Suicide Survivors' Mental Health and Grief Reactions: A Systematic Review of Controlled Studies

Abstract

There has been a debate over several decades whether suicide survivors experience more severe mental health consequences and grief reactions than those who have been bereaved through other causes of death. This is the first systematic review of suicide survivors' reactions compared with survivors after other modes of death. Studies were identified by searching the PsychINFO and MEDLINE databases. Forty-one studies met the eligibility criteria. A qualitative data analysis was performed. There were no significant differences between survivors of suicide and other bereaved groups regarding general mental health, depression, PTSD symptoms, anxiety, and suicidal behavior. The results regarding the overall level of grief are less clear, depending on whether general grief instruments or suicide-specific instruments are used. Considering specific grief variables, suicide survivors report higher levels of rejection, shame, stigma, need for concealing the cause of death, and blaming than all other survivor groups.

Full text

13Suicide and Life-Threatening Behavior 38(1) February 2008
2007 The American Association of Suicidology
Suicide Survivors’ Mental Health
and Grief Reactions: A Systematic Review
of Controlled Studies
Carl-Aksel Sveen,PhD,and Fredrik A.Walby,PhD
There has been a debate over several decades whether suicide survivors
experience more severe mental health consequences and grief reactions than those
who have been bereaved through other causes of death. This is the first systematic
review of suicide survivors’ reactions compared with survivors after other modes
of death. Studies were identified by searching the PsychINFO and MEDLINE
databases. Forty-one studies met the eligibility criteria. A qualitative data analysis
was performed. There were no significant differences between survivors of suicide
and other bereaved groups regarding general mental health, depression, PTSD
symptoms, anxiety, and suicidal behavior. The results regarding the overall level
of grief are less clear, depending on whether general grief instruments or suicide-
specific instruments are used. Considering specific grief variables, suicide survivors
report higher levels of rejection, shame, stigma, need for concealing the cause of
death, and blaming than all other survivor groups.
According to World Health Organization kinds of mourning and grief and reconstitu-
tive patterns: (1) those which accrue to deaths(WHO) estimates, approximately one million
people die from suicide every year (WHO, from heart, cancer, accident, disaster and the
like and (2) those which relate to the stigma-1999). Suicides not only affect the deceased,
but also have profound effects on spouses, tizing death of a loved one by suicide” (Cain,
1972, p. x).relatives, and friends. In 1972, Edwin S.
Shneidman, a pioneering authority in suicid- Since the early 1980s (Calhoun, Selby,
& Selby, 1982), several reviews of suicide sur-ology, stated: “there are essentially only two
vivors’ reactions have been published. Some
concluded that: “there is no empirical evi-
dence for the popular notion that survivors
Carl-Aksel Sveen and Fredrik Walby of suicide show more pathological reactions
are affiliated with the Suicide Research and Pre- and a more complicated and prolonged grief
vention Unit at the University of Oslo Faculty
of Medicine, Oslo, Norway, and the Department process than other survivor groups” (Van der
of Psychiatry at Diakonhjemmet Hospital in Wal, 1989, p. 166), or that: “there are many
Oslo. more similarities than differences between
We thank Professor Lars Mehlum, MD, suicide survivors and other bereaved groups”
PhD, for comments on an earlier draft of this pa- (McIntosh, 1993, p. 158). Other reviews have
per, and Kirsti Amundsen, MA, who assisted in
the search process. reached contrasting conclusions, stressing
Address correspondence to Carl-Aksel that suicide survivors’ reactions are different
Sveen, PhD, University of Oslo, Faculty of Med- or more severe than for other bereaved
icine, Suicide Research and Prevention Unit, groups (Ellenbogen & Gratton, 2001; Jor-
Sognsvannsveien 21, Bulding 12 NO-0320, Oslo, dan, 2001; Ness & Pfeffer, 1990).
Norway; E-mail: Fredrik.walby@medisin.uio.edu

14 Suicide Survivors’Reactions
The issue of the nature of the suicide ing keywords were used “in terms anywhere”:
(1) “Suicide” and “Survivo*” (466 hits); (2)survivors’ reactions compared to other survi-
vor groups has not been reconciled. Further- “Suicide” and “Bereave*” (434 hits); (3) “Sui-
cide” and “Acciden*” (705 hits); (4) “Suicide”more, a systematic review, applying explicit
search strategies and criteria, has yet to be and “Grief” (523 hits); (5) “Suicide,” “So-
cial,” and “Adjustment” (304 hits); (6) “Sui-done in this field of research. Hence, the ob-
jective of this study is to conduct a systematic cide” and “Sequelae” (73 hits); (7) “Suicide”
and “Aftermath” (71 hits); (8) “Suicide,”review of suicide survivors’ reactions com-
pared with survivors after other modes of “Sudden,” and “Death” (100 hits); (9) “Sui-
cide,” “Natural,” and “Death” (178 hits); (10)death.
“Suicide” and “Nonsuicidal” (84 hits).
In the MEDLINE search the follow-
ing keywords were used: (1) “Suicide” andMETHOD
“Survivo*” (348 hits); (2) “Suicide” and “Be-
reave*” (251 hits); (3) “in the title only” “Sui-Inclusion and Exclusion Criteria
for Studies cide” and “Acciden*” (121 hits).
As the search in MEDLINE did not
identify any additional studies compared withStudies were included if they met
seven criteria. (1) At least one bereaved con- the same searches in PsychINFO, and no ad-
ditional studies were identified in the Psychtrol group had been included. Studies that
only compared results with the normative INFO search by the use of the keywords
solely used in that search, no further searchsamples for the instruments used were ex-
cluded. (2) The reactions measured were the was executed in MEDLINE.
We identified several unpublished doc-survivors’ subjective experience and behavior,
thus excluding studies of social reactions and toral dissertations through the PsychINFO
search, and all the authors’ names were elec-attitudes toward survivors. Studies of bereave-
ment group interventions were excluded as tronically searched to see if their work was
published in peer-reviewed journals or inwell. (3) The suicide survivors group con-
tained no survivors of other modes of death. books. Further, the names of all authors
identified in studies in the original search(4) The same instruments were used for both
suicide survivors and controls, in the same or were electronically searched to identify addi-
tional work, and all reference lists in includedequivalent way. (5) Quantitative outcome
measures were used. (6) Publications report- papers were searched for relevant studies. In
addition, reference lists in the following un-ing different measures from the same study
were included, but if the same results were systematic reviews or bibliographies were
also manually searched: Calhoun et al. (1982);reported in several publications, only the one
with the most comprehensive report was in- Clark (2001); Ellenbogen and Gratton (2001);
Henley (1984); Jordan (2001); McIntoshcluded. (7) Results had to be published in En-
glish in peer-reviewed journals or in books. (1986, 1993, 1996); Murphy, Johnson, and
Lohan (2003); Ness and Pfeffer (1990); and
Van der Wal (1989).Search Strategy for Identification
of Studies The first author of this paper reviewed
the abstracts of all potential studies to deter-
mine whether they met the criteria for inclu-Eligible studies were identified by
searching the PsychINFO database from sion in the review. The complete text was
obtained of all studies that could not be ex-1872 through January 7, 2005, and the
MEDLINE database from 1951 through cluded based on the abstract alone. Difficult
cases were evaluated by discussion and reach-January 7, 2005.
In the PsychINFO search, the follow- ing consensus with the second author.

