Mortality prediction model for dementia patients can help guide end-of-life care decisions


A mortality prediction model for older adults with dementia can help clinicians frame discussions with patients and families regarding end-of-life care, such as home support and nursing homes. Additionally, the model can help doctors determine whether patients should continue with routine cancer screening or stop medications, such as insulin for people with type 2 diabetes – interventions that may harm more than help. .

In their study, published in JAMA internal medicine on September 26, 2022, researchers led by UC San Francisco followed 4,267 participants with probable dementia, who were not nursing home residents, in which 81% of them had died by the end of the follow-up period.

A prediction model was developed based on age, gender, body mass index, chronic conditions, smoking status, ability to walk multiple blocks and engage in activity vigorous. They also included the ability to perform activities of daily living, such as self-care, eating, and going to bed and getting up, as well as instrumental activities of daily living, such as preparing meals, l groceries, medication and money management. The model was found to be accurate in determining who lived and who died over a period of up to 10 years in approximately 75% of cases.

The participants, whose average age was 82 and of whom 12% were black and 69% were women, had been enrolled in the Health and Retirement Study, a nationally representative survey of adults over 50 years old. Their diagnosis of probable dementia was determined by a high rate accuracy algorithm. The findings were validated in a separate group of people enrolled in the National Health and Aging Trends Study.

The model can spark conversations about financial resources, treatment preferences

The prediction model can help guide discussions about the financial resources needed to support the person with dementia, said first author W. James Deardorff, MD, geriatrician at UCSF and the San Francisco VA Health Care System.

“An estimate of an individual’s prognosis can be an important factor in families’ financial planning, especially as many people with dementia require increased support at home and are eventually admitted to care homes. nurses,” he said, noting that previous studies have shown the average survival from the time of diagnosis to be between three and 12 years.

Additionally, people with limited life expectancies may wish to focus on quality of life and comfort, rather than trying to live as long as possible. This may cause them to forego certain interventions, such as cardiopulmonary resuscitation in the event of cardiac arrest. Information about an individual’s prognosis can help inform conversations about certain medical treatment preferences.”

W. James Deardorff, MD, geriatrician at UCSF

Factors linked to mortality include older age, male gender, body mass index below 18.5, past or current smoking status, chronic illnesses, difficulty walking multiple blocks, and performance of activities of daily living and instrumental activities of daily living.

Colonoscopies, insulin may no longer make sense

Of note, the authors suggest the model can help guide discussions between doctors and patients and their families about cancer screening, which may signal slower-growing malignancies that may not be life threatening. in danger before 10 to 15 years.

“For people with limited life expectancies, cancer screening, such as colonoscopy, may cause more harm – such as pain, bleeding – without living long enough to experience a mortality benefit” , said Deardorff.

Patients who also have type 2 diabetes may be advised to reduce or stop insulin, said lead author Sei Lee, MD, a professor in the UCSF Division of Geriatrics and principal investigator for the San Francisco VA Quality Scholars scholarship. “In younger patients, tight glycemic control reduces the risk of vision loss and kidney failure in 10 years. But in older patients, these benefits may not be realized, and the risks of hypoglycemia resulting from too much insulin can be very serious.”

In conjunction with the mortality prediction model, the researchers designed an individualized mortality risk calculator for patients with dementia, available online at ePrognosis. Other mortality calculators and cancer screening vs. risk vs. benefit calculators are available, along with videos to help physicians in their conversations with patients about life expectancy and goals of care.


Journal reference:

James Deardorff, W., et al. (2022) Development and external validation of a mortality prediction model for community-dwelling older adults with dementia. JAMA internal medicine.


Comments are closed.