UMEÅ UNIVERSITY MEDICAL DISSERTATIONS New Series No 139 From the Department of Psychiatry Umeå University, Umeå, Sweden
PSYCHOSOCIAL ASPECTS OF DEPRESSIVE DISORDERS
by
MARTIN EISEMANN
Umeå University 1985 ABSTRACT
Eisemann Martin: ’Psychosocial aspects of depressive disor ders’. From the Department of Psychiatry, Umeå University,
Umeå, Sweden.
The objective of this study was to elucidate the possible importance of factors from the social environment for the development of depression. As a theoretical framework, Eng el's biopsychosocial model (Engel, 1980), based on systems theory, has been applied. Proceeding from the single indivi dual (characterized by experience, personality, behaviour) as the highest level of the organismic hierarchy the follo wing system levels have been taken into account: dyads, family, community, culture-subculture.
The depressive patients (n=lll) showed to be living in a narrowed social network and to lack confiding relationships compared with a non-psychiatric control sample (n=98). The personality characteristics (e.g. anxiety, detachment, sus picion) of the patients were related to experienced loneli ness, contact difficulties, social network features and lei sure activities. By means of a discriminant analysis 83% of the subjects could be correctly classified. In a study of perceived parental rearing, depressives showed to have expe rienced lack of emotional warmth. As regards social class an overrepresentation of social class III in the subgroups of unipolar, bipolar and unspecified depression was observed.
Finally, implications for treatment are discussed in favour of a combination of drug and cognitive psychotherapy. Future research strategies are also suggested.
Key words: Depression, social environment, biopsychosocial model, social support, social networks. ”Psychosocial Aspects of Depressive Disorders”
Akademisk avhandling som med vederbörligt tillstånd av Rektorsämbetet vid Umeå Universitet för avläggande av doktorsexamen i medicinsk veten skap (dr med sc) i ämnet psykiatri kommer att offentligen försvaras i Psyki atriska klinikens föreläsningssal A, Regionsjukhuset, Umeå, fredagen den 29 mars 1985, kl 13.15.
av
Martin Eisemann
Umeå 1985 UMEÅ UNIVERSITY MEDICAL DISSERTATIONS New Series No 139 From the Department o f Psychiatry Umeå University, Umeå, Sweden
ABSTRACT
Eisemann Martin:’Psychosocial aspects of depressive disorders’. From the Department of Psychiatry, Umeå University, Umeå, Sweden.
The objective of this study was to elucidate the possible importance of fac tors from the social environment for the development of depression. As a theoretical framework, Engel’s biopsychosocial model (Engel, 1980), based on systems theory, has been applied. Proceeding from the single individual (characterized by experience, personality, behaviour) as the highest level of the organismic hierarchy the following system levels have been taken into account: dyads, family, community, culture-subculture. The depressive patients (n = 111) showed to be living in a narrowed social network and to lack confiding relationships compared with a non-psychiatric control sample (n = 98). The personality characteristics (e.g. anxiety, detach ment, suspicion) of the patients were related to experienced loneliness, con tact difficulties, social network features and leisure activities. By means of a discriminant analysis 83% of the subjects could be correctly classified. In a study of perceived parental rearing, depressives showed to have experienced lack of emotional warmth. As regards social class an overrepresentation of social class III in the subgroups of unipolar, bipolar and unspecified depres sion was observed. Finally, implications for treatment are discussed in favour of a combination of drug and cognitive psychotherapy. Future re search strategies are also suggested.
Key words:Depression, social environment, biopsychosocial model, social support, social networks. 3
To my mother and in memory of my father 4
'Nicht allein das Angeborene, sondern auch
das Erworbene ist der Mensch'
J W von Goethe 5
This doctoral dissertation is based on the following papers carried out at the Department of Psychiatry, Umeå Universi
ty:
I. Eisemann M. Contact difficulties and the experience of
loneliness in depressed patients and non-psychiatric
controls. Acta Psychiatr Scand 1984:70:160-165.
II. Eisemann M. Leisure activities of depressive patients.
Acta Psychiatr Scand 1984:69:45-51.
III. Eisemann M. The availability of confiding persons in
depressed patients. Acta Psychiatr Scand
1984:70:166-169.
IV. Eisemann M. The relationship of personality to social
network aspects and loneliness in depressed patients.