Sveen and Walby 15
Data Analysis Mental Health Variables
General mental health was measured in
Preliminary inspection of the obtained 12 studies. No significant differences were
studies revealed that a very large number of found in 19 of 20 measurements. Depression
measures, scales, and self-made question- was measured in 14 studies. Significant dif-
naires had been used. Given the substantial ferences were found in only four out of 23
heterogeneity of outcome measures in the measurements; in one of those studies, the
studies, including many measures with un- accident survivors were significantly more
known psychometric properties, we decided depressed than the suicide survivors. PTSD
that it would be inappropriate to apply a symptoms were measured 16 times in 12 stud-
meta-analytic approach to the data. The fact ies. Significant differences were found only
that confounding and selection bias may well twice. One study reported elevated symptom
lead to biased results in observational studies, levels for homicide survivors; the other re-
which in turn will be reflected in a meta- ported reduced symptom levels for SIDS sur-
analysis, represents an additional reason for vivors. Anxiety was measured in six studies.
our reluctance to use that method (Egger, Two out of eight measurements gave signifi-
Schneider, & Smith, 1998). Instead, we con- cant differences. Suicidal ideation was investi-
ducted a qualitative analysis of the data. The gated eight times in four studies, with non-
results are categorized under the concepts of significant results seven times, whereas suicide
“Mental Health Variables” and “Grief Vari- attempts were measured four times in one
ables” in order to enhance their clarity. Fur- study, always with nonsignificant results.
thermore, in order to manage the huge
amount of data, we decided to report only Grief Variables
results regarding the subscales depression
and anxiety when results from symptom Overall grief was measured in ten stud-
checklist subscales were reported, whereas ies. Of 14 measurements in these studies, sig-
results regarding other subscales are omitted, nificant differences were found three times.
both in the tables and in the text. When measured by more general grief in-
struments, such as the Texas Inventory of Grief
(TIG) and Inventory of Complicated Grief
RESULTS (ICG), no significant differences were found
regarding any of the ten measurements. In
contrast, when using the more suicide spe-Sixty-nine studies were obtained in full
text and assessed for relevance. Of these, 41 cific Grief Experience Questionnaire (GEQ),
significant differences were found in threewere judged eligible for inclusion (marked
with asterisks in the reference list). A con- out of four assessments when compared with
natural death survivors, and in two out ofdensed overview of the included studies is
presented in Table 1. Abbreviations of the in- three incidences when compared with acci-
dent survivors.struments are explained in the Appendix.
When outcomes have been measured at sev- Rejection has been measured six times.
Suicide survivors always reported a signifi-eral time points in a study, all measurements
are counted as a separate outcome in the fol- cantly higher level of rejection than survivors
of other modes of death. Shame or stigma waslowing sections in the text. When an out-
come is measured with different instruments measured 18 times across 12 studies. Signifi-
cant differences were found 15 times. Con-at the same point in time, both measure-
ments count as a separate outcome. Hence, cealment of the cause of death was measured in
two studies, both reporting significant differ-the number of measurements will not neces-
sarily be equal to the number of studies. ences. Guilt has been measured 15 times

TABLE 1
Included Studies (Longitudinal or Related Studies are Reported Together or in Adjacent Rows)
Authors Sample Recruitment Groups Time Measures Results
Bailley et al. Undergraduate In class. 34 SS 1M-29 Y GEQ, IES, TRIG, Total grief (GEQ), rejection, responsibil-
(1999) students who have 57 AS SMQ reg. recovery ity, unique reactions, shame, stigma, search
lost a friend or rel- 102 SNS and accept. for explanation; SS>AS, SNS, NAS. So-
ative. 12 % first- 157 NAS matic reactions, general reactions, social
degree family support, guilt, self-destructive behavior,
members. PTSD symptoms, grief (TRIG), accep-
tance, recovery; N.S.
Barrett & Spouses. Obituaries with 14 SS 2–4 Y GEQ, PIL. Total grief; SS >SNS, NAS. SS vs. AS;
Scott (1990) cause of death. 15 AS N.S. Rejection, unique reactions; SS >AS,
83% R.R. 15 SNS SNS, NAS. Shame; SS >SNS, NAS. Re-
13 NAS sponsibility, search for explanat., stigma;
SS >NAS. Loss of social support, guilt,
self-destructive behavior, somatic reactions,
general reactions; N.S. Purpose in life; N.S.
Cerel et al. Children 5–17 Y Obituaries/funeral 26 SS T1: 1 M GI, BAMO, CBCL, Anger (T2), shame (T2, T3), conduct disor-
(1999, 2000) and spouses. Reg- homes. 33% R.R. 332 NSB T2: 6 M CDI, CDRS-R, der (T0-T4), anxiety (T0-T1); SS >NSB.
ular contact with T3: 13 M CRTRS, DICA-R, Accept (T2, T3), relief (T1), family func-
deceased previous T4: 25 M DIDCA, FH-RDC, tioning (T0); SS <NSB. Depression, suicid-
2 Y. HAM-D, HEI, HSQ, ality (ideation and attempts), PTSD symp-
PDI, PHSCS. toms, mental health, psychosocial
functioning, social support; N.S.
Cleiren (1993) Spouses, parents, Official records, 91 SS T1: 4 M BDI, IES, LBS, SCS, T1: Preoccupation with the death; SS, AS >
siblings, or adult health services, 93 AS T2: 14 M SIS, SAS, interview ND. Guilt, isolation; SS >AS, ND. Search
children. Maxi- news reports, or 125 ND reg. loss, mental for meaning; SS >AS >ND. Difficulty
mum 5 from the obituaries. 66% health, social function- with detachment; AS >SS, ND. Relief; AS <
same family. R.R. 90% F.R. ing etc. SS, ND. Depression; AS >SS >ND. T2:
Need for emotional support; AS >SS. T1
& T2: All other loss-, health- (suicidal ide-
ation, PTSD symptoms), social function-
ing, and anger measures; N.S.

