WILLIAM M. PATTERSON, M.D.

HENRY H. DOHN, M.D.

JULIAN BIRD, M.A.. M.R.C.P., M.R.C. Psych.
GARY A. PATTERSON, M.S.

Evaluation of suicidal patients:

The SAD PERSONS scale
ABSTRACT: An easily learned scale utilizing a brief acronym identifying suicidal patients, the
(SAD PERSONS), and based on ten major risk factors, is evaluative factors are often subjec-
presented for assessing likelihood of a suicide attempt. A group tive and are sometimes applied in a
of medical students who were taught SAD PERSONS haphazard manner.
demonstrated a significantly greater ability to accurately evaluate Suicide is a major public health
problem. About 22,000 suicides are
and make recommendations for disposition of a low-risk and a
recorded annually in the United
high-risk patient, as judged by three experienced psychiatrists. States, making suicide the ninth
Conversely, a control group of students, who were not instructed overall cause of death.' It is the
in use of the scale, rated both patients at a higher risk for suicide, second most common cause of
and made dispositions accordingly. SAD PERSONS has a death on college campuses, the
positive influence on performance in evaluating suicidal patients. leading cause of death among
medical students, and the third
A consultation-liaison psychiatrist intent have been identified. Most highest cause of death among teen-
is called upon to evaluate a wide clinicians rely on their own experi- agers. The tragedy is compounded
variety of patients in a general hos- ence and frequently have their own by heavy economic losses, family
pital setting, including those criteria for making a judgment. turmoil, and psychological distress
thought to be suicidal. For the lat- They may include intuition, as well for the survivors.
ter, comprehensive evaluation in- as specific items such as a history of
volves weighing the risk factors re- previous attempts or knowing Risk factors
lated to suicidal potential. A whether the patient has an organ- The best intervention for suicide is
number of factors believed to have ized plan. Although clinical judg- prevention, which depends on
predictive value in terms of suicide ment is certainly important in prompt and thorough evaluation of
the potentially suicidal patient.
Since it is often the physician in the
Dr. Patlerson is associate professor of psychiatry, and at the time of this work Dr.
Doh" was chiefpsychiatric resident and Dr. Bird was visiting professor in psychiatry, office, emergency room, or general
all at the University ofAlabama School of Medicine. Mr. Patterson is director ofthe hospital who initially evaluates the
Office of Children's Affairs, Department of Health and Human Services, Columbia, suicidal patient, this is the physi-
S.c. Dr. Dohn is now at Eglin AFB, Florida and Dr. Bird has returned to London, cian who should be acquainted
England. Reprint requests to Dr. Patlerson at the Smolian Clinic, Room 210, with the following ten major risk
Department of Psychiatry, University Station, Birmingham, AL 35294. factors that predict the possibility