Acta Psychiatr Scand 1984:70:337-341.
V. Eisemann M. Depressed patients and non-psychiatric
controls: discriminant analysis on social environment
variables. Acta Psychiatr Scand 1985 (in the press).
VI. Eisemann M. Social class and social mobility of dep
ressed patients. Acta Psychiatr Scand (submitted).
VII. Perris C, Arrindell W, Eisemann M, Perris H, van der
Ende, von Knorring L. Perceived depriving parental
rearing and depression. Brit J Psychiatry 1985 (in the
press).
VIII. Eisemann M, Perris C, Perris H, von Knorring L. Per
ceived parental rearing practices in depressed
patients in relation to social class. Acta Psychiatr
Scand 1984:70:568-572. 6
CONTENTS
INTRODUCTION 7
Psychosocial factors in disease etiology 7
Theoretical framework 13
Aims of the present investigation 14
MATERIAL 17
Patients 17
Control subjects 18
METHOD 18
Diagnostic classification 18
Instruments 19
Procedure 23
Statistical methods 24
RESULTS 24
Paperi 24
Paper II 25
Paper III 26
Paper IV 26
Paper V 27
Paper VI 28
Paper VII 29
Paper VIII 30
GENERAL DISCUSSION AND CONCLUSIONS 31
Implications for treatment 33
Suggestions for future research 34
REFERENCES 36
ACKNOWLEDGEMENTS 46 7
INTRODUCTION
Depression can be regarded as one of the most common and most distressful illnesses afflicting human beings.
From a review of epidemiological investigations of affective disorders Boyd and Weissman (1981) estimated the lifetime risk for depression (subclinical depressive mood states and bipolar conditions excluded) at 8-12% for men and 20-26% for women in Western cultures. Figures from a 25 years prospec tive study on a total population in Southern Sweden by Hag- nell et al. (1982) demonstrate an increase in lifetime risk of depression during the last decades. For males in the
20-39 years group with a moderate or severe depression the risk of exhibiting the disorder was ten times higher during the period 1957-72 compared with the period 1947-57. Klerman
(1978) ascribed the increasing rate of depression partly to changes in the social environment, such as the breakdown of traditional social support systems (e.g. extended family), increased (often involuntary) geographical mobility, modern housing, and the stand back of religious institutions. Fur thermore, people have higher expectations and demands on the psychiatric services for relief from distressing symptoms.
The same author suspects that mankind in contrast to the post-war 'age of anxiety' has now entered an 'age of melancholy', precipitated by recent global events and repea ted 'doomsday prophecies' about overpopulation, famines, ecological destruction and the threat of a nuclear holocaust
(Grinspoon, 1982).
Psychosocial factors in disease etiology
The role of the social environment in disease etiology has been studied for quite a long time. Kraus and Lilienfeld
(1959) investigated the health consequences of the termina- 8
tion of important social relationships. In their study of a young widowed group they showed that widowers have a death rate three to five times higher than married men of the same age. However, it was not before the mid 1970's that the con cept of 'social support' was introduced by Cassel (1974,
1976). In his review Cassel (1976) presented strong evidence that social support had positive implications for health
status, and served as a buffer against the effects of psyc hosocial or physical stress. Cassel based his assumptions on
Frost's (1936) extended concept of host, agent and environ ment. He suggested that besides such general factors as nutrition, fatigue and overwork another category of environ mental factors should be taken into account which could
influence host susceptibility to disease agents from the environment, that is, 'the presence of other members of the same species'.
The primatologist Harlow and his co-workers (Harlow et al.,
1960, 1966, 1974) have shown that attachment is a powerful
'primary drive' in its own right, just as hunger and sex are. For example, they showed that totally isolated monkeys tended to exhibit a depressive type posture, including beha viour patterns such as huddling, rocking, and self-clutch ing. Mason (1964) found that changes in group membership and in quality of relationships are accompanied by significant neuroendocrine changes affecting the pituitary, the adreno cortical system, the thyroid and the gonads.