Demi (1984) Widows. Official records. 20 SS 12–21 M SAS, interview reg. Separated, worry, guilt, resentment; SS >
61% R.R. 20 NSB guilt, parenting etc. NSB. Victims age, previous marriages,
weakened parental role; SS <NSB. Overall
social adjustment; N.S.
Miles & Demi Parents who have Bereavement 62 SS 2 M–7 Y Open ended interview Guilt; SS: 92%, AS: 78%, ND: 71%. Sub-
(1991–1992) lost children groups. 32 AS reg. guilt. scales; Death Causation guilt; SS: 63%,
1–36 Y. 38 ND AS: 64%, ND: 26%. Childrearing guilt;
SS: 51%, AS: 52%, ND: 44%*. Otherwise
low scores. Guilt vs. loneliness; SS >AS, ND.
Demi & Miles Same as above, As above. 48% 59 SS 2 M–7 Y BHAS, HSCL. Emotional distress, physical health prob-
(1988) but AS and ND R.R. 61 NSB lems, anxiety, depression; N.S.
grouped as NSB.
Dyregrov et al. Parents from 140 Official records. 128 SS 14–15 M GHQ, ICG, IES, in- Psychological health, PTSD symptoms,
(2003) families. Chil- SS; 50% R.R. AS/ 68 AS terview reg. guilt, iso- complicated grief; N.S. between SS and
dren; 11–29 Y SIDS; 57 % R.R. 36 SIDS lation etc. AS. Psychological health, PTSD symp-
(SS), 0–1 Y toms; SS, AS >SIDS. Complicated grief;
(SIDS), 0–18 Y AS >SIDS. Isolation, guilt; N.S.
(AS).
Farberow et al. Spouses older Official records. 108 SS T1: 2 M BDI, BSI, MHSR, T1: Anxiety; SS >ND. Grief, mental
(1987, 1992a, than 55 Y. SS; 35% R.R. 199 ND T2: 6 M TIG, interview reg. so- health, depression (BDI and BSI subscale),
1992b) ND; 30% R.R. T3: 12 M cial support. emotional distress; N.S. T2: Grief, mental
T1–T4 SS; 66% T4: 30 M health, depression, emotional distress; N.S.
F.R. ND; 74% T3: Depression; SS >ND. Mental health;
F.R. SS <ND. Grief, emotional distress; N.S.
T4: Grief, mental health, depression, emo-
tional distress; N.S. T1–T4: Social sup-
port; SS <ND. Composite group measure-
ments: Anxiety, depression (BSI subscales);
N.S.
Flesch (1977) Spouses, adult Official records. 33 SS 4 W HHQ, interview reg. Mental health functioning; N.S.
children, siblings, 36 AS mental health.
or parents.
(continued)

TABLE 1
Continued
Authors Sample Recruitment Groups Time Measures Results
Grad & Zavas- Spouses. Official records. 30 SS T1: 2 M BDI, EPQ, SBS. T1: Grief items: Difficulty accepting
nik (1996, SS; 68/70% R.R./ 23 AS T2: 14 M death; SS, ND >AS. Unfairness, death
1999). F.R. AS; 56/87% 20 ND good solution; SS >AS. Shame; SS >AS,
R.R./F.R. ND; ND. Depression, personality (EPQ); N.S.
52/90% R.R./F.R. T2: Grief items: Sadness, lack of perspec-
tives, guilt, wish to damage oneself, ease
with new contacts, accept; AS >SS, ND.
Social isolation; ND >SS, AS. Depression,
personality (EPQ); N.S.
Harwood et al. Relatives, spouses, Official records or 46 SS 6–21 M GEQ, MADRS. Stigma, shame, rejection, unique reactions;
(2002) or friends. Mean hospitals, 41% 46 ND SS >ND. Total grief, somatic reactions,
age of the de- R.R. or less. general reactions, search for explanation,
ceased 72 Y. social support, guilt, responsibility, self-
Matched controls. destructive behavior, depression; N.S.
Kitson (2000) Widows. HS, SS, Official records. 85 SS 6 M BSI, SDS. Emotional distress, anxiety, depression
AS matched with 50% total R.R. 135 AS (SDS and BSI subscale); N.S.
SNS and NAS. 56 HS
167 SNS
106 NAS
Kovarsky Parents who have Bereavement 31 SS 0–7 M TRIG, UCLA LS, Self blame, blaming of the family, and
(1989) lost children groups. 44% R.R. 21 AS SMQ reg. blame. blaming of the doctor; SS >AS. Loneliness
15–29 Y. and grief; N.S.
McIntosh & Undergraduate In class. 79% R.R. 40 SS 1–34 Y IES, TRIG, UCLA Blaming; SS >AS, ND. Isolation; SS, ND
Kelly (1992) students who have 71 AS LS, SMQ reg. stigma, <AS. Relief; AS <SS <ND. Disbelief,
lost a friend, 63 ND guilt, blame, relief, sui- numbness; AS >SS, ND. Desire to under-
spouse or relative cidal ideation etc. stand why, stigma, shock; SS, AS >ND.
older than 12 Y. Grief, loneliness, suicidal ideation, PTSD
symptoms, recovery, guilt, thinking about
the death, sharing thoughts; N.S.

McNiel et al. Widows. Official records. 13 SS 15–18 M F-APGAR, FES, Blaming, lack of social support in the fam-
(1988) 48% R.R. 13 AS GHQ, LEQ, interview ily, guilt, hide cause of death (interview);
reg. family functioning SS >AS.* Social support, family function-
etc. ing, mental health, PTSD symptoms (ques-
tionnaire); N.S.
Murphy et al. Parents who have Official records. 41 SS T1: 4 M BSI, FACES III, MS, Accept of death; SS >AS, HS. Suicidal ide-
(1999, 2003a, attended be- 62% R.R. 67% 87 AS T2: 12 M TES, SMQ reg. sui- ation; HS >SS >AS. PTSD symptoms;
2003b); Lohan reavement inter- F.R. 35 HS T3: 24 M cidal ideation, accept HS >SS, AS (similar time course). Emo-
& Murphy vention pro- T4: 60 M of death etc. tional distress, marital satisfaction, family
(2002) grams. Children functioning, time before normal life; N.S.
12–28 Y. (comparisons between groups were made
only once as a composite)
Nelson & Parents and sib- Bereavement 34 SS 4 Y BFS. Family functioning; N.S.
Frantz (1996) lings. 24% single groups, and adver- 25 AS
parent families. tisements. 21 ND
Pennebaker & Spouses. AS not Official records. 1) 9 SS 0–1 Y SMQ reg. health, sup- Physical health problems, coping, confid-
O’Heeron involved in acci- 61% R.R. 2) 10 AS port, confiding. ing; N.S.
(1984) dent.
Pfeffer et al. Children 5–12 Y. Official records. 16 SS 0–18 M CBCL, CDI. Depression (total and 4/5 subscales); SS >
(2000) Knew cause of 65–70% r. r. 64 NAS NAS. Competence, behavior, self-esteem;
death. N.S.
Range & Cal- Undergraduate In class. 11 SS 0–22 M Interview reg. bereave- Lying about the cause of death; SS >AS,
houn (1990) students who have 17 AS ment. SNS, NA, HS. Treated differently; AS,
lost a friend or rel- 13 SNS SNS, HS >SS. Pressure to explain the
ative. 13 NAS death to others; SS, AS >SNS, NA, HS.
3 HS Change for the better; AS, HS >SNS, SS.
Well-being; N.S.
Range & Niss Undergraduate In class. 9 SS 2–13 Y HI, IES, ISEL, Perceived overall recovery; SS >AS. Unre-
(1990) students who have 17 AS MAACL, PRS, SMQ ality feeling; AS >SNS, NAS. PTSD symp-
lost a friend or rel- 17 SNS reg. social support. toms, social support, mood; N.S.
ative. 17 NAS
8HS
(continued)

