Innovative use of data helps patients with cancer avoid return hospital visits

Can medical record data help lower emergency department (ED) visits and hospital readmissions for some cancer patients? Yes, according to a study by ChristianaCare researchers published in the Journal of Clinical Pathways.

ChristianaCare Organizational Excellence, Oncology Nursing and Cancer Care Management teamed up to create a computer model based on variables from the patients medical record to identify at-risk patients prior to discharge.

In a 13-month study (July 2018 to October 2019), clinicians were able to focus interventions on these patients to help them avoid unnecessary return hospital visits.

The 30-day readmission rate saw a steady decline for patients classified as high-risk (by 16%) as well as for the total population (by 10%). Average hospital length of stay and number of ED visits also declined for both groups.

“Risk factor identification and early intervention are hallmarks of patient-centric clinical decision-making,” said Nicholas Petrelli, M.D., FACS, Bank of America endowed medical director at the Helen F. Graham Cancer Center & Research Institute. “Congratulations to our cancer care teams and performance improvement experts for their novel use of evidence-based predictive modeling to improve both the quality and value of care for our most vulnerable cancer patients.”

The team applied their model to clinical decision-making on 6E, the oncology unit at Christiana Hospital, to patients discharged to home and receiving care at three oncology practices at the Graham Cancer Center.

Participants included ChristianaCare Oncology Hematology, Gynecologic Oncology and Medical Oncology Hematology Consultants and ChristianaCare HomeHealth.

“This study confirmed for us that thoughtful discharge planning and targeted follow-up after discharge is critical to helping these acute-care patients avoid unnecessary trips back to the hospital,” said Tammy Brown, MSN, NEA-BC, OCN, director of Cancer Care Management.

The lead author on the study was Organizational Excellence consultant Kelsey Jarrold, MSISE. She and Binghamton University Graduate Research Associate Raghad Alkhawaldeh, Ph.D-c, led the design and implementation of a computer model that scores patients on a scale from low to medium to high risk for readmission based on 15 selected variables recorded in their electronic health record.

“Studies show that predicting risk scores during hospitalization is beneficial to avoiding readmissions,” Jarrold said. “One of our biggest challenges was to develop an accurate predictive model using available data at the time of clinical decision making.”

The final list of 15 variables (including medical history, clinical/test results, previous inpatient visits and discharge destination) was narrowed down from more than 100 variables initially identified after analyzing 6,336 Christiana Hospital oncology visits from 2013 to 2017.

Because of the complex nature of their illness, cancer patients in general are at high risk for hospital readmission. Baseline analysis revealed that patients discharged to a home health agency have a higher risk of readmission. In response, the team created broad interventions for all high-risk patients as well as a more targeted approach for home health patients.

“One of the most important things we learned was to use the data available to target our interventions and prioritize resources to focus on the patients who really might need support,” said 6E Nurse Manager Courtney Crannell, DNP, MSN, OCN, NE-BC.

Formalized day of discharge visits and coordinated, timed follow-up calls offered multiple opportunities for the care team to connect with patients and address any issues or needs that might arise, such as getting medications, connection with home services or equipment, understanding instructions and when to call the doctor.

Adding the provider’s office to the team opened one more avenue to address patient concerns, and confirmed a scheduled office visit within 10 days or sooner if the patient was not doing well. Among the baseline population, 43% of readmissions occurred within the first 10 days after discharge.

The team documented and shared results from patient contacts electronically. “We stopped working in silos and began to bring more continuity to our contacts with our patients,” said Karen Sites, MSN, RN, OCN, CCM, manager of Cancer Care Management. “This level of shared communication made it easier to address things that might cause a return to the hospital or escalate to readmission.”