Literature Review for PhD doctorate

International Journal of Law and Psychiatry 28 (2005) 207–221

Factors related to suicide in New York state prisons

Bruce B. Way a,T, Richard Miragliab, Donald A. Sawyera, Richard Beerc, John Eddya

a Central New York Psychiatric Center (CNYPC), Box 300, Marcy, NY 13034, United States

b OMH Bureau of Forensic Services (BFS), United States

c Bureau of Forensic Services (BFS), United States

Received 7 October 2003; received in revised form 26 July 2004; accepted 22 September 2004

Abstract

Objective: Examine factors related to prison suicides to aid prevention.

Method: Review the mental health records of all 76 suicides that occurred between 1993 and 2001 in New York

State Department of Correctional Services (NYSDOCS) prisons that had some contact with mental health services

during their incarceration. (This represented 84% of all NYSDOCS suicides.) Extract data from the psychological

autopsies for a sample of 40 of these suicides.

Results: Of the suicide victims with some mental health contact, 95% had a substance abuse history, 70%

displayed agitation or anxiety prior to the suicide, and 48% had a behavioral change. Common stressors preceding

the suicide were inmate-to-inmate conflict (50%), recent disciplinary action (42%), fear (40%), physical illness

(42%), and adverse information (65%) such as loss of good time or disruption of family/friendship relationships in

the community. Forty-one percent had received a mental health service within 3 days of the suicide. Compared to

the about 7200 inmates actively receiving mental health services in state prison, African-Americans and patients

with a Major Mood (Bi-polar or Major Depression) were under-represented. Adjustment Disorder, Schizophrenia,

and Personality Disorder diagnoses were over-represented. Suicide victims were more likely to have been

incarcerated for a violent crime.

Conclusion: Mental illness, anxiety/agitation, behavior change, stressors, history of substance abuse, and non-

African-American were important risk factors.

D 2005 Elsevier Inc. All rights reserved.

0160-2527/$ -

doi:10.1016/j.i

T Correspond E-mail add

see front matter D 2005 Elsevier Inc. All rights reserved.

jlp.2004.09.003

ing author. Tel.: +1 315 765 3825/736 8271.

ress: [email protected] (B.B. Way).

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221208

1. Background

Suicides by inmates in prison are distressful events. They become a center of attention for the treating

clinicians, the psychologists who conduct psychological autopsies, security staff, supervisors, oversight

agencies, prisoner advocacy groups, the media, and other inmates. They are a major trigger for lawsuits

(Metzner, 2002). The World Health Organization (WHO, 2000) issued a report on Preventing Suicide: A

Resource for Prison Officers, which begins with bjails and prisons are responsible for protecting the health and safety of their inmate populations, and the failure to do so, can be open to legal challenge.

Further, fueled by media interest, a suicide in a correctional facility can easily escalate into a political

scandal... It is within this context that correctional settings worldwide struggle with the problem of

preventing inmate suicide.Q (p. 5). Hayes (1995) in Prison Suicide: An Overview and Guide to Prevention, issued by the US Department of Justice, noted that suicide among un-sentenced or

remanded-for-trial inmate populations has received considerable research attention around the world, but

few studies have focused on longer-term convicted inmates. However, such research is greatly needed

since bstudies show that pretrial inmates differ from sentenced prisoners with respect to certain key risk factors for suicideQ (WHO, 2000, p. 6) and bprecipitating factors in suicidal behavior among prison inmates are somewhat unique and fester over time.Q (Hayes, 1995, p. 2).

Hayes notes that many correctional suicide studies offer limited insight because of their bexclusive focus on prison suicide rates,... absence of discussion regarding precipitating factors, and failure to

differentiate between prisons and jail suicidesQ (Hayes, 1995, p. 2). Several other authors (Fruehwald, Frottier, Eher, Gutierrez, & Ritter, 2000; Haycock, 1993; Morrison, 1996) have also noted the failure to

account for diverse correctional populations within and between studies.

In the United States, it is somewhat easier to distinguish incarcerated populations. Longer-term

sentenced inmates are mostly housed in bState PrisonsQ whereas bLocal JailsQ serve remand and short term sentenced inmates. In many other countries, however, this distinction does not exist, and the term

bprisonQ includes both remand and long-term sentenced inmates. For example, studies of prison suicides in Ireland (Dooley, 1997), England and Wales (Dooley, 1990), Greece (Spinellis & Themeli, 1997),

Finland (Joukamaa, 1997), Italy (Tatarelli, Mancinelli, Taggi, & Polidori, 1999), Austria (Fruehwald et

al., 2000), and Australia (McDonald & Thompson, 1993) include un-sentenced and sentenced prisoners,

and most do not distinguish these groups in their analyses.

Many prison studies focus on calculating and comparing suicide rates. These have less value,

however, as the rates are often calculated incorrectly, and it is difficult to establish an appropriate

comparison group. Studies use the average correctional facility daily census as the denominator,

instead of the more appropriate of ball offenders who were incarceratedQ in the time period. A leading expert in Correctional Litigation, Dr. Metzner (2002), has demonstrated how the use of average daily

census as opposed to ball at riskQ produces incorrect rates and comparisons. This is especially true when comparing institutions that serve bremandQ populations with short lengths of incarceration with bsentencedQ environments with much longer average lengths of incarceration. The correct bat riskQ denominator is disproportionately much greater than the daily census for remand than for sentenced

rates.