TABLE 1
Continued
Authors Sample Recruitment Groups Time Measures Results
Reed & Green- Spouses, friends, Official records, 92 SS 6–16 M SMQ reg. grief (6 Guilt, shame, rejection; SS >AS. Emo-
wald (1991) or relatives (23% via kin. 63% R.R. 105 AS scales). tional distress, Shock; SS <AS. Somatic
same family). complaints; N.S.
Reed (1993) Same as above, Same as above. 85 SS Same as above. Same as above (2 Expressive social support; SS >AS. Self-
but 16 friends ex- 96 AS scales). esteem, affective grief responses and cogni-
cluded. tive grief reactions, detachment; N.S.
Sherkat & As Reed & Green- Same as above. Not Same as above. Same as above (1 Depression; N.S.
Reed (1992) wald (1991), with- reported scale).
out 43 cases.
Reed (1998) As Reed (1993), Official records. 66 SS Same as above. Same as above (4 Separation anxiety; SS <AS. Rejection; SS
but only one from 79 AS scales). >AS. Shock-disbelief, depression; N.S.
each family.
Saunders Widows. None Official records. 4 SS 0–13 M Interview. Initiating of dating post bereavement; 3
(1981) separated at time 4 AS SS, 1 AS within 3 M, 3 AS, 1 SS, 1 HS
of bereavement. 6 ND within 3–6 M, 3 ND within 3–6 M, and 3
1 HS ND not dating by end of period.*
Seguin et al. Parents who have Official records. 30 SS T1: 4–8 M T1: BDI, GRS. T2: T1: Depression; SS >AS. Grief; N.S. T2:
(1995) lost children 70% R.R. 100% 30 AS T2: 7–11 M BDI, BS, FACES III, Shame, negative life events, negative im-
18–35 Y. In 20 F.R. LEI, SCL-90, SMQ pact on family adjustment (interview), lack
cases, parents reg. shame, interview of social support from immediate family,
from the same reg. family adjustment. lack of helpfulness from network; SS >AS.
family. Depression, grief, emotional distress, fam-
ily adaptation, positive impact on family ad-
justment; N.S.

Silverman et Undergraduate In class. 9 SS 16 M GEQ, GRQ, IES, Total grief, rejection, stigma, responsibil-
al. (1994– students who have 16 AS ISEL. ity, loss of social support, self-destructive
1995) lost a friend or rel- 9 SNS behavior, unique reactions; SS >AS, SNS,
ative during the 12 NAS NAS, HS. Search for explanation; SS >
last 5 Y. Deceased 9 HS NAS. Shame; SS >AS, HS. Somatic reac-
less than 70 Y. tions, general reactions, guilt; N.S. Inter-
personal support: total support; N.S, mate-
rial support; NAS >SS, AS, HS. PTSD
symptoms and recovery; N.S.
Stone (1972) Spouses. NSB Official records. 35 SS 1–2 Y GI. Sickness, stigma, blaming from others, mar-
without HS. 71% R.R. 5$ fee. 31 NSB riage problems, anger at deceased, guilt (es-
pecially salient); SS >NSB.
Thompson & Undergraduate In class. (800 10 SS 6 M IES, PSSS, SPOB. Prognosis; SS <NAS. Social contact before
Range (1990– students who have screened) 11 AS and after loss (change not measured); SS <
1991) lost a friend or rel- 12 NAS NAS, SNS, AS. Helpfulness from others
ative. 13 SNS (a) before the death; SS <NAS, SNS, (b)
after the death; SS <NAS, PTSD total
score; N.S. Subscale Avoidance, SS >NAS,
SNS. Subscale Intrusion; N.S.
Thompson & Undergraduate In class. (400 18 SS 7 M HUS, IES, ISEL, Unhelpful social support; SS >AS, NAS,
Range (1992– students who have screened) 13 AS MAACL, PRS, PSSS, SNS, HS. Belonging support; SS >AS,
1993) lost a friend or rel- 10 NAS HS. Blaming; AS >NAS, SNS. PTSD
ative. 10 SNS symptoms, mood; N.S.
5HS
Ulmer et al. Relatives or Funeral homes, 25 SS 1 M–7 Y IES, PIL, RFL, SL, Recovery; NAS, SNS >SS, AS. Fear of so-
(1991) friends. 44% in bereavement lec- 20 AS SPS, SMQ reg. recov- cial disapproval; AS >NAS, SNS. Purpose
counseling after tures/groups. 31% 30 SNS ery, fear of social dis- in life, reasons for living, PTSD symp-
bereavement. estimated R.R. 38 NAS approval etc. toms, overall social provisions, satisfaction
6 HS with life; N.S.
Smith et al. Same as above. Same as above. Same, Same as above. BA, SWBS, SMQ as Accept of loss; SS <NA, SNS, AS, HS.
(1991–1992) +1 SS, AS above. Finding meaning in the death; NAS,
SNS >SS, AS, HS.
Note. AS: Accident Survivors; F.R.: Follow-up rate; HS: Homicide Survivors; M: Months; NAS: Natural Anticipated Death Survivors; ND: Natural
Death Survivors; N.S.: Nonsignificant finding; NSB: Non-Suicide Bereaved Survivors; PTSD: Posttraumatic Stress Disorder; R.R.: Response rate; SMQ:
Self-Made Questionnaire; SS: Suicide Survivors; SIDS: Sudden Infant Death Syndrome Survivors; SNS: Sudden Natural Death Survivors; T0: Period prior
to bereavement; W: Weeks; Y: Years; *Significance level not reported; >: Significant difference, 5% level; <: Significant difference, 5% level.

