Abstract
Objective:
The authors describe risk factors and suicide rates during the year following discharge from mental health emergency department (ED) visits by adults with suicide attempts, suicidal ideation, or neither.
Methods:
National cohorts of patients with mental health ED visits for suicide attempts or self-harm (n=55,323), suicidal ideation (n=435,464), or other mental health visits (n=9,144,807) from 2008–2018 Medicaid data were followed for suicide for 1 year after discharge. Suicide rates per 100,000 person-years were determined from National Death Index data. Poisson regression models, adjusted for age, sex, and race/ethnicity, estimated suicide rate ratios (RR). Suicide standardized mortality ratios (SMRs) were estimated from National Vital Statistics System data.
Results:
Suicide rates per 100,000 person-years were 325.4 for suicide attempt or self-harm visits (RR=5.51, 95%CI:4.64–6.55), 156.6 for suicidal ideation visits (RR=2.59, 95%CI:2.34–2.87), and 57.0 for the other mental health ED visits (1.0, reference). Compared to expected suicide general population rates, SMRs were 18.2 (95%CI:13.0–23.4) for suicide attempt or self-harm patients, 10.6 (95%CI:9.0–12.2) for suicidal ideation patients, and 3.2 (95%CI:3.1–3.4) for other ED mental health patients. Among patients with suicide attempt ED visits in the 180 days before their index mental health ED visit, suicide rates per 100,000 person-years were 687.2 (95% CI:396.5–978.0) for attempt or self-harm visits, 397.4 (95%CI: 230.6–564.3) for ideation visits, and 328.4 (95%CI:241.5–415.4) for other mental health visits.
Conclusion:
In the year following discharge, emergency department patients with suicide attempts or self-harm, especially repeated attempts, have a high risk of suicide.
Keywords: Emergency mental health care
Emergency departments (EDs) can play a pivotal role in the management of patients at high suicide risk. ED clinicians can provide mental health assessments and crisis interventions, facilitate inpatient admission for patients at acute risk, and help connect patients who do not require hospital admission to appropriate follow-up care. The large proportion of people who die of suicide who have had ED visits within a year of their death (43.8%) underscores the potential role of EDs in suicide prevention.1
Several patient characteristics, such as male sex,2 older age,3 mood and substance use disorders,4,5 psychiatric inpatient care,6 and suicidal ideation or attempts7–10 are associated with increased suicide risk. Among patients with suicidal symptoms, use of violent methods11 and suicidal intent12 pose especially high risk. Given these risks, ED physicians who evaluate patients with suicidal and other mental health symptoms routinely grapple with difficult decisions concerning the appropriate level of care. These decisions have implications for patient safety and crisis recovery.
In the general population, suicide attempts are strongly associated with increased suicide risk.,13,14 Approximately one in twenty-five patients who attempt suicide will die of suicide within five years.15 The types of psychiatric disorders associated with suicide attempts also appear to influence future suicide risk16 and repeat suicide attempts may confer greater risk.17 However, less is known about the magnitude of suicide risk and clinical factors that influence risk of patients with suicidal ideation in the absence of attempts. In one study, ED patients presenting with suicidal ideation, but not attempts, had a one year suicide risk that was roughly sixteen times greater than general ED patients without suicidal symptoms.9
In evaluating suicide risk, it would be helpful to understand the effects of recent events on current risk. It is not known, for example, whether among mental health ED patients with suicidal ideation, those with recent prior ED visits with suicidal ideation are at increased suicide risk. Large cohort studies are needed to connect recent ED visit patterns to future suicide risk. Claims records from mental health ED visits and from the prior 6-months were used to inform suicide risk assessment. The analysis compared 1-year suicide death rates of visits for suicide attempts or self-harm, suicidal ideation, or neither. We further examined relationships of recent clinical diagnoses and ED visit patterns with suicidal risk across these three groups to test the hypothesis that repeated non-fatal attempts increase suicide risk. We focused on adults with Medicaid coverage, the largest behavioral health payment source in the US.18
Methods
Sources of Data
The study cohort was identified using 2008–2012 national Medicaid Analytic Extract data purchased from ResDAC (https://www.resdac.org). In addition to demographic data (age, sex, race/ethnicity) collected directly from the beneficiaries, the Medicaid Analytic Extract includes information collected from providers on dates, service setting, and clinical diagnoses of each clinical encounter. Each outpatient visit record includes up to 2 ICD-9-CM diagnoses and each inpatient discharge record includes up to 9 diagnoses. Dates and cause of death information were derived from linkage to data purchased from the National Death Index (NDI) (https://www.cdc.gov/nchs/data/factsheets/factsheet_ndi.htm), which provides a complete accounting of state-recorded deaths in the US and is the most complete resource available for tracing mortality in national samples.19
Cohort Assembly and Suicide Outcome
We identified all ED visits including mental disorder diagnoses (ICD-9-CM: 290–319) for adults aged 18–64 years. Three hierarchical visit groups were defined by: 1) codes for suicide and self-inflicted injury (E950.0-E959.8), hereafter, suicide attempt or self-harm visits; 2) suicidal ideation (V62.84), or 3) other mental health visits. Study visits were also required to have Medicaid eligibility for ≥180 days preceding the ED visit and, if admitted for inpatient care following the index ED visit, to have an inpatient stay of < 30 days and to be alive at hospital discharge. The outcome variable was suicide defined as ICD-10-CM X60-X84, Y87.0, or U03 as the underlying cause of death.
