Suicidal Behavior in Psychiatric Emergency Room Patients

Dirk M. Dhossche, MD, Department of Psychiatry, State University of New York at Stony Brook.

[South Med J 93(3):310-314, 2000. © 2000 Southern Medical Association]

Abstract

Background. Psychiatric emergency room (ER) patients are thought to be at increased risk of suicide. The prevalence and characteristics of suicidal behavior in a recent sample of patients who came to the ER for psychiatric evaluation were examined.
Methods. Charts of 311 consecutive psychiatric ER patients were reviewed. Suicidal behavior was considered present if current suicidal ideation or attempts within 24 hours of or during the emergency evaluation were noted in the chart.
Results. Suicidal behavior was present in 38% of the psychiatric ER patients. Younger age, white race, affective disorders in female patients, and substance abuse disorders in male patients were features of the suicidal group. Sex of the patient was not associated with suicidal behavior.
Conclusions. Suicidal behavior is prevalent in the psychiatric ER. Effective suicide prevention in this setting will hinge on finding more specific risk factors.

Introduction

Patients are often evaluated for suicidal ideation or attempts in the ER. The assessment of suicide risk has important implications for treatment, disposition, and medicolegal issues.[1] In a previous study,[2] psychiatric ER patients were found to be at increased risk of suicide compared with the general population.

I assessed demographic, psychosocial, and clinical features of ER patients with suicidal behavior, asking the following questions: What is the prevalence of suicidal behavior in a consecutive sample of psychiatric ER patients? How do demographic and clinical variables compare between patients with and without suicidal behavior? What are the determinants of psychiatric ward admission among ER patients with suicidal behavior?

Methods

Characteristics of the setting have been described previously.[3] In brief, the Comprehensive Psychiatric Emergency Program is located in University Hospital Stony Brook, Long Island, NY, and provides 24-hour care for a large, suburban community. The staff consists of attending psychiatrists, residents, nurses, and social workers. There is a short-term holding area for extended evaluations lasting up to 72 hours.

All consecutive cases evaluated in November 1992 were retrieved from the hospital computer system. If a patient had more than one visit to the psychiatric ER, only the first visit was included. The final sample consisted of 340 consecutive patients. Charts of 311 (91.5%) of these patients were reviewed. Demographics, primary diagnoses made by attending psychiatrists according to the Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition (DSM-III-R), and clinical information were recorded in a database. Suicidal behavior was defined as suicidal ideation or attempts as reported by the patients or informants (family members, police officers, mental health professionals, and physicians), within 24 hours of or during the ER visit. A history of suicide attempts was recorded separately.

Measures of substance abuse were primary DSM-III-R diagnoses of substance abuse disorders and results of toxicologic examination of urine. Urine was screened for a wide range of substances by immunoassay techniques, followed by confirmatory gas-liquid chromatography. An enzymatic reaction was used to detect alcohol. Urine toxicology was done in 155 (50%) of the 311 patients. Cases with missing toxicology were due to rapid turnover of patients, refusal by patients, and technical problems. Among the 49 patients (32%) whose toxicology results were positive for substances, 27 (17%) had results positive for alcohol. Cocaine was detected in 18 cases (12%) and cannabis in 9 cases (6%). Analyses were restricted to the two most common substances, ie, alcohol and cocaine.

The 95% confidence interval (CI) of the proportion of patients with suicidal behavior was calculated with the following formula:

equ (p = proportion, n = 311)

Chi2 Analyses, Fisher's exact tests, and two-tailed t tests were used for comparison between patients with and without suicidal behavior. Variables that were significantly associated with the dependent variable (ie, presence or absence of suicidal behavior or hospital admission versus nonadmission) at the .05 significance level were entered in a stepwise logistic regression model for multivariate adjustment. The final model was accepted only if backward and forward elimination of variables yielded the same results. All analyses were done with SPSS 6.1 software (SPSS Inc, Chicago, Ill). P < .05 was considered significant.

Results

Suicidal behavior was present in 117 (38%; 95% CI, 33% to 43%) of 311 consecutive psychiatric ER patients (Table 1). The distribution of age, race, employment status, and history of suicide attempts was different between patients with and without suicidal behavior. In the logistic regression, these variables were all retained in the final model (P < .05).