22 Suicide Survivors’Reactions
across 14 studies. In six incidences, suicide dent survivors than among the suicide survi-
vors. One out of three times, significant dif-survivors experienced a significantly higher
level of guilt than controls. In one study, the ferences between suicide survivors and natural
death survivors were found. Search for expla-accident survivors actually reported a higher
level of guilt than the suicide survivors. All nation has been measured seven times in six
studies. Compared with accident survivors,significant differences were measured within
18 months of bereavement. Differences were significantly higher levels in suicide survivors
were found twice. Compared with naturalreported relatively more often when inter-
views (three out of five times) rather than death survivors, significant differences were
found five times.questionnaires (three out of ten times) were
used. The related concept of responsibility has
also been measured in four studies. Three
studies reported significant differences com- DISCUSSION
pared with other bereaved controls. Blaming
has been measured in four studies. All of Methodological Issues
them reported that suicide survivors experi-
enced significantly higher levels of blaming. A multitude of samples, recruitment
methods, control groups, time passed sinceAnger at the deceased has been measured
seven times, in three studies. Significant dif- bereavement, and instruments have been
used in the studies included in this review.ferences were found twice. Loneliness or isola-
tion has been measured eight times across six A majority of the studies display significant
methodological limitations in one way or an-studies. Significant differences were found
three times. In one assessment, suicide survi- other. All studies are observational, and
thereby vulnerable to confounding variables.vors were worse off than accident survivors,
vice versa in another. In the third incidence, Differences between suicide-bereaved groups
and control groups from another mode ofnatural death survivors were worse off than
both suicide and accident survivors. Social death are often present, with variables such
as age of the deceased, age of the survivors,support has been measured 27 times across 17
studies. Significant differences were reported participation in intervention programs, or
mental health problems antedating exposure15 times. All six times when social support
was measured by interview methods, suicide being the most prominent. Another impor-
tant methodological issue is selection bias,survivors reported less social support than
other bereaved groups. Social adjustment, fam- specifically due to convenience sampling
based on bereavement groups or obituaries,ily functioning, or child behavior after bereave-
ment has been measured 20 times across low response rate, or low follow-up rate.
Stroebe and Stroebe (1989–1990) found thateight studies. Significant differences were
only found five times. Acceptance has been those who replied to bereavement surveys
were more likely to have experienced an ex-measured nine times in five studies. In five
out of nine measurements, acceptance of the pected death and to be healthier and more
recovered than nonrespondents. If this holdsdeath was significantly lower in suicide survi-
vors than accident survivors. The opposite true within the suicide survivor group, selec-
tion bias may thereby decrease differencesresult was found once. Relief has been mea-
sured seven times in three studies. No signifi- between survivor groups, making it more dif-
ficult to uncover significant differences.cant differences were found in four measure-
ments. Suicide survivors twice described more Most studies were explorative rather
than hypothesis-testing, and conducted manyrelief than accident survivors. Once they re-
ported less relief than nonsuicide survivors. comparisons among groups, making it more
likely to commit type I errors. CorrectionsShock (including numbness/disbelief) has
been measured five times. Three out of five for multiple comparisons were seldom used.
On the other hand, most studies have smalltimes, higher levels were found among acci-

Sveen and Walby 23
sample sizes, which may mask real differ- bers of suicide completers in adoption, twin,
and family studies (Brent & Mann, 2005).ences among groups (type II error). How-
ever, larger studies do not seem to find signif- Grief Variables. The results regarding
overall grief seem to be sensitive to the typesicant differences more often than smaller
ones. All in all, both type I and type II errors of instruments used. General grief instru-
ments do not uncover significant differences.are likely to have had an impact on the out-
comes of the studies in this review. Further- By using a more suicide-specific instrument,
the GEQ, significant differences were evi-more, several studies have not conducted
multivariate analysis when indicated. Non- dent, though not completely consistent, as
significant differences were found in threestandardized instruments with unknown reli-
ability or validity have been used in many of out of four incidences when compared with
natural death survivors. A particular featurethe studies. Measures of survivors’ state be-
fore exposure are always made retrospec- with the study that failed to uncover signifi-
cant differences by the GEQ, was that thetively. Many of the methodological short-
comings in this literature are difficult to mean age of the deceased was 72 years. It
may be that mode of death becomes less im-avoid. Selection bias and the possibility of
confounding variables will inherently be portant as the deceased approaches the ex-
pected lifespan. In this respect, it is worthpresent in this type of research, as there is no
randomized assignment to groups. noting that four of the measurements using
TIG also involved older subjects who had
lost their spouses (55 years or more). WhenMain Findings
compared with accident survivors, differences
were found in two out of three incidencesMental Health Variables. There were
very few significant differences between sur- when the GEQ was used. The third study re-
cruited respondents through obituaries,vivor groups regarding general mental health,
depression, PTSD symptoms, anxiety, and which may have led to selection bias, as the
suicide survivors experiencing the highestsuicidal ideation.
Approximately 90% of those who die levels of stigmatization probably would leave
out the cause of death in the obituary. Theby suicide have been suffering from one or
more mental disorders (Cavanagh, Carson, GEQ, with its 11 subscales, was developed to
include not only general grief reactions (so-Sharpe, & Lawrie, 2003), and many of these
disorders have a genetic component (Torg- matic reactions, general grief reactions), but
also reactions considered to be specific to sui-ersen, 1997). Thus there is an elevated prob-
ability that suicide survivors who are relatives cide bereavement (search for explanation,
loss of social support, stigmatization, guilt,of the deceased may have higher levels of
mental disorders antedating exposure. This is responsibility, shame, rejection, self-destruc-
tive behavior, and “unique” reactions) (Bar-also shown in several of the studies included
in the review. Taken together with the stress rett & Scott, 1989). Thus, it seems more
likely that significant differences will be de-that many suicide survivors have been ex-
posed to before bereavement, suicide survi- tected in studies where the instrument tar-
gets specific grief reactions, and is sensitivevors are expected to be more vulnerable than
other groups of survivors. Consequently, it is to variables considered important for suicide
survivors. This point is further supported bysomewhat surprising that higher levels of
group differences are not found. the fact that the only study that used both
general and specific instruments found differ-As only one study with limited statisti-
cal power provided data on suicide attempts, ences using the GEQ, but not by using the
Texas Revised Inventory of Grief (TRIG)the evidence is too limited to make infer-
ences. Increased risk of suicide attempt due (Bailley, Kral, & Dunham, 1999). Some cau-
tion regarding the applicability of the GEQto genetic, imitation, and contagion effects
are well documented among the family mem- is, however, necessary, as its psychometric

