·ii\
,; '·1
1
i.
;:
Regular Articles
The Vermont Longitudinal Study of Persons With
Severe Mental Illness, I: Methodology, Study Sample,
and Overall Status 32 Years Later
Courtenay M. Harding, Ph.D., George W. Brooks, M.D., Takamaru Ashikaga, Ph.D.,
John S. Strauss, M.D., and Alan Breier, M.D.
The authors report the latest findings from a
32-year longitudinal study of 269 back-ward patients
from Vermont State Hospital. This intact cohort
participated in a comprehensive rehabilitation
program and was released to the community in a
planned deinstitutionalization effort during the
mid-1950s. At their la-year follow-up mark, 70% of
these patients remained out of the hospital but many
were socially isolated and many were recidivists.
Twenty to 25 years after their index release, 262 of
these subjects were blindly assessed with structured
and reliable protocols. One-half to two-thirds of
them had achieved considerable improvement or
recovery, which corroborates recent findings from
Europe and elsewhere.
(Am J Psychiatry 198?; 144:718-726)
Understanding of the long-term course and outcome
of patients with prolonged psychiatric disorders
is often thwarted by patient and clinician mobility
(1, 2), short-term caseloads shaped by academic
training and service delivery systems (3, 4), the magnitude
of methodological hurdles (5-13), and disputes
Received Jan. 31, 1986; revised Ocr. 7, 1986; accepted Dec. 16,
1986. From the Department of Psychiatry, Yale University School of
Medicine, New Haven, Conn.; the College of Medicine and the
College of Engineering and Mathematics, University of Vermont,
Burlington; and the Clinical Neuroscience Branch, NIMH, Rockville,
Md. Address reprint requests to Dr. Harding, 150 CMHC, 34
Park St., New Haven, cr 06519."
Supported by NIMH grants MH-29575, MH-40607, and MH-
00340 and by Biomedical grant 50705429 from the College of
Medicine, University of Vermon~ '
Copyright © 1987 American Psychiatric Association.
718
over the classification of the disorders under study (10,
14-16). These conditions have produced sporadic,
contradictory data and untested assumptions that undercut
attempts to clarify the nature of psychiatric
illness, erode the ability to target treatment interventions,
and muddle efforts toward comprehensive 'public
policies.
It is possible, however, to generate a longitudinal
study that overcomes most of these obstacles (7, 17).
What is required is an intact cohort of patients,
selected for the established chronicity of their illness,
who are prospectively followed over many years, with
careful record keeping, structured and reliable protocols,
operational definitions, and standardized assessments
of psychopathology and psychosocial functioning.
The Vermont longitudinal study meets these
criteria.
Since the early 1950s, members of the Vermont
Longitudinal Research Project have been prospectively
following the course of an intact cohort of 269 patients
from the back wards of Vermont State Hospital (7, .'
17-26) in much the same manner as the catamnestic
studies of Manfred Bleuler at Burgholzli Hospital in
Switzerland (27-33). Known in the literature as The
Vermont Story cohort (26), the majority of these once
profoundly ill, severely disabled, long-stay patients
came from the sickest group in the hospital and met
, the DSM-I guidelines for the diagnosis of schizophrenia.
They participated in an innovative pioneering
rehabilitation program and were released to a hospitalrun
comprehensive community aftercare program between
1955 and 1965 (26).
Ten years after the inception of the program, we
conducted a follow-up study, which indicated that
two-thirds of--the cohort were not hospitalized but
were being maintained by heavy expenditures of clinAm
J Psychiatry 144:6, June 1987
•
• II•.J4$. O(WZ,.~ .• PX 40DWU4CQ Qit¢!QQ!P au ..2W ·4 ". §. Q.kit .x: .WQ.lUl.!CZLXXi.•.• "..}k~~
Am J Psychiatry 144:6, June 1987
HISTORY OF mE PROJECT
Phase I: The Rehabilitation Program in the 19505
In the early 1950s, one 6f us (G.W.B.) began to
study the efficacy of the then-new drug chlorpromazine
(36~. The back-ward"hopeless cases" were chosen as
subjects. Some patients responded well and were even~alJy
released. Other patients did not respond as well,
Ut the effect of the releases 'was to give fresh hope to
~~ the staff and the patients. In 1955 a multidis,~plmary
c~inical team, sponsored by the Vermont
~a.te Hospital and the Vocational Rehabilitation DiVl~lOn
of the Vermont State Department of Education,
::;:~ated a program of comprehensive rehabilitation
. community placement for those back-ward paOents
who had not improved sufficiently with chlorP2fomazine.
From January 1955 to December 196069 . ,
patle~ts who were considered among the most
~er.ely disabled and chronically mentally ill in the
ospltal were referred to the program (26). After the
ec
j,
.,
HARDING, BROOKS, ASHIKAGA, ET AL
At the time of selection, the group averaged 16 years'
duration of illness with an average of ten years of total
dis-ability and six years of continuous hospitalization. The
group members had from one to ten hospitalizations, with
a median of about two hospitalizations each. They had
completed from none to sixteen years of schooling, with a
median of about nine grades. Nearly all had been declared
financially incapable of paying anything for their own
care, and were committed to the hospital at State expense.
The group was, in other words, quite characteristic of the
schizophrenic group as outlined by Hollingshead and
Redlich (37). They were middle-aged, poorly educated,
lower-class individuals further impoverished by repeated
and prolonged hospitalizations. In addition, this group
had little social support. About five out of six were single,
divorced, widowed, or separated. They were seldom visited
by friends or relatives, and received very few packages
or letters. (26, p. 30)
At the time the subjects were selected} the research
team also described their presenting disabilities and
impairments: . . {h ~i.5
The patients, as a group, were very slow, concentrated) ~J
poorly, seemed confused and frequently had some impair- /
ment or distortion of recent or remote memory. They were (/\
touchy, suspicious, temperamental, unpredictable, and f).r1\over-dependent on others to make minor day-to-day deci- 'r! v \
sions for them. They had many peculiarities of appear- 1Ali
ance, speech, behavior, and a very constricted sense of / W1l.O
time, space, and other people so that their social judgment, I.
was inadequate. Very often they seemed to be goalless or,
if they had goals, they were quite unrealistic. They seemed
to lack initiative or concern about anything beyond their
immediate surroundings. Because of their very low socioeconomic
level and prolonged illness, they suffered from
profound poverty, inadequate educational opportunities,
anq:3 very limited experience in the world. .
