New Tool May Predict Mortality Risk for Older Adults With Dementia

Lara Salahi

September 29, 2022

A prognosis prediction model may help clinicians accurately predict death among older adults with dementia.

The tool, which covers six measures ― age, sex, body mass index (BMI), and physical fitness ― may help clinicians better frame discussions about treatment and end-of-life decisions with patients, according to the researchers, whose research was published September 26 in JAMA Internal Medicine.

"We hope that the mortality risk estimates from this model can be used by clinicians in a clinical context with patients and families to facilitate shared decision-making," said W. James Deardorff, MD, a geriatrician at the University of California, San Francisco (UCSF), and the San Francisco VA Health Care System, who helped conduct the study. "Given that there were no well-established prediction models that are specific to this community-dwelling population with dementia, we sought to develop one."

Approximately 6.5 million Americans aged 65 years or older have Alzheimer's disease or another form of dementia, according to the researchers. The median survival time from age at diagnosis varies widely, ranging from 3.3 to 11.7 years, they add. The ability to better predict which patients are more likely to survive longer with dementia could significantly help with financial planning, clinical choices, the use of long-term care, and other types of decisions for this population.

"For interventions that have immediate risks or burdens and delayed benefits, such as cancer screening and tight glycemic control in patients with diabetes, life expectancy can help differentiate between patients most likely to benefit vs most likely to be harmed by interventions," according to the researchers.

For the study, Deardorff and his colleagues at UCSF and the University of Michigan, Ann Arbor, followed for a 10-year period 4267 men and women who were considered to probably have dementia. The participants were enrolled in the Health and Retirement Study, a national survey of health and economic circumstances of adults aged 50 and older. All participants initially lived in the community outside of nursing homes. The average age of participants in the study was 82 years, and 69% were women.

The researchers validated their model with a separate group of individuals who were enrolled in the National Health and Aging Trends Study. Their model predicted with 75% accuracy which participants would die within the decade ― considered to be long term for this population.

Roughly one third of older adults with dementia who live outside of assisted living facilities live alone, according to the US Department of Health and Human Services. Another 30% live with others, and 27% live with a spouse or partner.

The model developed by the researchers accounts for age, sex, BMI, chronic conditions, smoke status, and level of physical activity.

"These factors are powerful predictors of mortality because increased functional impairments often represent the end result of progression of dementia and other chronic diseases," Deardorff told Medscape Medical News.

People with dementia commonly have symptoms such as hallucination or delusions, but the model does not consider those factors, Deardorff said. The factors the researchers included are ones more readily obtainable in clinical practice, the study authors write.

Nancy Schoenborn, MD, a geriatrician at Johns Hopkins University School of Medicine, Baltimore, says that in addition to providing a good degree of accuracy for mortality, the tool allows clinicians to mark how an individual's dementia is affecting their behavior and cognition during visits.

"When we have this type of tool to predict functional decline over time, that's where we find this helpful," said Schoenborn, who was not affiliated with the study. "That's been something that's been harder for clinicians to get a gestalt about."

Mara Schonberg, MD, MPH, associate professor of general medicine at Beth Israel Deaconess Medical Center, Boston, said that because the tool assesses some lifestyle factors, it may help motivate patients to take better care of themselves.

"Maybe someone will start exercising more vigorously so it could improve their life expectancy," Schonberg, who was not part of the study team, said.

A BMI of 18.5 or lower was associated with an increased risk of mortality, the study authors note. Older age was also a strong predictive factor. Women were at lower risk of earlier death within the 10-year period than men, the study found.

Schonberg said the prognostic tool is most beneficial when a clinical decision comes up that is dependent on a patient's living long enough to benefit from the intervention. It may also be helpful in nonclinical decision-making, such as making decisions regarding finances and where to live, she added.

"In primary care, we don't have a ton of time," Schonberg said. "So, it's probably not every patient we're doing this with, unless a patient asks."

According to Schonberg, the challenge some clinicians find, especially in the primary care setting, is how best to communicate regarding a prediction model, since it is not always precise.

"While these and other tools predict prognosis, we've now got to figure how to talk about it, not only with patients now but with caregivers," she said.

Deardorff's group said clinicians could use the new tool to assess whether they should recommend continuing certain treatments or interventions.

"Even if patients and caregivers do not wish to hear this information, the mortality risk estimates can still help clinicians make recommendations based on a patient's previously shared goals and treatment preferences," he said.

The researchers are now working on a prediction model to assess when an individual will become eligible to move into an assisted living setting.

"Providing patients and families with an estimate of time to nursing home eligibility may help with future planning," Deardorf said.

The research was supported by the National Institute on Aging. Schonberg and Schoenborn helped researchers to create the digital prognosis tool involved in the study, including Deardorff.

JAMA Intern Med. Published online September 26, 2022. Abstract

Lara Salahi is a journalist living near Boston.

For more news, 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.