High Dose Chlorpromazine Therapy

in Chronic Schizophrenia
Report of National Institute of Mental Health\p=m-\
Psychopharmacology Research Branch Collaborative Study Group
Robert F. Prien, PhD, and Jonathan O. Cole, MD, Washington, DC

THIS STUDY on the efficacy of high the treatment of chronic schizophrenia,

dose chlorpromazine treatment in chronic there is considerable confusion and contra
schizophrenia was developed under the Na- diction regarding matters of dosage, partic
tional Institute of Mental Health (NIMH) ularly the value of high dose schedules.
psychopharmacology program during 1964 Some investigators advocate the use of high
and was initiated in seven collaborating doses, while others contend that low doses
public mental hospitals early in 1965. The are just as effective and less dangerous.19 A
hospitals participating in this study were review of the literature reveals a surprising
Boston State Hospital, Boston; Broughton paucity of controlled studies comparing
State Hospital, Morganton, NC; Dorothea high and low doses in the chronic schizo
Dix State Hospital, Raleigh, NC; Kentucky phrenic patient. Several small studies in the
State Hospital, Danville, Ky; Manhattan mid-1950's compared different dose levels of
State Hospital, New York; St. Louis State chlorpromazine, but none of these studies
Hospital, St. Louis; and Springfield State used doses exceeding 600 mg/day. Only one
Hospital, Sykesville, Md. These hospitals comparative study reported in the literature
were selected to represent the entire urban- used relatively high doses of chlorproma
rural continum. Three hospitals admitted pa- zine. Schiele9 compared a high dose regi
tients exclusively from large urban centers, men of 1,400 to 3,000 mg with a low dose
two hospitals served areas which included regimen of 300 to 1,400 mg in 38 male
both urban and rural communities, and two chronic schizophrenics. While the results
hospitals served areas which were almost ex- suggested that the high dose treatment was
clusively rural. more effective, the authors acknowledged
that methodological considerations made any
Background interpretation of results tenuous.
There is also disagreement among inves
During the past 13 years, a large number tigators regarding optimum duration of
of studies have been published on the use of treatment. Several investigators stress the
chlorpromazine with chronic schizophrenic importance of administering high doses to
patients. While a majority of these studies chronic patients for several months.38-10
indicates that chlorpromazine is effective in They feel that many authors have drawn
Submitted for publication Sept 11, 1967. premature conclusions from experience with
From the Psychopharmacology Research Branch, high dose therapy administered for too short
National Institute of Mental Health (Drs. Prien a period of time.
and Cole), and the Biometric Laboratory, George
Washington University, Washington, DC (Dr. Prien). Thus, after more than a decade of re
Read before the 12th annual conference of the search on chlorpromazine, the question of
Veterans Administration Cooperative Studies in Psy- high dose efficacy remains essentially unan
chiatry, Denver, April 1, 1967.
Reprint requests to 1145 19th St NW, Washington, swered. This question has important thera
DC 20036 (Dr. Prien). peutic implications. In most public mental