Socio-demographic and Clinical Characteristics
The three study groups were classified by age (18–34; 35–44; 45–64 years), sex, and race/ethnicity (white, non-Hispanic; black, non-Hispanic; Hispanic; and other, non-Hispanic including American Indian/Alaskan Native, Asian, Native Hawaiian/Other Pacific Islander, more than one race, and unknown). Claims within 180 days preceding the index ED visit were used to characterize visits by codes with diagnoses for depressive, bipolar, anxiety, psychotic, personality including borderline as a subgroup, developmental, disruptive behavior, adjustment-related, substance use including alcohol and drug use, and other mental disorders (ICD codes in Supplementary Table 1). Visits were characterized by the presence of ≥1 inpatient mental health admissions within 180 days preceding the index ED visit. The visits were also hierarchically classified by ED visits in the prior 180 days with: 1) a suicide attempt or self-harm code, 2) a suicidal ideation code, or 3) a mental disorder diagnosis code.
Analysis
The analysis was performed in four stages. The first three were performed at the visit level and the fourth at the patient level. First, distributions of background clinical and demographic characteristics were derived for the three ED visit groups and the two suicidal symptom groups were compared to the other mental health visit group using risk ratios. Second, the three groups were followed forward until date of death or 365 days following ED discharge, whichever occurred first. Rates of suicide per 100,000 person-years were calculated. For each demographic and clinical group, age, sex, and race/ethnicity adjusted Poisson regression models estimated adjusted risk ratios with 95% confidence intervals with suicide as the dependent variable. Suicide attempt or self-harm visits and suicidal ideation visits were each compared to other mental health visits as the variable of interest. Third, survival curves of cumulative suicide risk over the first year were plotted separately for males and females and log-rank tests compared the three study groups. Finally, person-level age, sex, and race/ethnicity adjusted suicide standardized stratified mortality rate ratios (SMRs) were calculated using observed suicide deaths in the year following each patient’s first ED discharge. Expected annual suicide rates were derived from the US 2008–2012 general population using WONDER data.20 Separate suicide SMRs were derived for the three study groups overall and stratified by age, sex, and race/ethnicity.
In all visit-level analyses, generalized estimating equations adjusted for autocorrelations among visits from patients contributing multiple ED visits. All statistical analyses were performed with SAS 9.4 (SAS, Cary, NC). The University of Pennsylvania Institutional Review Board approved this study with a waiver of informed consent.
Results
Background characteristics
Most ED visits in each group were by women, adults ≤45 years, and white people (Table 1). The proportion of ED visits resulting in hospital admission was highest for visits presenting with suicide attempts or self-harm (46.6%), followed by suicidal ideation (27.6%), and other mental health visits (16.4%).
TABLE 1.
Characteristics of emergency department visits with mental disorder diagnoses. N=9,635,594
| Characteristic | 1 Suicide Attempta (N=55,323) % | 2 Suicidal Ideation (N=435,464) % | 3 No Attempt or Ideation (N=9,144,807) % | Risk Ratio of Attempt (95% CI) (No Attempt or Ideation Reference) | Risk Ratio of Ideation (95% CI) (No Attempt or Ideation Reference) |
|---|---|---|---|---|---|
| Sex | |||||
| 32.9 | 46.8 | 38.3 | -- | -- | |
| Female | 67.1 | 53.2 | 61.7 | 1.09(1.08–1.10) | 0.862(0.86–0.87) |
| Age, years | |||||
| 18–34 | 54.8 | 44.4 | 43.7 | -- | -- |
| 35–44 | 21.8 | 22.9 | 20.4 | 0.89(0.88–0.91) | 1.07(1.06–1.08) |
| 45–64 | 23.4 | 32.7 | 35.9 | 0.66(0.65–0.67) | 0.94(0.93–0.95) |
| Race/Ethnicity | |||||
| White, non-Hispanic | 64.7 | 56.5 | 55.9 | -- | -- |
| Black, non-Hispanic | 14.3 | 24.1 | 25.2 | 0.58(0.57–060) | 0.96(0.95–0.97) |
| Other, non-Hispanicb | 12.1 | 10.6 | 10.0 | 1.05(1.02–1.08) | 1.05(1.03–1.06) |
| Hispanic | 8.9 | 8.8 | 8.9 | 0.88(0.85–0.91) | 0.98(0.97–1.00) |
| Admitted to hospital | 46.6 | 27.6 | 16.4 | 2.85(2.82–2.88) | 1.69(1.68–1.70) |
| Any mental disorder diagnosisc | 78.4 | 86.3 | 76.2 | 1.03(1.02–1.03) | 1.13(1.129–1.13) |
| Depressive disorders | 50.5 | 60.6 | 34.5 | 1.46(1.45–1.48) | 1.76(1.75–1.76) |
| Bipolar disorders | 30.6 | 39.7 | 19.3 | 1.59(1.56–1.61) | 2.06(2.05–2.08) |
| Anxiety disorders | 34.1 | 34.1 | 25.3 | 1.35(1.33–1.37) | 1.35(1.34–1.36) |
| Psychotic disorders | 23.1 | 39.3 | 19.4 | 1.19(1.16–1.21) | 2.03(2.01–2.04) |
| Personality disorders | 11.5 | 14.5 | 4.6 | 2.51(2.42–2.60) | 3.15(3.10–3.20) |
| Borderline personality disorder | 7.2 | 6.5 | 1.9 | 3.87(3.69–4.06) | 3.49(3.42–3.56) |
| Developmental disorders | 5.3 | 8.2 | 7.3 | 0.73(0.70–0.77) | 1.13(1.10–1.15) |
| Disruptive behavior disorders | 10.3 | 11.0 | 6.5 | 1.58(1.52–1.63) | 1.70(1.67–1.72) |
| Adjustment-related disorders | 7.0 | 8.2 | 4.6 | 1.51(1.46–1.57) | 1.77(1.74–1.80) |
| Substance use disorder | 42.8 | 51.3 | 42.9 | 1.00(0.99–1.01) | 1.20(1.19–1.20) |
| Alcohol use disorder | 38.0 | 46.4 | 37.2 | 1.02(1.01–1.04) | 1.25(1.24–1.25) |
| Drug use disorder | 18.0 | 25.2 | 16.7 | 1.08(1.06–1.10) | 1.51(1.49–1.52) |
| Other mental disorders | 16.9 | 22.1 | 9.3 | 1.82(1.77–1.87) | 2.38(2.35–2.40) |
| Comorbid substance and other mental disorderc | 24.5 | 35.4 | 16.4 | 1.50(1.47–1.53) | 2.16(2.14–2.18) |
| Any inpatient mental health carec | 29.6 | 43.0 | 24.3 | 1.22(1.20–1.24) | 1.77(1.76–1.78) |
| Recent mental health ED visitsc, d | |||||
| Suicide attempt | 10.3 | 2.8 | 0.8 | 13.45(12.9–13.9) | 3.64(3.53–3.74) |
| Suicidal ideation | 10.9 | 30.0 | 5.5 | 2.01(1.96–2.06) | 5.49(5.45–5.54) |
| Mental disorder diagnosis | 31.0 | 31.8 | 47.7 | 0.65(0.64–0.66) | 0.67(0.66–0.67) |
Data from 2008–2012 Medicaid Analytic Extract.
