Evidence shows an elevated risk of suicide related for older adults diagnosed with depression. Depression – called major depressive disorder (MDD) – is the second most diagnosed mental health condition in the world after anxiety. Depression affects people from all cultures, all walks of life, and people in all age groups, from children to adults. With regards to depression in older adults, prevalence rates vary. In the U.S. in 2021, around five percent of adults over age 50 reported experiencing a major depressive episode in the past year. That’s just over five million people. In addition, in the U.S. in 2021, around three percent of adults over age 50 reported a major depressive episode with serious impairment in the past year. That’s close to three and half million people.
These facts and figures – which we retrieved from the 2021 National Survey on Drug Use and Health (2021 NSDUH) – concern us because one of the consequences of depression is suicide. Suicide among seniors with depression is a very real and serious issue. In fact, data from the 2021 NSDUH shows that in 2021, three-quarters of a million people over age 65 had serious thoughts of suicide and over one hundred thousand people over age 65 attempted suicide.
Those figures may be an underestimation. A study conducted in China among 750 seniors over age 65 showed that over twenty percent met clinical criteria for depression, but only one percent recognized that they had a depressive condition.
That prompts us to ask this series of questions:
Is depression more common among seniors than we think?
If so, are they at increased risk of suicide?
If they are, are there ways we can predict likelihood of suicide among the senior population?
That first question is difficult to answer, since nationwide surveys like the NSDUH rely on self-reporting to collect prevalence rates. However, a recent study may offer answers to the next two questions.
Predicting Suicide Risk in Older Adults With Depression
The study “Multimodal Brain Connectome-Based Prediction of Suicide Risk in People With Late-Life Depression,” published in February 2023, presents research on a new method of assessing suicide risk among solder adults with depression.
When depression occurs late in life, mental health professionals call it late-life depression (LLD). Evidence cited in this new paper indicates that over a third of people with LLD do not or cannot attain full remission after one to two standard courses of treatment. A standard treatment protocol for depression includes medication with an antidepressant, talk therapy/psychotherapy, and lifestyle changes like improving diet or increasing levels of activity. While this is effective for many patients, the data – as we state above – shows that this traditional approach does not work for everyone.
When depression treatment does not work, suicide risk increases: time and unsuccessful treatment attempts can exacerbate and escalate symptoms of depression.
These facts make assessing suicide risk among people with depression important. Existing research identifies several risk factors for suicide among seniors with depression.
Depression Risk Among Seniors: Primary Factors
- Previous suicidal ideation
- Previous suicidal behavior
- Financial crisis
- Impaired functional ability
- Due to physical illness
- Due to mental health conditions
- Substance/alcohol use
To identify these risk factors among seniors with depression, mental health professionals use a standard array of well-verified metrics, including:
- Patient Health Questionnaire-9 (PHQ9). This short assessment is common in hospitals and among primary care providers
- Ask Suicide-Screening Questions. This assessment is designed for children and adolescents
- SAFE-T. This assessment focuses on specific areas of suicide risk and offers treatment options
- The Columbia-Suicide Severity Rating Scale (C-SSRS). This assessment is common, and can be administered by people with no formal mental health training
- The Beck Scale for Suicidal Ideation (BSS). This is one of the most commonly used suicide assessment tools among mental health professionals, primarily because research has repeatedly verified its validity since publication in 1961.
These metrics provide useful information for treatment providers and helps them mitigate suicide risk among all populations, not only seniors. However, while these scales can identify suicide risk, none can accurately predict suicidality among seniors with depression. That’s what this new study is about: using a new method – brain scans – to assess and predict potential suicidal behavior in this at-risk demographic.
Brain Changes and Connectivity: The Use of Imaging to Identify Suicide Risk
Imaging technology has improved dramatically over the past two generations. The advent of MRI gave us the ability to see into the body and gather details on soft tissue structures, which was a significant improvement over x-ray technology, which allowed us to see bones and not much more. The recent improvements in MRI-type technology – PET scans, fMRI, and others – allow us to see inside the brains of humans while they’re engaged in cognitive tasks: that’s a huge step forward, because previously, studying the live human brain – while functioning – was impossible.