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 hospitals, the majority of chronic schizo
phrenics treated with low doses of phe
are
nothiazines. A presumed but unknown pro
portion have never received phenothiazine
therapy a very high dosage for a long pe
at
riod of time. The principal aim of this
study was to determine whether these
chronic schizophrenics would show major
improvement if high doses were adminis
tered in a carefully supervised manner over
a prolonged period of time.
Research Questions
This study was designed to respond to
the following questions:
1. Is a high dose of 2,000 mg of chlorpro
mazine administered over a prolonged peri
od of time more effective than the low dose
of 300 mg typically used in the treatment of
chronic schizophrenia?
2. Is high dose chlorpromazine more ad
vantageous than treatment routinely admin
istered to the chronic schizophrenic in the
public mental hospital?
3. Are the two dose levels of chlorproma
zine and the routine hospital treatment
more effective than an inert placebo?
4. Do the treatments operate differential
ly specific schizophrenic symptoms? For
on
example, is high dose more effective than
low dose in reducing delusional thinking?
5. Is there a type of patient who re
sponds to one treatment and not to anoth
er?
6. What proportion of chronic patients
show an exacerbation of symptoms follow
ing withdrawal of regular medication and
initiation of placebo?
7. Are there differences between treat
ments in the nature and/or frequency of
side effects?
Procedure
120 chronic schizophrenics,
Approximately
half male and half female, were selected at
each of the seven hospitals by the following
criteria:
1. A primary diagnosis of schizophrenia.
2. Age between 19 and 55.
3. Continuous hospitalization of at least two
years. Continuous hospitalization was defined
as no more than eight weeks out of the hospi
tal at any one time during the current stay.
Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015
4. No evidence of the following clinical dis
orders:
a. Organic brain disease, including loboto-
my.
b. Mental Deficiency (intelligence quo
tient below 70).
c. Medical conditions contraindicating use
of high doses.
In addition, patients were required to show
symptoms, signs, or behavior in one or more of
the following categories: perceptual disturb
ance, delusional thinking, thinking and speech
disorganization, withdrawal, mannerisms and
posturing, and lack of social competence.
Patients meeting these selection criteria
were randomly assigned to four treatment
groups: (1) high dose2,000 mg of chlorpro
mazine per day; (2) low dose300 mg of
chlorpromazine per day; (3) placebo; and (4)
routine-conventional hospital treatment. The
latter group consisted of patients who received
whatever medication or dose the hospital chose
to administer. The total number of patients in
each treatment group was as follows: high
dose, 208; low dose, 208; placebo, 212; and rou
tine treatment, 210.
All patients were rated on psychiatric, nurs
ing, and social work scales prior to entry into
the study. Patients were then observed on
their routine hospital medication for eight
weeks during which time physical and oph-
thalmological examinations were performed. At
the end of the eight week baseline period, rat
ings were repeated and patients who had been
assigned to high, low, and placebo groups were
shifted to study medication. The routine treat
ment group continued to receive hospital-de
termined medication. The high dose group re
ceived liquid concentrate chlorpromazine in
gradually increasing dosage according to a
predetermined schedule, reaching the maxi
mum level of 2,000 mg after 45 days. Patients
were maintained at that level for the remain
der of a 24-week period. Permanent dosage re
duction to 1,500 mg was permitted to control
for side effects. The low dose and placebo
group also received liquid concentrate, on a
double-blind basis, for a 24-week period.
Assessment of Clinical Change
Clinical change was assessed in three
First, the doctors made overall clini
ways.
cal judgments of severity of illness and de
gree of improvement. Second, specific psy
chopathology was rated by the doctors on
the basis of interviews and by nurses on the
basis of ward behavior. Third, readiness for
discharge was evaluated by social workers
 from interviews and observations of ward Scale (IMPS).This instrument developed
behavior. The assessment instruments used by Lorr et al13 consists of 81 symptom de
by the doctors, nurses, and social workers are scriptions completed by the psychiatrist or
described below. clinical psychologist on the basis of a one-
Discharge Readiness Inventory (DRI). hour interview. The scale has been factor
The DRI was developed by Hogarty et al11 analyzed to form ten independent subscales:
to evaluate release potential in the chronic 1. Excitement
schizophrenic patient. This scale was com 2. Hostile-Belligerence
pleted by the social worker twice for each 3. Paranoid Projection
patient, once prior to treatment and once at 4. Grandiose Expansiveness
the conclusion of treatment. The DRI con 5. Perceptual Distortions
sists of 62 items which have been factor ana 6. Anxious Intropunitiveness
lyzed to form four independent subscales. 7. Retardation and Apathy
Factor 1, Psychosocial Adequacy, measures 8. Disorientation
the level of patient adjustment within the 9. Motor Disturbances
hospital. Factor 2, Belligerence, measures 10. Conceptual Disorganization
the degree of hostility and obstreperous be Each patient was evaluated on the IMPS
havior. Factor 3, Community Adjustment prior to treatment and at eight-week inter
Potential, measures the patient's capacity vals during treatment.
for adjustment in the community. Factor 4, Nurse's Observation Scale for Inpatient
Manifest Psychopathology, measures the ex Evaluation (NOSIE).This scale, devel
tent to which the patient's behavior is dis oped by Honigfeld et al,14 consists of 80
rupted by hallucinations and delusions. items for evaluating the patient's behavior
on the ward. The NOSIE has been factor
A validation study on the DRI was con
ducted prior to the collaborative study to analyzed to form seven independent sub-
determine whether the DRI actually dis scales:
criminated between the discharge-ready and 1. Social Competence
2. Social Interest
nonready patient. At five of the collaborat 3. Cooperation
ing hospitals, a total of 206 discharge-ready 4. Personal Neatness
patients were evaluated on the DRI. These 5. Irritability
patients were chronic schizophrenics who 6. Manifest Psychosis
had been referred to social service for dis
charge and had been judged discharge- 7. Paranoid Depression
ready by three independent judges: the The NOSIE was completed by the re
referring psychiatrist, the social worker ac search nurse prior to treatment and at
cepting the referral, and the research social eight-week intervals during treatment.
worker. The discharge-ready patients were Global Assessments.Two global scales
compared with chronic schizophrenics meet were completed by the doctor at the conclu
ing collaborative study criteria. All four sion of the IMPS interview:
DRI factors significantly discriminated be 1. Global Improvement Scale: "Compared
tween the two groups of patients. Among to his condition at the start of treatment, how
the four DRI factors, Community Adjust much has the patient changed?" (1) very
ment Potential and Psychosocial Adequacy much improved; (2) much improved; (3)
discriminated best between the ready and slightly improved; (4) no change; (5) mini
nonready patients. The discharge-ready pa mally worse; (6) much worse; or (7) very
tient was best characterized by high posi much worse.
tive scores on these two factors. Complete 2. Global Severity of Illness Scale: "Con
results from this validation study are pre sidering your total clinical experience, how
sented elsewhere.12 What is important here is mentally ill is the patient at this time?" (1)
that all four DRI factors, particularly Com normal, not ill at all; (2) borderline men
munity Adjustment Potential and Psycho- tally ill; (3) mildly ill; (4) moderately ill;
social Adequacy, proved to be valid meas (5) markedly ill; (6) severely ill; or (7)
ures of discharge readiness. among the most extremely ill patients.
Inpatient Multidimensional Psychiatric In addition to the DRI, IMPS, NOSIE,

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 and global scales, a 40-item side effect
checklist was completed by the research
physician and the research nurse. This
checklist was completed prior to treatment
and at two-week intervals during treatment.
Side effects were rated slight, moderate, or
severe in intensity. Each patient was also
examined by an ophthalmologist prior to
treatment, after 12 weeks of treatment, and
at the conclusion of treatment.
Method of Analysis of Clinical Data
To avoid excessive length and complexity
this paper will deal only with comparisons in
volving the high dose groupie, high dose vs
low dose, high dose vs routine treatment, and
dose vs placebo. A subsequent paper will
high
deal with comparisons between the other treat
ment groups (eg, low dose vs placebo).
Each patient was evaluated on 23 criterion
measures: ten IMPS factors, seven NOSIE
factors, four DRI factors, and two global
scales. Except for the Global Improvement
Scale, criterion measures were analyzed by
analysis of covariance computed from a factori
al design involving four treatment groups, two
sexes, two age groups (under 40, over 40), and
two levels of current hospitalization (under ten
years, over ten years). The covariates were
pretreatment scores and the varites were
scores obtained when the patient left the
study. Both early terminators and patients
who completed the full 24 weeks of treatment
were included in the analysis. Each early ter
minator was evaluated at the time he was
dropped from the study and this score was
used as the patient's variate measure in the
analysis.
The convariance analyses showed either a
significant drug main effect or significant drug
interaction on all criterion measures. On the
basis of these analyses, three additional co-
variance analyses were computed for each cri
terion measure. Each analysis was based on
two treatments, two sexes, the two age levels,
and the two lengths of hospitalization. The
first analysis compared high dose and low dose,
the second compared high dose and routine
treatment, and the third compared high dose
and placebo. When an analysis showed a sig
nificant drug interaction, a multiple range
test15 was employed to determine the signifi
cance of differences in the individual compari
sons involved in the interaction. The 5% level
of statistical significance was used in all anal
yses reported in this paper.
On each criterion measure, 2-tailed t-tests
Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015
for correlated samples were used to determine
significance of change from pretreatment to
posttreatment. Analyses were performed for
each treatment group at each level of age and
hospitalization.
Results
Characteristics of the Sample at Start of
Study.A description of the background
characteristics of study patients is pre
sented in Table 1. The mean age of patients
was 41.6 years, 61% being over 40 years of
age. Total hospitalization ranged from 2 to
34 years with a mean of 14.5 years. The
mean duration of current hospitalization
was 13.1 years; 57% of the patients had
been hospitalized over ten years during the
current stay.
The analysis of the social characteristics
of the sample revealed that 84% of the men
and 58% of the women had never been
married. Twenty percent of the men an
40% of the women had never been em
ployed, while an additional 49% of the men
and 19% of the women had worked only at
unskilled jobs. Educational achievement was
relatively poor; only 17% of the men and
28% of the women had completed high
school.
Pretreatment symptomatology was ana
lyzed by age and length of current hospital
ization. The results showed that patients
hospitalized less than ten years were
markedly different from patients hospital
ized over ten years. The less hospitabzed
patients showed significantly more florid
schizophrenic symptomatology (conceptual,
delusional, and perceptual disturbance), so
cial competence, and social interest than the
patients hospitalized over ten years. The
latter group of patients were for the most
part relatively inactive, showed severe so
cial withdrawal, and exhibited little affect
either in the form of excitement or depres
sion.
Global Improvement Scale.Marked im
provement occurred in 10% of the high dose
group, 6% of the low dose group, 3% of the
placebo group, and 3% of the routine treat
ment group (for convenience, a patient rated
"much improved" or "very much improved"
on the global scale is referred to as markedly
improved). An additional 17% of the high
dose group, 12% of the low dose group, 8%
 Table 1.Background Characteristics of Study Patients The high dose treatment
was more effective than each
Characteristics Male Female Total
of the other treatments with
Age at admission Mean 40.8 42.4 41.6
to project SD 9.0 7.7 8.4 this subgroup of patients.
% over 39 58 66 62 Chi square analyses re
Age at first Mean 23.5 25.0 24.2 vealed that the high dose
hospitalization SD 5.9 6.8 6.4
Years current Mean 13.3 13.0 13.1
group had a significantly
hospitalization SD 8.5 8.7 8.6 higher proportion of marked
% over 10 56 54 55 ly improved patients than
Years total Mean 14.6 14.4 14.5 the routine treatment group
hospitalization SD 8.1 8.1 8.1
10 63 63 63 and placebo group. The high
% over
% over 20 26 26 26 dose also produced more
Race % White 89 90 89 marked improvement than
Marital status o married now 42 28 the low dose group, but this
or ever
only approached significance

