Picture this: Creating visual representations of symptom clusters may help oncologists identify and treat sentinel symptoms in patients who are receiving cancer therapies and potentially avoid costly and unnecessary acute hospitalizations.
Network visualizations – graphic representations of the frequency of individual symptoms and the strength of their co-occurrence with other symptoms – could be used by clinicians in urgent-care settings to determine whether a presenting symptom or cluster of symptoms could be safely managed in the outpatient setting or requires hospitalization, according to Bobby Daly, MD, and his colleagues at Memorial Sloan Kettering Cancer Center in New York.
“Uncontrolled symptoms are associated with unplanned acute care. Recognition of the complexity of symptom co-occurrence can drive improved management strategies,” the investigators wrote. The report was published in the Journal of Oncology Practice.
Preventing hospitalizations related to treatment toxicities is both sound clinical and financial practice.
“Patients who receive chemotherapy have, on average, one hospital admission and two emergency department visits per year, and 40%-50% of these stem from symptoms related to their treatment. Acute hospitalizations account for 48% of total cancer expenditures. Considerable regional variation exists, which suggests that these hospitalizations may be avoidable,” they wrote.
The Centers for Medicare & Medicaid Services has identified 10 potentially preventable conditions that may result in hospitalizations, and plans to measure the performance of cancer centers based on the frequency of urgent care visits and admissions for each condition.
The conditions (in alphabetical order) are: anemia, dehydration, diarrhea, emesis, fever, nausea, neutropenia, pain, pneumonia, and sepsis.
To get a better handle on symptom clusters in patients undergoing active cancer therapy, the investigators used an analytical method involving creation of network visualizations of symptom clusters.
“It has been established that physicians are uncomfortable with managing symptom clusters, and by better elucidating these clusters, we can improve management strategies,” Dr. Daly and his associates wrote.
They identified 23,341 unique patient visits to their center’s urgent care department in 2016, and included in their analysis only those patients who had received an intravenous chemotherapy agent, oral antineoplastic therapy, or immunotherapy within 30 days of presentation. They drew on the electronic health record (EHR) for information about the chief complaint and primary diagnosis for each visit.
They then created three patient cohorts for identifying and mapping potentially preventable symptom clusters.
Cohort 1 comprised patients who were receiving active treatment and presented to the urgent care center with one of the CMS-defined symptoms.
Cohort 2 comprised patients who were receiving active treatment and were evaluated with a clinician-developed definition of a potentially preventable symptom, defined as “a symptom that could be managed safely in the outpatient setting if the clinical team identified said symptom early and managed it proactively.”
Cohort 3 was similar to cohort 2, but included only those patients who started treatment for the first time.
The authors identified 6,429 presentations in cohort 1, and in this group of patients the network analysis identified two distinct symptom clusters centered around pain and fever.
In cohort 2, there were 5,731 visits and six symptom clusters centered around fever, nausea/emesis, fatigue, deep vein thrombosis, pain, and ascites.
In cohort 3, there were 1,154 visits and four symptom clusters centered around fever, nausea/emesis, fatigue, and deep vein thrombosis.
“By using EHR data fields for [urgent care center] primary diagnoses and chief complaints, we elicited symptom information and linked it to unplanned acute care. Unplanned acute care is of increasing importance to patients because it disrupts patients’ treatment trajectories as well as their lives outside the clinic. Use of the EHR also allowed for granular symptom data, such as a particular site of pain rather than just pain itself, which is absent from traditional sources of patient-reported symptom assessments,” Dr. Daly and his colleagues said.
The investigators recommend using data from the EHR in combination with “big data” artificial intelligence techniques to allow early detection of patients at risk for symptom clusters.
The study was supported by an ASCO Health Policy Fellowship to Dr. Daly. He disclosed employment and stock ownership with Quadrant Holdings Corporation, and stock ownership in CVS Health, Johnson & Johnson, McKesson, Lilly and the Walgreens Boots Alliance.
SOURCE: Daly B et al. J Oncol Practice, 2018 Jul 17. doi: 10.1200/JOP.18.00199.