24 Suicide Survivors’Reactions
properties have not yet been thoroughly in- Hence the evidence is too limited to make
inferences about a possible relationship be-vestigated. Although internal consistency is
acceptable (Barrett & Scott, 1989), the scale tween the concepts.
There is little support that suicide sur-structure seems disputable, as an eight-factor
solution was established by factor analysis of vivors experience anger, loneliness, or isola-
tion more often than other bereaved groups.the items, contrary to the original 11-factor
structure (Bailley, Dunham, & Kral, 2000). The results regarding social support
are conflicting. Social support has been oper-The evidence clearly indicates that sui-
cide survivors experience a significantly higher ationalized in many different ways between
studies, but that fact cannot explain the dis-level of rejection compared with all other be-
reaved groups, including accident survivors. crepancy in the results. However, the results
seem sensitive to the choice of data collectionIt seems that suicide survivors commonly ex-
perience suicide as an act of intentional rejec- methods, as significant differences are uncov-
ered relatively more often by interview thantion.
The evidence clearly shows that shame by questionnaire.
A clear majority of studies do not re-and stigma are especially salient for suicide
survivors, which in turn may lead to conceal- port significant differences regarding social
adjustment, family functioning, or child be-ment of the cause of death. Due to the rela-
tive scarcity of measurements, the evidence havior, regardless of the relationship with the
deceased. It is, however, noteworthy that inregarding concealment of the cause of death
is somewhat weaker than the evidence re- both studies where survey methods and struc-
tured interview methods were used, suicide-garding shame and stigma.
There is no clear answer to the ques- bereaved parents and widows, respectively,
reported a higher level of negative impact ontion of whether suicide survivors experience a
higher level of guilt/responsibility than other family functioning than accident-bereaved
controls when interview methods were used,survivors. There may be a weak tendency to-
ward relatively higher levels of guilt/respon- whereas no differences were uncovered by
survey methods. Hence, the results regardingsibility for suicide survivors than other survi-
vors in the first 18 months, with increasing family functioning may also be sensitive to
the method of data collection.difficulty in finding significant differences
thereafter. One should also bear in mind that Relief and acceptance are related con-
cepts that can be depicted as different pointsthe results seem to depend on the method of
data collection. The modest amount of sig- on the same continuum, ranging from pro-
test to relief, with acceptance somewhere innificant differences may be somewhat sur-
prising, but points to the important fact that between. The results regarding both accep-
tance and relief are contradictory. Althoughother survivors also report relatively high
levels of guilt/responsibility. Another reason it does not reconcile the results, it is interest-
ing to note that the subjects in the study thatfor the difficulty in uncovering clear trends
could be that the concept of guilt is differ- found a significantly higher level of accep-
tance among suicide survivors compared withently operationalized among studies.
One might suspect that blaming would accident survivors had attended a bereave-
ment intervention program, possibly affect-affect the survivors’ experience of guilt/re-
sponsibility. Therefore it is somewhat puz- ing the survivors’ acceptance of the death
(Murphy, Johnson, Wu, Fan, & Lohan, 2003a).zling that the results regarding the former
concept are less ambiguous. To inquire into Several authors who have investigated
relief, acceptance, shock, or search for expla-the relationship between blaming and guilt,
both concepts should be measured in the nation report evidence supporting the notion
that suicide survivors are a heterogeneoussame study. This is only reported three times
in the literature, of which two found signifi- group (Bailley et al., 1999; Grad & Zavasnik,
1996, 1999; McIntosh & Kelly, 1992; Reed,cant differences regarding both concepts.

Sveen and Walby 25
1993; Seguin, Lesage, & Kiely, 1995). Some lication bias is present. According to criteria,
unpublished doctoral dissertations and stud-experience a profound sense of grief, whereas
others also experience relief. As discussed, ies have not been evaluated in this review. As
the debate in the literature as to whether sui-approximately 90% of those who die by sui-
cide have been suffering from one or more cide survivors’ grief is different and more se-
vere than that of other bereaved groups hasmental disorders (Cavanagh et al., 2003).
When suffering from serious mental disor- been ongoing, and negative findings might be
considered interesting, one might argue thatders, the deceased may have exposed those
close to them to many difficulties and worry, the level of publication bias is negligible. On
the other hand, the debate has been some-including suicide threats or attempts, often
for years. The element of relief for the sui- what theoretically driven, especially in the
early years, which may have created a pres-cide survivors seems to be tied to the fact that
the difficult times are finally over, for both sure to keep nonsignificant results “in the
drawer.”the survivors and the deceased. The level of
shock/unexpectedness is likely to be rela-
tively low for this group of survivors, which
may lead to an attenuation of their grief ex- CONCLUSION
perience. Compared with this subgroup of
suicide survivors, accident survivors may ex- The evidence collected in this system-
atic review gives no support for the existenceperience higher levels not only of shock and
disbelief, but also of other grief reactions as of significant differences between survivors of
suicide and other bereaved groups regardingwell. On the other hand, when the suicide is
unexpected, and difficult to explain and ac- mental health variables, including general
mental health, depression, PTSD symptoms,cept, the survivors’ grief reactions may be
more severe or different. If the existence of anxiety, and suicidal behavior. As to the over-
all level of grief, the results are less clear, re-distinct subgroups of suicide survivors can be
further substantiated, it may explain some of gardless of which control groups are used.
When general grief instruments are used,the controversy in the literature; whereas one
subgroup of suicide survivors experiences significant differences are never found. In
contrast, when suicide-specific instrumentsmore severe or different reactions than other
bereaved groups, the other subgroup does are used, significant differences dominate.
When specific aspects of the suicide survi-not. When both subgroups are pooled, dif-
ferences compared with other groups of sur- vors’ grief are considered, the evidence clearly
shows significant differences compared withvivors may be leveled out.
Finally, there seem to be no differences all other survivor groups regarding the fol-
lowing variables; rejection, shame, stigma,in findings between recent and older studies.
concealing the cause of death, and blaming.
It is important to emphasize that theLimitations
reviewed outcome variables are found to be
sensitive to the research methods used. Sig-Even though we have conducted a sys-
tematic review, some studies may have es- nificant differences between suicide survivors
and survivors after other modes of death arecaped our attention. First, there is no guaran-
tee that the search strategy identified all more often found when interview methods
rather than survey methods are used. Fur-eligible studies. Second, all eligible studies
may not be identifiable from their abstracts. thermore, most studies had methodological
shortcomings: lack of clear hypotheses, lackThird, in most cases only one author con-
ducted the abstract evaluation procedure, of multivariate statistical approaches, conve-
nience sampling, low response rate, unstan-leaving the possibility that some studies may
have been erroneously omitted. dardized or too general instruments, and
small samples.It is difficult to ascertain whether pub-

26 Suicide Survivors’Reactions
FUTURE RESEARCH enhancing the possibility of reaching those
survivors who are presumed to have the most
severe reactions; (4) use standardized instru-Future research in this area should: (1)
test theoretically-derived hypotheses, includ- ments with adequate psychometric properties
enabling future meta-analysis; (5) develop in-ing assessing specific aspects of the suicide
survivors’ experiences simultaneously, in or- struments that are specifically aimed at in-
vestigating variables associated with suicideder to reach a better understanding of both
the complexity and the interrelatedness of survivors’ reactions; (6) study the possible ex-
istence of subgroups with different reactionstheir reactions (e.g., among such variables as
blaming, guilt, and relief ); (2) use sufficient among suicide survivors (e.g., by addressing
variables such as relief and expectedness); andsample size and multivariate statistical ap-
proaches; (3) use survivors after other forms (7) study survivors that are not well studied
(e.g., children).of sudden death as control group; (4) focus
on selection and response rate, thereby also
APPENDIX
Abbreviations Regarding the Instruments Utilized in the Studies
Abbreviation Name Comments
BA Belief in Afterlife
BAMO Behavioral, Anxiety, Mood, Other Symptoms
BDI Beck Depression Inventory
BFS Bloom’s Family Scales Family functioning
BHAS Bereavement Health Assessment Scale
BS Bereavement Scale Earlier losses
BSI Brief Symptom Inventory Symptom checklist
CBCL Child Behavior Checklist
CDI Children’s Depression Inventory
CDRS-R Children’s Depression Rating Scale
CRTRS Conners Revised Teacher Rating Scale Behavioral problems in school
DICA-R Diagnostic Interview for Children and Structured interview regarding psychiatric di-
Adolescents agnoses
DIDCA Diagnostic Interview for Depression in Structured interview regarding depression
Children and Adolescents
EPQ Eysenck Personality Questionnaire Extraversion, Neuroticism
FACES III Family Adaptability and Cohesion Eval- Family functioning
uation Scale
F-APGAR Adaptation, Partnership, Growth, Af- Family functioning
fection, Resolve
FES Family Environment Scale Social support in the family
FH-RDC Family History Research Diagnostic Psychiatric diagnoses in the family
Criteria
GEQ Grief Experience Questionnaire Overall grief and 11 subscales
GHQ General Health Questionnaire
GI Grief Inventory (2 different scales) Cerel et al. (1999): Structured interview,
Stone (1972): Questionnaire
GRQ Grief Recovery Questions Perceived recovery
GRS Grief Reaction Schedule
HAM-D The Hamilton Rating Scale for De-
pression
HEI Home Environment Interview Company with family (quantity and quality)