Studies often compare incarcerated suicide rates with free-community rates, but according to Metzner

(2002) this is methodically incorrect. Instead, the community reference group should be adjusted to

match the suicide vulnerability profile of the inmates. Examples of attempts for better community

reference groups have been: Opiate Drug Users (Gore, 1999) and different ethnic groups (McDonald &

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221 209

Thompson, 1993). Gore states: bstatistical approaches to estimating the expected number of suicides in prison should account for the high prevalence of drug dependence among inmatesQ (p. 50).

Controlling for all the variables that are related to suicide in order to find the right community reference

group is difficult and may be impossible. Not only is a community referent difficult, Kennedy and Hormant

(1988) argue that even comparing correctional institutions is problematic since each has different inmate

characteristics and physical plants. Haycock (1993) supports this idea by discussing that bconfinement is not everywhere equally suicidogenic...Q (p. 129). Therefore, perhaps we should give up our rate comparison exercises, and instead focus on stressors and risk factors as recommended by Hayes (1995).

2. Purpose of study

The objective of the current study is to explore stressors and risk factors associated with suicide

among sentenced prison inmates as recommended by Hayes (1995).

3. Method

The New York State Department of Correctional Services (NYSDOCS, 2002a) operates 71 prisons. It

had about 67,000 inmates under-custody on January 1, 2002 and receives about 25,000 admissions each

year. NYSDOCS serves only inmates with sentences of 1 year or more. Inmates are not new to custody

having been incarcerated for many months in a local jail before transfer to State prison. Hill, Smith,

Sawyer, and Miraglia (2002) reports an average of 10 months in the New York City jail prior to transfer

to state custody.

In NYSDOCS facilities, all mental health services are provided under the auspices of the NYS Office

of Mental Health’s Central NY Psychiatric Center (CNYPC). About 7200 inmates were actively

receiving mental health services on 1/1/02, and there are about 4500 admissions annually.

Data was extracted from the mental health treatment charts of all completed suicides between 1993

and 2001 (n =76) that had contact with Mental Health services at anytime during their current prison

commitment. The 76 comprised 84% of all suicides that occurred in NYSDOCS in the period (2002b).

In addition to the charts, data was extracted from a random sample of 40 psychological autopsy reports.

A sample of 40 of the 76 was considered of adequate size. After the data was extracted client identifiers

were removed. CNYPC staff routinely conduct psychological autopsies for suicides. Current policy

requires that a NYS licensed psychologist (Ph.D.) complete the autopsy.

Upon entry into the New York State prison the inmate spends a few weeks in a breception prisonQ during which all inmates are screened for mental illness. This includes personal contact, review of record

information, and searches of several databases. If the inmate is admitted to mental health services a

psychiatrist interviews the patient, reviews the charts, and creates a psychiatric diagnosis. Psychiatrists

are encouraged to review and clarify diagnoses during treatment, and diagnoses may be modified at any

time due to subsequent assessment or a change in the patient’s mental status. Diagnoses are recorded in

the patientTs mental health chart, as well as in an automated database. The database also contains demographic variables.

Comparisons are made to the patients on the MH Caseload on 1/1/02 and all the NYSDOCS inmates

in custody on 1/1/02. While all the inmates in NYSDOCS and all the inmates on the Mental Health

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221210

caseload during this nine-year period would be preferable comparison groups, reports on these groups

were not available.

For many of the comparisons between the patients that committed suicide and the MH caseload or the

NYSDOCS inmates, dichotomous 2�2 Chi-square tests were conducted. These all had one degree of freedom. These include: White vs. non-white, African-American vs. non-African-American, Hispanic

vs. non-Hispanic, each psychiatric diagnosis, violent vs. non-violent commitment crime, murder/

manslaughter vs. non-murder/manslaughter, and occurred in a maximum security prison vs. non-

maximum security prison. Comparisons of group means used the T-test. These tests would have N �1 degrees of freedom for the smaller of the two groups. P values less than .05 were considered statistically

significant. P values of less than .01, and less than .001 are also listed on the charts.

4. Results

Table 1 displays the characteristics of the 76 suicides that had contact with mental health services

in the time period, and the two comparison groups. The suicide group was significantly younger

Table 1

Characteristics of suicides in NYS prison that had any contact with mental health services 1993–2001

Suicides

N =76 (%)

MH caseload

N =7236 (%)

NYSDOCS population

N =69,152 (%)

Age

15–19 6.6 2.2 1.4

20–24 9.2 11.1 16.3

25–29 21.1 11.9 16.6

30–34 26.3 16.8 18.2

35–39 15.8 19.9 18.3

40–44 11.8 17.8 13.8

45–49 6.6 10.4 7.6

50+ 2.6 9.9 7.8

Total 100.0 100.0 100.0

Age mean 32.8 37.1*** 34.6*

Gender

Female 4.0 14.0* 4.0

Ethnicity

White 36.8 29.4 17.5***

Black 23.7 44.6*** 50.8***

Hispanic 36.8 23.9 29.9

Other 5.3 2.1 1.8

Total 100.0 100.0 100.0

Primary or secondary diagnosis

Bi-polar or major depression 9.2 21.2*

Schizophrenia 38.2 21.7***

Adjustment 27.6 6.6***

Impulse 2.6 4.2

Personality 39.5 22.6***

Substance 18.4 24.8

Comparisons with suicide victims *p b.05, **p b.01, ***p b.001.