APRIL 1983 • VOL 24 • NO 4 343

Suicidal patients

of eventual suicide. with the occurrence of cirrhosis. by suicide notes. In fact, over two
Sex. Women attempt suicide Later studies7 corroborate these thirds of the suicides in one studyl4
three times more frequently than findings and point up the impor- had communicated suicidal ideas,
men, but men actually kill them- tance of alcoholism in the suicide 41% specifically stating their inten-
selves more than three times as attempter. tions.
often as do women.1.2 Rational thinking loss. Any psy- The extent of planning of the
Age. There is a bimodal distribu- chosis (ie, schizophrenia, manic- attempt and the lethality of the
tion ofsuicide with peaks in the late depressive illness, organic brain plan are both factors requiring
teenage and elderly years. 1 The syndrome) presents a hazard to the greater degrees of organization.
high-risk category proposed by patient if judgment and rational Rosenl 5 studied a large group of
Tuckman and Youngman 3 includes thought are severely impaired and suicide attempters and found that
45 or more years of age and being if he or she is delusional or hallu- such organization meant that the
male as important risk factors. Pa- cinating. It is difficult to establish serious attempter had a higher sui-
tients 19 or younger or 45 or older rapport and trust with these indi- cide risk than the less serious one.
are thus considered at higher risk viduals, not to mention the ever No spouse. Divorced, widowed,
for suicide. present danger of their hearing separated, or single patients have a
Depression. Affective disorders voices with instructions to hurt or higher risk of suicide than married
and their relation to suicide have kill themselves. Several authors2.7 patients, especially those with chil-
been examined extensively by have reported the presence of psy- dren to reinforce the marriage. '·3
Guze and Robins4 in their review of chosis as a risk factor for suicide. Sickness. The role of physical ill-
17 studies. They showed that affec- ness is somewhat inconsistent in
tive disorder is a significant factor studies 1 of suicidal patients, but
in 12% to 60% of suicides. The Twenty-five percent to 50% chronic, debilitating, and severe ill-
average suicide rate was 30 times ofpersons who kill ness is thought to be a definite risk
greater in depressives compared themselves have previously factor. 2 Suicidal patients often seek
with the general population. They attempted to do so. medical attention prior to their act,
also suggested that about 15% of with illness often being directly in-
primary depressives die by suicide. volved in the outcome. 16
OthersS·8 have also found a diagno- Social supports lacking. The sui- Other factors. Other characteris-
sis of depression to be a good pre- cidal patient often lacks significant tics have been mentioned as suicide
dictor of suicidal behavior. others (relatives, friends), employ- risk factors, but they seem to play
Previous attempt. Numerous ment at a meaningful job, and reli- only a minor role in determining
studies9 demonstrate that 25% to gious supports. Individually these overall risk. Such items include
50% of persons who kill themselves supports may not be statistically race,1 geographic region of resi-
have previously attempted to do so. significant, but collectively they dence,' religion, 1 occupation,1 drug
In fact, suicide rate by previous represent a major and important abuse,' defenselessness,17 season,18
attempters is up to 64 times higher factor to assess in suicidal patients. and genetic predisposition.1 9 How-
than the overall rate in the general Many authors l.3 comment on this, ever, they are inconsistent, less well
population. 10.1 I A prior attempt not especially in regard to recent loss of documented, and of less practical
only increases the risk in the year a social support. importance to the average physi-
following the attempt, but also in- Organized plan. The person with cian who is evaluating the poten-
creases the subsequent lifetime sui- a well-delineated, specific plan en- tially suicidal patient.
cide rate. 5 visaging a lethal available method While it is recognized that a phy-
Ethanol abuse. Early studies6. '2 is at a far greater risk than sician cannot absolutely predict
indicated an increased suicide rate others. n'3 The degree of organiza- whether a patient will commit sui-
in alcoholics. In fact, it is estimat- tion is often shown by direct or cide, he is regularly expected to
ed '2 that 15% of all alcoholics com- indirect communication of suicidal make decisions based on such pre-
mit suicide, the suicide often com- intent, by seeking medical care or dictions. Many nonpsychiatrists are
ing as a late complication of chron- counseling prior to the attempt, by frightened by or are uncomfortable
ic alcoholism and correlating well "getting one's affairs in order," and with patients who verbalize or act

344 PSYCHOSOMATICS
out suicidal intent, and they are prediction. There are various other SAD PERSONS scale
anxious to delegate this decision- scales to aid in suicidal assess- As a device to assist memory in
making responsibility to someone ment,22.26 but many of them are too recalling a set of facts, the acronym
else. Psychiatrists, who are fre- complex or cumbersome for practi- is often used for teaching purposes
quently faced with this problem, cal and routine use. They also do in medicine. One such acronym de-
tend to approach it with more con- not readily lend themselves to the veloped and utilized very success-
fidence and less apprehension than teaching of the medical students fully on our consultation-liaison
their non psychiatric colleagues. and nonpsychiatric physicians with service consists of the two words
However, psychiatrists also often whom the liaison psychiatrist SAD PERSONS (Table 1), each
find these judgments difficult. For comes in contact. letter of which denotes one of the
example, the severely psychotically Many brief articles can be found ten major risk factors mentioned
depressed, elderly male who has a in the medical literature to aid the earlier.
chronic illness and expresses a de- physician in evaluation of the sui- This device is readily accepted in
sire to kill himself is much less cidal patient. What these articles a teaching context, is easy to learn,
problematic than the 30-year-old lack is a simple, clear-cut, and and gives the student or resident a
histrionic woman who has gestured practical guide to suicide assess- sense of confidence in decision-
before and is threatening to do so ment. The suicidal patient is prone making. It also ensures that he or
again. to produce adverse emotional reac- she has considered each factor ac-
tions in the attending physician,27.28 cepted as important in assessing
Assessment scales hence hampering objectivity and suicide potential. Lastly, it serves as
A consultation-liaison service in a thoroughness, and so it is essential a bridge to the development of a
teaching hospital has a major re- that an objective and complete base of clinical experience in deal-
sponsibility for teaching medical method for the physician to evalu- ing with the suicidal patient. The
students as well as psychiatric and ate these frequent, and sometimes total score is used in decision-mak-
nonpsychiatric residents. One com- upsetting, patients be devised. ing (Table 2), and is intended to
mon teaching task concerns the correlate with clinical assessment
evaluation of suicide intent or de- and disposition.
termination of the probability of One important exception applies
suicide in a particular patient. to this device. Factors can pair up
Clinical intuition can be important in a critical fashion that overrides
in a physician's thought process, the guidelines and requires addi-
but it is very difficult to teach it to tional clinical judgment. For exam-
others. We therefore are left with ple, a 14-year-old girl who at-
developing a way to teach such tempted to hang herself "because
assessment in an effective manner. the devil came and told me to"
Numerous clinical scales are might score only three points on the
available for assessing the suicide scale. This case would obviously
potential of patients.'·3.2o.26 Tuck- require closer scrutiny because of
man and Youngman3 devised a 17- the aggregation of critical factors.
factor system for identifying high-
and low-risk patients. Litman and Validation of the scale
associates 13 studied profiles of tele- Two groups of third-year medical
phone callers to the Los Angeles students were asked, based on clin-
Suicide Prevention Center to de- ical judgment, to assess the inclina-
velop a 15-item scale of suicidal One point Is scored for each tion toward suicide of two patients
risk. Beck and associates 21 prepared factor deemed present. The appearing in videotape interviews.
the Scale for Suicidal Ideation, total score thus ranges from 0 All students were at the same point
(very little ) to 10 (very high
which involves 19 items. Weisman of training during their clinical ro-
risk).
and Worden 25 developed the Risk- tation in psychiatry. An investiga-
Rescue Rating Scale for suicide tional group of 36 students was
(continued)
APRIL 1983 • VOL 24 • NO 4 345
 RestfuI~
alert awakening G
I.;
~ 15mg
30mg
capsules Suicidal patients