Sveen and Walby 27
Abbreviation Name Comments
HHQ Health History Questionnaire Mental health
HI Help Inventory Received help and support
HSCL Hopkins Symptom Checklist Symptom checklist
HSQ Health/Sickness Questionnaire Checklist regarding physical symptoms
HUS Helpful/Unhelpful Support Facilitativeness of social support
ICG Inventory of Complicated Grief
IES Impact of Event Scale PTSD symptoms, 2 subscales; avoidance and
intrusion
ISEL Interpersonal Support Evaluation List Perceived social support
LBS Leiden Bereavement Scale Level of detachment from the deceased
LEI Life Events Inventory Life events after exposure, after Holmes and
Rahe
LEQ Life Event Questionnaire Same as IES
MAACL Multiple Affect Adjective Checklist
MADRS Montgomery A
˚sberg Depression Rating
Scale
MHSR Mental Health Self Rating One scale/item
MS Marital Satisfaction
PDI Psychiatric Diagnostic Interview
PHSCS Piers-Harris Self Concept Scale Self esteem
PIL Purpose In Life Test Meaning and purpose etc.
PRS Perceived Recovery Scale
PSSS Perceived Social Support Scale
SAS Social Activity Scale or Social Adjust- Cleiren (1993): Social activities, Demi (1984):
ment Scale Social adjustment
SBS Slovene Bereavement Scale
SCL-90 Symptom Checklist (90 items)
SCS Sense of Control Scale
SDS Self-Rating Depression Scale
SIS Social Integration Scale Social functioning
SL Satisfaction with Life Scale
SPOB Scale for Prediction of Outcome after Recovery, accept, meaning etc.
Bereavement
SPS Social Provision Scale Social support
SWBS Spiritual Well-Being Scale
TES Traumatic Experience Scale PTSD symptoms
TRIG/TIG Texas (Revised) Inventory of Grief Overall grief, 7 items (TIG)
UCLA LS University of California Los Angeles
Loneliness Scale
REFERENCES
proposition. Suicide and Life-Threatening Behavior,
Studies analyzed for this review are indicated with 29, 256–271.
an asterisk. Barrett,T.W., & Scott,T.B. (1989).
Development of the Grief Experience Question-
Bailley,S.E., Dunham,K., & Kral,M.J.naire. Suicide and Life-Threatening Behavior,19,
(2000). Factor structure of the Grief Experience 201–215.
Questionnaire (GEQ). Death Studies, 24, 721–738. *Barrett,T.W., & Scott,T.B. (1990).
*Bailley,S.E., Kral,M.J., & Dunham,Suicide bereavement and recovery patterns com-
K. (1999). Survivors of suicide do grieve differ- pared with non-suicide bereavement patterns. Sui-
cide and Life-Threatening Behavior,20, 1–15.ently: Empirical support for a common sense

28 Suicide Survivors’Reactions
Brent,D.A., & Mann,J.J. (2005). Family spouses of older suicides. Journal of Gerontology:
Psychological Sciences,47, 357–366.
genetic studies, suicide, and suicidal behavior.
*Farberow,N.L., Gallagher-Thompson,
American Journal of Medical Genetics. Part C: Semi-
D., Gilewski,M., & Thompson,L. (1992b). The
nars in Medical Genetics, 15, 13–24.
role of social supports in the bereavement process
Cain,A.C.(Ed.). (1972). Survivors of sui- of surviving spouses of suicide and natural deaths.
cide. Springfield, IL: Charles C. Thomas.
Suicide and Life-Threatening Behavior,22, 107–124.
Calhoun,L.G., Selby,J.W., & Selby,*Flesch,R. (1977). Mental health and be-
L.E. (1982). The psychological aftermath of sui- reavement by accident or suicide: A preliminary
cide: An analysis of current evidence. Clinical Psy- report. In B. L. Danto & A. H. Kutcher (Eds.),
chology Review, 2, 409–420.
Suicide and bereavement (pp. 128–137). New York:
Cavanagh,J.T.O., Carson,A.J., Sharpe,Information Corporation.
M.,&Lawrie,S.M. (2003). Psychological au- *Grad,O.T., & Zavasnik,A. (1996). Sim-
topsy studies of suicide: A systematic review. Psy- ilarities and differences in the process of bereave-
chological Medicine,33, 395–405. ment after suicide and after traffic fatalities in
*Cerel,J., Fristad,M., Weller,E., & Slovenia. Omega,33, 243–251.
Weller,R. (1999). Suicide-bereaved children and *Grad,O.T., & Zavasnik,A. (1999). Phe-
adolescents: A controlled longitudinal examina- nomenology of bereavement process after suicide,
tion. Journal of the American Academy of Child and traffic accident and terminal illness (in spouses).
Adolescent Psychiatry,38, 672–679. Archives of Suicide Research,5, 157–172.
*Cerel,J., Fristad,M., Weller,E., & *Harwood,D., Hawton,K., Hope,T., &
Weller,R. (2000). Suicide-bereaved children and Jacoby,R. (2002). The grief experiences and
adolescents: II. Parental and family functioning. needs of bereaved relatives and friends of older
Journal of the American Academy of Child and Ado- people dying through suicide: A descriptive and
lescent Psychiatry,39, 437–444. case-control study. Journal of Affective Disorders,
Clark,S. (2001). Bereavement after sui- 72, 185–194.
cide—How far have we come and where do we go Henley,S.H.A. (1984). Bereavement fol-
from here? Crisis,22, 102–108. lowing suicide: A review of the literature. Current
*Cleiren,M. (1993). Bereavement and Psychological Research and Reviews,3, 53–61.
adaptation: A comparative study of the aftermath Jordan,J.R. (2001). Is suicide bereave-
of death. Washington, DC: Hemisphere Publish- ment different? A reassessment of the literature.
ing Corp. Suicide and Life-Threatening Behavior,31, 91–102.
*Demi,A.S. (1984). Social adjustment of *Kitson,G.C. (2000). Adjustment to vio-
widows after a sudden death: Suicide and non- lent and natural deaths in later and earlier life for
suicide survivors compared. Death Education,8Black and White widows. Journal of Gerontology:
(Suppl.), 91–111. Social Sciences,55B, 341–351.
*Demi,A.S.,&Miles,M.S. (1988). Sui- *Kovarsky,R.S. (1989). Loneliness and
cide bereaved parents: Emotional distress and disturbed grief: A comparison of parents who lost
physical health problems. Death Studies,12, 297– a child to suicide or accidental death. Archives of
307. Psychiatric Nursing,3, 86–96.
*Dyregrov,K., Nordanger,D., & Dyre- *Lohan,J., & Murphy,S.A. (2002). Fam-
grov,A. (2003). Predictors of psychosocial dis- ily functioning and family typology after an ado-
tress after suicide, SIDS and accidents. Death lescent or young adult’s sudden violent death.
Studies,27, 143–165. Journal of Family Nursing,8, 32–49.
Egger,M., Schneider,M.,&Smith,G.D.McIntosh,J.L. (1986). Survivors of sui-
(1998). Spurious precision? Meta-analysis of ob- cide: A comprehensive bibliography. Omega,16,
servational studies. British Medical Journal,316,335–370.
140–144. McIntosh,J.L. (1993). Control group
Ellenbogen,S., & Gratton,F. (2001). studies of suicide survivors: A review and critique.
Do they suffer more? Reflections on research Suicide and Life-Threatening Behavior,23, 146–161.
comparing suicide survivors to other survivors. McIntosh,J.L. (1996). Survivors of sui-
Suicide and Life-Threatening Behavior,31, 83–90. cide: A comprehensive bibliography update, 1986–
*Farberow,N.L., Gallagher,D.E., 1995. Omega,32, 147–175.
Gilewski,M.J., & Thompson,L.W. (1987). An *McIntosh,J.L., & Kelly,L.D. (1992).
examination of the early impact of bereavement Survivors’ reactions: Suicide vs. other causes. Cri-
on psychological distress in survivors of suicide. sis,13, 82–93.
The Gerontologist,27, 592–598. *McNiel,D.A., Hatcher,C., & Reubin,
*Farberow,N.L., Gallagher-Thompson,R. (1988). Family survivors of suicide and acciden-
D., Gilewski,M., & Thompson,L. (1992a). tal death: Consequences for widows. Suicide and
Life-Threatening Behavior,18, 137–148.Changes in grief and mental health of bereaved