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221 211

than the total mental health group (t =4.3, df =75, p b.001) and the total prison population. (t =1.8,

df =75, p b.5) Three (4%) of the 76 suicide victims were female, which is significantly smaller than

the MH caseload 14% (x 2 =6.4, df =1, p b.05), but the same as the NYSDOCS percentage. African-

Americans were significantly less likely to commit suicide than non-African-Americans on the

mental health caseload (x 2 =13.3, df =1, p b.001) or in the total NYSDOCS population (x

2 =22.3,

df =1, p b.001). White suicide victims were not significantly over-represented compared to other

mental health recipients, but were over-represented compared to total prison population. (x 2 =19.6,

df =1, p b.001).

The primary or secondary diagnoses of Major Mood (Bi-polar or Major Depression) were

significantly under-represented among the suicide victims compared to the MH caseload (x 2 =6.5,

df =1, p b.05), while Schizophrenia (x 2 =11.9, df =1, p b.001), Adjustment Disorder (x

2 =52.4, df =1,

p b.001) and Personality Disorder (x 2 =12.2, df =1, p b.001) were significantly over-represented.

Nearly 3/4 of those committing suicide were currently receiving mental health treatment at the time

of the suicide (see Table 2), about 1/2 were on psychiatric medications, and 41% were seen by MH staff

within 3 days of the suicide.

As seen in Table 3, hanging was the most common method (86%) while 8% were due to a drug

overdose. Twenty-one percent of the suicides were discovered between 6 am and 9 am as the

correctional officer conducted his/her rounds. Sixteen percent left a suicide note, and 34% made suicidal

statements. Fifty-two percent of the suicides occurred in the general inmate population, 23% in

disciplinary housing, 9% in Intermediate Care Program Cell (long-term housing for mentally ill

inmates), 1% in a mental health crisis observation cell, 7% in administrative segregated housing for

protective custody, and 8% in other housing units.

Seventy-one percent of the suicide victims were incarcerated for violent offenses or crimes against

persons (see Table 4). Crime data for the total MH caseload is not available, but 71% (x 2 =6.1,

df =1, p b.05) is significantly higher than the overall NYSDOCS (2002a) distribution. Twenty-seven

Table 2

Mental health characteristics of suicides

Number

Currently receiving mental health services 73.7

Psychiatric medications

Yes 47.4

Refusing 11.8

No 40.8

Total 100.0

Time since last MH service

1–3 days 40.9

4–7 days 15.2

1–2 weeks 9.1

2–4 weeks 9.1

1–3 months 6.1

2–6 months 7.6

6–12 months 4.5

12 months plus 7.6

Total 100.0

Table 3

Suicide characteristics

Percent

Suicide method

Suffocation 2.8

Overdose 8.3

Laceration 2.8

Hanging 86.1

Total 100.0

Time of discovery

12–3 am 17.1

3–6 am 11.4

6–9 am 21.4

9 am–12 pm 8.6

12–3 pm 12.9

3–6 pm 8.6

6–9 pm 12.9

9–12 am 8.6

Total 100.0

Suicide note left 15.8

Suicide ideation 34.2

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221212

percent of the suicides were incarcerated for killing another person (Murder/Manslaughter) as

compared to 18% in the whole NYSDOCS (2002a) population. This difference approached statistical

significance (x 2 =3.44, df =1, p =.07). All 3 female suicide victims were committed for drug

offenses. Thirteen percent had a minimum sentence of 20 or more years, which is only slightly

higher than the 11% overall prison population. Controlling for prison type by examining only

Table 4

Commitment crime

Suicide victims

N =76 (%)

Total NYSDOCS

N =67,332 (%)

Attempted murder 1.3 2.7

Murder 17.3 10.6

Manslaughter 8.0 5.1

Rape/sodomy 12.0 6.0

Assault 13.3 5.1

Robbery 10.7 23.1

Kidnap 4.0 .7

Arson 1.3 .3

Poss. weapon 2.7 3.0

Violent subtotal 70.6T 56.6 Property (e.g. burglary) 12.0 13.6

Drugs 17.3 28.7

Other 0.0 1.1

Total 100.0 100.0

T p b.05.

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221 213

inmates in the maximum-security prisons also produced no significant difference on minimum

sentence length.

Thirty-six percent of the NYSDOCS under custody population was confined to a maximum-security

facility, while 83% of the suicides occurred in these settings. (x 2 =118.5, df =1, p b.001) Cells in

maximum-security prisons are single cells, whereas medium and minimum-security prisons have

dormitories and double cells.

4.1. Data extracted from the psychological autopsies

For this sample of 40 of the 76, 55% had a history of inpatient psychiatric treatment in the community

prior to incarceration, 52% had prior suicide attempts, and 95% had a substance abuse history. Seventy

percent were experiencing anxiety or agitation, and 48% displayed behavior dysfunction or change in

usual behavior.

As for precipitating stressors, 42% of the suicide victims had a recent disciplinary action, 42% had

physical illness, 50% were involved in inmate-to-inmate conflict, and 40% had fear of physical harm.