Inspired by Lorenz's work on imprinting (Lorenz, 1945),
Bowlby (1969, 1977) simultanously and independently, pointed to the contribution of the environment to an individual's psychological development and well-being. He proposed an
'attachment theory' to describe the need of human beings to establish and to maintain affectional bonds to other people as well as to explain the variety of psychological impair- 9
ment such as depression, emotional detachment and anxiety in
connection with loss and separation. Bowlby sees a strong
causal relationship between an individual's attachment expe
rience with his parents during infancy and his later capaci
ty to form affectional bonds. He also believes that when
social support, in the form of an attachment figure, is
available early in life, people become more self-reliant and
capable to function as supports for others.
Already at the end of the last century Durkheim (1883) noti
ced that 'social development rests on no other variable fac
tor than the number of individuals in relationship and their
material and moral closeness'. In their comprehensive fol
low-up study of more than 6000 Alameda County residents,
Berkman and Syme (1979) applied a social network index in
cluding the number of close friends and relatives and fre
quency contacts. The age-adjusted mortality rate for the
most isolated men was 2.3 times higher than for those with
the most social contacts. The corresponding figure for women was even higher, i.e. 2.8 times.
In his review of a variety of studies covering the whole
life cycle, Cobb (1976) found strong evidence that social
support can protect people from a number of pathological
states, e.g. low birth weight, arthritis, tuberculosis,
alcoholism and depression. According to Cobb, 'social sup
port begins in utero, is best recognized at the maternal breast, and is communicated in a variety of ways, but espe
cially in the way the baby is held (supported). As life progresses, support is derived increasingly from other mem bers of the family, then from peers at work and in the com munity, and in case of special need from a member of the helping professions. As life's end approaches, social sup port ... is again derived mostly from members of the family'. 10
Even politicians attached importance to the concept of 'so
cial support'. In the USA, the President's Commission on
Mental Health Task Panel on Community Support Systems (1978)
concluded from a review of more than 200 studies that it was
desirable to 'strengthen the natural networks to which peop
le belong and on which they depend' in order to reduce vul nerability and risk and to increase social competence and
sense of community.
When studying 'social support' the majority of current re
search workers apply the concept of social networks instead of the more restricted concept of dyadic relationships. The
term social network was introduced by a British anthropolo gist when studying a Norwegian fishing village (Barnes,
1954). Barnes used concepts from mathematical graph theory
to describe the individual's social field. As the interest
in social networks has increased, so has the diversity of
operational definitions. Cohen and Solokovsky (1978) defined
social network relationships as 'all links within the prece
ding year with a frequency of at least once a month'. Tols-
dorf's (1976) criteria were that members of a social network
'must know each other by name, they must have an ongoing personal relationship and they must have some contact at
least once a year'. Shulman (1975) asked the respondents to
'name the people outside his or her household that he or she
feels closest to'. Henderson and co-workers (1978) used the
term 'primary group' - originally introduced by Cooley
(1909) - which they characterized as 'those with whom one has interaction and committment'. The concept of social net work presents one way to examine the total social field wit hin which the individual is living. According to Caplan
(1974) social networks 'operate by mobilizing the resources of the individual, enabling him to master emotional bur dens'. Social networks enable people to give and receive love and affection, to be dependent and be depended on. 11
Being part of a social network gives one self-esteem, raises one's self-confidence and maintains in this way psychologi cal well-being and ameliorates psychological and/or physio logical stress.
During the last ten years a series of studies have appeared comparing the nature of social relationships of psychiatric patients with those of controls. In their study of working class women, Brown et al. (1975) identified the 'lack of an intimate confiding relationship' as one of four factors which were associated with increased neurotic morbidity.
When comparing the intra- and extrafamilial contacts of dep ressive patients with those of somatically ill patients,
Blöschl (1976) found that the former reported a significant ly lower number of contact-persons outside the family. Hen derson et al. (1980) reported an association between the lack of available attachment figures (e.g. spouse, parents, siblings) and neurotic symptomatology. In a study by Mine field (1979) depressed women were found to report fewer con fidants than did controls. However, the comparability of the findings from these studies is hampered by the utilization of badly defined diagnostic groups. In addition, most stu dies have been focused on neurotic patients.
One of the determinants of social behaviour which in this context is so far largely unexamined, are individual per sonality characteristics. They have important implications when discussing causality in studies of social networks. One of the common interpretations of the relationship between social network déficiences and psychological dysfunction is that the former causes the latter. However, Henderson et al.