These patients also suffered a high incidence of chronic
physical disabilities. Their psychomotor performance in a
wide variety of tests was impaired so that their reaction
times were prolonged and their ability to perform any type
of skilled .or precise activiry was impaired. They suffered
an increased incidence of many degtnerative and chronic I
diseases, including tuberculosis and malignant tumors. I
719 I
j
're-education
of the staff, a program was constructed in
collaboration with the patients that consisted of drug
treatment, open-ward care in homelike conditions,
group therapy, graded privileges, activity therapy,
industrial therapy, vocational counseling, and self-help
groups. .
In the community treatment component, the same
clinical team went into the community and established
halfway houses and outpatient clinics, found job openings,
made job placements, and linked patients to
natural support n~rworks. In that era of custodial care,
before the advent of community mental health centers
and the later deinstitutionalization movement, this
cOmprehensive-program wa-~ considered unusual and
innovative.
The average age of the subjects was 40 years. The
group was described by their clinical team in The
Vermont Story as follows:
Mrr,.••TIS' '27
-=,:.,
Jt;..effort, time, and money (34). Many of these
pitients were socially isolated while living and workiJ:Jg
in other institution-like settings. This follow-up
period was similar to that charted by many longitudiualstudie.s
of similar patient groups, studies that have
heavUy influenced our ideas about the long-term
oourse and outcome of people with severe and prolonged
psychiatric disorders. The Vermont study and
other longer studies provide evidence that some psychiatric
illnesses require longer time periods to acquire
a more complete and accurate picture of course and
ootcome.
lneurdfort to. reassess.me outcome of the Vermont
cohort over a longer period ;£ti~e, we were aDreto
account for all but seven members of the original
cohort (97%) in the early 1980s. This situation gave us
the opporrunity to find out whether these subjects were
stiD as disabled as they had been 20 ·or more years
earlier. We conducted both a structured cross-sectional
assessment of the subjects' current status across a wide
range of characteristics and a retrospective documentation
of what had happened to members of this
cohan in the intervening years. Adding these new data
to the prospectively gathered information from the
earlier hospitalization and rehabilitation program, we
have been able to provide a more comprehensive
picture of the long-term course of schizophrenia and
other severe psychiatric disorders.
The specific focus of this report from the Vermont
longitudinal Research Project is a description of the
sample, methodology, and design of the project and
documentation of the' long-term outcome for the cohan
as a whole. A companion paper in·this issue of the
Journal examines the long-term outcome of those
subjects within the larger cohort who were rediagnosed
as meeting the newer DSM-Ill criteria fer
~zophrenia (35).
I
L .....
METHOD
11 4
7 3
N %
168 68
61 25
Subjects Remaining After
Exclusion of 22 With
Organic Disorders
(N=247)'
13
7
71
178
Total
Group
Measures
Batteries of structured instruments were used for
collecting data. They included interview schedules and
record abstraction protocols.
The Vermont Community Questionnaire (VCQ)
was a battery of interview instruments designed to ..
document and assess a subject'S history and funetion-t:
ing in a wide variety of areas across time. Fifteen
established scales were combined to create the VCQ'
and to acquire such a data base (7). The field inter- ...~'.'
viewers were blind to hospital records and diagnostic ,
information about each subject. .~
The VCQ consisted of two structured interviews, i
each with standardized probes, ratings, and computer,l
coding. The VCQ-cross-sectional interview (VCQ-C) '1]
assessed current status; the VCQ-Iongitudinal inter- :1
Am J Psychiatry 144:6, June 198·
'Twenty-two subjects classified as having.organic disorders according
to DSM-III criteria at the index admission were excluded from
the .data analyses. '
bMost of these subjects were interviewed' and gave considerable
information, but they refused to sign the forms permitting use of
their data.
Phase III: The 20- to 25-Year Follow- Up Study
We recently completed our latest follow-up of the
original 269 subjects. We have follow-up data on the
22 subjects with organic disorders but removed them
from our ongoing analyses to make our study comparable
to others in the field. Table 1 reveals the cohort's
current status.
The catamnestic period of the subjects ranges from
22 to 62 years, with an average of 32 years, which
makes this study one of the longest ever conducted.
The subjects who were still alive at follow-up (N=168)
were divided nearly evenly between the sexes (81 men iand 87 women): The mean age was 59 years (range::::~
38-83 years), with two-thirds of the group 55 years -_
old or older. The year of birth of the subjects ranged,~
from 1897 to 1942. {
The remainder of this report focuses on the method
and results of the long-term follow-up study of the
subjects who were still alive and could be interviewed.
Alive and interviewed
Deceased; family
interviewed
Alive; refused
parricipationb
Could not be located
Subjecrs'
Follow-Up
Status
J.~
TABLE 1. Status at 20- to 25-Year Follow-Up of 269 Chronic '5
Psychiatric Patients in the Vermont Study *·t
c_:=·"e;.dp
There was a very high incidence of needs in such areas as
dental care, visual corrections, and hearing aids. Many
patients, especially after prolonged phenothiazine treatment,
were obese. Care of the feet had been neglected. In
our experience, there also seemed to be a high incidence of
chronic skin disorders including eczema-dermatitis, tinea,
and psoriasis. (26, p. 31)
720
Phase II: The .1965 Follow-Up Study
In 1965, after 5 years of the hospital-based rehabilitation
program and 5 years during which the p'rimary
focus was the community component, Deane and
Brooks (34) conducted a follow-up study of these
patients. They found that two-thirds of the cohort
c()!:ll~.i?e.. _r:n.ai.n!ail1~c!. in, ,the SQqlll1~Qity if.5l-!ftic,it;nt
transitional facilities and adequate aftercare were provided.