Diagnoses and toxicology results are shown for men and women separately in Tables 2 and 3. Among patients with suicidal behavior, affective disorders were the most frequent diagnosis in both sexes (Table 2). However, substance abuse disorders were more frequent in men than women. Detection of alcohol or any substance was associated with suicidal behavior, but only in men (Table 3).

Among patients with suicidal behavior, being admitted to the hospital was associated with age and diagnosis in univariate and multi-variate analyses (Table 4). No such association was found for sex, race, presence of previous suicide attempts, history of substance abuse, toxicology, who brought the patient to the ER, marital status, or employment status.

Discussion

Main Findings

Suicidal behavior was a frequent problem in psychiatric ER patients. Younger age, white race, affective disorders in women, and substance abuse disorders in men, were features of psychiatric ER patients with suicidal behavior. Increasing age and diagnostic assessments of affective and psychotic disorders were associated with psychiatric ward admission among patients with suicidal behavior.

Strengths and Limitations

Although the sample size of 311 patients is comparable to that of other studies, the number of patients is still relatively small for detailed analyses of subgroups of patients. Thus, findings should be considered preliminary, and replication in larger studies is necessary.

Data obtained from chart entries may be subject to underreporting of suicidal behavior.[4] Information for this study was extracted from a semi-structured questionnaire that was systematically completed by nurses in face-to-face contact with patients and from informants' information, police reports, and clinical assessments. This reduces the chance that a major source of underreporting was due to incomplete assessments. Moreover, the rate of 38% is comparable to rates in other ER studies.[5,6]

Higher rates of suicidal behavior have been reported in anonymous surveys compared to nonanonymous or face-to-face assessments.[7,8] The value of detailed assessments of suicidal behavior is beyond dispute. Yet, the clinical superiority of research versus clinical assessments, interview versus questionnaire, and face-to-face contact versus anonymous reporting (which lacks clinical value), remains unsettled. The present standard of care rests on thorough clinical assessments. More research is warranted on reliable methods to assess suicidality.

Presence of suicidal behavior was recorded without differentiating between suicidal ideation versus attempts. This may be interpreted as a limitation, since previous studies have shown that there are several differences between suicide ideators and attempters.[9] However, studies focusing on one type of suicidality may be too restrictive and may not cover the whole population at risk of suicide. Women are usually over-represented in samples of attempted suicide, but there are more men than women in most suicide samples.[10] Also, only about one third (range, 22% to 44%) of people who die by suicide are known to have made previous attempts.[9] Therefore, it is argued that once suicidal behavior has become the focus of clinical attention, a "lumping" approach is reasonable because there is little evidence that suicidal ideators versus attempters have different prognoses and are at different risks of suicide in ER populations. Further studies are warranted to examine different types and aspects of suicidality, including "aborted" attempts, seriousness of the attempt, or intensity of suicidal preoccupations.[11-13]

Selection factors in ER populations need to be considered. Previous studies[14,15] have suggested that people with low socioeconomic functioning, people who are chronic mentally ill, substance abusers, single people, and men are overrepresented in psychiatric ER samples. Another factor may be important in this particular study setting; ie, people with medically serious suicide attempts may be underrepresented in the study sample. Such patients are directly admitted to the wards and are not evaluated in the psychiatric ER, but they are seen by the psychiatric consultation service after emergency treatment and stabilization on the intensive care unit. This may provide a possible explanation for the absence of sex differences in suicidal behavior due to a higher medical ward admission rate of women with overdoses. These factors limit the representativeness of the study sample and generalizability of findings.

The associations between toxicology results and suicidal behavior may be artifacts, since only half of the consecutive patients (155 of 311) had toxicologic screening of urine. However, urine toxicology was done in as many patients with suicidal behavior (50%) as without suicidal behavior (49%). This suggests that there was no obvious bias in the subgroup in which toxicologic examination of urine was done. In addition, comparison of suicidal people with and without the urine toxicology showed no differences in age, race, marital status, history of suicide attempts, or diagnosis. However, more suicidal men than women were screened toxicologically (71% versus 50%, P < .05). This may underlie the apparent association between positive toxicologic results and suicidality in men. However, the associations between suicidal behavior and toxicology results in men are similar to findings in primary diagnoses of substance abuse disorders and are consistent with other reports of sex differences in substance abuse disorders, as discussed below.

