Eleven Ascension hospitals will participate in testing a computerized method of scanning content within electronic health records that will automatically recommend patients for palliative care.
The National Institutes of Health commissioned the study. Ascension is working with the University of Pennsylvania, which will analyze data from the hospitals over three years to determine the accuracy and value of the system. Results are expected in 2019.
Hollar
Lynn Hollar, Ascension's senior clinical analyst at the start of the project, said participants began work in fall 2014 to develop a way to analyze information in electronic health records for patients with any of three chronic illness diagnoses — end-stage renal disease, chronic obstructive pulmonary disease or advanced dementia. Upon a patient's admission to a hospital, the algorithm will search his or her health records for specific histories or complicating factors.
If primary and secondary criteria are met, the program will automatically generate a consult order for the patient's doctor to consider. The order is set with a start time of 3 p.m. the next day, and an alert sent to the patient's physician or physicians — usually giving doctors about 24 hours to respond. If the physician wants to "opt out" the doctor can remove the order.
If the doctor doesn't opt out by removing the order within a day, the consult order will alert a palliative care team to visit the patient.
A small team of information technology associates within Ascension have been working with Sarah Hill, system manager of palliative care initiatives at St. Louis-based Ascension. Hill said the goal of the initiative is to increase the opportunity for more patients to benefit from palliative care.
"Unfortunately, there are still misconceptions that palliative care is only for those at the end of life or dying, which is a huge disservice to those who could greatly benefit from palliative care," Hill said. The algorithm removes such impediments.
Hill
She said the participating hospitals all use the electronic health records system of Cerner Corporation of Kansas City, Mo., which provides uniformity in the test. They also have strong palliative care teams. Administrators at all 11 hospitals volunteered to take part.
Beginning this month, Columbia St. Mary's Hospital Ozaukee in suburban Milwaukee will be the first hospital to have the consult orders electronically placed into patient charts. Another hospital will be added every six to eight weeks.
David Schneck, director of EHR programs, clinical information systems at Ascension Information Systems, said the data extraction used to identify patients in the study is a long series of "if-thens." It begins with the three basic categories of chronic illness, further defined by the secondary eligibility criteria.
Variables assessed by the algorithm include the number of readmissions, use of surgically placed feeding tubes, oxygen dependency and residence in long-term care. When a consult order is issued, data from 28 measures are sent to the university for compilation and analysis, she said.
Hill said the goals are to better identify patients who can avoid expensive readmissions to hospitals for treatment of chronic illness and receive care that fits their needs. She speaks with empathy, having watched her grandmother suffer prolonged illness before her death.
"Part of what turned me down this career path was seeing firsthand the benefits to patients and their families of receiving palliative care, particularly for people like my grandmother who were not necessarily dying but who were experiencing multiple chronic conditions and generally just a failure to thrive," Hill said. "I want to make sure that other patients receive that chance that my grandmother did not."