Doylestown Health is using a platform developed by XSOLIS to help nurses stay on top of fast-changing patient conditions and ensure each patient is properly classified and cared for.
For all their advantages, EHRs aren't so good at detecting changes in patient status that separate those being observed and those who end up admitted. Nursing staff are often left to comb and click through records in an endless game of catch-up.
But at Doylestown Health, AI and algorithmic technology are delivering this in a more efficient manner.
The suburban Philadelphia healthcare network, centered around an independent 270-bed hospital, is using predictive analytics technology from XSOLIS to improve medical utilization management. In the first six months of use, officials say they've improved observation rates by 20% and observation to inpatient conversion rates by 37%. And three years later, the initial return on investment of 4.6x has now improved to 7.3x.
Mary Beth Mitchell, MSN, RN, CPHQ, CCM, SSBB, senior executive director of care transformation strategies at Doylestown Health, oversaw this transformation, as well as hospice/palliative care and clinical documentation improvement, while heading the hospital's case management department.
Mary Beth Mitchell, MSN, RN, CPHQ, CCM, SSBB, senior executive director of care transformation strategies at Doylestown Health. Photo courtesy Doylestown Health.
Mitchell says hospitals would like to be able to admit all presenting patients, but payers insist on observation status as a less-costly alternative based on how sick the patient is. That usually does not last more than 48 hours.
"We are required contractually to review and assure that we have the patient in the appropriate status, so that when we bill the insurer, we're billing appropriate," she says.
Utilization review (UR) nurses must review every patient who comes in and is placed in a bed, whether they're on observation status or inpatient status, to be sure they are in the right status, Mitchell says. These nurses create patient synopses that are sent to the payer, who then can agree or disagree with the status assigned to the patient by the hospital.
Prior to adopting the XSOLIS technology platform, those nurses would, on a daily basis, start at one end of the 270-patient roster, either by payer or by floor, and work their way through to the other end, one chart at a time, to look for changes in patient status that rise to the threshold of changing status from observation to inpatient or vice versa, Mitchell says.
"I could look at a chart in the morning, and the patient looks appropriate for observation," she says. "But during the course of the day, lots of stuff happens to patients. But [UR nurses] are not going to look at that chart again till the next day, because this is a manual process."
Some hospitals start with certain diagnoses, but they're still guessing what they will find in those particular charts, Mitchell says.
The technology platform "assigns a severity for us, and through their AI platform [we] are able to use that severity to predict that the patient should be inpatient or observation status," she says.
The technology continually combs through each chart, looking for events entered by clinicians and notifying UR nurses when those events rise to the level of suggesting a change in status, Mitchell says.
"It's almost like an assistant, re-reviewing your charts constantly," she says.
Since UR nurses typically work on a Monday-through-Friday schedule, the technology is particularly useful in catching changes in patient status late on Fridays, also alerting those nurses about changes over the weekend when they arrive Monday morning, she says.
Unlike the presentation of data in EHRs, where less relevant data is often a distracting presence for UR nurses, the technology highlights key measures.
"When I'm going through an EHR, I have to click in and out of every tab," Mitchell says. "I have to look at every medication the patient is on. I really don't want to sift through things that aren't meaningful. [The technology] boils that down. For the medication lists, we only see what's considered notable meds."
The XSOLIS platform presents synopses of the recommended status changes to UR nurses, who can snip them and send them to payers via electronic fax or other means, Mitchell says.
The technology also accounts for traditional Medicare's standards for admissions and the fact that most private payers use one of two criteria – Milliman or Interqual.
One drawback is that this process can reduce the UR nurse's role to being a box-checker Mitchell says. But using the right technology can restore their ability to practice to the top of their license by allowing them to consider multiple diagnoses for a patient.
"The nursing staff loves this, because they're getting to use their clinical skills," she says. "It's more fulfilling to do their job."
Mitchell says healthcare organizations should thoroughly examine and test the technology platform before putting it into use. Different vendors and products offer different pathways and goals, making it vital to ensure that one platform can fit seamlessly into a health system's workflow and meet the needs of administrators and staff.
"We asked for data," she says. "We asked to speak with other hospitals. Were they actually seeing this make a difference? How are they utilizing it? By the time we made the decision, we felt pretty comfortable this was going to help us accomplish what we needed it to accomplish."
"It's really important in this day and age for hospitals to learn to leverage technology to their advantage," Mitchell adds. "Any time you do something manually, somebody's going to miss something. We leverage the technology to help us."
“It's really important in this day and age for hospitals to learn to leverage technology to their advantage. Any time you do something manually, somebody's going to miss something.”
— Mary Beth Mitchell, MSN, RN, CPHQ, CCM, SSBB, senior executive director of care transformation strategies at Doylestown Health
Scott Mace is a contributing writer for HealthLeaders.
Technology is replacing manual processes for establishing patient status that were often prone to error.
An AI algorithm assigns a severity score to each patient and predicts which patients under observation are prone to switching to inpatient status.
The platform's constant re-review of charts also helps nurses consider the effects of multiple diagnoses on patient status.