Sixty-five percent of the inmates had just received badverse information.Q

5. Discussion

We will compare our percentages only with studies that have focused on sentenced prison inmates in

an attempt to remedy a criticism in the literature. Salive, Smith, and Brewer (1989) studied all (n =37)

inmates deaths classified as suicide in Maryland state prison between 1979 and 1987, and compared

them with MarylandTs general population. Anno (1985) examined 38 inmate suicides that occurred in Texas prisons between 1980 and 1985. He, Felthous, Holzer, Nathan, and Veasey (2001) reviewed all

suicides in 20 of the 107 prison units in Texas in 1996–1997 (n =25). Dooley (1990) examined 295

suicides that occurred in England and Wales prisons between 1972 and 1987 of which 140 were

sentenced inmates. Fruehwald et al. (2000) studied 220 prison suicides in Austria between 1975 and

1997 of which 97 were sentenced. Anno (1985), Dooley (1990), He et al. (2001), and Fruehwald et al.

(2000) compared suicide victims with other sentenced inmates.

5.1. Demographics

We found that suicide victims were younger than other inmates, but Anno (1985) and He et al.

(2001) found no age difference in comparison with other Texas prison inmates. We found that

African-Americans were under-represented compared to other ethnic groups receiving mental health

services, and compared to the overall prison population. He et al. (2001) also found African-

American were under represented among suicide victims. Anno (1985) found that white and

Hispanic inmates were over-represented. Female suicides were under represented in our study

compared to those receiving mental health services whereas, the percentage was similar to the

NYSDOCS proportion. He et al. (2001) also found that female suicides were similar in proportion to

the prison population.

The general literature on suicide supports our ethnicity results, but reports that males are

disproportionately more likely to commit suicide than females (Appleby, 1992; Earle, Forquer, Volo,

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221214

& McDonnell, 1994). It is unknown why sentenced women inmates were as likely as men to commit

suicide in prison, but there has been a recent upswing in research on female inmates that could reveal

reasons for this difference. Regarding age, some studies report younger ages (Appleby, 1992; Earle et al.,

1994) are more at risk for suicide, but others (Busch, Fawcett, & Jacobs, 2003) report older ages. It is

unknown why our age result differed from other prison studies.

5.2. Mental illness

We found that 84% of all suicides in NYSDOCS had been on the active mental health caseload at

some point during their incarceration, and 74% of these 76 inmates were currently receiving services.

While our 84% is higher than other sentenced inmate studies (He et al., 2001—76%, Anno, 1985—

58%) the relationship between suicide and mental illness is well established. Fawcett (2001), discussing

suicide in the free community, reports that 90% of suicides have a mental illness.

5.3. Diagnosis

Only one sentenced prison study was found that reported diagnosis of suicide victims (He et al.,

2001). In terms of how a diagnosis is created they only state, bthat the screening, diagnostic, and subsequent evaluation procedures are rather thorough.Q (p. 900). We also are very thorough in our evaluation and diagnostic process as noted above, and it is unknown why some of our diagnostic results

differ with He et al. (2001). Further research is needed to clarify.

5.4. Schizophrenia/psychotic disorder

He et al. (2001) reported that 44% of the suicides had a psychotic diagnosis. This is similar to our

38% figure. The relationship between psychosis and suicide is well established (Appleby, 1992; Roy,

1982). It may be symptoms of bactive psychosisQ that are more important, however, rather than a Psychosis diagnosis. This could explain why Earle et al. (1994) found in a study of over 16,000

psychiatric outpatients in state operated programs that significantly fewer suicide victims (32%) had a

primary diagnosis of schizophrenia than all outpatients (47%). Since most of their patients with

Schizophrenia had been released from inpatient care, it is more likely they were medication compliant

and did not have had active psychosis recently. Investigation of the relationship between recent

psychotic symptoms and prison suicide, not just diagnosis, is warranted.

5.5. Major depression or bi-polar

Only 9% of the suicide victims in our study had a Major Depression or Bi-polar diagnosis.

Compared to our overall MH caseload, these diagnoses were under-represented. This is divergent with

He et al. (2001) who found that 64% of the suicides had a Mood disorder. We culled the four suicide

victims with a Dysthymia or Depressive NOS diagnosis, but including these does not explain the large

difference. Our result is also different than much of the general suicide literature (Earle et al., 1994;

Morgan & Priest, 1991; Roy, 1982) that has found an association between depressive illness and

suicide. A possible explanation may have to do with prior violence. Apter et al. (1991) compared

samples of violent and nonviolent psychiatric inpatients that had attempted suicide. They found that

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221 215

the two groups were similar on most suicide correlates, but for the violent group there was no

correlation between depression and suicide. Depression was correlated with suicide attempts for non-

violent inpatients. Further, Ramsay, Gray, and White (2001) in a review of suicides in a maximum-

security hospital reported that none of the 14 inmates that committed suicide had clear depressive

episodes. Presumably many of the 14 inmates were in the maximum-security hospital due to violence.

Perhaps our finding of so few Major Mood disorders is related to the large percentage (71%) of

suicide victims who were incarcerated for a violent crime. Our findings, together with the findings of

Apter et al. (1991), suggest that a Major Mood disorder may be a less important suicide risk factor for

violent offenders.

Often associated with depression is the concept of hopelessness, which in the general literature has

a key relationship with suicide (Appleby, 1992; Beck, Steer, Kovacs, & Garrison, 1985; Hayes, 1995).

We did not make an assessment of hopelessness, but future studies of prison suicide should include

this.