Includes suicide attempts and self-harm events.
Includes American Indian/Alaskan Native, Asian, Native Hawaiian/Other Pacific Islander, unknown, and more than one race.
During 180 days prior to the index emergency department visit.
Subgroups defined hierarchically.
Based on non-overlapping confidence intervals, the groups differed with respect to clinical diagnoses during the 180 days preceding the index ED visits. Depressive, bipolar, anxiety, psychotic, and especially personality disorders were diagnosed more commonly among suicide attempt or self-harm visits and suicidal ideation visits than among other mental health visits. During this 180 day period, substance use disorders were diagnosed in approximately half of suicidal ideation visits (51.8%) and in lower percentages of suicide attempt or self-harm (42.8%) and other mental health (42.9%) visits. In this period, a larger percentage of suicide attempt or self-harm visits (10.3%) than suicidal ideation visits (2.8%) or other mental health visits (0.8%) had ≥1 recent prior ED visits with a suicide attempt. Also in this period, a larger proportion of suicidal ideation visits (30.0%) than attempt or self-harm visits (10.9%) or other mental health visits (5.5%) had recent prior ED visits with suicidal ideation.
Risk of suicide
The suicide rate per 100,000 person-years during the first year following discharge was highest for suicide attempt or self-harm visits (325.4), intermediate for suicidal ideation visits (156.6), and lowest for other mental health visits (57.0) (Table 2). As compared to the other mental health mental health visits, the age, sex, race/ethnicity adjusted suicide risk ratio was higher for suicide attempt or self-harm visits (5.51, 95% CI:4.64–6.55) than for suicidal ideation visits (2.59, 95% CI:2.34–2.87). A similar pattern was found across most sex, age, and race/ethnicity groups. Suicide risk was consistently higher among visits by males than females, and visits by white and Hispanic adults than by black adults. Among suicide attempt or self-harm visits, the suicide rate per 100,000 person-years was higher for visits by patients aged 45–64 years than 18–34 years.
TABLE 2.
Annualized rates of suicide per 100,000 person-years for emergency department visits during the first year following hospital discharge (N=9,635,594)
| Characteristic | Suicide rate per 100,000 person-yearsa | Adjusted Risk Ratio of Attempt (95% CI) (No Attempt or Ideation Reference) | Adjusted Risk Ratio of Ideation (95% CI) (No Attempt or Ideation Reference) | ||
|---|---|---|---|---|---|
| Suicide Attemptb (N=55,323) | Suicidal Ideation (N=435,464) | No Attempt or Ideation (N=9,144,807) | |||
| Overall | 325.4(269.0–381.7) | 156.6 (139.8–173.5) | 57.0(53.4–60.0) | 5.51(4.64–6.55) | 2.59(2.34–2.87) |
| Sex | |||||
| Male | 516.1(395.6–636.5) | 183.9(156.3–211.5) | 79.9(73.1–86.7) | 5.79(4.58–7.32) | 2.29(1.99–2.63) |
| Female | 231.8(172.1–291.4) | 132.6(112.3–152.8) | 42.8(38.6–47.0) | 5.19(4.02–6.70) | 3.06(2.64–3.55) |
| Age, years | |||||
| 18–34 | 211.1(155.5–266.7) | 134.3(112.1–156.6) | 46.8(41.7–51.8) | 4.36(3.35–5.68) | 2.61(2.22–3.07) |
| 35–44 | 405.5(258.1–552.9) | 162.5(127.7–197.2) | 61.8(53.7–69.8) | 5.98(4.21–8.51) | 2.55(2.06–3.16) |
| 45–64 | 518.5(369.6–667.5) | 182.8(149.2–216.5) | 66.9(60.1–73.6) | 6.72(5.03–8.97) | 2.59(2.19–3.06) |
| Race/Ethnicity | |||||
| White, non-Hispanic | 388.6(315.4–461.8) | 200.8(176.3–225.4) | 75.1(69.6–80.5) | 5.45(4.52–6.58) | 2.54(2.27–2.85) |
| Black, non-Hispanic | 88.4(0–177.6) | 49.6(29.1–70.1) | 20.9(15.7–26.1) | 4.58(1.63–12.87) | 2.16(1.42–3.27) |
| Other, non-Hispanicc | 267.7(119.3–416.2) | 181.5(120.7–242.2) | 60.5(49.2–71.9) | 4.68(2.67–8.17) | 2.96(2.12–4.13) |
| Hispanic | 325.8(87.0–564.6) | 135.3(79.5–191.1) | 42.1(29.0–55.1) | 8.50(4.07–17.73) | 3.18(2.13–4.73) |
| Admitted to Hospital | 399.2(309.8–488.5) | 166.2(135.6–196.7) | 61.4(55.4–67.3) | 5.88(4.63–7.47) | 2.61(2.15–3.17) |