(temazepam)
A more appropriate half-life taught use of the SAD PERSONS o to 10 rating scale, and the other
scale during a lecture on emergency representing 1 to 4 for the four
One3(}.mg capsule. h.s.-usua/adundosage.
psychiatry given by the chief psy- available dispositions. The 0 to 10
One l>mg capsule. h.s.-recommended initialdosaga chiatry resident, and a control scale was analyzed with a t test for
lor elderlyandlor debilitated patients.
IIIDICATIOIIS AIID USAO.: Restoril· (temaze- group of 21 students was given ba- significance to determine if use of
pam) is indicated lor the reliel of insomnia associated sically the same lecture but without the SAD PERSONS scale would
with complaints of difficulty in falling asleep. Irequent
nocturnal awakenings. and lor earty morning awaken- the acronym. have a positive impact on a stu-
ings. Since insomnia is often transient and intermit-
tent. the prolonged administration of Restoril is gen- One week following the lecture, dent's clinical judgment when eval-
erally not necessary or recommended. Restoril has
been employed lor sleep maintenance lor up to 35 each group viewed the two video- uating for suicide potential.
conseCutive nights 01 drug administration in sleep lab- tapes. One involved a 29-year-old A comparison between the mean
oratory studies.
The possibility that the insomnia may be related to a woman judged by three psychia- ratings by the control and inves-
condition lor which there is more specilic treatment
should beconsidered. trists to be at low risk for suicide. tigational groups for the low-risk
COIIT..AIIIDICATIOIIS: Benzodiazepines may The second showed a 29-year-old patient yielded a significant dif-
cause letal damage when administered during preg-
nancy. An increased risk 01 congenital malformations depressed man, determined by the ference (P<.O 1), with the control
associated with the use 01 diazepam and chlordiaz-
epoxide during the lirst trimester 01 pregnancy has same three psychiatrists to be at group tending to rate this woman as
been suggested in several studies. Also. ingestion 01
therapeutic doses 01 benzodiazepine hypnotics during
significant risk for suicide. The more seriously suicidal than did the
the last weeks 01 pregnancy has resulted in neonatal three psychiatrists were experi- investigational group (mean rat-
CNS depression. Restoril is contraindicated in preg-
nant women. Consider a possibility 01 pregnancy enced faculty members who also ings of 5.10 vs 3.06, respectively).
when instituting therapy or whether patient intends to
become pregnant.
viewed the videotapes, and rated The rating of 3.06 differs little from
WA..III110S: Patients receiving Restoril (temaze- the low-risk and high-risk individ- the 3 arrived at by the three psychi-
pam) should be cautioned about possible combined
effects with alcohol and other CNS depressants. uals as scoring 3 and 5 respectively atrists in their own separate assess-
P ...CAUTIOIIS: In elderly and lor debilitated
patients. it is recommended that initial dosage be lim-
on the SAD PERSONS scale. ment.
ited to 15 mg. The usual precautions are indicated lor Following each videotape, each A significant difference (P<.OI)
severely depressed patients or those in whom there is
any evidence 01 latent depression; it should be recog- group rated the patient for suicide was also obtained with the ratings
nized that suicidal tendencies may be present and pro- risk on a scale of 0 to 10, and for the high-risk man. With a con-
tectivemeasures may be necessary.
If Restorit is to be combined with other drugs hav- selected one of the four available trol-group mean of 8.14 and an
ing known hypnotic properties or CNS-depressant
effects. due consideration should be given to potential dispositions listed in the guidelines investigational-group mean of 4.94,
additiveeffects.
In/ormation lor Patients: Patients receiving Restoril in Table 2. the control group again rated the
should be cautioned about possible combined effects
with alcohol and other CNS depressants. Patients
patient at greater risk than did the
should be cautioned not to operate machinery or drive Results investigational group even though
a motor vehicle. They should be advised ot the possi-
bility 01 disturbed nocturnal sleep tor the tirst or sec- The data generated by the inves- the latter gave this at-risk patient a
ond night alter discontinuing the drug. tigational and control groups were high enough rating, very close to
Laboratory Tests: The usual precautions should be
observed in patients with impaired renal or hepatic separated into two independent the 5 given by the three psychia-
tunction. Abnormal liver tunction tests as wetl as blood
dyscrasias have been reported with benzodiazepines. measures with one representing the trists. The same trends were evident
Pregnancy: Pregnancy Category X. See Contraindica-
tions.
Pediatric Use: Safety and effectiveness in children
below the age 01 18years have not been established.
ADV."S. . . . ACTIOIIS: The most common
adverse reactions _e drowsiness. dizziness and leth-
argy. Other side effects include conlusion. euphoria Table 2-Guldellne. for Action with the Scale
and relaxed leeling. Less commonly reported were
weakness. anorexia and diarrhea. Rarely reported
were tremor. ataxia. lack 01 concentration. loss 01
equilibrium. lalling and palpitations. And rarely Total poIn" PropoHd cllnlce' .ctlon
reported were hallucinalions. horizontal nystagmus
and paradoxical reactions. including excitement, stim-
ulation and hyperactivity. oto 2 send home with follow up
Restoril is a controlled substance in Schedule IV.
Caution must be exercised in addiction-prone individ- 3 to 4 Close follow up;
uals or those who might increase dosage.
DOSAO. AIID ADMIIIISTRATIOII: Adults: 30 mg
Consider hospitalization
usual dosage belore retiring; 15 mg may suffice in
some. Elderly andlor debililated: 15 mg recommended 5 to 6 Strongly consider hospitalization.
initially until individual response is determined. depending on confidence in
SUPPU.D: Restoril (temazepam) capsules-15 mg
maroon and pink. imprinted "ResToR1L 15 mg"; the follow-up arrangement
30 mg. maroon and blue. imprinted "ResTORIL
30 mg". Packages 01100. 500 and ControlPak· pack- 7 to 10 Hospitalize or commit
ages 01 25 capsules (continuous reverse-numbered roll
01 sealed blisters). (RES-Z2 11/1/81)
Be/ore prescribing. see package insert lor /ull producl
in/ormation.

