August 20, 2015
2 min read
Save

Diagnostic model predicts impending death in patients with cancer

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Researchers have developed a diagnostic model based on two objective bedside physical signs that predicts impending death within 3 days for patients with cancer.

Using two variables as a foundation for the model — a palliative performance scale (PPS) and the drooping of nasolabial folds, or the folds that run from the nose to the corner of the mouth — David Hui, MD, MSc, assistant professor in the department of palliative care and rehabilitation medicine in the division of cancer medicine at The University of Texas MD Anderson Cancer Center in Houston, and colleagues developed this model to help clinicians formulate their diagnosis of impending death.

David Hui

David Hui

“The ability to recognize that a patient has entered the final days of life with high accuracy has important implications for end-of-life care,” Hui told HemOnc Today. “This model takes advantage of the sensitivity of a high palliative performance status to rule out impending death, and the high specificity of drooping of nasolabial fold to rule in impending death.  It has approximately 80% accuracy and is applicable anytime during admission.” 

The researchers documented 62 physical signs every 12 hours from admission to either death or discharge for 357 patients with advanced cancer who were admitted to acute palliative care units between April 5 and July 6, 2010 at MD Anderson Cancer Center and between January 27 and June 1, 2011 at Barretos Cancer Hospital in Brazil.

Ninety percent of the patients had complete data for all 62 signs. Researchers also used the PPS, a validated 11-point scale that ranges from 0% (death) to 100% (completely asymptomatic) based on the patient’s function, oral intake and cognitive status.

The 3-day mortality rate was 24% on admission.

The final model had four terminal leaves. The first consisted of a PPS score of 20% or lower with a presence of drooping nasolabial folds and was associated with a 94% 3-day mortality rate. The second, a PPS score of 20% or lower and no drooping folds, had a 3-day mortality rate of 42%. The third leaf included a PPS score between 30% and 60% and was associated with a 16% 3-day mortality rate, and the fourth leaf — a PPS score of 70% or higher — had a 3% 3-day mortality rate.

The model is applicable at time of admission with an 81% diagnostic accuracy for the original tree and 80% for cross-validation. The model had a diagnostic accuracy range of 79% to 84% for additional days in the acute palliative care unit.

The researchers acknowledged only a small percentage of patients had a high risk for death and some who did not have drooping nasolabial folds but had multiple other late signs potentially could have been misclassified as low risk.

To address this limitation, the investigators created a second model in which the number of late signs present was included as a variable. However, this variable did not change the overall structure of the first model and helped identify a subgroup of patients who had an intermediate probability of 3-day mortality (ie, PPS ≤ 20% and drooping absent and PPS 30% to 60%). Patients with two or more late signs had a higher 3-day mortality rate.

“To our knowledge, this is the first diagnostic model for impending death,” Hui said. “Further validation of this model in other settings is needed. We are also in the process of developing educational videos so clinicians and family members can be more familiar with these tell-tale signs.” – by Anthony SanFilippo

For more information:

David Hui, MD, MSc, can be reached at the department of palliative care and rehabilitation medicine, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Unit 1414, Houston, TX 77030; email: dhui@mdanderson.org.

Disclosure: The researchers report no relevant financial disclosures.