| Any mental disorder dis.d | 373.4(305.3–441.5) | 165.3(146.3–184.4) | 67.0(62.3–71.7) | 5.30(4.42–6.36) | 2.36(2.11–2.63) |
| Depressive disorders | 407.9(317.2–498.7) | 176.1(151.9–200.3) | 93.7(85.4–102.1) | 4.19(3.36–5.22) | 1.80(1.58–2.04) |
| Bipolar disorder | 443.2(328.1–558.2) | 184.4(152.3–216.4) | 101.3(89.3–113.3) | 4.20(3.25–5.43) | 1.79(1.52–2.10) |
| Anxiety disorders | 429.4(315.4–543.5) | 207.2(172.3–242.1) | 92.0(82.6–101.5) | 4.54(3.49–5.91) | 2.13(1.82–2.49) |
| Psychotic disorders | 430.9(292.8–568.9) | 151.9(123.5–180.3) | 95.3(83.9–106.7) | 3.80(2.77–5.23) | 1.51(1.27–1.80) |
| Personality disorders | 313.5(151.0–476.0) | 193.4(143.9–243.0) | 162.4(129.9–194.9) | 1.87(1.12–3.11) | 1.21(0.95–1.56) |
| Borderline personality dis. | 324.3(100.3–548.3) | 186.3(121.2–251.3) | 187.5(128.2–246.7) | 1.70(0.87–3.33) | 1.00(0.68–1.47) |
| Developmental disorders | 203.9(15.6–392.2) | 115.3(59.9–170.7) | 30.8(20.2–41.4) | 6.47(2.44–17.17) | 3.73(2.24–6.24) |
| Disruptive behavior disorders | 317.4(155.4–479.4) | 104.2(70.0–138.4) | 64.1(50.8–77.4) | 4.86(3.03–7.81) | 1.60(1.17–2.18) |
| Adjustment-related disorders | 491.7(248.8–734.6) | 140.3(92.3–188.3) | 88.3(67.2–109.3) | 5.45(3.21–9.24) | 1.52(1.08–2.15) |
| Substance use disorder | 443.4(338.6–548.1) | 175.8(149.8–201.8) | 81.2(74.0–88.3) | 5.05(4.00–6.39) | 2.17(1.83–2.42) |
| Alcohol use disorder | 442.0(330.1–553.9) | 166.8(140.2–193.4) | 81.7(73.9–89.4) | 5.02(3.90–6.45) | 1.99(1.72–2.32) |
| Drug use disorder | 601.0(419.5–782.5) | 199.4(159.6–239.2) | 98.5(85.8–111.2) | 5.42(4.04–7.26) | 2.03(1.69–2.42) |
| Other mental disorders | 555.3(352.9–757.8) | 182.6(139.7–225.4) | 118.4(101.0–135.8) | 4.32(3.05–6.13) | 1.55(1.25–1.91) |
| Substance & other mental dis.d | 530.7(371.6–689.9) | 186.4(153.4–219.3) | 127.4(112.6–142.1) | 3.78(2.81–5.10) | 1.48(1.25–1.75) |
| Inpatient mental health cared | 513.4(384.6–642.1) | 183.6(153.6–213.6) | 100.8(90.9–111.6) | 4.51(3.49–5.82) | 1.79(1.52–2.09) |
| Recent ED visitsd, e | |||||
| Suicide attempt ED visits | 687.2(396.5–978.0) | 397.4(230.6–564.3) | 328.4(241.5–415.4) | 2.10(1.47–3.00) | 1.22(0.86–1.73) |
| No suicide attempt ED visits | 284.0(236.5–331.5) | 149.7(133.6–165.9) | 55.0(51.4–58.5) | 5.04(4.24–5.99) | 2.58(2.32–2.86) |
| Suicidal ideation | 429.3(264.6–594.0) | 167.1(131.8–202.5) | 153.0(127.5–178.5) | 2.57(1.75–3.79) | 1.14(0.94–1.37) |
| No suicide ideation ED visits | 312.6(253.9–371.2) | 152.1(137.0–167.2) | 51.5(28.1–54.9) | 5.93(4.91–7.16) | 2.81(2.53–3.12) |
| Mental disorder ED visits | 361.7(271.8–451.6) | 170.5(148.3–192.7) | 62.0(56.2–67.8) | 5.52(4.25–7.15) | 2.62(2.27–3.03) |
| No mental disorder ED visits | 309.0(242.2–375.9) | 150.1(129.8–170.5) | 52.5(48.9–56.1) | 5.69(4.62–7.00) | 2.65(2.35–2.98) |
Data from 2008–2012 Medicaid Analytic Extract and National Death Index. Results from a series of Poisson regressions adjusted for age, sex, and race/ethnicity.
The analysis includes 47 suicide deaths in the suicide attempt group 171 in the suicidal ideation group and 1,275 in the no attempt or ideation group.
Includes suicide attempts and self-harm events.
Includes American Indian/Alaskan Native, Asian, Native Hawaiian/Other Pacific Islander, unknown, and more than one race.
During 180 days prior to the index emergency department visit.
Subgroups defined hierarchically.
Within each group, the highest suicide rates occurred among visits ≥1 prior suicide attempt or self-harm ED visits in the preceding 180 days. For these groups, suicide rates per 100,000 person-years were 687.2 for suicide attempt or self-harm visits, 397.4 for suicidal ideation visits, and 328.4 for other mental health visits. Among groups defined by recent clinical mental health diagnoses, the highest suicide rates per 100,000 person-years in the suicide attempt or self-harm visits group were with drug use disorders (601.0), the highest in the suicidal ideation group were with anxiety disorders (207.2), and the highest in the other mental health group were with borderline personality disorder (187.2).