5.6. Adjustment disorder

Twenty-eight percent of the suicide victims in our study had an Adjustment Disorder diagnosis and

this diagnosis was over-represented compared to other mental health recipients. Our percentage is similar

to the 20% reported by He et al. (2001). Adjustment disorder is not reported as an important risk factor in

the general suicide literature. Within prison environments, however, it may be more important. It is

probably linked with the frequency of stressors found in the suicide victim charts, which are discussed

below.

5.7. Personality disorder

Almost 40% of the suicide victims in our study had a Personality Disorder diagnosis; it was over-

represented compared to the mental health caseload. He et al. (2001) reported that 56% of the suicides

had this diagnosis. Anti-social Personality Disorder is much more common in prison settings than in

other environments (Rotter, Way, Steinbacher, Sawyer & Smith, 2002), and it is linked with violence,

impulsiveness, and manipulative behavior. It is suggested that these individuals attempt suicide for gain,

but it is clear that things can go wrong, and any gesture or threat should not be considered lightly (Hayes,

1995). We did not measure impulsivity in our study, but it is common in Personality Disorder, and

Fawcett (2001) identified impulsivity as an important risk factor with suicide. Future research should

include this variable.

5.8. Substance abuse

Substance abuse is a major risk factor for suicide in the general suicide literature (Appleby, 1992;

Earle et al., 1994) and our clinicians found that 95% of the suicide victims had this history in their

records. Although our result is substantially higher than the 68% found by He et al. (2001), history of

substance abuse should be considered a major risk factor.

Only a small percentage of our suicide victims, however, had a substance abuse diagnosis. This could

be an artifact of our automated database that permits only two diagnoses, or more probably it is

substantially under diagnosed.

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221216

5.9. Anxiety/agitation

We report that the majority (70%) of the suicide victims were agitated or anxious, and one-half had a

change in behavior preceding the suicide. No prior sentenced prisoner study has documented this issue,

although Anno (1985) anecdotally reported bizarre behavior prior to suicide.

While our finding is new in the prison suicide research, Fawcett (2001) wrote, brecent research has shown that impulsiveness and severe anxiety, panic attacks, and agitation co-morbid with depression are

often immediate suicide risk factors...Q (p. 94). Further, recently Busch et al. (2003) found that 79% of inpatient suicides had severe or extreme anxiety and/or agitation as measured by SADS. Due to its high

prevalence, when screening for suicide risk, clinicians should carefully assess high levels of agitation

and anxiety, and consider recent behavioral changes.

5.10. Suicide ideation

We found that only 34% expressed suicide ideation while He et al. (2001) reported a much higher 72%.

Fawcett (2001) comments, bresearch shows that suicidal ideation is often not communicated to professionals or is denied by patients just prior to suicide and, when present, it is often useful not as an

immediate risk factor, but as a chronic factorQ (p. 94). Robins (1981) reported that 69% of the suicides in their study expressed ideation to family, friends, or co-workers, but only 18% told a helping professional.

Earle et al. (1994) reported that 73%, and Busch et al. (2003) found that 77% denied ideation at last contact.

The general suicide literature strongly suggests that we should not trust a denial of suicide ideation to

our clinician. In our assessment of suicide we should contact inmates in neighboring cells, correction

officers on the housing unit, and visiting family members, and then have them alert mental health staff if

changes occur and they overhear an inmate discussing suicide.

5.11. Prior suicide attempts

In our study, 52% had a prior suicide attempt. This percentage is between the 64% found by He et al.

(2001) and the 45% reported by Anno (1985). While commonly associated in the general literature

(Appleby, 1992), Fawcett reports data from a NIMH study that showed prior attempts were no more

frequent among those that did and did not complete suicide within a year. Therefore, lack of prior

attempt behavior should not preclude a judgment of suicide risk.

5.12. Prison stressors

There were frequent significant stressors among the suicide victims—recent disciplinary action,

inmate-to-inmate conflict, fear of physical harm, physical illness, and recent adverse information. Where

comparable, our results are similar to He et al. (2001). They reported 48% and we found that 50% were

involved in inmate-to-inmate conflict. We report that 42% had a physical illness while He et al. (2001)

reported 52%. Also, Anno (1985) found that many suicide victims had just received bad news such as

the death of a family member. Earle et al. (1994) reported that 55% of suicide victims had a physical

health problem.

Most of the issues we focused on, however, were acute stresses. We did not examine or measure

cumulative stress over time. Ramsay et al. (2001) in a review of suicides in a maximum-security

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221 217

hospital, however, reported continued significant subjective stress over many years in the suicide

victims. Also, Hayes (1995) and Bonner (1992) suggest that stress in sentenced inmates grows and

becomes overwhelming. Stress over time should be considered for inclusion in future studies.

5.13. Commitment crime

We found that 71% of the suicide victims were committed for a violent crime, and that this was

significantly higher than the NYSDOCS under custody population. Our figure is higher than the 62%

reported by Salive et al. (1989), the 58% found by Anno (1985), and the 44% found by He et al. (2001).

Dooley (1990) reported that sentenced prisoners in England and Wales who committed suicide were

more likely to be convicted of a violent or sexual offense, and 16% of suicide victims were murderers as

compared to 4% of the total sentenced population. We found that 27% were convicted of murder/

manslaughter, which approach statistical significance with a p value of .07.