Seventy percent of the subjects were out of the
hospital at that follow-up: 30% had been discharg~d
and had never returned, and 40% had been readmitted
at some time but had been discharged again. Of the
30% of the subjects who were in the hospital at
follow-up, 20% had been readmitted and had stayed,
and 10% had never been discharged. The average
number of readmissions for the recidivists in the
cohort was 1.98.
Other findings indicated that being female, schizophrenic,
chronically ill, and married during some part
of one's life were important predictors of good functioning
at follow-up. Age at first admission did not
predict which patients would do better. At the 1965
follow-up, most subjects were single (60%), used
community care facilities primarily for socializing, had
a tendency to replace the institution with sheltered
employment (e.g., a job as a cook in a nursing home,
with bed and board), and maintained substantial contact
with rehabilitation workers.
Thus, 5-10 years after release from the rehabilitation
program, 70% of the patients were out of the
hospital, which was considered remarkable at the time
because they had been expected to live out their lives in
the ~ospital. However, the study concluded with a
warmng:
Implicit in our findings is the fact that any plan for
rehabilitation of the chronic patient be conceived as longterm,
since all of our evidence suggests that the commitment
necessary to the chronic mental patient has no
foreseeable end, and that unless constant attention be
given to the chronic patient, the end result may be simply
that he is out of the hospital, but operating at a high level
of inadequacy and a low level of employment. (34, pp. ii,
iii)
It is at this point that most follow-up studies stop
and most programs are discontinued. Thus, most of
our understanding of the long-term outcome for severe
mental disorder is derived from such shorter-term'
data. The question asked in the present study was, Do
these patients still continue to display such impairment
and disabilities 20 to 25 years later, as predicted earlier
by our own research team?
VERMONT LONGITUDINAL STUDY, I
. j: T ;: :I rrrwmt::tSE': "TV
HARDING, BROOKS, ASHlKAGA, ET AL
1
...
ew (VCQ-L) documented retrospective data over the
:eceding 20-25 years.
The major areas of functioning covered by the
CQ-C were residence, work, finances, intimate rela:lIiships,
family information, social support system,
·pical weekly activities, basic self-care, utilization of
eatmentlsocial services, contact with the criminal
Istice system, community involvement, degree of satfaction,
environmental stressors, competence, and
5ychopathology. Approximately 1 hour and 15 mintes
were required to administer the interview's 135
llestions. The rater section of the interview battery
:2maip~~~8_~Mi~ional items based on observation
Ild other informati~~-gath~~edfi:omthe "su-bJects and
leir environment.
The second interview (the VCQ-L) required 75
linutes and was held within 1 week of the first. The
'CQ-L had 156 questions that documented status and
vents during the preceding 20 years in a year-by-year
)llow-back procedure utilizing a modified Meyer/
leighton Life Chart (38). This chart provided a graph:
overview of each subject's life and was completed
rith a set of structured probes, codes, and protocols
reated for this project. The Life Chart was a large,
ned sheet of paper vertically separated into years,
rom 1982 at the top to 1955 at the bottom, and
orizontally separated into 10 outcome areas. These:
lomains included residence, hospitalization, work,
Duree of income, important personal relationships,
,eaths of important people, other life events, use of
ommunity support systems, physical health, and med:atioils.
A Life Chart was completed for each subject,
vho worked with the interviewer on the chart spread
lut on a table.
Each of the field interviewers had had 5-8 years of
Irevious clinical experience with a range of clients who
vere deinstitutionalized and labeled "chronic paients"
by their community mental health clinics.
A small instrument called the Verinform was deigned
to verify the interview information by asking a
'ariety of informants about the subjects' current status
md historical data. The Verinform was used for
nterviews with general practitioners, aftercare or vo:ational
rehabilitation counselors; family, or friends..,hoever
knew the subject well.
The Hospital Record Review Form was designed to
lrovide a standardized method for recording data
:rom Vermont State Hospital records. The interview
nstrument known as the Psychiatric and Personal
-Iistory Schedule from the World Health Organization
WHO) Collaborative Project on Determinants of Out:ome
of Severe Mental Disorders (39), was converted
:rom an interview format to a standardized form for
lbstracting record information on psychiatric history
n a systematic and structured manner. The new fornat
maintained the coded answers but assigned them
:0 document five different time periods: first admis;ion,
episodes between first and Index admission,
ndex admission (the admission preceding entry into
:he rehabilitation program-the only common denom-
4mJ Psychiatry 144:6, June 1987
inator across all subjects and designated arbitrarily as
index for research definition purposes), life history,
and episodes during the years in the community after
index release. In addition, the WHO signs and symptoms
checklist was augmented with Strauss's Case
Record Rating Scale (40), the Strauss-Carpenter Prognostic
Scale (41), and the Global Assessment Scale
(42). The record reviewer was blind to all outcome and
interview data. She was a clinical psychologist with
several years of experience with Vermont State Hospital
records.
The component instruments in each battery that
used the work of others had been tested extensively for
rdia&lllry-andvallCIiry"oy-ilieii 6ilginators~-However,
groups of individual questions from each classic scale
were taken from their original context and interwoven
with questions from other instruments (e.g., the questions
on social relationships from all instruments were
put together to make a more natural interview sequence);
therefore, reliability studies of the entire VCQ
and Hospital Record Review Form batteries were
deemed essential.
Initially, the VCQ-C and VCQ-L interviews were
each field-tested with a wide range of community
people matched in age to the cohort. These consultants
critiqued the appropriateness of the questions vis-a-vis
their life experiences and suggested that we add questions
about powerlessness, disability income, medication
compliance, and the increasing number of deaths
of people in their personal support systems. This
strategy led to improvement in the battery's ability to
tap relevant issues for people in the age ranges to be
assessed and improvement in its construct validity.
Changes and deletions of items honed it to the size
used in the follow-up study.
