ED Docs Can Identify Patients Likely to Die Within 1 Month

Ricki Lewis, PhD

September 13, 2019

Emergency department (ED) physicians who answered the question "Would you be surprised if your patient died in the next 1 month?" could identify patients over age 65 years who indeed died within a month, according to results of a study published online September 13 in JAMA Network Open.

A palliative care plan can ease the end of life for patients and cut costs, but identifying which patients are in the last months of their lives is not an exact science. One strategy is for a provider to answer the "surprise question:" "Would you be surprised if your patient died in the next 1 month?"

An ED is a particularly likely place to find patients nearing the end of life. Studies have shown about three quarters of older adults with serious health conditions visit the ED during the final 6 months of life, and such visits may signal a turning point towards accelerated physical decline.

To assess how effective the surprise question is in predicting 1-month mortality, Kei Ouchi, MD, MPH, Department of Emergency Medicine, Brigham and Women's Hospital, Boston, Massachusetts, and colleagues compared the predictions to actual deaths reported in the National Death Index. The prospective cohort study, conducted at a medical center in Portland, Maine, included consecutive patients aged 65 years or older admitted to hospital from the ED between January 1, 2014, and December 31, 2015. Physicians answered the question via the electronic medical record when admitting patients.

Researchers analyzed data from 10,737 older adults who visited the ED a total of 16,223 times during the study period and were admitted to hospital. Within a month of admission, 893 (8.3%) had died.

The ED physicians responded positively to the surprise question for 2104 (19.6%) patients, of whom 533 actually died.

A multivariable analysis, which controlled for age, sex, race, diagnosis on admission, and comorbidities, showed that the odds of death at 1 month were higher among patients identified by physicians who answered yes to the surprise question (odds ratio, 2.4; 95% CI: 2.2 - 2.7; P < .001).

ED physicians were also able to identify patients likely to die by 6 and 12 months.

Although helpful, the surprise question wasn't a perfect predictor. Sensitivity was 20%, specificity was 93%, and positive and negative predictive values were 43% and 82%, respectively. Accuracy was 78%.

"The diagnostic test characteristic of the surprise question alone was poor, which makes it a poor screening tool for identifying patients with high risk of 1-month mortality," the researchers write.

The study extends the value of the surprise question beyond clinical settings in which physicians have a relationship with the patient and are aware of pre-existing risks. In the ED, the question may help rapidly identify patients who might benefit from a palliative approach, the researchers conclude, calling the strategy valuable and easy to use.

The overestimation of 1-month mortality would not adversely impact patients, the researchers add, citing two recent studies that demonstrate no harm in initiating discussion of serious, life-limiting conditions and palliative care earlier.

In an invited commentary, James Downar, MDCM, MHSc, division of palliative care, Department of Medicine, University of Ottawa, Ontario, Canada, and colleagues suggest using "big data" could improve the accuracy and value of the surprise question.

They note that just as video streaming services consult big data to make viewing suggestions to subscribers and pharmaceutical companies use big data to target drugs to healthcare consumers, researchers could use data from thousands or millions of older patients admitted from the ED to help physicians rapidly identify patients likely to die within a month.

Downar and colleagues explain how in Canada the Risk Evaluation for Support: Predictions for Elder life in the Community Tool (RESPECT) is an algorithm built on existing assessments of the needs of individuals in home care and residential long-term care settings used to predict survival. Patients or their family members can self-administer RESPECT. Physicians can use it to identify patients at increased risk of death within a month and link the determination to palliative services.

"No identification tool can overcome some of the other barriers to the provision of palliative care, including a lack of palliative resources or a lack of skill and comfort among nonspecialist clinicians, but overcoming the challenge of timely identification would represent a substantial achievement in the struggle to provide a palliative approach to care for individuals who would benefit from it," the commentators write.

Study limitations include extension to other populations outside the white, urban study population and lack of detail about patients' diagnoses and course of illness in the ED and over the 1-month period.

Two commentary authors have reported receiving grants from the Canadian Frailty Network to study an automated system (mHOMR) for identifying inpatients at elevated risk of 1-year mortality and one commentary author has received grants to develop RESPECT. One study coauthor has reported receiving fees from UpToDate for editing the palliative care section. The other researchers have reported no relevant financial relationships.

JAMA Network Open 2019;2:e1911139, e1911146. Full text, Editorial

Follow Medscape on Facebook, Twitter, Instagram, and YouTube.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: