Are Wishes for Death or Suicidal Ideation Symptoms of Depression in Older Adults?

Clinicians may question whether thoughts of being better off dead are normal consequences of aging rather than symptoms of depression. In this study, we examine whether thoughts of suicidal ideation are as strongly linked to depression severity in older adults as they are in other age groups in a large geographically and ethnically diverse sample.

METHODS

Design

Cross-sectional cohort study.

Participants

A total of 509,945 adults 18 years and older outpatients from four large integrated health systems in the Mental Health Research Network (MHRN) completed 1.2 million Patient Health Questionnaires (PHQ) and had data available to calculate Charlson Comorbidity Index scores from 1/1/2010 through 12/31/2012.

Measurements

The PHQ8 was used to estimate depression severity, while suicidal ideation was measured using the 9 th item of the PHQ. Data were abstracted from a Virtual Data Warehouse used to standardize data definitions across healthcare systems.

RESULTS

In older adult patients, suicidal ideation was very strongly associated with depression severity. Older adults who had at least moderately severe depression, defined as a PHQ8 ≥15, were 48 times more likely (95% CI: 42.8 to 53.8) to report suicidal ideation than those with minimal or mild symptoms of depression (PHQ8 score <10) after adjustment for all other variables in the model, including medical comorbidity burden.

CONCLUSIONS

Depression severity was by far the strongest predictor of suicidal ideation in older adult patients. Older patients who report suicidal ideation should be screened for depression.

INTRODUCTION

Despite consistent evidence that the risk of suicide death is higher in older adult patients than other age groups (Juurlink, Herrmann, Szalai, Kopp, & Redelmeier, 2004), clinicians may wonder whether thoughts of death in older patients are related to depression or instead are a natural consequence of aging. It is important to make this distinction, as thoughts of death related to depression significantly increase one’s risk of suicide and are amenable to treatment. However, it may be that older adults think about death more as they perceive their future time as more limited, and that this thinking is a normal part of aging (Lang and Carstensen, 2002) Along these lines, a workgroup for the American Association of Geriatric Psychiatry (AAGP) stated that clinicians have only a crude understanding of how to differentiate thoughts of death that are expected phenomena of aging from thoughts of death that indicate psychopathology.(Szanto et al., 2013) The AAGP workgroup lamented a lack of tools to determine when thoughts of death in older populations deserve intervention.

A first step toward making this distinction is to examine whether thoughts of death or suicide show the same relationship to depressive symptoms in older adults as in younger people. If the association between thoughts of death and depression severity is as robust in older patients as it is in younger patients, this would suggest that suicidal ideation or thoughts of death in older people should generally be treated as symptoms of depression, as they would at younger ages.

The Patient Health Questionnaire (PHQ) was recently recommended as a tool for screening for and monitoring depression in U.S. healthcare systems(“HEDIS Depression Measures Specified for Electronic Clinical Data Systems,”) and for screening for suicidal ideation(“Zero Suicide.,”). Here, we use data from a large sample of outpatients completing the PHQ to determine whether thoughts of death are as strongly associated with depression severity in older adults as it is in other age groups in a large diverse outpatient sample of patients who have and have not been previously diagnosed with depression. Depression severity is assessed using the first 8 questions of the PHQ, collective known as the PHQ8. Thoughts of death are assessed via the 9 th item of the PHQ. We specifically evaluate whether thoughts of death is as strongly associated with depression severity in older patients as it is in younger patients in a large diverse outpatient sample of outpatients.

METHODS

DATA SOURCES

The Mental Health Research Network (MHRN) is a consortium of public-domain research centers affiliated with 13 large not-for-profit integrated healthcare systems across the U.S. Four MHRN healthcare systems (HealthPartners, Kaiser Permanente Colorado, Kaiser Permanente Southern California and Kaiser Permanente Washington) were routinely using the Patient Health Questionnaire (PHQ) to screen for and monitor depression in 2010-2012 and contributed data for this study. These organizations provide comprehensive medical care to approximately 4.4 million members and patients in Minnesota, Wisconsin, Colorado, Idaho, California and Washington. Samples for mixed-model healthcare systems (HealthPartners and Kaiser Permanente Washington) were limited to members who were also patients to ensure availability of complete electronic health record (EHR) data.

EHR, insurance claim and other administrative data were organized into a Virtual Data Warehouse (VDW), a federated data system that facilitates sharing of de-identified data by using common data definitions and formats while keeping protected health information at each healthcare system.(Ross TR, 2014) Each local Institutional Review Board approved all study procedures and granted waivers of consent to use de-identified data for research.

STUDY SAMPLE

The study sample included patients 18 and older who completed a PHQ during an outpatient encounter between 1/1/2010 and 12/31/2012 and were enrolled in the health plan at the time of the PHQ completion. Patients could contribute multiple PHQs to the sample if completed in separate encounters.