Each field battery was then subjected to two sets of
interrater trials (trial 1, N=21 pairs; trial 2, N=18
pairs). The sets were completed 6 months apart to test
for the degree of change in raters' assessments during
the intervening time period. Both raters attended a
cross-sectional and a longitudinal interview for each of
the test subjects. Each rater scored the interview independently,
and the pairs of ratings were then compared
for concordance. The kappa coefficients from the first
and second sets of trials are shown in table 2.
The Hospital Record Review, with 1,800 items, was
divided into its five subsections and also subjected to
trials of interrater agreement berween the one reviewer
who It:ft the team and the one who subsequently joined
the project. Kappa coefficients ranged from .40
(p<.Ol) to .95 (p<.OOl), and all were significant.
In summary, on the basis of the evidence presented,
we rendered the conclusion that these instruments
were moderately reliable and contained an acceptable
degree of face and construct validity.
Procedure
Each subject still living at the time of follow-up who
was willing to participate was given two interviews
721
VERMONT LONGITUDINAL STUDY, I
RESULTS
TABLE 2. Kappa Levels for Interrater Agreement on Trials of
Instrument Batteries Given to Vermont Study Subjects
Overall Psychological and Social Functioning
The Global Assessment Scale (GAS) (42) was chosen
to provide a single score that would capture the
essence of the subjects' psychological and social func- ~
rioning. Scores on this scale showed that 68% (N=tmJ.
114) of the study sample were functioning above the
cutoff point of 61 designated by the authors as "some
mild symptoms (e.g., depressive mood or mild insomnia)
or some difficulties in several areas of functioning,
but generally functioning pretty well, has some mean~
had not completed high school. However, an eighthgrade
education was considered to be the norm before
the 19405 in Vermont (43).
Nineteen percent (N=32) of the 168 subjects were
currently married, and seven percent (N= 11) were
widowed. Fifty-one percent (N=86) were still single,
and 23% (N=39) were divorced or separated.
Eighty-eight percent of the subjects (N= 148) 'lived
in residential and rural neighborhoods rather than
industrial or commercial areas. Fifty percent (N=81)
lived in independent housing (house, apartment, mobile
home, or rooming house), and 40% (N=64) lived
in boarding homes. (This information was coded for
an N of 161.) Five single middle-aged men were
currently in the hospital, seven were in level II nursing
homes (InsiifutlonsfOr inatvtdlia1s, im:luding the mentally
retarded, who do not require 24-hour nursing
care but who do require care above the level of roOm
and board), and four were in other settings. Of the
subjects receiving boarding home care, seven seemed
capable of living independently, often assisting the
boarding home operator and taking responsibility for
management of the home. An additional 23 were
actively involved in activities within the house, at the
local community mental health center, or in the communityand
were self-motivated. Longitudinal patterns
of residence revealed an average of two readmissions
for the group since release from the iridex hospitalization.
The average total length of stay was 2 years or
less.
Twenty-six percent (N=44) of the 168 subjects were
employed; half of them were classified as working in
unskilled jobs. Thirty-three percent (N=56) were unemployed,
8% (N= 14) were volunteers, and 5% (N=
8) were housewives. Due to the advanced ages in the
sample, an additional 26% (N=44) were classified as
elderly, widowed, or retired. Solid information was
unavailable on four (2%) of the subjects for this rating.
OUf findings also indicated that 85% of the sample
had a gross income of less than $10,000 a year. In
1982 Vermont ranked 36th in the nation for per capita
income, with an average of $9,979 (44). Using an
assessment from the Community Care Schedule by
Schwartz et aI. (45) and budget sheets outlining expenses
and income, we rated 77% of the subjects as
having an adequate income; the schedule's definition
of adequate was that the "amount of money received
will cover the subject's basic needs comfortably."
.963
.59b
,973
,70b
Kappa
.963
,SOb
First Trial Second Trial
Field instruments
Vermont Community Questionnaire
Cross-sectional
Interview
Rater section
Longitudinal
Interview
Rater section
Hospital Record Review Form
First admission
Episodes between first
aruLi.nd~J( "~[t1ission
Index admission
Episodes after discharge
Topical life histoty
Life Chan (experimental)
Demographic Data for the Cohort at Follow-Up
Fifty-one percent of the 168 subjects who were alive,
did not have an organic disorder, and were interviewed
(approximately 1 week apart) at his or her place of
residence by one of the two interviewers. In addition,
two or three people who knew the subject well were
interviewed in a structured protocol to verify current
status and historical data. These people included relatives,
general practitioners, counselors, clinicians, and
friends. All but 17 subjects (10%) resided in Vermont.
The subjects who lived elsewhere were interviewed
with the same protocol. Strict attention was paid to the
protection of patients' rights such as privacy, confidentiality,
refusal, and informed consent.
Relatives, friends, and caregivers of the deceased
members of the original cohort were also interviewed.
A structured protocol documented the lives and the
levels of functioning of these subjects until the time of
their deaths. The inclusion of these data provided a
more balanced view of the long-term course of severe
psychiatric disorders than has been available in past
studies, which have relied on data from survivors only.
A separate report will be devoted to the deceased
subjects.
This group of back-ward patients represented the
most severely ill group from Vermont's only state
hospital. Two to three decades after a comprehensive
rehabilitation program and a planned deinstitutional·
ization, one-half to two-thirds of these patients were
rated as considerably improved or recovered. The
findings also showed a wide variation in many areas of
functioning for these patients.
3 p<,OOl.
bp<.Ol.
Insuument
..
722 Am JPsychiatry 144:6, June 1987
'111
IiURE 1. Global Assessment Seale Scores of 168 Subjects in the
nnont Study Who Were Alive and Were Interviewed at Follow-Up
40,-
30
OMen
~Women
20
o 0- 11- 21- 31- 41- 51- 61- 71- 81- 91-
10 20 30 40 50 60 70 80 90 100
+-- Poor --+ +-- Fair --+ +---- Good ---+
GLOBAL ASSESSMENT SCALE SCORE
ngful relationships and most untrained people would
lot consider him sick." A 2 x2 (GAS by Sex) chi,quare
test with Yates' correction revealed no signifi:ant
differences between the sexes in level of functionng
(X2
=0.13, df=l, p=.72), but it should be noted
:hat 68% of those with a GAS score between 31 and
70 were men, while 62% of those with a GAS score
Jetween 71 and 90 were women (see figure 1).