Education % attending 43 54 48
high school at the 0.1 level. When all de
% completing 17 28 22 grees of improvement (slight
high school plus marked) were included
Work history % never employed 20 40 29
% employed only 49 19 35
in the analysis, the high dose
in unskilled jobs group showed a significant
ly higher incidence of im-
of the placebo group, and 14% of the routine provement than each of the other treat
treatment group were found to be slightly ments.
improved. Younger, More Hospitalized Patients
Table 2 presents global ratings by treat (Under 40, Hospitalized Over Ten Years).
ment group, age, and length of hospitaliza This subgroup of patients also appeared to
tion. The major results from Table 2 are benefit from the high dose treatment. How
summarized below. ever, slight improvement accounted for prac
Younger, Less Hospitalized Patients (Un tically all the difference between the high
der 40, Hospitalized Under Ten Years). dose group and each of the other treatments.

Table 2.Change in Global Psychiatric State by Treatment Group,

Age, and Length of Current Hospitalization*
Years
Age in Currently Treatment Markedly Slightly Total No Total
Years in Hospital Group Nf Improved Improved Improved Change Worsened

Under 40 Under 10 High 53 10(19%) 10(19%) 20(38%) 23(43%) 10(19%)

Low 46 3 (7%) 2 (4%) 5(ll%)t 30(65%) 11(24%)
Placebo 35 1 (3%)t 3 (9%) 4(11%); 9(26%) 22(63%);
Routine 61 1 (2%) 9(15%) 10(16%)t 41(67%) 10(16%)
Under 40 Over 10 High 23 1 (4%) 7 (30%) 8(35%) 12(52%) 3(13%)
Low 32 2 (6%) 3 (9%) 5(16%) 18(56%) 9(28%)
Placebo 24 1 (4%) 1 (4%); 6(25%) 17 (71%);
Routine 21 3(14%) 3(14%) 12(57%) 6(29%)
Over 40 Under 10 High 37 5(13%) 8(22%) 13(35%) 19(51%) 5(14%)
Low 33 5(15%) 7(21%) 12(36%) 15(45%) 6 (18%)
Placebo 32 3 (9%) 2 (6%) 5(16%) 15(47%) 12(38%);
Routine 29 2 (7%) 9(31%) 11(38%) 17(59%) 1 (3%)
Over 40 Over 10 High 78 4 (5%) 7 (9%) 11(14%) 56(72%) 11(14%)
Low 77 2 (3%) 10(13%) 12(16%) 52(68%) 13(17%)
Placebo 98 1 (1%) 9 (9%) 10(10%) 60(61%) 28(28%);
Routine 83 3 (4%) 6 (7%) 9(11%) 68(82%) 6 (7%)
*
Change in global psychiatric state is determined from the Global Improvement Scale comparing the patient's
clinical condition at termination with his clinical condition prior to treatment.
t Data are presented only for patients who were evaluated by the same rater at pretreatment and termination.
; The difference between this group and the high dose group is significant at the P=0.05 level (chi square
analysis).

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 Table 3.Pretreatment and Posttreatment Ratings on the Global Severity of Illness Scale
Years Pretreatment Severity of Illness Posttreatment Severity of Illness
Currently
Age in in Treatment Marked Moder Border Marked Moder- Border-
Years Hospital Group N* Worse ate
or Mild line Worse
or ate Mild line
Under 40 Under 10 High 53 70% 25% 6% 60% 15% 17% 8%
Low 46 67% 28% 4% 78% 11 4%
Placebo 35 74% 23% 3% 91%
Routine 61 72% 15% 11% 2% 69% 18% 11% 2%
Under 40 Over 10 High 23 70% 17% 9% 4% 70% 22% 4% 4%
Low 32 72% 25% 3% 0 78% 19% 3%
Placebo 24 83% 17% 92% 8%
Routine 21 90% 90% 10%
Over 40 Under 10 High 37 70% 11% 5% 57% 24% 14% 5%
Low 33 55% 27% 12% 6% 52% 24% 17% 6%
Placebo 32 59% 31% 10% 78% 12% 6% 3%
Routine 29 58% 31% 5% 5% 45% 38% 12% 5%
Over 40 Over 10 High 78 83% 5% 73% 21 1%
Low 77 73% 25% 74% 19 7%
Placebo 98 74% 18% 2% 86% 11% 3% 1%
Routine 83 12% 6% li 92% 6% 2% 0