MEASURES

Patient Demographics

Age and self-reported gender and race/ethnicity were recorded in healthcare systems’ EHRs and incorporated into the VDW. Race and ethnicity categories were mutually exclusive as per national guidelines.(Institute of Medicine. Ethnicity, and Language Data: Standardization for Health Care Quality Improvement., 2009) Patients self-reporting Hispanic ethnicity were considered Hispanic regardless of self-reported race.(Taylor, 2012)

Depression Diagnosis and Treatment

All patients who were receiving an FDA-approved medication for depression or engaged in psychotherapy were included in these analyses. All FDA-approved medications for depression, with the exception of trazodone, which is used primarily for insomnia, were considered treatment for depression (this list is available at www.mhresearchnetwork.org). Psychotherapy treatment was defined using standardized procedural terminology codes for diagnostic interviews and assessments and individual psychotherapy. Appointments less than 30 minutes or clearly designated as medication management only were not considered psychotherapy.

To determine whether results may differ in a sample of patients truly screened for depression using the PHQ, we repeated our main analysis in a subset of patients who had no diagnosis of depression (ICD-9 diagnosis 296.2, 296.3, 300.4 or 311) or treatment for depression (antidepressant prescription or psychotherapy treatment) in the previous 5 years. All diagnosis and treatment data were extracted from the VDW.

Depression Severity and Suicidal Ideation

All four healthcare systems recommended use of the PHQ for assessment of patient depression severity at initial and follow-up visits for depression, but recommendations for its use varied between healthcare systems and between clinics. The PHQ was recorded in each system’s EHR. The first 8 items of the PHQ (Huang, Chung, Kroenke, Delucchi, & Spitzer, 2006; Kroenke, Spitzer, & Williams, 2001) (collectively called the PHQ8 (Kroenke et al., 2009)) ask about frequency of depressive symptoms in the past two weeks. Item 9 of the PHQ asks “over the last two weeks, how often have you been bothered by thoughts that you would be better off dead, or of hurting yourself in some way?” Response options for all items of the PHQ are “not at all” (0 points), “several days” (1 point), “more than half the days” (2 points) or “nearly every day” (3 points). Scores of 1-4 on the PHQ8 indicate minimal depression, 5-9 mild, 10-14 moderate, 15-19 moderately severe, and 20-27 severe.(Kroenke, et al., 2001) (Kroenke, et al., 2009) Suicidal ideation was defined as a score on item 9 of the PHQ >0.

Charlson Comorbidity Index Score

Overall burden of medical comorbidity was calculated using the Charlson Comorbidity Index Score.(Charlson, Pompei, Ales, & MacKenzie, 1987) The Charlson is comprised of 19 categories of comorbid medical conditions, obtained from ICD-9 diagnosis codes and each weighted based on the adjusted risk of post-discharge mortality. Charlson scores were divided into three groups: 0, 1, and >1, with higher scores indicating greater burden of morbidity and higher likelihood of 10-year mortality.

Analyses

Percentages and fully adjusted multivariate logistic regression models were used to describe the relationship between each covariate and suicidal ideation. Models were adjusted for all covariates as well as the health system. Wald tests(Wald, 1945) were used to calculate p-values for the association between each covariate and suicidal ideation. Logistic regression models estimated the proportion and relative odds of reporting suicidal ideation on item 9 of the PHQ as a function of age (in four categories) and PHQ8 score (in four categories), both in the overall sample and in those patients without diagnosis or treatment of depression. All logistic regression models were estimated using generalized estimating equations with a log link. In cases where there were multiple observations nested within patients, an independent correlation structure was used. All analyses were performed using SAS version 9.3.

RESULTS

Patient Characteristics

A total of 1,228,308 PHQs were completed by 509,945 patients in 2010 through 2012 ( Table 1 ). Seventy-two percent of PHQs were completed by women, and 17% of PHQs (203,668) were completed by patients over the age of 65. PHQs were most frequently completed by patients who self-identified as non-Hispanic white (n=801,715) or Hispanic (n=195,292), followed by patients who self-identified as non-Hispanic black (n=81,159), Asian (n=60,513), Native American/Alaskan Native (n=14,400) or Native Hawaiian/Pacific Islander (n=5948). One-third of PHQs indicated that patients were experiencing no or minimal symptoms of depression (PHQ8 score of 0-4), while 25% indicated mild symptoms of depression (PHQ8=5-9), 19% indicated moderate symptoms (PHQ8=10-14) and 20% indicated moderately-severe or severe symptoms of depression (PHQ8≥15). Most PHQs were completed by patients in mental health (41%) or primary care settings (36%), with another 23% completed by patients in other settings, such as health education or specialty care clinics. Most patients (67%) had very low comorbidity burden as indicated by a Charlson score of zero.