In order to describe more of the individual components
that went into the assessment of overall functioning,
the Levels of Function Scale (41) was used. On
this scale, subjects are scored from 0 (poor) to 4 (best)
on nine items of interest; the reliability of the scale has
been demonstrated. A product-moment correlation
revealed that the overall total score was highly correlated
(r=.88) with the GAS score just reported. Table
3 summarizes.the findings from the Levels of Function
Scale. Individual areas of functioning were restored for
one-half to four-fifths of this group. Because of the
wide variation in outcome functioning at follow-up
within specific subjects, the global rating of slight or no
impairment was given to only 55% (N=92) of the
cohort. No impairment was rated for subjects who
were asymptomatic and living independently, had
close relationships, were employed or were otherwise
productive citizens, were able to care for themselves,
and led full lives in general. Other subjects did well in
some areas of functioning but not so well in others.
Theirs was a very mixed picture on a continuum
weighted toward dysfunction.
DISCUSSION
Current assumptions about the long-term course of
schizophrenia and other severe mental illnesses include
the idea that people with repeated episodes are at best
likely to achieve marginal levels of functioning over
time. Heterogeneity of outcome is expected, with a
Am J Psychiatry 144:6, June 1987
HARDING, BROOKS, ASHIKAGA, ET AL
TABLE 3. Results From the Strauss·Carpenter Levels of Function
Scale for the 168 Subjects of the Vermont Study Who Were Alive
and Interviewed
Area of Functioning N %
Not in hospital in past year 140 83
Met with friends every week or ·two 111 66
Had one or more moderately to
very dose friends 128 76
Employed in past year" 79 47
Displayed slight or no symptoms 121 72
Able to meet basic needs 133 79
Led moderate to very full life 128 76
Slight or no impairment in overall
function 92 55
"Qual[ryof work could not be rated; IssUes of confidentiality
prevented visits to subjects' work sites.
dichotomized split between "process" and "reactive"
patients (46), good premorbid and poor premorbid
functioning (47), or type I and type II illness (48), or
the familiar breakdown into "one-third get better,
one-third stay the same, and one-third get worse." The
findings from the Vermont cohort, drawn from the
most chronically ill patients (the lowest third of the
hospital), revealed that over one-half of these once
profoundly ill, long-stay patients had achieved a much
higher level of functioning than had been predicted by
our own research team during the early days of the
patients' community tenure. Their achievement is even
more remarkable given their original levels of chronic-
ity.
These findings hold for other subsets of the cohort as
well. For example, our companion paper in this issue
(35), about the subjects who were rediagnosed as
meeting the DSM-III criteria for schizophrenia, describes
similar proportions of restored and hetewgeneous
functioning, as does our forthcoming paper
about the outcome of the deceased subjects before they
died. Another paper will delineate the reduction by
46% of the number of individual subjects who currently
use the public mental health system in Verinont.
It appears that they have left the formal system and
turned to natural community supports over time.
Our findings of heterogeneity of outcome, with
significant improvement or recovery for half the cohort,
corroborate the results of four other long-term
follow-up studies conducted within the last 15 years:
Manfred Bleuler's 23-year study of 208 patients at
Burgholzli Hospital in Zurich (32), Ciampi and MulIer's
37-year study of 289 patients in Lausanne (49),
Huber and colleagues' 22-year follow-up study of 502
subjects in Bonn (50), and the "Iowa 500" study by
Tsuang et al. (51). The studies from Europe have not
been seriously regarded by some investigators because
of such methodological difficulties as the lack of reliable
diagnostic criteria, the number of deceased and
missing subjects at follow-up (especially in the Lausanne
and Bonn studies), and the use of less structured
. clinical interviews to assess psychopathology and acquire
outcome data (9). However, the Iowa and Ver-
723
.""".'-'..._...-.... -'---'--~~'-~-""""'---""""'--!II'8I~I!I!!lI"II!I_iI!!!'!!l!I!!"_!Ill!!l!II ""
VER,MONT LONGITUDINAL STUDY, I
724
i~
a
tl
'"tl
IT
tl
p
tl
\(
b
a
\\
a
If
..~
c·
b
o
h
6
h
ri
J
5
\
B
R
n
C
.\
l
B
s
I·
I'
.\
L
Am J Psychiatry 144:6, June 19
broken ties with family and community, as well as
impoverished financial and educational backgrounds.
In addition, many subjects behaved violently or in
other unacceptable ways in their communities and
were often brought to the hospital by local police. Few
of these subjects used street drugs, but some abused
alcohol. .
Our findings should be generalizable to similar
long-stay patients, estimated by Minkoff (56) and ; (
others to number about 1.1 million in the United I'States. The results of the study could be considered If.
conservative for the long-term course of severe mental l'"
illness because anyone who is less severel
h
y ill, wlho has ~:
spent less time institutionalized, and w 0 is ess se- ~.
vere1y h;iI}9}c.aPPe:d may have a better chance of long- .~
term recovery. However,' tn-euse .. of 'street drugs, •
imultiple short-term hospitalizations, and the predom- ..~:'il.
inance of DSM-III axis II diagnoses add new complex- ~
ities to the current picture of prolonged psychiatric :'1;
disorder, especially schizophrenia, for subgroups of I
younger patients; these are currently being assessed in
the elegant studies of Test et al. (57) in Wisconsin. .~.
Another aspect of the question of generalizability is ;i;.
whether our subjects' improvement was an artifact of
,'"the quiet, benign rural environment. Zubin (58):;
pointed out, "Where in the world would you get the •
natural history of schizophrenia if not under the best
circumstances? You wouldn't want to go to the ghetto,
where people are suffering from a lot of other sources
of distress, to see whether the outcome is good there..
You want to go to the best place, where triggering·
mechanisms are at a minimum" (p. 407).