for patients who evaluated by the rater at pretreatment and termination

Data are presented only were same

Marked improvement occurred in only 4%
of the high dose patients.
Older, Less Hospitalized Patients (Over
40, Hospitalized Under Ten Years).The
greatest improvement on low dose and rou
tine treatment occurred with this subgroup
of patients. The high dose group also
showed a high incidence of improvement
with these patients but did not differ signif
icantly from the routine treatment and low
dose groups.
Older, More Hospitalized Patients (Over
40, Hospitalized Over Ten Years).This
high dose treatment was least effective with
this subgroup of patients. Only 14% of the
high dose patients improved as compared to
16% of the low dose patients and 10% of
the placebo patients.
There is one other noteworthy finding
from the Global scale. In the placebo group,
clinical change was clearly a function of
age: 66% of the patients under 40 were
found to be worse as contrasted with only
31% of the patients over 40.
Global Severity of Illness Scale.The co-
variance analyses of the Global Severity of
Illness Scale showed that high dose was sig
nificantly more effective than low dose and
routine treatment only with younger, less
hospitalized patients. High dose was signifi
cantly more effective than placebo at all
levels of age and hospitalization.
When pre-post change scores were tested
for significance it was found that the only
significant improvement in global severity
of illness occurred among younger, less hos
pitalized patients on high dose. Significant
worsening occurred with the placebo group
at all levels of age and hospitalization.
Despite the significant improvement ef
fected in younger, less hospitalized high
dose patients, one might ask to what extent
these patients were still mentally ill after
treatment. Table 3 gives the distribution of
pretreatment and posttreatment scores on
the Global Severity of Illness Scale. It may
be seen that improvement in younger, less
hospitalized patients consisted mainly of a
shift from marked and moderate illness to
mild and borderline illness: 25% were rated
mildly or borderline mentally ill after treat
ment as compared to 6% prior to treat
ment.
Changes in Specific Symptoms and Be
haviors.The DRI, IMPS, and NOSIE
provide measures of 21 specific symptoms
and behaviors. The results from these three
instruments correspond very closely to re
sults from the global scales. Briefly, the re
sults may be summarized as follows:
1. On all three scales high dose was
found to be the most effective treatment for
younger, less hospitalized patients. Signifi
cant treatment differences for this subgroup
of patients are shown in Table 4. It may be
seen that high dose was significantly more

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 effective than low dose and placebo in 18 of
21 symptom areas. High dose was signifi
cantly superior to routine treatment in nine
areas, including the important DRI meas
ures of Community Adjustment Potential
and Psychosocial Adequacy. In no symptom
area was high dose less effective than the
other treatments.
Table 5 gives the difference between pre
and posttreatment scores for younger, less
hospitalized patients. Significant difference
scores are shown with an asterisk. High dose
patients showed significant improvement of
15 of the 21 measures. Even on the other
six measures, change was in the direction of
improvement. The only other significant im
provement occurred with routine treatment
in the areas of Excitement (IMPS), Hostile-
Belligerence (IMPS), Manifest Psychosis
(NOSIE), and Paranoid Depression,
(NOSIE). Placebo and low dose patients
showed significant worsening or no change
on all measures.
2. The results with high dose were less
impressive among older patients and patients
hospitalized over ten years. High dose was
significantly superior to low dose and rou
tine treatment in only five symptom areas:
Excitement (IMPS), Paranoid Projection
(IMPS), Conceptual Disorganization
(IMPS), Motor Disturbance (IMPS), and
Irritability (NOSIE). High dose was signif
icantly more effective than placebo in most
symptom areas. However, this was due more
to placebo worsening than to high dose
improvement.
3. Routine treatment was particularly ef
fective with patients over 40 having less
than ten years of hospitalization. With this
subgroup of patients, routine treatment was
significantly better than high dose in sev
eral important areas including Communi
ty Adjustment Potential (DRI), Psychoso
cial Adequacy (DRI), Social Competence
(NOSIE), and Retardation and Apathy
(IMPS).
4. The results on Retardation and Apa
thy deserve special mention. Based on the
chlorpromazine literature, it was expected
that high dose would be less effective than
the other treatments in reducing symptoms

Table 4.DRI, IMPS, and NOSIE Measures*

Groups Being Measures Showing a Significant

Compared Difference Between Groups (P<0.05)t
High dose group Psychosocial Adequacy (DRI) Motor Disturbances (IMPS)
more improved Belligerence; (DRI) Conceptual Disorganization (IMPS)
than low dose group Community Adjustment (DRI) Social Competence (NOSIE)
Manifest Psychopathology (DRI) Social Interest; (NOSIE)
Excitement (IMPS) Cooperation (NOSIE)
Hostile-Belligerence (IMPS) Personal Neatness (NOSIE)
Paranoid Projection (IMPS) Irritability (NOSIE)
Perceptual Distortions (IMPS) Manifest Psychosis (NOSIE)
Disorientation (IMPS) Paranoid Depression (NOSIE)
High dose group Psychosocial Adequacy (DRI) Conceptual Disorganization (IMPS)
more improved Community Adjustment (DRI) Social Competence (NOSIE)
than routine Paranoid Projection (IMPS) Cooperation (NOSIE)
treatment group Disorientation (IMPS) Irritability (NOSIE)
Motor Disturbances (IMPS)
High dose group Psychosocial Adequacy (DRI) Motor Disturbances (IMPS)
more improved Belligerence (DRI) Conceptual Disorganization (IMPS)
than placebo Community Adjustment (DRI) Social Competence (NOSIE)
Manifest Psychopathology (DRI) Social Interest (NOSIE)
Excitement (IMPS) Cooperation (NOSIE)
Hostile-Belligerence (IMPS) Personal Neatness (NOSIE)
Paranoid Projection (IMPS) Irritability (NOSIE)
Perceptual Distortions (IMPS) Manifest Psychosis (NOSIE)
Disorientation (IMPS) Paranoid Depression (NOSIE)
High dose group
less improved
than any other
treatment group
*
Showing a significant difference between high dose and low dose, high dose and routine treatment, and high
dose and placebo for younger, less hospitalized patients. Younger, Ie ss hospitalized patients are patients under
age 40 who have been hospitalized less than ten years.
t Treatment groups were compared by analysis of covariance which adjusts for pretreatment level.
.; Difference significant only for female patients.

