A dementia mortality prediction model may help doctors frame discussions with patients and families regarding end-of-life care alternatives such as at-home care and nursing facilities. Furthermore, the approach may help clinicians decide whether to continue routine cancer screening or withdraw drugs, such as insulin for Type 2 diabetes patients – measures that may cause more harm than good.
Researchers led by UC San Francisco followed 4,267 participants with probable dementia who were not residents of nursing homes in their study, which was published in JAMA Internal Medicine on September 26, 2022, and found that 81% of them had died by the end of the follow-up period. A prediction model was created based on a person’s age, gender, BMI, chronic conditions, smoking status, and ability to walk several blocks and engage in vigorous activity.
They also included the ability to perform daily living activities such as personal care, eating, and getting in and out of bed, as well as instrumental daily living activities such as meal preparation, grocery shopping, medication management, and money management. In approximately 75% of cases, the model correctly predicted who lived and who died over a 10-year period.
The participants, who were enrolled in the Health and Retirement Study, a nationally representative survey of adults over 50, had an average age of 82, 12% were Black, and 69% were female. A high accuracy algorithm determined their probable dementia diagnosis. The findings were validated in a separate group of people who took part in the National Health and Aging Trends Study.
Model May Prompt Conversations About Financial Resources, Treatment Preferences of Dementia
According to the first author W. James Deardorff, MD, a geriatrician at UCSF and the San Francisco VA Health Care System, the prediction model can help guide discussions about what financial resources are required to support individuals with dementia.
“An estimate of an individual’s prognosis can be an important factor in financial planning for families, especially as many people with dementia require increased support at home and are eventually admitted to nursing homes,” he said, noting that previous studies have shown the average survival time from time of diagnosis to be between three and 12 years.
“Additionally, individuals with limited life expectancy may wish to focus on the quality of life and being comfortable, rather than trying to live as long as possible. This may lead them to forego certain interventions, such as cardiopulmonary resuscitation in the event of a cardiac arrest. Information about an individual’s prognosis can help inform conversations about certain medical treatment preferences,” he said.
Older age, male sex, body mass index less than 18.5, former or current smoking status, chronic diseases, difficulty walking several blocks, and performing activities of daily living and instrumental activities of daily living are all associated with death.
Colonoscopies, Insulin May No Longer Make Sense
“An estimate of an individual’s prognosis can be an important factor in financial planning for families, especially because many people with dementia require increased support at home and are eventually admitted to nursing homes,” he said, noting that previous studies have shown the average survival time from time of diagnosis to be between three and 12 years.
Patients with Type 2 diabetes may be advised to reduce or discontinue insulin, according to senior author Sei Lee, MD, UCSF professor of geriatrics and senior scholar for the San Francisco VA Quality Scholars fellowship. “Tight glycemic control in younger patients reduces the risk of vision loss and kidney failure 10 years later.” However, in older patients, these benefits may not be realised, and the risks of low sugars caused by too much insulin can be severe.”
In addition to the death prediction model, the researchers developed a customised mortality risk calculator for dementia patients, which is available on the ePrognosis website.
Other mortality calculators and risk-versus-benefit cancer screening calculators are available, as are movies to assist clinicians in discussing life expectancy and treatment goals with patients.