For most clinical characteristics, the adjusted suicide risk ratios were significantly higher for suicide attempt or self-harm and suicidal ideation visits than other mental health visits (reference group). Among visits with recent personality disorder diagnoses, however, the suicide risk ratios were not significantly higher for suicidal ideation than other mental health visits. Among suicidal ideation and other mental health visits, a similar pattern was also observed for ED visits with recent suicide attempt or self-harm and suicidal ideation.
Timing of suicide risk
For males (Figure 1) and females (Figure 2), the cumulative 1-year suicide risks were highest for suicide attempt or self-harm visits, intermediate for suicide ideation visits, and lowest for the other mental health visits. All six pairwise comparisons between the three study groups were significant (p<.0001). Across all groups, 20.4% of the suicide deaths occurred within 30 days, 33.5% within 60 days, 42.2% within 90 days, and 73.3% within 180 days of discharge. The percentage of suicide deaths that occurred within 180 days of discharge was 89.4% among suicide attempt or self-harm visits, 78.9% for suicide ideation visits, and 71.9% for other mental health visits.
Figure 1.
Cumulative suicide risk of male patients during the 365 days following discharge from a mental health emergency department visit
Figure 2.
Cumulative suicide risk of female patients during the 365 days following discharge from a mental health emergency department visit
Suicide risk in relation to general population
In relation to the general population, the demographically standardized suicide SMR for the first year following discharge was 18.2 (95%CI:13.0–23.4) for suicide attempt or self-harm patients, 10.6 (95%CI:9.0–12.2) for suicidal ideation patients, and 3.2 (95%CI:3.1–3.4) for other mental health patients (Table 3). Among ED patients without attempts or ideation, the suicide SMR was 3.9 (95%CI:3.6–4.2) for females and 2.9 (95%CI:2.7–3.1) for males.
TABLE 3.
Suicide rates per 100,000 person-years and suicide standardized mortality ratios in a cohort of emergency department patients with mental disorder diagnoses in the first year following discharge
| Suicidal Attempt or Self-Harm Event | Suicide Ideation | Other Mental Health Patients | ||||
|---|---|---|---|---|---|---|
| Observed Suicide Ratea | SMR (95% CI) | Observed Suicide Ratea | SMR (95% CI) | Observed Suicide Ratea | SMR (95% CI) | |
| Total | 209.3 | 18.2 (13.0–23.4) | 136.3 | 10.6 (9.0–12.2) | 38.4 | 3.2 (3.1–3.4) |
| Sex | ||||||
| Male | 391.8 | 16.5 (10.3–22.7) | 205.6 | 8.9 (7.2–10.6) | 65.3 | 2.9 (2.7–3.1) |
| Female | 128.5 | 21.2 (11.9–30.4) | 88.7 | 14.8 (11.2–18.4) | 23.3 | 3.9 (3.6–4.2) |
| Age group | ||||||
| 18–34 | 114.1 | 11.4 (5.8–17.0) | 111.7 | 9.8 (7.6–12.0) | 29.1 | 2.9 (2.6–3.1) |
| 35–44 | 343.3 | 26.0 (12.8–39.1) | 190.8 | 13.6 (9.8–17.5) | 47.4 | 3.7 (3.3,−4.1) |
| 45–64 | 393.7 | 26.6 (13.6–39.7) | 145.9 | 9.4 (6.7–12.1) | 48.0 | 3.4 (3.1,−3.7) |
| Race | ||||||
| White | 280.8 | 18.8(12.8–24.8) | 175.4 | 10.3 (8.5–12.0) | 52.6 | 3.3 (3.1–3.5) |
| Black | 0 | 0(0.0–0.0) | 40.4 | 5.9 (2.4–9.4) | 12.0 | 1.9 (1.5–2.3) |
| Hispanic | 233.9 | 44.3(8.8–79.7) | 114.6 | 18.3 (9.1–27.6) | 19.4 | 3.2 (2.4–3.9) |
| Others | 112.7 | 13.5(0–28.7) | 142.7 | 15.4 (8.5–22.3) | 42.7 | 4.8 (4.0–5.6) |
Data from 2008–2012 Medicaid Analytic Extract, National Death Index, and the National Vital Statistics System Mortality multiple cause-of-death files. SMRs standardized by age, sex, and race/ethnicity.
Suicide rates reported in 100,000 person-years.
Abbreviation: SMR denotes standardized mortality rate ratio.
Discussion
Mental health ED patients were at significantly greater suicide risk than the general population. This risk, which was concentrated during the first months following discharge, was highest for patients with suicide attempts or self-harm, intermediate for suicidal ideation, and lowest, though still elevated, for other mental health problems. Within each group, attempt or self-harm visits in the past 6-months emerged as the strongest suicide risk factor. Recent suicide attempt or self-harm history was associated with more than doubling suicide risk following suicide attempt or self-harm visits and suicidal ideation visits and increasing by nearly six-fold suicide risk following other mental health visits. These patterns underscore the critical importance of evaluating all emergency department mental health patients for a recent history of suicide attempts or self-harm.
History of suicidal behavior
A lifetime history of multiple attempts has been reported to increase suicide over long-term follow-up.21–23 The present findings extend this research by documenting increased suicide risk associated with a recent prior suicide attempt or self-harm including among current visits for attempts or self-harm. This finding contradicts clinical lore that people presenting with repeated self-harm are not at increased risk for suicide.