Table 1

Patient characteristics and thoughts of self-harm assessed by the PHQ. 509,945 patients completed 1,228,308 PHQs. Item 9 of the PHQ asks about thoughts of being better off dead or of hurting oneself over the last two weeks. Response options include “not at all” (Q9=0), “several days” (Q9=1), “more than half the days” (Q9=2), or “nearly every day” (Q9=3). Reported suicidal ideation is defined as PHQ item 9>0. Odds ratios are fully adjusted for all table covariates and the care system.

Suicidal Ideation

In patients of all ages, including older adults, the strongest predictor of suicidal ideation was depression severity: the more severe a patient’s depression, the greater the likelihood that he or she reported suicidal ideation ( Figure 1a ). Older adult patients who had at least moderately severe depression (PHQ8≥15) were 48 times more likely to report nearly daily suicidal ideation (95% CI: 42.8 to 53.8) than those with minimal or mild symptoms of depression ( Figure 1b ). While the strength of the association between frequent thoughts of self-harm and PHQ8 score was mildly weaker in older patients compared to other age groups (relative odds of frequent suicidal ideation in patients with PHQ8 scores ≥15 ranged from 48- to 140- times the odds for those with PHQ8 scores <10), it was still very robust at 48-fold. When specifically testing for a linear interaction comparing older adults to other age groups, the interaction was highly significant (Wald statistic = 122.1, df=1, p<0.0001) but the actual magnitude of the effect was quite modest. The coefficient for the interaction term was 0.96 (95% CI: 0.95-0.97), indicating that the effect of each 1-point increase in PHQ8 score on the odds of frequent suicidal ideation was 96% as large in patients over the age of 65 compared to patients under age 65. Depression severity was a vastly larger determinant of suicidal ideation than was age.

(a) Proportion of encounters in which patients reported thoughts of death or self-harm “nearly every day” according to PHQ8 scores, stratified by age (N=1,228,308).

(b) Relative odds of a patient having indicated thoughts of death or self-harm “nearly every day” according to PHQ8 scores, stratified by age (N=1,228,308).

Across all ages, PHQ8 scores of 5-9 were associated with a 7-fold increased adjusted odds of suicidal ideation compared to PHQ8 scores of 0-4 (95% CI 6.8-7.3), while scores of 10-14 had a 19-fold increased adjusted odds of suicidal ideation (95% CI 18.6-19.7) and scores of 15 and over had a 57-fold increased adjusted odds of suicidal ideation (95% CI 55.1-58.5, p <.001) (Table 1 ). On 83% of PHQs, patients did not report thoughts of suicide, while such thoughts were reported as occurring several days in the past two weeks on 11% of PHQs, more than half the days on 4% of PHQs, and nearly every day on 2% of PHQs.

Finally, we examined the relationship between depression severity and suicidal ideation in a sub-sample of patients who had no current or past diagnoses or treatment of depression ( Figure 2a ). The relationship between depression severity and suicidal ideation was essentially the same, with patients of all ages at highly increased risk of suicidal ideation if their depression was moderately-severe or severe. Figure 2b demonstrates that even in patients who had no history of depression diagnosis or treatment, the vast majority of patients reporting nearly daily suicidal ideation had significant symptoms of depression, most with at least moderately severe depression.

(a) Proportion of encounters where patients reported thoughts of death or self-harm “nearly every day” by PHQ8 scores, stratified by age. Sample limited to patients with no current or past depression diagnosis or treatment (N=289,040).

(b) PHQ8 score among patients aged 65 or older reporting thoughts of death or self-harm “nearly every day”, stratified by depression diagnosis/treatment status (N= 4,601).

DISCUSSION

Similar to other age groups, older adults with moderately-severe to severe depression were 48 times more likely to report suicidal ideation than other older adults in this large, diverse outpatient sample. The relationship between depression severity and suicidal ideation was very strong across all age groups, and depression severity was by far the strongest predictor of suicidal ideation for all age groups.

Our findings suggest that thoughts of death or suicidal ideation experienced by older people are essentially as strongly related to symptoms of depression as they are in other age groups, and that these thoughts should not be considered normal consequences of aging. Our finding is consistent with several previous studies(Alexopoulos, Bruce, Hull, Sirey, & Kakuma, 1999; Barnow and Linden, 2000; Gensichen, Teising, Konig, Gerlach, & Petersen, 2010), but may refute others that suggest that thoughts of death may indicate natural preparations for death (Szanto, et al., 2013) or that item 9 of the PHQ confuses death ideation with suicide ideation.(Heisel, Duberstein, Lyness, & Feldman, 2010) Our results indicate that the close relationship between depression severity and thoughts of death is similar in all age groups and suggests that this relationship is not somehow unique in older adults. Of concern, our previous work found that older adults who report thoughts of death on the PHQ are at significantly increased risk for both suicide attempts and deaths.(Rossom et al., 2017) Taken together, these results indicate that patients of any age – including older adult patients – who report suicidal ideation should be evaluated for depression and treated as indicated, and should be considered at significantly increased risk of suicide. Thoughts or wishes for death in older patients should not be attributed to normal a normal development stage, but instead to depression, similar to other patients, and should be seen as a significant risk factor for self-harm.