5.14. Long sentences

We report that suicide victims did not have longer minimum sentences than other inmates; this is

divergent with other studies. Salive et al. (1989) reported that a larger percentage (24%) of the suicide

victims were serving life sentences. Fruehwald et al. (2000) found that among sentenced prisoners in

Austria, 35% of suicides had a sentence of 5 or more years as compared to 17.5% overall of the

sentenced inmates. Dooley (1990) found in England and Wales that a disproportionate number of suicide

victim were serving long sentences (4 years or more) and 25% were life sentences. It is unknown why

we did not find this relationship. Longer or a life sentence could be a chronic stressor and merits

continued investigation.

5.15. Method

Eighty-six percent of the suicides were hangings, which is similar to He et al. (2001)—76%, Salive et

al. (1989)—86%, and Anno (1985)—89% findings. Only 6 of the victims in our study overdosed

although 36 were on psychiatric medications at the time of the suicide. This suggests that our nursing

staff prevented the accumulation or bcheekingQ of un-taken medications, and/or overdose is not the preferred method. Overdose has a higher probability of resuscitation. Hanging is the common method in

institutional settings. Perez-Carceles, Inigo, Luna, and Osuna (2001) reported that 94% of 34 suicides in

a prison hospital in Spain were hangings.

5.16. Location–isolation and single cells

There has been considerable controversy over the use of isolation cells for suicidal inmates. Hayes

(1995) in his US Department of Justice report states: ba primary recommendation, based chiefly on overwhelming consistent research, is that isolation should be avoided whenever possibleQ (p. 7). Bonner (1992) also recommends against isolation, and Liebling and Hall (1993) state that placing potentially

suicidal inmates in seclusion in prison strip cells is ba practice to be ashamed ofQ. Felthous (1997), however, discussed that the literature does not distinguish between two very different environments of

isolation. Isolation with limited social contact/deprivation versus isolation with close observation by

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221218

staff. bThe problem is not so much the use of single cells or separating vulnerable inmates; the lethal risk comes from leaving suicidal inmates unattended and with materials that can be used for self-destruction.Q (p. 289).

We found that suicide victims were over-represented in maximum-security prisons. In New York

State all inmates in maximum-security prisons are in single cells, whereas inmates in medium and

minimum prisons are housed in double cells or dormitories. Anno (1985) reported that 97% and He et

al. (2001) reported that 76% of suicides occurred in single cells. In psychiatric hospitals, suicides

almost always occur when patient is alone. Perez-Carceles et al. (2001) found that in a maximum-

security psychiatric hospital 94% of the suicide victims were alone. Similarly, in the free community

living alone is often associated with increased suicide risk (Appleby, 1992). Clearly a single cell is a

suicide risk.

The majority of suicides in our study, while in a single cell, did not occur in an bisolationQ cell. Fifty- two percent of the victims lived in General Population, and 9% lived in Intermediate Care Programs. In

these housing units inmates participate in many activities such as recreation, school, work, mental health

programming, substance abuse programming and religious services.

We did find that 23% of the suicides occurred in Special Housing Units (disciplinary confinement

cells) in which inmates are in their cells 23 h/day. While we donTt have total number of inmates in this type of cell over the 9 years, based on census, SHU has only 4% of the NYSDOCS inmates. This

disproportion indicates that SHU housing is also a suicide risk. However, inmates identified as suicidal

are not placed in SHU cells. They are placed there due to serious disciplinary infractions such as an

assault on a correctional officer. Inmates identified as suicidal, regardless of disciplinary infractions, are

immediately transferred to Mental Heath Observation Cells as discussed below. For all inmates in SHU,

mental health therapists provide cell side visits 5 days a week. For those SHU inmates on the active

mental health caseload, therapists provide at least 2 treatment sessions a month in private out of the SHU

cell space. Our psychiatrist also has a private session with the patient each month. A nurse visits SHU

patients receiving medications once or twice a day. We have also established joint committees with

NYSDOCS at the maximum-security prisons to bi-weekly review new SHU admissions, as well as

continually monitor mental health recipients. Depending on clinical needs, the committee may move a

mentally ill individual from SHU cells to alternative housing.

5.17. Continuous observation—mental health observation cells

When an inmate, regardless of current housing unit, makes a suicide threat or gesture she/he is

transferred to a Mental Health Observation Cell. In our study, only one suicide occurred in the 9 years in

the 56 MH crisis observation cells operated by CNYPC. This is an extremely small percentage of 3000

annual admissions to these cells (Kahkejian, Way, & Nash, 2004). In these cells, mats replace blankets,

and the inmate is under constant observation. These stays in observation cells are very short in duration

with a median of 4 days. Those having a continuing need for mental health service are sent to the

inpatient hospital.

While constant observation is provided in our observation cells, the literature reports that 15- or 30-

min checks are not enough. Busch et al. (2003) found that 42% of the suicides in an inpatient psychiatric

hospital occurred when the patients were on 15-min checks and another 20% when on 30-min checks.

Even under constant observation, vigilance is important. Kennedy, Whittington, and White (1995) report

two suicides that were committed beneath blankets while staff continuously watched.

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221 219

5.18. Recently seen by mental health services

We found that 41% had a clinical service within 3 days of the suicide, and another 15% in 4 to 7 days.