A recent history of ED visits with suicidal ideation was also associated with increased suicide risk. Among mental health ED patients without current suicidal or self-harm symptoms, recent suicidal ideation visit was associated with a nearly 3-fold greater suicide risk. However, a recent suicidal ideation visit did not appear to significantly increase risk among patients presenting with suicide ideation. This suggests a recent history of suicidal ideation operates differently on risk in the presence and absence of current suicidal ideation and highlights the importance of evaluating mental health ED patients without current suicidal symptoms for recent prior suicidal ideation visits.
Borderline personality disorder and drug use disorders
Among other mental health visits, a history of borderline personality diagnosis was associated with suicide risk over three times higher than overall mental health visits without suicidal symptoms. Even in the absence of overt suicidal symptoms, ED patients with a history of borderline personality diagnoses should be evaluated carefully for suicide risk. In the context of persistently elevated risk, some ED clinicians24 and mental health clinicians25 may develop negative views of patients with borderline personality disorder that may lead to inadequate safety assessments or compromise efforts to contact outpatient providers or social supports.26 However, brief training workshops have improved mental health clinicians’ attitudes and perceived competence in treating patients with borderline personality disorder.27,28
Healthcare professionals also tend to hold negative views of patients with drug use disorders that can result in lower personal engagement.29 Although drug use disorders are established risk factors for suicide,30 less is known about the effects of drug use disorders on suicide risk among suicidal patients. Among patients presenting with suicide attempts or self-harm, we found that patients with recent drug use disorder diagnoses had nearly twice the suicide risk. Prior research has provided inconsistent support for substance use disorders as a risk factor for suicide among acute care patients with suicidal symptoms.9,11,21 In evaluating this issue, it may be important to consider whether substance use disorder is diagnosed in the ED by symptoms of intoxication or withdrawal that may identify different patients than those who are diagnosed with substance use disorders in other settings. Drug intoxication can also complicate ED suicide risk assessment by impairing the patient’s responses to clinical questions.31 As a result, drug use disorder diagnoses derived from an ED visit might be less strongly related to future suicide risk than diagnoses from clinical evaluations preceding ED visits as in the present study.
Anxiety disorders and suicidal ideation
Among suicidal ideation visits, anxiety disorders were the diagnostic group with the highest suicide risk. Retrospective epidemiological research reveals significant associations of panic disorder and social anxiety disorder with nonfatal suicide attempts.32 Detailed clinical studies are needed to examine whether and how anxiety disorders contribute to transitions from suicidal ideation to suicidal behavior. Anxiety sensitivity is common in panic disorder and other anxiety disorders33,34 and may contribute to suicide risk.35 Longitudinal psychological research is needed to evaluate associations between anxiety disorders, anxiety sensitivity, suicidal ideation, and suicide.
Clinical implications
Although the current study focused on patients presenting to emergency departments, patients with suicidal symptoms in all clinical settings should receive careful mental health evaluations with a focus on modifiable intervention targets. According to a recent meta-analysis, annual suicide rates did not significantly differ between patients with suicidal thoughts or behaviors discharged from EDs or general hospital inpatient units.36
The concentration of suicide risk during the first few months following ED discharge has important clinical implications for short-term management. A thorough assessment should be provided of risk and protective factors of ED mental health patients including evaluation of suicidal intent.37 Some single encounter suicide prevention interventions significantly improve linkage to follow-up care and reduce subsequent suicide attempts.38 Common components include brief contacts involving telephone calls, post-cards, letters, or text messages reminders urging follow-up care;39 care coordination involving scheduling of outpatient appointments and reducing barriers to appointment attendance; and safety planning including helping patients identify personal warning signs of an impending suicidal crisis, internal coping strategies, social supports, and available mental health professionals, as well as providing patients with lethal means counseling.40 A variety of brief therapeutic interventions have also demonstrated promise including problem solving therapy for patients with repeated self-harm,41 follow-up telephone calls providing case management and supportive counseling,42 and follow-up text messages expressing care and concern.43
Limitations
This study has several limitations. First, the suicidal ideation codes44 and mental disorder diagnoses are based on routine clinical assessments and were not subjected to independent expert validation. Because suicidal symptoms are sometimes not detected in EDs, they are not consistently captured in the ED claims record.45 Second, different results might have been obtained if the analysis included more recent data, a wider age range, and uninsured or privately insured ED patients. Among ED patients with suicide attempts or self-harm, for example, Medicaid patients are significantly more likely than privately insured patients to have recently received inpatient psychiatric care and diagnosed with depression, schizophrenia, or bipolar disorder.46 Third, no information was available concerning several suicide risk factors such as a family history of suicide, lifetime personal history of attempted suicide, proximal stressful life events, or access to firearms or other lethal means.47 Fourth, stigma,48 low autopsy rates,49 and forensic uncertainty50 may result in underreporting or misclassification of suicide in the NDI data. Fifth, some visits classified as suicide attempts or self-harm events may have been for non-suicidal self-harm, though prior research revealed accurate documentation of suicidal intent.51 Finally, person-level comparisons to population norms tend to underweight visits by people with more ED visits who may be a higher risk group.
Conclusions
ED physicians commonly face clinical uncertainties in evaluating suicidal risk and in determining which mental health patients can be safely discharged to the community. In making these clinical judgments, ED physicians should bear in mind that suicide risk following ED discharge is substantially higher for mental health patients with current or recent ED visits for suicide attempts or self-harm events. Even among ED mental health patients without current suicidal symptoms, a recent suicide attempt or self-harm event markedly increased suicide risk. Learning which ED mental health patients pose high short-term suicide risk will hopefully improve clinical judgments concerning the appropriate level of follow-up care.
Supplementary Material
Clinical Points.
Assessing suicide risk in emergency department patients with mental health symptoms can be challenging.
Evaluate patients carefully for current as well as recent past suicidal symptoms.
Patients with a current suicide attempt and a prior attempt in the last 6 months appear to be at especially high risk for dying of suicide shortly after their emergency visit.