(1978) suggested another plausible alternative. They pro posed 'that patients are deficient in their primary group because they are deficient in the social skills necessary to establish and maintain it. Their personality attributes 12
might lead independently both to neurotic symptoms and to primary group deficiency'.
One of the determinants which are generally accepted as important for the shaping of an individual's personality structure, is parental rearing practices (Erikson, 1950;
Newman & Newman, 1975; Albright, 1978; Murphy & Moriarty,
1976). Becker (1964) found that warmth and permissiveness result in an active, sociable, creative, independent and friendly personality. Parental hostility and strictness, on the other hand, promote neurotic symptoms and a socially withdrawn personality. In one of our studies on perceived parental rearing behaviour in depressed patients (Perris et al., 1983a) we found significant correlations between power assertive practices (punitive, overprotective, shaming, guilt engendering), lack of emotional warmth and aspects of socialization, anxiety and suspicion. In another study (Per ris et al., 1983b) we investigated the personality traits in a series of depressed patients with a family loading for affective disorders and in sporadic depressives, classified according to the definitions given by Winokur (1978, 1979).
The group with heredity loading differred significantly from the sporadic group on several personality aspects, sugges ting an interaction between family loading with affective disorders and personality and perceived rearing practices.
Another factor from the social environment which has fre quently been associated with depression, is social class.
However, the results reported in different studies have pro ved to be inconsistent. Manic-depressive illness was first reported to be more frequent amongst the upper and middle classes (Stern-Piper, 1925). However, more recent findings do not support such a relationship (Weissman & Myers, 1978).
Epidemiological survey studies report a somewhat higher pre valence of non-bipolar depression in lower social strata 13
(Weissman & Myers, 1978). Brown et al. (1975) reported hig her rates of clinical depression among working-class women than among upper class women. In the study by Weissman et al. (1978), however, life time rates were higher in the upper social classes. The authors suggest that this is due to a longer duration of symptoms in the lower social class as a result of poor access to treatment. To exclude the pos sibility that results might be influenced by a decline in social class as a consequence of the depressive disorder,
Birtchnell (1971) suggested studying the parental social class. This approach makes it possible to study the social mobility of the patients. So far, the findings on the impact of social class on depressive disorders are rather inconsis tent. The usage of different and often poorly defined diag nostic criteria and definitions make comparisons difficult.
Theoretical framework
Some sixty years ago the biologist Paul Weiss (1925, 1977)
introduced the concept of systems theory in biology in an attempt to harmonize empirical holism and practical reduc-
tionism. He elaborated the basic phenomenon that nature is built up as a hierarchically arranged continuum in which
less complex units are superposed by more complex and larger units (systems). This theory was futher developed by von
Bertalanffy (1968) and by Engel (1980). The latter used sys tems theory as a basis for his biopsychosocial model of medicine that allows for the missing dimensions in biomedi cal model. The biopsychosocial model is 'per definitionem' a way of understanding disease phenomena which have to be seen as a dynamic product processed by the interaction of biolo gical, psychological and sociocultural factors with each other and with the biophysical and psychosocial environment.
In the past, emphasis was put on either biological, beha vioural, sociological, psychoanalytical or existental 14 aspects, depending on which school one belonged to. The pre sent model rejects the view that illness is caused by one single factor. Body and mind are seen as one entity. The human being as a whole is more than an addition of the parts. Figure 1 shows the hierarchical structure of nature, which in fact is made up of two hierarchies, i.e. with the individual (person) at the same time being the highest level of the organismic level and the lowest unit of the social hierarchy. Engel (1980) emphasizes the distinctive qualities and relationships of each level (system) which make methods and rules for study unique for that level, necessary. Figure
2 points to another important concept of living systems theory. Living systems are highly open and allow a steady flow of inputs and outputs of information across the boun daries. The figure also illustrates the complexity of the relationships between the internal and external environment.
The interaction between the biological (organismic) and psychosocial system makes them both vulnerable for disrup tion. The vulnerability is varying with time since it depends on the actual status of the systems involved. This concept of individual vulnerability has been the theoretical framework of the research on depression carried out at our
Department. Perris & Perris (1984) point to the probability that some of the interacting factors might 'have been ope rating long before the onset of the depressive breakdown'.
They also stress the implications of the concept of indivi dual vulnerability for treatment planning.