Sveen and Walby 29
*Miles,M.S.,&Demi,A.S. (1991–1992). *Reed,M.D. (1998). Predicting grief
A comparison of guilt in bereaved parents whose symptomatology among the suddenly bereaved.
children died by suicide, accident or chronic dis- Suicide and Life-Threatening Behavior,28, 285–301.
ease. Omega, 24, 203–215. *Reed,M.D.,&Greenwald,J.Y. (1991).
*Murphy,S.A., Braun,T., Tillery,L., Survivor-victim status, attachment and sudden
Cain,K.C., Johnson,L.C., & Beaton,R.D.death bereavement. Suicide and Life-Threatening
(1999). PTSD among bereaved parents following Behavior, 21, 385–401.
the violent deaths of their 12- to 28-year-old chil- *Saunders,J.M. (1981). A process of be-
dren: A longitudinal prospective analysis. Journal reavement resolution: Uncoupled identity. Western
of Traumatic Stress,12, 273–291. Journal of Nursing Research,3, 319–332.
Murphy,S.A., Johnson,L.C., & Lohan,*Seguin,M., Lesage,A., & Kiely,M.C.
J. (2003). Challenging the myths about parents’ (1995). Parental bereavement after suicide and ac-
adjustment after the sudden, violent death of a cident: A comparative study. Suicide and Life-
child. Journal of Nursing Scholarship,35, 359–364. Threatening Behavior,25, 489–498.
*Murphy,S.A., Johnson,L.C., Wu,L., *Sherkat,D.E., & Reed,M.D. (1992).
Fan,J.J., & Lohan,J. (2003a). Bereaved parents’ The effects of religion and social support on self-
outcomes 4 to 60 months after their children’s esteem and depression among the suddenly be-
death, by accident, suicide or homicide: A com- reaved. Social Indicators Research,26, 259–275.
parative study demonstrating differences. Death *Silverman,E., Range,L., & Over-
Studies, 27, 39–61. holser,J. (1994–1995). Bereavement from suicide
*Murphy,S.A., Tapper,V.J., Johnson,as compared to other forms of bereavement.
L.C., & Lohan,J. (2003b). Suicide ideation Omega,30, 41–51.
among parents bereaved by the violent deaths of *Smith,P.C., Range,L.M., & Ulmer,A.
their children. Issues in Mental Health Nursing,24,(1991–1992). Belief in afterlife as a buffer in sui-
5–25. cidal and other bereavement. Omega,24, 217–225.
*Nelson,B., & Frantz,T. (1996). Family *Stone,H.W. (1972). Suicide and grief.
interactions of suicide survivors and survivors of Philadelphia: Fortress Press.
non-suicidal death. Omega,33, 131–146. Stroebe,M.S., & Stroebe,W. (1989–
Ness,D.E., & Pfeffer,C.R. (1990). Se- 1990). Who participates in bereavement research?
quelae of bereavement resulting from suicide. A review and empirical study. Omega,20, 1–29.
American Journal of Psychiatry,147, 279–285. *Thompson,K.E., & Range L.M. (1990–
*Pennebaker,J.W., & O’Heeron,R.C.1991). Recent bereavement from suicide and other
(1984). Confiding in others and illness rate among deaths: Can people imagine it as it really is?
spouses of suicide and accidental-death victims. Omega,22, 249–259.
Journal of Abnormal Psychology,93, 473–476. *Thompson,K.E., & Range L.M. (1992–
*Pfeffer,C.R., Karus,D., Siegel,K., & 1993). Bereavement following suicide and other
Jiang,H. (2000). Child survivors of parental death deaths: Why support attempts fail. Omega,26,
from cancer or suicide: Depressive and behavioral 61–70.
outcomes. Psycho-Oncology,9, 1–10. Torgersen,S. (1997). Genetic basis and
*Range,L.M., & Calhoun,L.G. (1990). psychopathology. In S. M. Turner & M. Hersen
Responses following suicide and other types of (Eds.), Adult psychopathology and diagnosis (pp. 58–
death: The perspective of the bereaved. Omega,85). New York: Wiley.
21, 311–320. *Ulmer,A., Range,L.M., & Smith,P.C.
*Range,L.M., & Niss,N.M. (1990). (1991). Purpose in life: A moderator of recovery
Long-term bereavement from suicide, homicide, from bereavement. Omega,23, 279–289.
accidents and natural deaths. Death Studies,14,Van der Wal,J. (1989). The aftermath of
423–433. suicide: A review of empirical evidence. Omega,
*Reed,M.D. (1993). Sudden death and
20, 149–171.
bereavement outcomes: The impact of resources
World Health Organization. (1999).
on grief symptomatology and detachment. Suicide
and Life-Threatening Behavior,23, 204–220. Figures and facts about suicide. Geneva.