Our findings of this connection between depression severity and suicidal ideation stand in apparent contrast to a survey which found that older New Yorkers were more likely to identify psychosocial stressors rather than depression as the reason for their suicidal ideation.(“Suicidal thoughts among elderly driven by physical health socioeconomic vulnerabilities., ” 2014) This was touted by the press, including a blog on the New York Times(Graham, August 1, 2014), as evidence that older adults considering suicide do not cite depression as the primary reason. However, this depiction presents a false dichotomy, as depression and psychosocial issues are of course by no means mutually exclusive. Additionally, while it may be easier for an individual to identify immediate stressors contributing to their thoughts of self-harm, these same psychosocial issues often contribute to and worsen the severity of depression. This seems to be particularly relevant for older adults, as older adults who died by suicide were more likely to have experienced serious relationship stressors, family discord, social isolation, employment changes and financial problems in the year or two preceding their death than others who died by suicide.(Beautrais, 2002; Duberstein, Conwell, Conner, Eberly, & Caine, 2004; Fiske, Wetherell, & Gatz, 2009; Rubenowitz, Waern, Wilhelmson, & Allebeck, 2001) Further, insight into depression can vary, with some patients not recognizing their anhedonia, irritability or feelings of emptiness as potential symptoms of depression, instead mistakenly thinking that depression only causes feelings of sadness.(Alexopoulos, 2005) Ultimately, our findings highlight the important distinction between association (suicidal ideation is very strongly associated with depression) and attribution (both suicidal ideation and depression may be attributed to specific life circumstances). Our study finds depression severity to be an extremely robust predictor of suicidal ideation regardless of age.

Prior studies have found that rates of suicidal ideation in older adult primary care patients range from 1% to 7%.(Callahan, Hendrie, Nienaber, & Tierney, 1996; Lish et al., 1996; Olfson et al., 2000) Our study found older adults reported such thoughts in 15% of encounters where PHQ9s were completed, but it should be noted that in addition to primary care clinics, some patients in our study were treated in mental health and specialty clinics, where rates of reported suicidal ideation were higher. In our sample, older adults experienced suicidal ideation at the same rate as 18-29 year-olds and 30-44 year-olds (15%), while patients aged 45-64 experienced a higher rate of suicidal ideation (20%). Given the high suicide death rate among older adults, with only 2-4 suicide attempts per suicide death in older patients compared to 8-40 suicide attempts for each suicide death in the general population,(Crosby, Cheltenham, & Sacks, 1999) the fact that older adults experienced suicidal ideation at rates similar to other ages is concerning and reinforces the need to address depression and suicidal ideation in this population.

There are several potential limitations to our data and our study. One is a limitation inherent to the PHQ, in that this tool was designed to screen for depression and assess its severity, not assess suicide risk per se. We do not have a measure of suicidal ideation beyond this single item assessment. We also do not have data on alternate depression screens, such as the Geriatric Depression Scale, and accordingly are not saying the PHQ is better than other screens, but rather that our data do not support concerns that item 9 of the PHQ has some very different meaning in older people or needs to be interpreted differently in older patients. This is a cross-sectional study, providing only a snapshot of associations between suicidal ideation and depression severity. Our sample included adults who were seen in outpatient clinics and completed a PHQ, likely biasing our sample towards patients who had a history of or who were at increased risk of depression and/or suicidal ideation, potentially limiting the generalizability of our findings to the general population. However, when we limited our analyses to patients without diagnoses or treatment of depression, this did not significantly alter our findings. Despite this, we do acknowledge that the use of the PHQ in these healthcare systems would likely have selected a population of older adults at higher risk for depression, and we cannot assess the frequency or correlates of suicidal ideation among a population of older adults not screened for depression. Overall, we think these limitations are outweighed by the opportunity to examine depression and suicidal ideation in a very large diverse sample of patients.

The relationship between depression severity and suicidal ideation is extremely robust in older adults. Thoughts of self-harm or suicide are not natural or inevitable consequences of aging, but are strongly associated with depression, a treatable illness. All patients – including older patients – who experience suicidal ideation should be screened for depression, and both suicidal ideation and depression should be treated as indicated.

Acknowledgments

Supported by NIMH Cooperative Agreement U19MH092201.

Sponsor’s Role: NIMH did not participate in the design, methods, data collection, analysis or preparation of this paper