RES·183-2 A Pt8maceulkal DIvision

SANDOZ, INC.
East Hanover. NJ 07936
348 PSYCHOSOMATICS
for the disposition ratings as well. as confirmed by three experienced SAD PERSONS is a simple ac-
psychiatrists. Likewise, they dem- ronym that can be taught to medi-
Discussion onstrated the ability to select an cal students in one lecture. It is easy
An interesting problem encoun- appropriate disposition based on to remember and positively influ-
tered in conducting theinvestiga- evaluation of these risk factors. In ences a medical student's assess-
tion was that because of the popu- general, they were more conserva- ment of a potentially suicidal pa-
larity and widespread use of the tive than the control group; there- tient. Its application in a teaching
scale among faculty and housestaff fore, they would tend to hospitalize program can contribute to in-
since its introduction only 18 only those persons judged to be at creased clinical acumen in a com-
months ago, it was difficult to pre- high risk. The control group rated plex diagnostic area. 0
vent potential control-group stu- both the low-risk and high-risk
dents from hearing about it. subjects as more at hazard, and
Students in the investigational therefore would tend to elect possi- We gratefully acknowledge the assis-
group more accurately assessed bly unnecessary hospitalization for tance of Drs. Arthur Freeman III and
suicidal risk factors in two patients, low-risk persons as well. Steven Cohen-Cole.