The impact of rural life on the course of illness is:
very complex. In 1982 Vermont was a community of
518,846 people scattered in clumps across 9,273;
square miles (59, 60), with a density of 56 people pen:
square mile. Vermonters tend to pride themselves on
Yankee individuality and independence (61, 62); therefore
it is possible to regulate one's social distance anq·
to find a niche sometimes for people with eccentric
behaviors. Despite its ecological niches and visuaJ,.:
beauty, however, Vermont registers fairly high rates of
suicide (63), alcoholism, and incest (64), as well as ..
unemployment, poverty, and long, difficult winters. ,to
Families and patients alike experience the pain of ~I
stigma and social rejection. Therefore, it is not clear:;{i
whether the primary environment has been the decid- ;I:
ing factor in the improvement process, but perhaps the: .; ,
secondary effects of a rural environment, such as small", '
numbers and stability, have been helpful. As we said
earlier in describing phase I of the project's history,
same five members of the clinical research team w
with patients to the community and were available f
support and clinical care over a 10-year period wh
the subjects were both in and OUt of the hospital.
continuity of contact persisted during the second 1
years, as people changed their roles from clinicians
patients to friends and neighbors. (It also may expl .
the 97% rate of original subjects found and the 5 .
refusal rate.)
mont studies have improved on many of these meth.
,odobgical deficiencies, and the findings are nearly
identical to those of the European studies.
Thus, of the 1,300 subjects in the five studies who
were assessed two or three decades later, more than
one-half were found to have considerably improved or
recovered (52). Such similarities in the findings across
studies seem to override the differences. Together, the
data give evidence that, contrary to the expected
downward and deteriorating course for schizophrepia
or for other severe and chronic psychiatric disorders,
symptoms can be ameliorated over time and functioning
can be restored. Further support for the heterogeneity
of patients' outcome functioning is supplied by
many other- shorte!'iel-low-up.studies as well (41, 49,
53-55).
Although the Vermont Longitudinal Research Project
achieved a more rigorous design than many earlier
efforts, the study still has several shortcomings. In
1955, the original selection of the subjects depended
on referrals of patients by the staff to the rehabilitation
program. The original investigators then restricted
entry into the research cohort to the most chronic
patients only.
The project was primarily aimed toward establishing
a treatment program to deinstitutionalize the patients
remaining in the back wards. This program
represented one of the first attempts at deinstitutionalization
in the United States and was one of the
few that was carefully planned. The study became a
follow-up study as time and contacts continued. The
standardization of assessments and the structured protocols
were instituted as the study proceeded. Thus,
the study was not initially planned as a 32-year
follow-up but ended up as such.
There were no subjects who could become a control
group in the Vermont State Hospital, inasmuch as the
entire most severely ill third of the patients in the
state's only hospital (excluding the geriatric population)
was selected to participate in the program. For
years an appropriate comparison sample was sought,
and recently a study was undertaken in the state of
Maine that matched each of the Vermont study's
subjects by age, sex, diagnosis, and level of chronicity
with a patient from the Augusta Mental Health Institute.
These new subjects are being interviewed with the
same protocols and instruments to determine their life
courses and illness trajectories. The primary difference
between the two samples is the presence or absence of
rehabilitation efforts. It is hoped that the difference in
outcome between rehabilitated and nonrehabilitated
subjects will be helpful in the study of mediating
factors in long~term course.
As we have noted, the Vermont subjects were selected
from the most disabled patients in the hospital.
We have portrayed the severity of their handicaps, the'
length of their institutionalization, their lack of response
to phenothiazines after an average of 2V2 years
of psychotropic treatment, the back-ward hopeless
atmosphere in which they had lived for years, the
I,
I
I f
I
I
, '
Further; in any discussion of the effect of rural life, it
; important to point out that the data from the Bonn
nd Zurich studies came from industrialized cities and
Jat Lausanne is a medium-sized city. Only Iowa and
rermont are clearly rural localities, but the trends in
lle data across all five studies with different environ:lents
are similar.
The knowledge gained from our study and others
hat there is a wide range of long-term outcomes
Irovides an impetus to continue the search, begun by
he investigators in the three European studies, for
ongitudinal patterns of course of illness demonstrated
lYsuhgmups QL patientLWh~5=hi~y~ _c"onsid~~~~le
mprovement or recovery and those who do not.
\.dditional questions to be asked now are: When in the
:ourse of their illness did those patients who improved
)egin to do so? Are there any predictors of future
mtcome status? Many older concepts of predictors
lave not been as strong as oncethought (32, 41, 65"":
,7, and our companion paper in this issue), and we
lave begun investigations into these important ques-
:ions.
The answers will begin to reshape our psychological
md biological concepts of severe men,tal illness and the
way in which service delivery systems and treatments
are designed.
ACKNOWLEDGMENTS
The following people contributed to this phase of the project:
design and methodology: Brendan Maher, Ph.D.; the late Robert
Shapiro, M.D.; Bonnie Spring, Ph.D.; Joseph L. Fleiss, Ph.D.; Jane
Murphy, Ph.D.; Joseph M. Tobin, M.D.; Lee Robins, Ph.D.; Leona
Bachrach, Ph.D.; Edward Zigler, Ph.D.; Stanley Herr, J.D.; and Jon
Rolf, Ph.D.; additional aid with instrumentation: William Woodruff,
M.D.; Alan Gelenberg, M.D.; Gerard Hogarty, M.S.W.; Paula
Clayton, M.D.; Janet Mikkelsen, M.S.W.; and Thomas McGlashan,
M.D.; data collection: Paul D. Landed, M.S.W.; Carmine M.