Funding/support:
This research was supported by a grant (R01 MH107452) from the National Institute of Mental Health.
Role of the sponsor:
The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation or approval of the manuscript.
Footnotes
Potential conflicts of interest: Dr. Marcus reports receipt of consulting fees from Allergan. None of the other authors report conflicts of interest.
References
- 1.Ahmedani BK, Westphal J, Autio K, et al. , Variation in patterns of health care before suicide: a population case control study. Prev Med 2019;127:105796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.World Health Organization, Suicide world health statistics data visualizations dashboard, year 2016. https://apps.who.int/gho/data/node.sdg.3-4-viz-2?lang=en. Accessed March 15, 2021.
- 3.World Health Organization, Global health observatory data repository, Suicide rate estimates, crude, 10-year age groups, estimated by country. https://apps.who.int/gho/data/node.main.MHSUICIDE10YEARAGEGROUPS?lang=e Accessed March 15, 2021
- 4.Chesney E, Goodwin GM, Fazel S. Risks of all-cause and suicide mortality in mental disorders: a meta-review. World Psychiatry 2014;13:153–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Olfson M, Wall M, Wang S, et al. Short-term Suicide Risk After Psychiatric Hospital Discharge. JAMA Psychiatry.2016;73(11):1119–1126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Chung DT, Ryan CJ, Hadzi-Pavlovic D, et al. Suicide rates after discharge from psychiatric facilities: a systematic review and meta-analysis. JAMA Psychiatry 2017;74(7):694–702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chakravarthy B, Hoonpongsimanont W, Anderson CL, et al. Depression, suicidal ideation, and suicidal attempt presenting to the emergency department: differences between these cohorts. West J Emerg Med 2014;15(2):211–216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Olfson M, Wall M, Wang S, et al. Suicide following deliberate self-harm. Am J Psychiatry 2017;174(8):765–774. [DOI] [PubMed] [Google Scholar]
- 9.Goldman-Mellor S, Olfson M, Lidon-Moyano C, et al. Association of suicide and other mortality with emergency department presentation. JAMA Network Open. 2019;2(12):e1917571–e1917571. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Crandall C, Fullerton-Gleason L, Aguero R, et al. Subsequent suicide mortality among emergency department patients seen for suicidal behavior. Acad Emerg Med 2006;13:435–442. [DOI] [PubMed] [Google Scholar]
- 11.Runeson B, Haglund A, Lichtenstein P, et al. Suicide risk after nonfatal self-harm: a national cohort study, 2000–2008. J Clin Psychiatry. 2016;77(2):240–6. [DOI] [PubMed] [Google Scholar]
- 12.Harriss L, Hawton K, Zahl D. Value of measuring suicidal intent in the assessment of people attending hospital following self-poisoning and self-injury. Br J Psychiatry 2005;186:60–66. [DOI] [PubMed] [Google Scholar]
- 13.Franklin JC, Ribeiro JD, Fox KR, et al. Risk factors for suicidal thoughts and behaviors: a meta-analysis of 50 years of research. Psychol Bull 2016;46:1–46. [DOI] [PubMed] [Google Scholar]
- 14.Fazel S, Runeson B. Suicide. New Eng J Med 2020;382:266–274. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Carroll R, Metcalfe C, Gunnell D. Hospital presenting self-harm and risk of fatal and non-fatal repetition: systematic review and meta-analysis. PLoS One 2014;9(2):e89944. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tidemalm D, Langstrom N, Lichtenstein P, et al. , Risk of suicide after suicide attempt according to coexisting psychiatric disorder: Swedish cohort study with long term follow-up. BMJ 2008;337(3):a2205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Chan MKY, Bhatti H, Stockton S, et al. , Predicting suicide following self-harm: systematic review of risk factors and risk scales. Br J Psychiatry 2016;209:277–283. [DOI] [PubMed] [Google Scholar]
- 18.Substance Abuse and Mental Health Services Administration. Projections of National Expenditures for Treatment of Mental and Substance Use Disorders, 2010–2020. HHS Publication No. SMA-14–4883 Rockville, MD: Substance Abuse and Mental Health Services Administration, 2014 [Google Scholar]
- 19.Wojcik NC, Huebner WW, Jorgensen G. Strategies for using the National Death Index and the Social Security Administration for death ascertainment in large occupational cohort mortality studies. Am J Epidemiol 2010;172(4):469–477 [DOI] [PubMed] [Google Scholar]
- 20.CDC WONDER, Multiple causes of death, 1999–2018. https://wonder.cdc.gov/mcd-icd10.html
- 21.Zahl DL, Hawton K. Repetition of deliberate self-harm and subsequent suicide risk: Long-term follow-up study of 11,583 patients. Br J Psychiatry 2004;185: 70–75. [DOI] [PubMed] [Google Scholar]
- 22.Cooper J, Kapur M, Webb R, et al. , Suicide after deliberate self-harm: a 4-year cohort study. Am J Psychiatry 2005;162:297–303. [DOI] [PubMed] [Google Scholar]
- 23.Suokas J, Suomien K, Isometsa E, et al. , Long-term risk factors for suicide mortality after attempted suicide – findings of a 14-year follow-up study. Acta Psychiatr Scand 2011;104:117–121. [DOI] [PubMed] [Google Scholar]
- 24.Treloar AJC, Lewis AJ. Professional attitudes towards deliberate self-harm in patients with borderline personality disorder. Austr New Z J Psychiatry 2008;42:578–584. [DOI] [PubMed] [Google Scholar]
- 25.Black DW, Pfol B, Blum N, et al. , Attitudes toward borderline personality disorder: a survey of 706 mental health clinicians. CNS Spectr 2011;16:67–74. [DOI] [PubMed] [Google Scholar]