REFERENCES 10. Peterson A, Arvad G. Kindler A: Epidemiolog- 19. Stallone F. Dunner DL. Ahearn J. et al: Statis-
1. Resnick HLP: Suicide. in Kaplan HI, Freed- ical differences between white and nonwhite tical predictions 01 suicide in depressives.
man AM, and sadock BJ (eds): Comprehen- suicide attempts. Am J Psychiatry 130:1071- ComprPsychiatry21:381-387, t980.
sive Textbook of Psychiatry. ed 3. Bal1imore. 1076.1973. 20. Bedrosian RC, Beck AT: Cognitive aspects 01
Williams & Wilkins. 1980. pp 2085-2098. 11. Ettlinger RW: Suicides in a group of patients suicidal behavior. Suicide Life Threat Behav
2. Farberow NL. Lilman RE: Suicide prevention. who had previously attempted suicide. Acta 9:87-96.1979.
in Resnick HLP, Ruben HL (eds): Emergency Psychiatr Scand 40:363-378. 1964. 21. Beck AT. Kovacs J. Weissman A: Assess·
Psychiatric Care. Bowie. Md, Charles Press. 12. Porterfield AL: Ecological correlates 01 alco- ment 01 suicidal intention: The scale 01 sui·
1975. pp 103-188. holism. Societal Probs 5:326-338, 1958. cidal ideation. J Consult Clin PsychO/4 7:343-
3. Tuckman J. Youngman WF: A scale lor as- 13. Litman RE. Farberow NL, Wold CI, et al: 352. 1979
sessing suicide risk 01 attempted suicidllS. J Prediction models of suicidal behaviors. in 22. Pierce OW: Suicidal intent in sell-injury. Br J
ClinPsycho/24:17-19.1968. Beck AT. Resnick HLP. Lettieri OJ (eds): Psychiatry 130:377-385.1977.
4. Guze S. Robins E: Suicide and primary affec- Prediction of Suicide. Bowie, Md. Charles 23. Buglass D. Horton J: A scale lor predicting
tive disorder. Br J Psychiatry 117:437-483. Press, 1974. pp 141-159. subsequent suicidal behavior. Br J Psychiatry
1970. 14. Robins E. Gassner S. Kayes J. et al: The 124:573-578, 1974.
5. Avery D. Winokur G: Suicide, attempted sui· communication of suicidal intent: A study 01 24. Farberow NL. McKinnon 0: A suicide predic·
cide and relapse rates in depression. Arch 134 consecutive cases of successful (com- tion schedule for neuropsychiatric hospital
Gen Psychiatry 35:749-753, 1978. pleted) suicide. Am J Psychiatry 115:724- patients. J Nerv Ment Dis 158:408-419. 1974.
6. Robins E. Murphy GE. Wilkinson RH. et al: 733. 1959. 25. Weisman AD. Worden JW: Risk·rescue rating
Some clinical considerations in the preven· 15. Rosen DH: The serious suicide attempt. in suicide assessment. Arch Gen Psychiatry
tion 01 suicide based on a study of 134 JAMA 235:2105-2109.1976. 26:553-560,1972.
successlul suicides. Am J Public Health 16. Dorpat T, Anderson WF. Ripley HS: The rela· 26. Bagley C. Greer S: Clinical and social predic·
49:888-899. 1959. tionship of physical illness to suicide. in Res- tors of repeated attempted suicide: A mul·
7. Miles CP: Conditions predisposing to suicide: nick HLP (ed): Suicidal Behaviors. Boston. tivariate analysis. Br J Psychiatry 119:515-
A review. J Nerv Ment Disease 164:231-246. Little Brown, 1968. pp 209-219. 521,1971.
1977. 17. Pokorny AD, Kaplan HB: Suicide following 27. Dressler OM. Prusoff B. Mark H. et al: Clinical
8. Minkoff K, Bergman E. Beck AT, et al: Hope- psychiatric hospitalization. J Nerv Ment Dis altitudes toward the suicide attempter. J Nerv
lessness. depression and attempted suicide. 162:119-125.1976. Mental Dis 160:146-155. 1975.
Am J Psychiatry 130:455-459.1973. 18. Wenz FV: Effect 01 seasons and sociological 28. Mallsberger JT. Buie DH: Countertrans-
9. Patel NS: Attempted and completed suicide. variables 'on suicidal behavior. Public Health terence hate in the treatment of the suicidal
Mad Sci Law 14:273-279. 1974. Rep 92:233-239, 1977. patient. Arch Gen Psychiatry 30:625-633.
1974.

APRIL 1983 • VOL 24 • NO 4 349