These cases were not detected as having a high suicide risk and, therefore, not sent to a Mental Health

Observation cell. A recent clinical contact just prior to a suicide is not an uncommon finding in free-

community psychiatric practice. Earle et al. (1994) found that 84% of the suicide victims in mental

health outpatient programs were seen within 2 weeks by their therapist. Roy (1982) reported that 90% of

the suicides among psychiatric outpatients were within 1-month of their last outpatient appointment.

While our mental health staff are seeing many of the suicide victims, we are not detecting their high

risk. We provide over 1/2 million clinical contacts a year, and distinguishing these inmates is difficult.

Morgan and Priest (1991) even in a psychiatric hospital setting found that bin the majority the seriousness of the suicide risk was either not recognized or precautions to prevent it failed.Q (p. 373). Fawcett (2001) comments: bit is clear, even from the limited data we have, that this acute high risk state is often not noted or detected by treating clinicians...Q (p. 95). Along this same line, Hendin, Maltsberger, Lipschitz, Haas, and Kyle (2001) in a review of suicides receiving treatment in private practice comment:

bproblems in communication between patient and therapist, often originating in therapeutic anxiety over the patientTs possible suicide, were identified as factors interfering with crisis recognition.Q (p. 169).

Confounding suicide risk recognition, the patientTs clinical condition may be improving. Goh, Salmons, and Whittington (1989) reported that 40% of psychiatric inpatients were described by

clinicians as bimprovingQ, just before suicide took place. Morgan and Priest (1991) found that 81% of the suicides had misleading clinical improvement, 52% with significant improvement.

Increasing our ability to detect high risk is difficult, however. Many different instruments do exist

with varying attributes (See a review of twenty different instruments by Range and Knott, 1997).

However, Fawcett (2001) and recently Rudd (2003) point towards the need for instruments to improve

by focusing on the acute or proximal risk factors as oppose to the more chronic factors. Rudd (2003)

stated, bfurther study might reveal that we can translate many risk factors into meaningful and easy to understand warning signs for intervention and educational programs.Q (p. 99).

In addition, the WHO report (2000) suggests conversations with inmates at critical periods after

adverse information is shared as a way to detect high risk. We do contact inmates when parole is denied,

but contact in other situations could be provided.

Different than improving the detection of risk is to bhardenQ the environment to prevent suicide. This is one of our approaches in our inpatient hospital, to remove all physical features that could be used for

hanging, as well as use only dormitory housing. We have not had a suicide in the hospital since at least

1980 when a new automated data collection system was implemented.

Monitoring of patients in SHU units could be improved electronically with motion detectors to alert

staff when an inmate gets out of bed at night. Such motion detectors have been used in hospital settings.

Further, perhaps, some maximum-security inmates could be double bunked, or observation improved.

6. Summary and future work

In summary, the major risk factors we found are: 1) agitation or recent change of behavior, 2) prison

stressors, 3) history of substance abuse, 4) mental illness and/or mental health services, 5) Schizophrenia,

Personality Disorder, or Adjustment Disorder, 6) single cell, and 7) violent commitment crime.

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221220

In addition to the suggestions in the Discussion Section, future work should examine various sub-

populations of suicide victims such as Schizophrenia and Adjustment disorders. Also, comparisons with

inmates that do not complete suicides are needed. It would be important to know about how non-suicidal

inmates experience the various prison stressors discussed above. Studying the interaction between

stressors and various vulnerabilities such as mental illness is very important. Finally, incorporating long-

term sentenced prison suicides from many states and countries into a single data base with a standard

format would greatly enhance our understanding and prevention efforts.

Acknowledgements

The authors wish to thank the following: Bruce Bradigan, MA, John Culkin, MA, Larry Farago,

M.D., Joe Himmelsbach, Ph.D., Jurgen Karker, Ph.D., Beatrice Kovasznay, M.D., Joyce Richardson,

CSW, Robin Nash, Thomas Ryan, Ph.D., Al Shimkunas, Ph.D., and Hal Smith.

References

Anno, B. J. (1985). Patterns of suicide in the Texas department of corrections. 1980–1985. Journal of Prison and Jail Health, 5,

82–93.

Appleby, L. (1992). Suicide in psychiatric patients: Risk and prevention. British Journal of Psychiatry, 161, 749–758.

Apter, A., Kotler, M., Sevy, S., Plutchik, R., Brown, S. L., Foster, H., et al. (1991). Correlates of risk of suicide in violent and

nonviolent psychiatric patients. American Journal of Psychiatry, 148, 883–887.

Beck, A. T., Steer, R. A., Kovacs, M., & Garrison, B. (1985). Hopelessness and eventual suicide: A 10-year prospective study

of patients hospitalized with suicidal ideation. American Journal of Psychiatry, 142, 559–563.

Bonner, R. L. (1992). Isolation, seclusion, and psychological vulnerability as risk factor for suicide behind bars. In R. W. Maris,

A. L. Berman, J. T. Maltsberger, & R. I. Yufit (Eds.), Assessment and prediction of suicide (pp. 398–419). New York7

Guildford Press.

Busch, K. A., Fawcett, J., & Jacobs, D. G. (2003). Journal of Clinical Psychiatry, 64, 14–19.

Dooley, E. (1990). Prison suicide in England and Wales 1972–87. British Journal of Psychiatry, 156, 40–45.

Dooley, E. (1997). Prison suicide—politics and prevention: A view from Ireland. Crisis, 18, 185–189.