These advances mean we now know which brain areas are associated with suicide risk. This knowledge is essential for the study we discuss in this article: to create a predictive assessment for suicidality among seniors with depression, researchers need to know where to look. In this study, they chose to examine “functional and structural network connectivity” in the brain with connectome-predictive based modeling (CPM). Previous research verifies that CPM can predict brain and behavior associations among older adults, and this study leverages CPM predictability to assess suicide risk in senior adults.
To create and verify a predictive suicide risk-assessment for senior adults with depression, researchers recruited a group of seniors with depression and divided them into three groups:
Non-suicidal group:
- Patients with no history of suicidal ideation or behavior
Ideation/plan group:
- Patients with a history of suicidal ideation but no suicide attempts
Suicide attempt group:
- Patients with a history of at least one suicide attempt
Next, researchers administered standard, question-based, suicide risk assessments to patients in all three groups. Then, researchers collected real-time functional and structural connectivity data from brain networks associated with depression using fMRI (functional MRI) technology. Finally, researchers collated all the data, applied advanced statistical analysis, and evaluated the predictive potential of the brain scan data in relation to the data from the standard assessments and the relative suicidality of the patients in the study.
Can Brain Scans Predict Suicidality in Seniors with Depression?
The research team knew one thing going in: one assessment, one metric, or one test was unlikely to predict suicidality with any accuracy. Therefore, they created six assessment profiles or models, and determined which assessment profile/model showed the best predictive power. Here are the models they created.
1. Model A included:
- Age
- Sex
- Education
- Onset of LLD
- Duration of LLD
2. Model B included:
- Standard assessment scores only
3. Model C included:
- Functional positive network strengths of assessments
- Negative network strength of assessments
- Structural positive network strengths of assessment
4. Model D included:
- Standard assessments
- Functional connectivity profiles from brain scans
5. Model E included:
- Standard assessments
- Structural connectivity profiles from brain scans
6. Model F included:
- Standard assessments
- Functional connectivity profiles from brain scans
- Structural connectivity profiles from brain scans
7. Model G included:
- All of the above
The researchers then used the data to compare seniors with depression who had never thought about or committed suicide with seniors who had considered suicide or attempted suicide to determine which model had the most powerful – i.e. reliable – predictive capability.
- When comparing people who’d never considered suicide with people who’d attempted suicide:
- Model D showed the greatest accuracy, or the ability confirm the presence of suicidality among people with known suicidality
- Models D and E showed the greatest sensitivity, i.e. the ability to predict the presence of suicidality among groups where suicidality is not known
- Models D and E showed the greatest specificity, or the ability to predict the absence of suicidality among groups where suicidality is not known
- When comparing people who’d considered suicide with people who’d attempted suicide:
- Model D showed the greatest accuracy
- Models D showed the greatest sensitivity
- Models D and F showed the greatest specificity
- When comparing people who’d considered suicide with people who’d never thought about or attempted suicide:
- Model G showed the greatest accuracy
- Models G showed the greatest sensitivity
- Models G showed the greatest specificity
Multimodal Assessment and Suicide Risk in Older Adults with Depression
First, demographic information alone had no predictive power. Second, standard assessments alone had no predictive power. Third, brain scans alone had no predictive power. Fourth, standard assessments plus functional connectivity data from bran scans had more power than standard assessments plus structural connectivity data form brain scans. Fifth, among people with lower risk, i.e. people who’d never considered suicide and people who’d considered but neither made a plan nor attempted suicide, an all-of-the-above approach to assessment had the best predictive power.
Where does that leave us?
With three important facts/data points:
- Single assessments have less predictive power than multiple assessments
- Brain scans that show functional connectivity are more predictive than brain scans that show structural connectivity
- A combination of standard assessments + functional connectivity data have the most predictive power
We can also sum up these three points with one sentence: a multimodal approach to assessing suicide risk that includes brain scans can help us effectively predict suicidality in seniors with depression. This is an important development in the treatment of depression in our older population, and especially among older people with treatment-resistant depression who are at increased risk of suicide. If we can use a multimodal approach to assessing suicide risk that includes data on functional connectivity in the brain, we can work to prevent suicide in seniors with depression. While access to this technology may be a limiting factor, understanding the importance of multimodal assessment is an important new detail in our understanding of depression among seniors, which we can apply in our work with our older patients every day.