Aims of the investigation
The objective of the present investigation was to elucidate the possible importance of a set of factors from the social hierarchy (according to Engel, 1980) for the development of depression. Proceeding from the single individual (characte rized by experience, personality, behaviour) as the highest 15
SYSTEMS HIERARCHY
(LEVELS OF ORGANIZATION)
BIOSPHERE ? SOCIETY - NATION $ CULTURE - SUBCULTURE * COMMUNITY t FAMILY t DYADS *______PERSON
(experience & behaviour) ------5------NERVOUS SYSTEM * ORGANS/ORGAN SYSTEMS % TISSUES I CELLS t ORGANELLES % MOLECULES t ATOMS % SUBATOMIC PARTICLES
Figure 1. Hierarchy of natural systems. Adapted from Engel
GL, Am J Psychiatry 1980:137:537 16
BIOSPHERE SOCIETY-NATION CULTURE-SUBCULTURE (age, sex, cl ass, -s COMMUNITY (health care, work, n> FAMILY (nuclear, extended DYADS (doctor, PERSON (experience CO NERVOUS SYSTEM Cl ORGAN/ORGAN SYSTEMS TISSUE
CELL CO o o ORGANELLE 3 o MOLE CULE
Figure 2. Continuum of natural systems. Adapted from Engel
GL, Am J Psychiatry 1980:137:537. 17
level of the organismic hierarchy, the following system levels have been taken into account: dyads (spouse, friend), family (children, relatives), community (acquaintances, lei sure) , culture-subculture (social class, parental rearing).
The specific hypotheses which have formed the bases for the different investigations are described in connection with the summaries of these studies. Within the theoretical fra mework of systems theory and within our concept of indivi dual vulnerability we attempted a) to identify eventual
interactions between different factors related to the occur rence of depressive disorders, b) to investigate whether socio-environmental factors differ in various diagnostic groups, and c) to discuss the implication of environmental
factors for the treatment of depressive patients.
MATERIAL
A. Patients : The basic series comprised 111 patients (41 male and 70 female) with a mean age of 46.3 years (range
21-67 years) who were treated as hospitalized or out-pa
tients at the Department of Psychiatry, Umeå University.
They were consecutive patients. All patients consented to participate in a comprehensive study of depression in its biological, clinical and psychosocial aspects. Exclusion criteria were the following: older than 67 years, alcohol or
drug problems as primary reason for treatment, a diagnosis of schizophrenia or cycloid psychosis, depression related to diagnosed brain damage.
For the studies on perceived parental rearing behaviour,
data on some additional patients (total n=141) were availab
le (papers VII and VIII). Due to missing data on some variables the Ns are differing. 18
B. Control subjects: The non-psychiatric controls were collected among consecutive patients consulting a primary care unit for minor somatic disorders. Subjects with a psyc hiatric diagnosis in their past history were excluded from the group. The control group consists of 43 males and 54 females ((with a mean age of 43.3 +-11.7 years (range 20 -
70 years). For the study of parental rearing behaviour a series comprising of 205 healthy individuals collected on a previous occasion (Arrindell et al., 1985) has been used.
METHOD
Diagnostic classification
The hospital records of the patients participating in the comprehensive study of depression were scrutinized by two experienced psychiatrists who classified the patients accor ding to current classification systems of affective disor ders - DSM III (American Psychiatric Association, 1980),
ICD-9 (WHO, 1977), Feighner et a l 's diagnostic criteria
(Feighner et al., 1972), Multi-Aspects Classification Model
"MACM" (Ottosson & Perris, 1973; Eisemann et al., 1980). For the purpose of the present investigations the classification of affective disorders used at our Department for research purposes was applied. The diagnostic subgroups taken into account were: unipolar, bipolar, neurotic, unspecified dep ression according to the definitions given by Perris (1973) and by d ’Elia et al. (1974). These definitions might be sum marized as follows:
a. Unipolar depression: patients who had suffered from
recurrent episodes of depression (generally at least
three), mostly of a psychotic severity with a symptom-
free interval between the episodes. 19
b. Bipolar depression: patients who had previously suffered
from at least one episode of depression and one of mania
and who were currently depressed when entering the study.
c. Neurotic-reactive depression (RND): patients characte
rized by a vulnerable personality structure and a prone
ness to depressive reactions in case of adverse external
events; mostly of non-psychotic severity.
d. Unspecified depressive disorder (NUD): patients who suf
fered from a depressive disorder either psychotic or
non-psychotic but who did not fulfil the criteria for
inclusion into any of the aforementioned groups.