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 Table 5.Difference Between Pretreatment and
Posttreatment Means for Younger, Less
during the third and fourth
Patients in Four Treatment
months of the study. The
Hospitalized Groups
routine treatment group
Mean Difference Score showed no clear pattern of
Symptom or Behavior High Low Placebo Routine change. Clinical change on

DRI routine treatment was high

Psychosocial Adequacy +12.6* 8.3 34.6* +0.5 ly variable. Some patients
Community Adjustment + 6.6* 7.3* 21.4* 0.2 who showed improvement
Belligerence + 2.1 4.3* 8.0* +1.2 at two months were deteri
Manifest Psychosis + 2.3* 0.7 2.6* +1.3 orated at six months and

NOSIE

Social Competence + 6.2* 4.3 18.9* +1.3 vice versa.

Social Interest + 0.5 8.7 24.1* 0.5 Early Terminations.
Personal Neatness + 3.0* 2.5* 8.5* 1.0 Table 6 shows the number
Cooperation + 3.7* 4.6* 12.1*

0.8 of patients in each treatment

Irritability + 8.5* +5.0 17.2* +3.0 group who were dropped
Manifest Psychosis_+ 4.7* 1.0 9.1* +2.4* from the study because of
Paranoid Depression 0.2 1.6* +1.9* side effects or deteriorated

+ 2.0*
behavior. The placebo group

IMPS
Excitement_+ 7.4* 0.0 16.0* +5.8*
+6.0*
had the highest percentage
Hostile-Belligerence_+ 8.3* 3.7 13.7*
ofdropouts (38%), followed
Paranoid Projection + 4.1* 4.5 11.3* 1.6
Grandiose-Expansiveness + 0.2 2.1 4.0 +2.0 by the high dose group
PerceptuarpTstortions_+ 6.8* 2.9

9.3* +4.0 (25%), and the low dose

Anxious 5.4 +3.9 0.2 +0.9 group (15%). Terminations

Intropunitiveness' +
Retardation and Apathy_+ 1.4 3.9

7.1 due to side effects were most

2.9
Disorientation 7.3* 5.5 19.5*
frequent in the high dose
+0.6

+
Motor Disturbances + 4.7* 1.0 21.4*
group, while the placebo
+0.5
13.0* +2.4
Conceptual Disorganization + 3.8
group had the highest inci
1.6

Younger, less hospitalized patients are patients under age 40 who have dence of dropouts due to
been hospitalized less than ten years.
Mean Difference Score: A "plus" difference indicates improvement from deteriorated behavior. In ad
pretreatment to posttreatment, and a "minus" difference indicates worsen dition to the terminations
ing.
*
Difference score significant at 0.05.
=
listed in Table 6, approxi
mately 3% of the high dose
measured by this factor (ie, slowed motor patients and 2% of the low dose patients
behavior, reduced general reactivity, and went home on trial visit, while another 4%
apathy). It was therefore surprising to find to 6% of the patients in each treatment group
that with younger, less hospitalized patients, were terminated for administrative reasons.
high dose actually reduced Retardation and Time of Early Termination.Of the 51
Apathy more than the other treatments (al high dose patients dropped from the study,
though the difference between high dose and 25% were terminated during the first eight
the other treatments was significant at only weeks, 45% during the second eight weeks,
the 0.1 level). At the other levels of age and and 30% during the last eight weeks. The
hospitalization, the results with high dose time pattern of dropouts for the placebo
were consistent with the literature; the high group was similar to that for the high dose
dose group showed a significant increase in group. Of the 82 placebo dropouts 24% oc
Retardation and Apathy and was the least curred during the first eight weeks, 56%
effective of the four treatments. during the second eight weeks, and 20%
Time of Improvement.Preliminary anal during the last eight weeks. The 31 low
ysis of the trends of clinical change over the dose dropouts were evenly distributed across
six-month treatment period indicated that the three eight-week periods. Time of ter
most high dose patients reached a plateau of mination was not related to sex, age, or
improvement by four months. Patients on length uf hospitalization in any of the three
low dose generally achieved maximum im treatment groups.
provement within two months, while the Side Effects.Table 7 presents the inci
greatest deterioration on placebo occurred dence of side effects judged to be moderate

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 Table 6.Percent of Treatment-Related
Terminations in Three Treatment Groups
Photosensitivity Reactions.More than
20% of the patients in the high dose group
Treatment Group had photosensitivity reactions as compared
to 2% of the patients in the other treatment
High Low Placebo
Reason for (N 208) (N 208) (N 212)
= = =
groups. There was a significant relationship
Termination % % % between photosensitivity reactions and eye
changes in the high dose group. Approxi
Side effects
Cardiovascular 6 11 mately 75% of the patients with photosen
Neurological 3 0 0 sitivity reactions developed eye opacities in
Ophthalmologic contrast to only 24% without photosensitiv
Weight loss ity reactions.
Other Cardiovascular Effects.Moderate to se
Total side effects 20 vere cardiovascular reactions (dizziness, syn
Deteriorated cope, hypotension) occurred in 26% of the
Behavior 5 13 33
Total terminations 25 15 38
high dose patients, 9% of the low dose
patients, and 8% of the placebo patients.
Cardiovascular side effects were particular
or severe in each study treatment. The high ly severe in older women on high dose med
dose group had the highest percentage of ication. Approximately 14% of the high
patients with at least one moderate to se dose women over 40 were terminated for
vere side effect (61%), followed by the pla cardiovascular disturbances, while another
cebo group (41%), low dose group (34%), 21% had their dose reduced at some time
and routine treatment group (15%). The during the study to control for severe cardio
side effects showing the greatest difference vascular symptoms.
between groups are listed below. Seizures.Twelve patients had a total of
Eye Change.Only a brief description of 17 convulsive seizures during the treatment
eye change will be presented here. A more period. Eleven of these patients were in the
detailed report is presented in another pa high dose group and one was in the low
per.16 Ocular change occurred in 35% of the dose group.
high dose group, 12% of the low dose Drowsiness.Excessive drowsiness was
group, 8% of the placebo group, and 5% of the most common side effect in the high
the routine treatment group. As can be seen dose group, occurring in 38% of the pa
in Table 7, the distribution of stellate tients. Drowsiness was most severe during
(grade 2) lens opacities was approximately the first three weeks on maximum dose.
the same for the four treatment groups. Dis- When severe drowsiness persisted, reduc
ciform (grade 1) lens opacities, however, tion of the 2,000 mg dose to 1,500 or 1,600
occurred much more frequently in the high mg generally reduced the drowsiness to a
dose group. A surprising finding was the oc tolerable level.
currence of anterior corneal change in 26% Extrapyramidal Reactions.The high
of the high dose group. This change some dose group had the highest incidence of
times appeared as diffuse bedewing, usual dystonic reactions and parkinson-like symp
ly whitish but occasionally brownish. The toms (tremor, rigidity), while the placebo
change usually appeared only in the inferi group had the highest incidence of what
or two-thirds of the cornea and was often was reported as akathisia. While the inci
identical in appearance with corneal change dence of akathisia and parkinsonism did
seen with chloroquine toxicity. There is am not differ greatly among high, low, and pla
ple evidence that the change is reversible. cebo groups, the reactions occurring with
At nine months after conclusion of the high dose patients were often more severe
study, approximately 50% of the corneal and required more extensive use of antipar-
opacities had improved markedly or disap kinsonian medication. Thirty-one percent of
peared completely. There was no report of the high dose patients, 12% of the low
reduction in visual acuity which could be dose patients, and 12% of the placebo pa
attributed to either lenticular or corneal tients received antiparkinsonian medication
changes. at some time during the study. It should be

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 Table 7.Percent of Patients in Each Treatment Group In addition to the above
With Side Effects ol at Least Moderate Severity
side effects, dry mouth, ex
Treatment Group cessive salivation, constipa
High Low Placebo Routine
tion, and rashes also oc
(N 208)
=
(N 208)
=
(N=212) (N 213)
= curred most frequently in
_Side Effect*_%_%_%^ the high dose group.
Central Nervous System One death occurred dur
Akathisia 9 9 16
Dystonie reaction 9 12 ing the course of treatment.
Parkinson reaction 16 8 13 The patient was a 37-year-
Seizures 6 10 old woman on 1,700 mg of
Autonomie Nervous System chlorpromazine. The patient
Dizziness, faintness 19 died after a grand mal seiz
Syncope ure, reported to be the third
Dry mouth seizure in a three-hour peri
Nasal congestion
od. Autopsy findings were
Constipation
Diarrhea essentially normal. Ventric
disturbance ular fibrillation during a
Urinary
Nausea, vomiting convulsive seizure is a pos
Salivation sible explanation.
Allergic Reactions Side Effects by Age and
Rashes, itching Sex.A detailed report of
Photosensitivity 22
side effects in various age
Other
Drowsiness 38 15 14 and sex groups will be pre
Peripheral edema 3 11 sented in another paper.
Blurred vision 10 1 Briefly, the relationship be
Eye opacities tween treatment, age, and
Lens-grade If 19
sex may be summarized as
Lens-grade 2 follows. In the high dose
Anterior cornea 26
opacities
Total with 35 12 group, the greatest frequen
Side effects requiring cy and severity of side ef
antiparkinson medication 31 12 12 27 fects occurred in older wom
* Refers either to adverse reactions which occurred for the first time dur- en, followed by older men,
ing treatment or to reactions which became more severe after treatment was
started.
younger men, and younger
t Opacities disciform in shape with no accumulation along the sutures women, in that order. In the
were classified as grade 1. Stellate shaped opacities were classified as
grade 2.
placebo group, younger pa
tients, particularly men, had
noted, however, that the percentage of high the highest incidence of moderate to severe
dose patients on antiparkinsonian medica reactions, while in the low dose group and
tion did not differ significantly from the per routine treatment group, there was no ap
centage of routine treatment patients receiv parent sex or age effect.
ing antiparkinsonian drugs during the treat Dosage Reduction.There were 146 high
ment period (see Table 7). dose patients ( 70% of the total sample) who
Table 8.Percent of Improved and Unimproved High Dose
completed the full 24 weeks
Patients Having No Moderate to Severe Side Effects* of treatment. Of these, 108
Percent With No Serious Side Effects
(75%) were on the maxi
mum dose of 2,000 mg at
Years Improved Unimproved Total the end of the study. Only
Age in Currently Patients Patients Patients 12 patients finished the
Years in Hospital % % %
Under 40_Under 10_45_48_47
study on doses below 1,600
Under 40_Over 10_50_47_49
mg. Cardiovascular disturb
Over 40_Under 10_46_33_38 ance, particularly in older
Over 40_Over 10_36_30_31 women, was the principal
cause of dosage reduction.
* Based on Global
Improvement Scale.

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 Occurrence of Side Effects and Clinical the areas of changePsychosocial Adequacy
Change.There was no apparent relation and Community Adjustment Potentialrep
ship between occurrence of side effects and resent the most characteristic features of
clinical improvement. Table 8 gives the per discharge-ready patients. What remains to be
cent of improved and unimproved high dose answered is whether these improved patients
patients who had no moderate to severe are in fact able to leave the hospital. Further
side effects. The criteria for improvement analyses are planned to determine the extent
used in this table was the Global Improve to which improved study patients resemble
ment Scale. The other scales provided simi the discharge-ready patients seen in the DRI
lar results. validation study.
These findings, however, do not justify a
Comment blanket endorsement of high dose treatment
for all younger, less hospitalized patients.
The results indicate that high dose is sig Approximately one-fifth of the population
nificantly more effective than low dose, pla of schizophrenics under 40 hospitalized less
cebo, and routine treatment with younger, than ten years were screened from the study
less hospitalized patients (ie, patients under prior to treatment because of medical condi
40 with less than ten years current hospital tions contraindicating the use of high doses
ization). With older or longer hospitalized of chlorpromazine. Even with the "clean"
patients, high dose appears to offer no great sample used in the study, approximately
advantage over low dose and routine treat half of the patients had at least one moder
ment. ate to severe side effect during treatment.
The range of symptoms affected by high Ten percent showed side effects serious
dose chlorpromazine among younger, less enough to warrant termination from the
is
hospitalized patients was impressive. There study. Thus, high dose treatment not un the
was significant improvement in 15 of the 21 dertaken without some risk. However,
symptom areas measured by the IMPS, fact that 45% of the younger, less hospital
had no
DRI, and NOSIE. Even in the other six ized patients showing improvementindicates
symptom areas, change (though nonsignifi moderate to severe side effects
cant) was in the direction of improvement. that high dose can be both safe and clinical
This wide range of improvement suggests ly effective. Hopefully, further refinements
that the effectiveness of high doses extends of study data will provide greater sensitivi
beyond mere "tranquilization" or sedation. ty in describing the characteristics of pa
In particular, the improvement in the key tients for whom high dose chlorpromazine
treatment regimen.
schizophrenic areas of delusional thinking, is the appropriate
disorientation Among older or longer hospitalized pa
hallucinations, and suggests
that high dose chlorpromazine is an effec tients, high dose was more effective than
tive antipsychotic treatment with younger, low dose and routine treatment in a
number
of symptom (including Paranoid Pro
less hospitalized patients. areas