Prevalence of Suicidal Behavior

The prevalence of suicidal behavior of 38% in this study is comparable with prevalence rates in other reports. In 343 consecutive psychiatric ER patients,[5] suicidal ideation and/or suicide attempts were present in 39%. In another ER study[6] of 294 consecutive patients, 53% had suicidal behavior. Naturally, high rates of suicidal behavior are expected in acute-care settings. Recent studies[16,17] in outpatient settings have also shown high rates of suicidal behavior.

Characteristics of Patients With Suicidal Behavior

High rates of unemployment, being single, previous suicide attempts, previous substance abuse, affective disorders, and substance abuse disorders have been reported in samples of people who have attempted suicide and those who have died by suicide.[9,18,19] In the present study, sex of the patient was not associated with suicidal behavior. This is contrary to several studies[7,9,18] reporting a greater tendency among women toward suicidal behavior. Recent studies[20,21] in suicide attempters, however, have reported sex ratios closer to 1:1. Alternatively, the absence of a sex difference may be artifactual, since women with overdoses may be admitted directly to the medical wards. It is also important to emphasize the differences between people who attempt suicide and people who die by suicide. In the latter group, there is a preponderance of men, older age, and firearms as the method of self-harm.

Lower rates of suicidal behavior in blacks have been reported in the Epidemiologic Catchment Area Surveys.[21] This is contrary to an earlier report[22] suggesting that suicide attempts were more frequent in black ER patients compared with whites. Lower rates of suicidal behavior are congruent with lower suicide rates in blacks, however.[23]

Previous reports[9,24-26] have suggested that substance abuse is a risk factor for depression and suicidal behavior. In the present study, substance abuse correlated with suicidal behavior, particularly in men. A previous finding[3] that suicidality is increased disproportionately in men with recent cocaine use compared with women was not a finding in this study. This may be due to the small number of patients in the current sample whose urine was positive for cocaine. However, detection of substances was more frequent in men with suicidal behavior (Table 3). Overall, findings support recent studies[27-29] suggesting that neuropsychiatric effects of substance abuse are sex-specific.

An increased risk of suicide has been shown in psychiatric ER patients compared with the general population.[2] Risk factors included male sex; white race; affective, schizophrenic, and substance abuse diagnoses; and repeat visits. However, these are not specific enough to predict individual cases of suicide. Further studies are needed to assess more specific risk factors for suicide in this population.

Conclusions

This study found a high rate of suicidal behavior in psychiatric ER patients. This is expected in crisis patients with high rates of diagnosable psychiatric disorders. Age and diagnoses were important factors that clinicians considered when admitting patients with suicidal behavior to psychiatric wards. Among patients with suicidal behavior, more specific risk factors for suicide are needed to plan cost-effective preventive measures.

This study was supported by the United States Public Health Service (grant No. MH-00747).

Reprint requests to Dirk M. Dhossche, MD, University of South Alabama College of Medicine, Department of Psychiatry, 2451 Fillingim St, Mobile, AL 36617-2293.

Table 1. Comparison of Sociodemographic and Clinical Variables in Psychiatric Emergency Room Patients With and Without Suicidal Behavior

  Patients With
Suicidal Behavior
(n = 117)
No. (%)		 Patients Without
Suicidal Behavior
(n = 194)
No. (%)
Age, yr*
   <30 66 (56) 75 (39)
   30-44 35 (30) 85 (44)
   45-59 10 (8) 25 (13)
   >/=60 6 (5) 9 (5)
Sex
   Male 71 (61) 126 (65)
   Female 46 (39) 68 (35)
Race*
   White 100 (85) 146 (75)
   Black 17 (14) 48 (25)
Marital status
   Married 20 (17) 24 (12)
   Not married 97 (83) 170 (88)
Employment status†
   Employed 52 (44) 53 (27)
   Unemployed 65 (56) 141 (73)
Previous suicide attempts**
   Yes 32 (27) 13 (7)
   No 85 (73) 181 (93)
Hospital admission status
   Admitted 48 (41) 78 (40)
   Not admitted 69 (59) 116 (60)

* P < .05.
† P < .005.
** P < .001.