Consalvo, Ph.D.; Janet Wakefield, Ph.D.; William Deane, Ph.D.;
Barbara Curtis, R.N.; and Robert Lager, B.A.; data management:
Susan Childers, A.C.S.W.; Lori Witham; Mary Ellen Fortini, Ph.D.;
Sandi Tower; Andrea Pierce; Mary Noonan; Dorothy Myer; and
Joanne Gobrecht; manuscript review: Luc Ciompi, Prof.Dr.Med.;
Prof. John Cooper; Boris Astrachan, M.D.; Malcolm ~. Bow.ers, ]r.,
M.D.; Richard Musty, Ph.D.; George Albee, Ph.D.; Thomas
Achenbach, Ph.D.; Paul Carling, Ph.D.; Lawrence Gordon, Ph.D.;
and Frederick Schmidt, Ph.D.; and manuscript preparation: Nancy
L. Ryan.
REFERENCES
1. Bachrach LL: The concept of young adult chronic psychiatric
patients: questions from a research perspective. Hosp Community
Psychiatry 1984; 35:573-580
2. Community Mental Health Centers and Psychiatrists. Washington,
DC, and Rockville, Md, American Psychiatric Association
and National Council of Community Mental Health Centers,
1985
3. Neilsen AC III, Stein Ll, Talbott JA, et al: Encouraging psychiarrists
to work with chronic patients: opportunities and limitations
of residency education. Hosp Community Psychiatry
1981; 32:767-775
4. Schwartz SR, Goldfinger SM: The new chronic patient: clinical
characteristics of an emerging subgroup. Hosp Community
Psychiatry 1981; 32:470-474
5. Bachrach LL: A note on some recent studies of released mental
Am J Psychiatry 144:6, June 1987
HARDING, BROOKS, ASHIKAGA, ET AL
hospital patients in the communiry. Am] Psychiatry 1976; 133:
73-75
6. Bellak L: Dementia Praecox. New York, Grune & Stratton,
1948
7. Harding CM, Brooks GW: Life assessment of a cohort of
chronic schizophrenics discharged twenty years ago, in The
Handbook of Longitudinal Research, volli. Edited by Mednick
S, Harway M, Finello K. New York, Praeger, 1984
8. Langfeldt G: The Prognosis in Schizophrenia. Acta Psychiatr
Neurol Scand (Suppl) 1956; 110
9. Shapiro R, Shader R: Selective review of results of previous
follow-up studies of schizophrenia and other psychoses, in
Schizophrenia: An International Follow-up Study. By the World
Health Organization. New York, John Wiley & Sons, 1979
10. Stephens JH: Long-term course and prognosis in schizophrenia.
Semin- P~hiat.ff- 19.7.0;- U6A.-485 "
11. Strauss ]S, Carpenter WT: The prediction of outcome in
schizophrenia, I: characteristics of outcome. Arch Gen Psychiatry
1972; 27:739-746
12. Stromgren E: Recent studies of prognosis and outcome in
mental disorders, in Comparative Epidemiology of the Mental
Disorders. Edited by Hoch P, Zubin ]. New York, Grune &
Stranon, 1961 _
13. Zubin J, Salzinger S, Burdock EI, et al: A biometric approach to
prognosis in schizophrenia. Ibid
14. Carpe"nter WT, Heinrichs DW, Wagman AMI: On the heterogeneity
of schizophrenia, in Controversies in Schizophrenia:
Changes and Constancies. Edited by Alpert M. New York,
Guilford Press, 1985
15. Vaillant GE: The disadvantages of DSM-III outweigh its advantages.
Am] Psychiatry 1984; 141:542-545
16. Cooper ]E, Kendell RE, Gurland BJ, et al: Psychiatric Diagnosis
in New York and London: A Comparative Study of Mental
Hospital Admissions. New York, Oxford University Press,
1972
17. Harding CM, Brooks GW: Longitudinal assessment for a
cohort of chronic schizophrenics discharged twenty years ago.
Psychiatr J Univ Onawa 1980; 5:274-278
18. Brooks GW: Opening a rehabilitation house, in Rehabilitation
of the Mentally III. Edited by Greenblatt M, Simon B. Washington,
DC, American Association for the Advancement of
Science, 1959
19. Brooks GW: Rehabilitation of hospitalized chronic schizophrenic
patients, in Chronic Schizophrenia. Edited by Appleby
L, Scher J, Cumming J. Chicago, Free Press, -1960
20. Brooks GW: Motivation for work in psychiatric rehabilitation.
Dis Nerv Syst 1961; 22:129-132
21. Brooks GW: Rural community influences and supports in a
rehabilitation program for state hospital patients, in Mental
Patients in Transition. Edited by Greenblatt M, Levinson DJ;
K1erman GL. Springfield, III, Charles C Thomas, 1961
22. Brooks GW, Deane WN: Attitudes of released chronic schizophrenic
patients concerning illness and recovery as revealed by
a structured post-hospital interview. ] C1in Psychol 1960; 16:
259-264
23. Brooks GW, Deane WN: The chronic mental patient in the
community. Dis Nerv Syst 1965; 26:85-90
24. Brooks GW, Deane WN, Lagor RC, et al: Varieties of family
participation in the rehabilitation of released chronic schizophrenic
patients. ] Nerv Menr Dis 1963; 136:432-444
25. Brooks GW, Deane WN, Laqueur HP: Fifteen years of work
therapy. Dis Nerv Syst (Suppl) 1970; 31:161-165
26. Chittick RA, Brooks GW, Irons FS, et al: The Vermont Story.
Burlington, Vt, Queen City Printers, 1961
27. Bleuler M: Research and changes in concepts in the study of
schizophrenia, 1941-1950. Bull Isaac Ray Medical Library
1955; J:1-132
28. Bleuler M: Conception of schizophrenia within the last fifty
years and today. Proc R Soc Med 1963; 56:945-952
29. Bleuler M: A 23-year longitudinal study of 208 schizophrenics
and impressions in regard to the nature of schizophrenia, in The
Transmission of Schizophrenia. Edited by Rosenthal D, Kery SS.
Oxford, England, Pergamon Press, 1968
725
VERMONT LONGITUDINAL STUDY, I
" '
, .
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
Bleuler M: The genesis and nature of schizophrenia. Psychiatry
Digest 1969; 30: 17-26
Bleuler M: The long-term course of the schizophrenic psychoses.
