Intermountain’s COPD Risk Scores Help Identify Patients Who Need Hospice

Clinicians at Salt Lake City-based health system Intermountain Healthcare have developed risk scores for patients suffering from chronic obstructive pulmonary disease (COPD) that can help health care providers identify which patients need hospice care versus curative treatments.

While other risk stratification models to predict mortality exist, this is the first that allows for identification of high-risk COPD patients on a system level, according to the developers.

Intermountain physicians and staff considered a long list of COPD patient characteristics — including lab test results, diagnoses, lung function testing among other variables — and applied an algorithm to determine which patients were most alike and group them into clusters. Through a process called recursive partitioning, the risk score developers narrowed down the number of variables to six that are necessary to effectively stratify patient risk. 


The result was a decision tree using the six variables to determine the patient’s risk level, called a LIVE score. Scores range from low risk (LIVE 5) to high risk (LIVE 1). The developers then validated their findings using cohorts of Intermountain patients as well as external cohorts in four health systems, including the University of Chicago hospitals and U.S. Department of Veteran’s Affairs medical facilities.

“A high risk live score (LIVE 1 and 2) have very high mortality,” Denitza Blagev, M.D., medical director for quality, specialty based care and associate professor of medicine at the University of Utah, told Hospice News. “Interestingly, the LIVE score is laboratory based, consistent with the observation that for many COPD patients it’s the severity of comorbidities that drives the mortality and morbidity risk. Thus, for high risk patients (LIVE 1 and 2) who have a diagnosis of COPD the mortality is much higher than for low risk patients with COPD.”

More than 114,300 patients who died in hospice during 2017 had a respiratory condition as a principal diagnosis, up from 122,000 in 2017 — a 19.7% increase, according to the National Hospice and Palliative Care Organization. Hospice patients with a respiratory diagnosis tend to have longer lengths of stay than patients with other conditions, an average of 74.9 days and a median of 20, compared to an average of 48 and a median of 19 for cancer patients. Patients with respiratory conditions represent 10.9% of all Medicare hospice spending.


Patients who Intermountain placed in the highest risk groups, LIVE 1 and LIVE 2, have an approximately 40% chance of dying within one year, according to a study Blagev and her co-authors intend to publish in the near future. The palliative care referral rate among those patients is also close to 40%. 

“These findings suggest that clinicians on the ground recognize the high risk of these patients and refer them to palliative care,” Blagev said. “These findings also suggest that a routinely calculated LIVE Score that is high risk in these patients with COPD may identify a small subset of COPD patients that are at the highest risk and may be candidates for evaluation for hospice and palliative care services.”

Blagev recommends assessing patients’ risk scores at admission or at the time of diagnosis. The researchers published their initial findings in Frontiers in Medicine, and they plan to publish several additional research papers as their work progresses.

Companies featured in this article: