped us ensure the loop was being closed and there was a mechanism to track provider responsibility, progress and outcomes in one place," he said.
MEETING THE CHALLENGE
Griffin Health embedded the Inflo technology into its existing care navigation processes. The platform began by analyzing radiology reports, looking for language or patterns that indicated a potentially significant finding, like a lung nodule. Once identified, the platform organized patients into workflows based on the type of follow-up needed and the timing for the next step in care.
"The platform automated provider outreach to ensure an order for the recommended next step was made in the EHR, monitored patient engagement, and determined if and when the follow-up took place," Liu explained. "With this information now centralized in one place, our care coordinators could begin using the system to monitor and identify any missed follow-up imaging studies, identify barriers, and contact patients directly when necessary.
"The system gave them a central dashboard that tracked the status of each case: initial finding, recommended follow-up, whether that appointment had been made, and if the patient ultimately had the follow-up study performed," he added.
RESULTS
First of all, using the technology increased the closure rate by 50%.
"For just incidental findings – which comprise nodules and other findings not explicitly evaluated – we realized a 50% improvement in closure rates," Liu reported. "That means more patients are getting the care they need, and fewer are falling through the cracks. From a risk management and patient safety perspective, that shift is huge."
Second, the follow-up completion rate is up 17%.
"Our follow-up completion rate for flagged patients rose 17%," he said. "That represents a meaningful jump in real patients being seen, diagnosed and treated sooner. In some cases, it's the difference between catching a disease early and missing a window for intervention."
Third, patients have been enrolled in a lung cancer screening program based on artificial intelligence recommendations.
"Inflo Health helped us identify and enroll 18 patients in our lung cancer screening program – patients we might not have otherwise engaged at that critical moment," he explained. "These are real lives, potentially extended or saved because an AI system surfaced the right data at the right time, and our teams were equipped to act on it."
ADVICE FOR OTHERS
Be clear about what problem you're trying to solve – and what success looks like, Liu advised.
"For us, the priority was closing the loop on follow-up care," he explained. "We needed to track findings; but more important, we needed a system that created actionable information. Whatever technology you choose, make sure it doesn't just identify problems or produce more data – it needs to make it easier for people to do the right thing.
"Second, don't underestimate the importance of working with providers and members of the care team to thoughtfully integrate new tools into existing workflows, when possible," he continued. "The best AI in the world won't make a difference if your teams don't use it – or if it isn't integrated into the way providers and staff do their work."
In Griffin Health's case, success hinged on choosing a platform that could work with its care coordinators and clinicians.
"Good technology should always support underlying processes and workflows, not dictate how providers and staff deliver care to patients," he concluded.
Follow Bill's HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.
WATCH NOW: Chief AI Officers require a deep understanding of the technologies and clinical ops