Table 2. Primary Psychiatric Diagnoses in Psychiatric Emergency Room Patients With and Without Suicidal Behavior, by Sex

Primary Diagnoses Females* (n = 114) Males† (n = 197)
With Suicidal Behavior
(n = 46)
No. (%)		 Without Suicidal Behavior
(n = 68)
No. (%)		 With Suicidal Behavior(n = 71)
No. (%)		 Without Suicidal Behavior
(n = 126)
No. (%)
Adjustment disorder 9 (20) 11 (16) 12 (17) 8 (6)
Affective disorder** 26 (56) 20 (29) 21 (30) 14 (11)
Psychotic disorder‡ 5 (11) 19 (28) 10 (14) 52 (41)
Substance abuse disorder 3 (6) 3 (4) 15 (21) 17 (13)
Other or no disorder 3 (6) 15 (22) 13 (18) 35 (28)

* Chi2 = 13.4, df = 4, P = .01.
† Chi2 = 27.7, df = 4, P < .001.
** This category includes major depression, dysthymia, and bipolar disorder.
‡ This category includes schizophrenia, schizoaffective disorders, schizophreniform disorders, and atypical psychoses.

Table 3. Toxicologic Results in 155 Psychiatric Emergency Room Patients With and Without Suicidal Behavior, by Sex

  Females (n = 45) Males (n = 110)
With Suicidal Behavior
(n = 17)
No. (%)		 Without Suicidal Behavior
(n = 28)
No. (%)		 With Suicidal Behavior
(n = 42)
No. (%)		 Without Suicidal Behavior
(n = 68)
No. (%)
Any drug 4 (23) 5 (18) 23* (55) 17* (25)
Alcohol 3 (18) 1 (4) 17† (40) 6† (9)
Cocaine 2 (12) 2 (7) 8 (19) 6 (9)

* Chi2 = 9.9, df = 1, P = .002.
† Chi2 = 15.7, df = 1, P < .001.

Table 4. Comparison of Sociodemographic Variables* and Diagnostic Information in Suicidal Psychiatric Emergency Room Patients (n = 117) Who Were or Were Not Admitted to the Hospital

  Suicidal Patients Who Were Admitted
(n = 48)
No. (%)		 Suicidal Patients Who
Were Not Admitted
(n = 69)
No. (%)
Age†
   <30 21(44) 45 (65)
   30-44 16 (33) 19 (28)
   45-59 6 (13) 4 (6)
   >/=60 5 (10) 1 (1)
Sex
   Male 29 (60) 42 (61)
   Female 19 (40) 27 (39)
Race
   White 42 (87) 58 (84)
   Black 6 (13) 11 (16)
Primary diagnosis**
   Adjustment disorder 2 (4) 19 (28)
   Affective disorder‡ 29 (60) 18 (26)
   Psychotic disorder§ 8 (17) 7 (10)
   Substance abuse disorder 5 (10) 13 (19)
   Other or no diagnosis 4 (8) 12 (17)

* Marital status, employment status, and history of suicide attempts were not associated with incidence of hospital admission.
† P < .05.
** P < .001.
‡ This category includes major depression, dysthymia, and bipolar disorder.
§ This category includes schizophrenia, schizoaffective disorders, schizophreniform disorders, and atypical psychoses.