- 26.Hong H Borderline personality disorder in the emergency department: good psychiatric management. Harv Rev Psychiatry 2016;24(5):357–366. [DOI] [PubMed] [Google Scholar]
- 27.Treloar AJ. Effectiveness of education programs in changing clinicians’ attitudes toward treating borderline personality disorder. Psychiatr Serv. 2009;60(8):1128–1131. [DOI] [PubMed] [Google Scholar]
- 28.Shanks C, Pfohl B, Blum N, et al. Can negative attitudes toward patients with borderline personality disorder be changed? The effect of attending a STEPPS workshop. J Personality Dis 2011;25(6):806–811. [DOI] [PubMed] [Google Scholar]
- 29.Van Boekel LC, Brouwers EP, Van Weeghel J, et al. Stigma among health professionals towards patients with substance use disorders and its consequences for healthcare delivery. Drug Alc Depend 2013;131(1):23–35. [DOI] [PubMed] [Google Scholar]
- 30.Poorolajal J, Haghtalab T, Farhadi M, et al. Substance use disorder and risk of suicidal ideation, suicide attempt, and suicide death: a meta-analysis. J Pub Health 2016;38(3):e292–e291. [DOI] [PubMed] [Google Scholar]
- 31.Betz ME, Boudreaux ED. Managing suicidal patients in the emergency department. Ann Emerg Med 2016;67(2):276–282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Nock MK, Hwang I, Sampson NA, et al. Mental disorders, comorbidity and suicidal behavior: results from the National Comorbidity Survey Replication. Molec Psychiatry 2010;15:868–876. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.McNally RJ. Anxiety sensitivity and panic disorder. Biol Psych 2002;52(10):938–946. [DOI] [PubMed] [Google Scholar]
- 34.Reiss S, Peterson RA, Gursky DM, et al. Anxiety sensitivity, anxiety frequency, and the prediction of fearfulness. Behav Res Therapy 1986;24(1):1–8. [DOI] [PubMed] [Google Scholar]
- 35.Stanley IH, Boffa JW, Rogers ML, et al. , Anxiety sensitivity and suicidal ideation/suicide risk: a meta-analysis. J Consult Clin Psychol 2018;86(110)46–960. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Wang M, Swaraj S, Chung D, et al. Meta-analysis of suicide rates among people discharged from non-psychiatric settings after presentation with suicidal thoughts and behaviors. Acta Psychiatr Scand 2019;139:472–483. [DOI] [PubMed] [Google Scholar]
- 37.Bolton JM, Gunnell D, Turecki G. Suicide risk assessment and intervention for people with mental illness. BMJ 2015; 351:h4978. [DOI] [PubMed] [Google Scholar]
- 38.Doupnik SK, Rudd B, Schmutte T, et al. , Association of suicide prevention interventions with subsequent suicide attempts, linkage to follow-up care, and depression symptoms for acute care settings: a systematic review and meta-analysis. JAMA Psych 2020;77(10):1021–1030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Milner AJ, Carter G, Pirkis J, et al. Letters, green cards, telephone calls and postcards: systematic and meta-analytic review of brief contact interventions for reducing self-harm, suicide attempts and suicide. Br J Psychiatry 2015;206:184–90. [DOI] [PubMed] [Google Scholar]
- 40.Stanley B, Brown GK, Brenner LA, et al. Comparison of the safety planning intervention with follow-up vs usual care of suicidal patients treated in the emergency department. JAMA Psychiatry. 2018; 75(9):894–900. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Hatcher S, Sharon C, Parag V, et al. , Problem-solving therapy for people who present to hospital with self-harm: Zelen randomized controlled trial. Br J Psychiatry 2011;199:310–316. [DOI] [PubMed] [Google Scholar]
- 42.Miller IW, Camargo CA Jr, Arias SA, et al. ED-SAFE Investigators. Suicide prevention in an emergency department population: the ED-SAFE Study. JAMA Psychiatry. 2017;74(6):563–570. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Comtois KA, Kerbrat AH, DeCou CR, et al. Effect of augmenting standard care for military personnel with brief caring text messages for suicide prevention: a randomized clinical trial. JAMA Psychiatry. 2019;76(5):474–483. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Classen C, Larkin G. Occult suicidality in an emergency department population. Br J Psychiatry 2005;186(4):352–353. [DOI] [PubMed] [Google Scholar]
- 45.Stanley B, Currier GW, Chesin M, et al. Suicidal behavior and non-suicidal self-injury in emergency departments underestimated in administrative claims data. Crisis 2018;39:318–325. [DOI] [PubMed] [Google Scholar]
- 46.Marcus SC, Bridge JA, Olfson M. Payment source and emergency management of deliberate self-harm. Am J Public Health 2012;102(6):1145–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Owens PL, Fingar KR, Heslin KC, et al. Emergency department visits related to suicidal ideation, 2006–2013. HCUP statistical brief #220. Rockville, MD: Agency for Healthcare Research and Quality; 2017. www.hcup-us.ahrq.gov/reports/statbriefs/sb220-Suicidal-Ideation-ED-Visits.jsp. [PubMed] [Google Scholar]
- 48.Tollefson IM, Hem E, Ekeberg O. The reliability of suicide statistics: a systematic review. BMC Psychiatry 2012;12:9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Kapusta ND, Tran US, Rockett IR, et al. Declining autopsy rates and suicide misclassification; a cross-national analysis of 35 countries. Arch Gen Psychiatry 2011;68(1):1050–1057. [DOI] [PubMed] [Google Scholar]
- 50.Rockett IR, Hobbs GR, Wu D, et al. Variable classification of drug-intoxication suicides across US States: a Partial artifact of forensics?. PLoS ONE. 2015;10(8):e0135296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Simon GE, Savarino J. Suicide attempts among patients starting depression treatment with medications or psychotherapy. Am J Psychiatry 2007;164:1029–1034. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.