Psychol Med 1974; 4;244-254
Bleuler M: The Schizophrenic Disorders: Long·Term Patient
and Family Studies. Translated by Clemens SM. New Haven,
Yale University Press, 1978
Bleuler M: Schizophrenic deterioration. Br J Psychiatry 1983;
143:78-79
Deane WN, Brooks GW: Five·Year Follow-Up of Chronic
Hospitalized Patients. Waterbury, Vermont State Hospital, Sept
1967
Harding CM, Brooks GW, Ashikaga T, et al: The Vermont
longitudinal study of persons with severe mental illness, 1I:
long-term outcome of subjects who retrospectively met DSM·
III criteria for schizophrenia. Am JPsychiatry 1:987; 144:727-
735
fu"~<:>~__G~:EEE:!i~nc~ with the use of chlorpromazine and
reserpine in psychiatry. NTngITMed 1956; 1'54:T119""-1123
Hollingshead AB, Redlich FC: Social Class and Mental Illness:
A Community Study. New York, John Wiley & Sons, 1958
Leighton AH, Leighton DC: Gregorio, the Hand-Trembler: A
Psychobiological Personality Study of a Navaho Indian. Peabody
Museum Papers. Cambridge, Mass, Harvard University
Press, 1949
World Health Organization: Collaborative Project on Determinancs
of Outcome of Severe Mental Disorders {1977-1979):
Research Protocols. Geneva, WHO, Aug 1978
Strauss JS, Harder DW: The Case Record Rating Scale. Psychiatry
Res 1981; 4:333-345
Strauss JS, Carpenter WT Jr: Prediction of outcome in schizo·
phrenia, III: five-year outcome and its predierors. Arch Gen
Psychiatry 1977; 34:159-163
Endicott J, Spitzer RL, Fleiss JL, et al: The Global Assessment
Scale: a procedure for measuring overall severity of psychiatric
disturbance. Arch Gen Psychiatry 1976; 33:766-771
Bureau of the Census: Statistical Abstraers of the United States
(1944-1945). Washington, DC, US Department of Commer~e,
1945, pp 223, 231
Summary of US Census Data for Vermont Cities and Towns.
Montpelier, Agency of Development and Community Affairs,
Stare of Vermonr, 1982
Schwartz C, Muller C, Spitzer RL, et al: Community Care
Schedule (CCS). New York, New York State Psychiatric Institute,
1977
Garmezy N: Process and reactive schizophrenia: some conceptions
and issues, in The Role and Methodology of Classification
in Psychiatry and Psychopathology: NIMH Public Health Service
Publication 1584. Edited by Katz MM, Cole J, Batton WE.
Washington, DC, Government Printing Office, 1965
PhilHps L: Case history data and prognosis in schizophrenia. J
Nerv Ment Dis 1953; 117:515-525
Crowe T: Molecular pathology of schizophrenia: more than
one disease process? Br Med J 1980; 12: 16-68
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
Ciompi L, Miiller C: Lebensweq und Alter der Schizophrenen.
Eine katamnestische lonzeitstudies bis ins Senium. Berlin,
Springer Verlag, 1976
Huber G, Gross G, Schiittler R: Schizophrenie: Verlaufs und
socialpsychiarrische Langzeit unter suchungen an den 1945 bis
1959 in Bonn hospitalisierten schizophrenen Kranken: Monographien
aus dem Gesamtgebiete der Psychiattie. Bd 21. Berlin,
Springer Verlag, 1979
Tsuang MT, Woolson RF, Fleming JA: Long-term outcome of
major psychoses, I: schizophrenia and affective disorders compared
with psychiatrically symptom-free surgical conditions.
Arch Gen Psychiatry 1979; 36:1295-1301
Harding CM, Zubin J, Strauss JS: Chronicity in schizophrenia:
faer, partial fact, or artifact? Hosp Community Psychiatry (in
press)
Gardos G, Cole JO, LaBrie RA: A 12-year follow-up study of
chronic schizophrenics. Hasp Community Psychiatry 1982; 33:
!nlT-9lr4
Strauss JS, Kokes RF, RitzIer BA, et al: Patterns of disorder in
first admission psychiarric patients. J Nerv Ment Dis 1978; 166:
611-625
World Health Organization: The International Pilot Study of
Schizophrenia. Geneva, WHO, 1973
Minkoff K: A map of chronic patients, in The Chronic Mental
Patient. Edited by Talbott JA. Washington, DC, American
Psychiatric Association, 1978
Test MA, Knoedler WH, Allness DJ: The long-term rreatment
of young schizophrenics in a community support program. New
Directions for Mental Health Services 1985; 26:17-27
Zubin J: General discussion, Controversies in Schizophrenia:
Changes and Constancies. Edited by Alpert M. New York,
Guilford Press, 1985
Bureau of the Census: Local population estimates-Vermont
(1982), in Current Population Reports, Series P-26, Number
82-45-SC, US Department of Commerce. Washington, DC, US
Government Printing Office, 1984
Arnold TW: Land area and population history in Vermont Data
Series. Burlington, University of Vermont, Center for Rural 1"
Studies, 1982 ,~
Hill RN: Yankee Kingdom: Vermont and New Hampshire. I
New York, Harper & Brothers, 1960 ' .;;j;
Merrill PH: Vermont Under Four Flags: A History of the Green ~
Mountain State (1635-1975). Barre, Vt, Northlight Studio !
Press, 1975 .
Statistics. Burlington, Vermont State Health Department, 1982
Division of Planning and Rehabilitation Services: Statistics.
Waterbury, Vermont Department of Human Services, 1982
Ciompi L: Catamnestic long-term study on the course of life and ~,
aging in schizophrenia. Schizophr Bull 1980; 6:606-608
Huber G, Gross G, Schiirtler R, et al: Longitudinal studies of
schizophrenic patients. Schizophr Bull 1980; 6:592-605
Vaillant GE: A 10-year follow-up of remirting schizophrenics.
Schizophr Bull 1978; 4(11):78-85
726 Am J Psychiatry 144:6, June 1