References

  1. Bristol JH, Giller E, Docherty JP: Trends in emergency psychiatry in the last two decades. Am J Psychiatr y 1981; 138:623-628
  2. Hillard JR, Ramm D, Zung WWK, et al: Suicide in a psychiatric emergency room population. Am J Psychiatry 1983; 140:459-462
  3. Dhossche D, Rubinstein J: Drug detection in a suburban psychiatric emergency room. Ann Clin Psychiatry 1996; 8:59-69
  4. Malone KM, Szanto K, Corbitt EM, et al: Clinical assessment versus research methods in the assessment of suicidal behavior. Am J Psychiatry 1995; 152:1601-1607
  5. Szuster R, Schanbacher B, McCann S: Characteristics of psychiatric emergency room patients with alcohol or drug-induced disorders. Hosp Community Psychiatry 1990; 41:1342-1345
  6. Breslow RE, Klinger BI, Erickson BJ: Acute intoxication and substance abuse among patients presenting to a psychiatric emergency room. Gen Hosp Psychiatry 1996; 18:183-191
  7. Meehan PJ, Lamb JA, Saltzman LE, et al: Attempted suicide among young adults: progress toward a meaningful estimate of prevalence. Am J Psychiatry 1992; 149:41-44
  8. Safer DJ: Self-reported suicide attempts by adolescents. Ann Clin Psychiatry 1997; 9:263-269
  9. Hawton K, Catalan J: Attempted Suicide: A Practical Guide to Its Nature and Management. Oxford, Oxford University Press, 1987, pp 3-25
  10. Murphy GE: Why women are less likely than men to commit suicide. Compr Psychiatry 1998; 39:165-175
  11. Marzuk P, Tardiff LK, Portera L, et al: The prevalence of aborted suicide attempts among psychiatric in-patients. Acta Psychiatr Scand 1997; 96:492-496
  12. Elliott AJ, Pages KP, Russo J, et al: A profile of medically serious suicide attempts. J Clin Psychiatry 1996; 57:567-571
  13. Beck AT, Brown GK, Steer RA, et al: Suicide ideation at its worst point:a predictor of eventual suicide in psychiatric outpatients. Suicide Life Threat Behav 1999; 29:1-9
  14. Dhossche D, Ghani SO: A study on recidivism in the psychiatric emergency room. Ann Clin Psychiatry 1998; 10:59-67
  15. Huffine CL, Craig TJ: Social factors in the utilization of an urban psychiatric emergency service. Arch Gen Psychiatry 1974; 30:249-255
  16. Asnis GM, Friedman TA, Sanderson WC, et al: Suicidal behavior in adult psychiatric outpatients. I: Description and prevalence. Am J Psychiatry 1993; 150:108-112
  17. Zisook S, Goff A, Sledge P, et al: Reported suicidal behavior and current suicidal ideation in a psychiatric outpatient clinic. Ann Clin Psychiatry 1994; 6:27-31
  18. Weissman MM: The epidemiology of suicide attempts, 1960 to 1971. Arch Gen Psychiatry 1994; 30:737-746
  19. Rich CL, Young D, Fowler RC: San Diego Suicide Study: I. Young vs old subjects. Arch Gen Psychiatry 1986; 43:577-582
  20. van Heeringen C, Jannes C: Recent changes in the age- and gender-specific rates of attempted suicide in Gent. Soc Psychiatry Psychiatr Epidemiol 1993; 28:66-70
  21. Petronis KR, Samuels JF, Moscicki EK, et al: An epidemiologic investigation of potential risk factors for suicide attempts. Soc Psychiatry Psychiatr Epidemiol 1990; 25:193-199
  22. Pederson AM, Awad GA, Kindler AR: Epidemiological differences between white and nonwhite suicide attempters. Am J Psychiatry 1993; 130:1071-1076
  23. Vital Statistics of the United States, 1992. Vol 2, Part A. Washington, DC, US Department of Health and Human Services, 1992
  24. Cornelius JR, Salloum IM, Mezzich J, et al: Disproportionate suicidality in patients with comorbid major depression and alcoholism. Am J Psychiatry 1995; 152:358-364
  25. Felts MW, Chenier T, Barnes R: Drug use and suicide ideation and behavior among North Carolina Public School students. Am J Public Health 1992; 82:870-872
  26. Pages KP, Russo JE, Roy-Byrne PP, et al: Determinants of suicidal behavior: the role of substance use disorders. J Clin Psychiatry 1997; 58:510-515
  27. Brady KT, Grice DE, Dustan L, et al: Gender differences in substance abuse disorders. Am J Psychiatry 1993; 150:1707-1711
  28. Dhossche D: Excess suicidality in male psychiatric emergency room patients with positive toxicology for cocaine (Abstract). 7th European Symposium on Suicide and Suicidal Behavior. Programme and Abstracts 1998; Abstract 0-105
  29. Chang L, Ernst T, Strickland T, et al: Gender effects on persistent cerebral metabolite changes in the frontal lobes of abstinent cocaine users. Am J Psychiatry 1999; 156:716-722