Earle, K. A., Forquer, S. L., Volo, A. M., & McDonnell, P. (1994). Characteristics of outpatient suicides. Hospital and

Community Psychiatry, 45, 123–126.

Fawcett, J. (2001). Treating impulsive and anxiety in the suicidal patient. Annals of the New York Academy of Sciences, 93,

94–105.

Felthous, A. R. (1997). Does bisolationQ cause jail suicides? Journal of American Academy of Psychiatry and Law, 25, 285–294.

Fruehwald, S., Frottier, P., Eher, R., Gutierrez, K., & Ritter, K. (2000). Prison suicides in Austria, 1975–1997. Suicide and Life-

Threating Behavior, 30, 360–369.

Goh, S. E., Salmons, P. H., & Whittington, R. M. (1989). Hospital suicides: Are they preventable? Profile of the psychiatric

hospital suicide. British Journal of Psychiatry, 154, 247–249.

Gore, S. M. (1999). Suicide in prisons: Reflections of the communities served or exacerbated risk? British Journal of

Psychiatry, 175, 50–55.

Haycock, J. (1993). Comparative suicide rates in different types of involuntary confinement. Medical Science and the Law, 33,

128–138.

Hayes, L. M. (1995). Prison suicide: An overview and guide to prevention. Washington DC7 US Justice Department, National

Institute of Corrections.

He, X., Felthous, A. R., Holzer, C. E., Nathan, P., & Veasey, S. (2001). Factors in prison suicide: One-year study in Texas.

Journal of Forensic Sciences, 46, 896–901.

B.B. Way et al. / International Journal of Law and Psychiatry 28 (2005) 207–221 221

Hendin, H., Maltsberger, J. T., Lipschitz, A., Haas, A. P., & Kyle, J. (2001). Recognizing and responding to a suicide crisis.

Annuals of New York Academy of Sciences, 932, 169–187.

Hill, M., Smith, H., Sawyer, D., & Miraglia, R. (2002). Forensic diagnostic unit: Reception and classification center at

downstate correctional facility. Marcy NY7 Central New York Psychiatric Center.

Joukamaa, M. (1997). Prison suicide in Finland, 1969–1992. Forensic Science International, 89, 167–174.

Kahkejian, D., Way, B. B., & Nash, R. (2004). Mental health crisis observation cell and dormitory bed usage. Marcy NY7

Central New York Psychiatric Center.

Kennedy, D. B., & Hormant, R. J. (1988). Predicting custodial suicides: Problems with the use of profiles. Justice Quarterly, 5,

441–456.

Kennedy, N. M., Whittington, R. M., & White, A. C. (1995). Suicide by self-strangulation whilst under observation. Medical

Science and the Law, 35, 174–177.

Liebling, A., & Hall, P. (1993). Seclusion in prison strip cells. British Medical Journal, 307, 399–400.

McDonald, D., & Thompson, N. J. (1993). Australian deaths in custody, 1980–1989. 2 Causes. Medical Journal of Australia,

159, 581–585.

Metzner, J. L. (2002). Class action litigation in correctional psychiatry. Journal of American Academy of Psychiatry and Law,

30, 19–29.

Morgan, H. G., & Priest, P. (1991). Suicide and other unexpected deaths among psychiatric inpatients. British Journal of

Psychiatry, 158, 368–374.

Morrison, S. (1996). Custodial suicide in Australia: A comparative study of different populations. Medical Science and the Law,

36, 167–177.

New York State Department of Correctional Services. (2002a). HUB system: Profile of inmate population under custody on

January 1, 2002. Albany, NY7 New York State Department of Correctional Services.

New York State Department of Correctional Services. (2002b). Inmate suicide report: 1993–2001. Albany, NY7 New York State

Department of Correctional Services.

Perez-Carceles, M. D., Inigo, C., Luna, A., & Osuna, E. (2001). Mortality in maximum-security psychiatric hospital patients.

Forensic Science International, 119, 279–283.

Ramsay, L., Gray, C., & White, T. (2001). A review of suicide within the State Hospital, Carstairs 1972–1996. Medical Science

and the Law, 41, 97–101.

Range, L. M., & Knott, E. C. (1997). Twenty suicide assessment instruments: Evaluation and recommendations. Death Studies,

21, 25–58.

Robins, E. (1981). Final months: A study of the lives of 134 persons who committed suicide. Cambridge7 Oxford University

Press.

Rotter, M., Way, B., Steinbacher, M., Sawyer, D., & Smith, H. (2002). Personality disorders in prison: Aren’t they all antisocial?

Psychiatric Quarterly, 73, 337–349.

Roy, A. (1982). Risk factors for suicide in psychiatric patients. Archives of General Psychiatry, 39, 1089–1095.

Rudd, M. D. (2003). Warning signs for suicide? Suicide and Life-Threatening Behavior, 33, 99–100.

Salive, M. E., Smith, G. S., & Brewer, T. F. (1989). Suicide mortality in the Maryland state prison system, 1979 through 1987.

Journal of the American Medical Association, 262, 365–369.

Spinellis, C. D., & Themeli, O. (1997). Suicide in Greek prisons: 1977 to 1996. Crisis, 18, 152–156.

Tatarelli, R., Mancinelli, I., Taggi, F., & Polidori, G. (1999). Prison suicides in Italy in 1996–1997. European Psychiatry, 14,

109–110.

WHO. (2000). Preventing suicide: A resource for prison officers. World Health Organization7 Geneva.