A basic question that may be asked of all jection, Conceptual Disorganization,

Motor
drug studies reporting symptomatic im Disturbance, and Irritability). Despite this
improvement, high dose showed
provement among chronic schizophrenics is symptomatic low dose and
whether the treatment has resulted in im no significant superiority over

provement in practical aspects of the pa routine treatmentDRI on either the global scales
tient's adjustment, namely those areas of or the important factors of Community
Adjustment Potential and Psychosocial Ade
behavior which most affect the patient's whether
the hospital and adjust in quacy. In light of this, it is doubtful
ability to leave
older
the community. Our results indicate that the clinical benefits of high dose with
younger, less hospitalized patients on high or longer hospitalized patients justify the
dose, in addition to showing generalized high risk of side effects. This is particularly
symptom reduction, also show significant true with older patients where the risk of
improvement in areas of social adjustment side effects greatest. is
known to be related to release. This im One interesting finding was that patients
provement is particularly meaningful since hospitalized under ten years had signifi-

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 cantly more florid symptomatology (ie, per likely explanation for this discrepancy be
ceptual, delusional, and conceptual disturb tween interview and ward findings is that
ance) and showed significantly more social the nurses, who were continually exposed to
competence than patients hospitalized over the patient on the ward, were more sensi
ten years. This finding supports a theory of tive to signs of behavioral deterioration
ten advanced by the proponents of social than the interview rater who only saw the
psychiatry. It is argued that the acute psy patient once every two months. The social
chotic symptoms which characterize the ear workers, who also had extensive contact
ly stages of schizophrenic illness subside with the patients on the ward, also de
over time,17 only to be replaced by deficien scribed the older patients as consistently
cies in social competence fostered by the worse in most symptom areas. Only 26% of
"total institution."1822 Lack of contact with the older placebo patients were actually
the community, minimal challenge to social withdrawn from the study because of deteri
and interpersonal skills, and the "custodial" orated behavior. However, a sizable number
atmosphere of the chronic ward all contrib of patients remaining in the study showed
ute to this progressive deterioration. (Other noticeable regression on the ward and it is
tenable explanations are the existence of a doubtful that they would have remained off
progressive "organic" impairment of brain medication had they not been involved in a
function or selection release of less im study. Even if the relapse rate of 26% is
paired members of the chronic schizophren taken at face value, it would still seem to
ic population.) The result is that patients high to justify long-term drug withdrawal
hospitalized for long periods of time may as a treatment policy with older schizophren
suffer more from the debilitating effects of ic patients.
"institutionalization" than from the prima Some investigators advocate treatment
ry effects of the schizophrenic illness. This programs for chronic patients which involve
may explain why only a few of the study periodic short-term withdrawal of phenothia
patients hospitalized more than ten years zine medication.23'25 While this study was
showed marked improvement. It may be un not designed to evaluate the effectiveness of
reasonable to expect phenothiazine thera these "intermittent chemotherapy" programs,
pies to undo the effects of many years of so our placebo results do indicate that when
cial deprivation and provide the patients such programs are used, the drug-free period
with the social and personal resources which should not exceed two months. After this
they lack. period, the probability of relapse and deteri
This study leaves no doubt that chlorpro oration sharply increases.
mazine side effects are related to both dose The treatment prescribed by the hospital
and age. High dose produced more frequent physician (ie, routine treatment) was most
and serious side effects than low dose at all effective with patients over 40 having less
levels of age and hospitalization. Among pa than ten years hospitalization. Most of the
tients on high dose, older patients, particular study patients benefiting from routine treat
ly women, showed more severe cardiovascu ment had been first hospitalized late in life
lar reactions, drowsiness, and parkinsonism (ie, after age 30). Many showed cyclic pat
than younger patients. The incidence of side terns of illness and had been discharged a
effects was not related to length of hospitali number of times prior to the study.
zation or weight. There was also no signifi Results from the routine treatment group
cant relationship between occurrence of side should be interpreted with a certain amount
effects and clinical improvement. of caution. Ward evaluations were not done
The results on placebo present a mixed under doubleblind conditions. The research
picture. On the doctor's interview scales, the nurses supervised the administration of study
majority of older patients on placebo medication and were thus aware of which
showed no measurable deterioration over patients were in the routine treatment group.
the six-month study period. On the ward Patients in the routine treatment group were
the results were far less positive. The nurses also known to the physician evaluating the
rated the majority of older placebo patients side effects. Despite these limitations, the
as worse in most symptom areas. The most routine treatment group was a highly valu-

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 able control group which allowed us to com
pare study treatments with treatments the
physicians considered best for the patient.
There was no difference between treat
ment groups in the number of patients ac
tually leaving the hospital. Only a few pa
tients in each treatment group (2% to 4%)
were discharged or granted a trial visit dur
ing or immediately following treatment. It
is questionable, however, whether discharge
can be regarded as a meaningful criteria of
improvement. It has been argued that hos
pital and community deficiencies are often
more important than clinical condition in
keeping the schizophrenic patient in the
hospital.26 The success or failure of dis years hospitalization.
charge seems to be particularly dependent
on the relationship between the chronic pa
tient and his family or other key figures in progress.
the community.27 It is apparent that drug
effectiveness cannot be measured by dis Health Service grants MH-10292, MH-11384, MH-
charge rate, if this, in turn, is significantly 10332, from the National Institute of Mental
influenced by such indirect criteria. The use Health, and Public Health Service contract SA-
of valid and reliable rating instruments 43-ph-3064.
such as the DRI and IMPS appears to be
the best method available for evaluating
effects of psychopharmacologic agents in
chronic schizophrenic patients.
Summary
In a collaborative study involving 838
chronic schizophrenics in seven public men
tal hospitals, a high dose regimen of chlor
promazine (2,000 mg/day) was compared
with a low dose chlorpromazine regimen
(300 mg/day), a placebo, and routine hos
pital treatment. Treatment followed a dou
ble-blind procedure for 24 weeks. Clinical
evaluations using five rating scales provided
23 criteria of change. The following results
were obtained:
1. High dose was significantly more effec
tive than low dose, placebo, and routine
treatment with patients under 40 currently
hospitalized less than ten years. High dose
improvement with these younger, less hospi
talized patients included change in a wide
range of schizophrenic symptoms and behav
iors. Particularly significant was the improve
ment in Psychosocial Adequacy and Com
munity Adjustment Potentialtwo areas of
social adjustment known to be related to
release.
2. Among older or longer hospitalized
patients, high dose appeared to offer no
great advantage over low dose or routine
treatment.
3. High dose produced a significantly
greater number of serious side effects than
each of the other treatments. Side effects
from high dose were particularly severe in
older patients (ie, over 40 years of age).
4. Placebo patients showed significant
worsening on practially all criteria of change.
Greatest deterioration occurred during the
third and fourth months of treatment.
5. Routine treatment was most effective
among patients over 40 with less than ten
Additional analyses relating to prediction
of outcome on high dose and placebo are in
This investigation was supported by Public
10496, MH-10989, MH-11046, MH-11047, and MH-
Data were contributed by
the following
individu
als and their staff from the seven participating hos
pitals: Edwin M. Davidson, MD, and Thomas Melvin M.
Kayce, MD, Boston State Hospital; A.
Smith, MD, Broughton State Hospital; Myron G.
Sandifer, MD, and William J. Buffaloe, MD, Doro
thea Dix State Hospital; William J. Kernohan, MD,
and Winston T. Wilson, PhD, Kentucky State Hos
pital; Herman C. Denber, MD, and Edwin Meshel,
MD, Manhattan State Hospital; Robert R. Knowles,
MD, BS, DPM, and Cecil G. Baker, MD, St. Louis
State Hospital; Martin Gross, MD, and Arthur Man-
del, MD, Springfield State Hospital.
Roland R. Bonato, PhD, Frances H. Fleming,
MN, and Naomi F. Hartwick of the Biometrie Lab
oratory, George Washington University, and Solo
mon C. Goldberg, PhD, and Gerard E. Hogarty,
MSW, Psychopharmacology Research Branch, Na
tional Institute of Mental Health, assisted in this
study.
References
1. Ayd, F.: Large Doses of Chlorpromazine in
the Treatment of Psychiatric Patients, Dis Nerv
Syst 16:146-149, 1955.
2. Delay, J., and Deniker, P.: Chlorpromazine
and Neuroleptic Treatments in Psychiatry, J Clin
Exp Psychopath 17:19-24, 1956.
3. Forrest, F., et al: Drug Maintenance Problems
of Rehabilitated Mental Patients: The Current
Drug-Dosage "Merry-Go-Round," Amer J Psychiat
121:33-40, 1964.
4. Kinross-Wright, V.: The Intensive Chlorpro-
mazine Treatment of Schizophrenia, Psychiat Res
Rep Amer Psychiat Assoc 1:53-62, 1955.
5. Lieberman, D.: Tranquilizers in Psychiatry:
Chlorpromazine Therapy, Brit J Med 1:512-513,
1956.

Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015

 6. Mendelsohn, R.; Penmon, A.; and Schiele, B.:
Massive Chlorpromazine Therapy, Psychiat Quart
33:55-76, 1959.
7. Pollack, B.: Preliminary Report on 500 Pa-
tients Treated With Thorazine at Rochester State
Hospital, Psychiat Quart 29:439-456, 1955.
8. Rosati, D.: Prolonged High Dosage Ataractic
Medication in Chronic Schizophrenia, Brit J Psy-
chiat 110:61-63, 1964.
9. Schiele, B., et al: Comparison of Low and
High Dosage Procedures in Chlorpromazine Thera-
py, Psychiat Quart 33:252-259, 1959.
10. Winkleman, W.: A Clinical and Socio-Cultur-
al Study of 200 Psychiatric Patients Started on
Chlorpromazine 10\m=1/2\ Years Ago, Amer J Psychiat
120:861-869, 1964.
11. Hogarty, G.: Discharge Readiness: The Com-
ponents of a Casework Judgment, Soc Casework
47:165-171, 1966.
12. Hogarty, G.: The Discharge Readiness Inven-
tory: Validity as an Outcome Measure in the
Treatment of Chronic Schizophrenia, read before
the Fifth Annual Meeting of the American College
of Neuropsychopharmacology, San Juan, PR, 1966.
13. Lorr, M., et al: Inpatient Multidimensional
Psychiatric Scale: Manual, Palo Alto, Calif: Consult-
ing Psychologist Press, 1962.
14. Honigfeld, G., and Klett, C.: The Nurse's Ob-
servation Scale for Inpatient Evaluation (NOSIE):
A New Scale for Measuring Improvement in Chron-
ic Schizophrenia, J Clin Psychol 21:65-71, 1965.
15. Winer, B.: Statistical Principles in Experi-
mental Design, New York: McGraw Hill Book Co.,
Inc., 1962.
16. Delong, S.; Prien, R.; and Cole, J.: Opacities
Downloaded From: http://archpsyc.jamanetwork.com/ by a New York University User on 06/22/2015
During Prolonged High Dose Chlorpromazine
Therapy, Arch Ophthal, to be published.
17. Casey, J., et al: Combined Drug Treatment
of Chronic Schizophrenia, Amer J Psychiat
117:997-1002, 1961.
18. Barton, R.: Institutional Neurosis, Bristol,
England: John Wright & Sons, Ltd., 1959.
19. Brown, G., and Wing, J.: A Comparative
Clinical and Social Survey of Three Mental Hospi-
tals, Social Rev Monogr No. 5, 1962.
20. Goffman, E.: Asylums: Essays on the Social
Situation of Mental Patients and Other Inmates,
New York: Doubleday & Co., Inc., 1961.
21. Martin, D.: Institutionalization, Lancet 2:1188-
1192, 1955.
22. Stanton, A., and Schwartz, N.: The Mental
Hospital: A Study of Institutional Participation in
Psychiatric Illness and Treatment, New York: Bas-
ic Books, Inc., Publishers, 1954.
23. Good, W.; Sterling, M.; and Holtzman, W.:
Termination of Chlorpromazine With Schizophrenic
Patients, Amer J Psychiat 115:443-448, 1959.
24. Olson, G., and Peterson, D.: Sudden Removal
of Tranquilizing Drugs From Chronic Psychiatric
Patients, J Nero Ment Dis 131:252-255, 1960.
25. Rothstein, C.: An Evaluation of the Effects
of Discontinuation of Chlorpromazine, New Eng J
Med 2:67-69, 1960.
26. Hogarty, G.: Hospital Differences in the Re-
tention and Release of "Discharge-Ready" Chronic
Schizophrenics, Special Studies Section, Psycho-
pharmacology Research Branch, NIMH (prepubli-
cation report).
27. Gordon, H., and Groth, C.: Mental Patients
Wanting to Stay in the Hospital: Attitudes, Arch
Gen Psychiat 4:124-130, 1961.