Table 1. The series
n sex ratio (M/F) age (years)
Unipolar 33 0.6 50.9 +- 11.2
Bipolar 22 0.6 43.5 +- 12.9
NUD 30 0.6 47.0 +- 13.3
RND 25 0.5 40.8 +- 15.3
Controls 98 0.7 43.3 +- 11.7
Instruments
For the different parts of the thesis the following instru ments were used:
1. Assessment of leisure activities
For the assessment of leisure activities a specially con structed 'Leisure Activity Inventory' was administered. This inventory covers the following aspects: type of activity, frequency, place, presence of other people, and satisfaction 20 with the leisure activity.
2. Assessment of the social network, contact difficulties
and loneliness
For the assessment of the social network a specially con structed and self-administered inventory was used. The fif teen questions of the inventory concern: sociodemographic data, availability of confiding persons, contacts with a) family members (not living in the same household), b) other relatives c) friends, and d) acquaintances. The subjects were asked to report ’regular' contacts, i.e. people they meet regularly at least on a monthly basis. The extent of contact difficulties and loneliness was measured by means of visual analogue scales which formed the last section of the questionnaire.
3. Assessment of personality
Two inventories measuring different aspects of personality were applied to the patient sample:
a) The Cesaree Marke Personality Schedule (CMPS) (Cesarek &
Marke, 1968) is intended to measure eleven of the psychoge nic needs suggested by Murray (Murray, 1938) and is based upon 165 questions to be answered 'yes’ or 'no'. Psychogenic needs according to Murray are possibly secondary to primary viscerogenic needs such as hunger and thirst. The measured variables are:
(1) Achievement: need to accomplish something difficult and to rival and surpass others; (2) Affiliation: need to please and win affection of cathected objects and to adhere and remain loyal to friends; (3) Aggression: need to revenge an injury, impulsive aggression and irritability; (4) Defense of status: need to maintain self-esteem by support and app- 21
rovai from others; (5) Guilt feeling: guilt feelings and superego conflicts; (6) Dominance: need to dominate and to lead others; (7) Exhibition: need to expose oneself, to be in the center, to be noticed; (8) Autonomy: need for autono my and independence; (9) Nurturance: need to help, nurse and take care of others; (10) Order: need for order, cleaniness and planning; (11) Succorance: need to be helped, nursed, supported and consoled. In earlier studies the CMPS has pro ved to give valuable, reliable and valid information (St- randman, 1978).
b) The second inventory used has been constructed for research purposes at the Karolinska Institute in Stockholm
(Schalling, 1978; Schalling & Åsberg, 1981). The inventory is named KSP (Karolinska Scales of Personality) and is com posed of subscales from personality schemes in use, e.g. the
Sensation Seeking Scale (SSS) (Zuckerman, 1971) and the
Impulsiveness-Venturesomeness-Empathy (IVE) inventory (Ey senck & Eysenck, 1978). The inventory comprises 135 ques tions grouped in the following 15 subscales: psychic anxie ty, somatic anxiety, muscular tension, social desirability, impulsiveness, monotony avoidance, distance preference, psychasthenia, socialization, indirect aggression, verbal aggression, irritability, suspicion, guilt and inhibition of aggression. In an earlier study when patients were investi gated when depressed and when recovered from their depressi ve episode (Perris et al., 1979) it was demonstrated that most of the subscales were fairly state independent.
4. Psychopathological rating
For the clinical assessment each patient was rated by means of the Comprehensive Psychopathological Rating Scale (CPRS)
(Åsberg et al., 1978). For the purpose of the present study the Depression Subscale Score and the scores for two princi 22
pal factors derived from the subscale have been used (Perris et al., 1984). Factor I corresponds to the core symptoms of the depressive syndrome and Factor II to a component of retardation.
5. Assessment of memories of perceived parental rearing
practices
For this purpose the EMBU ('Egna Minnen Beträffande Upp- forstran') (Perris et al., 1980), an 81-item questionnaire to be answered according to a 4-point scale, separately for the mother and the father, was used. This instrument, now available in ten different languages, has been successfully applied by other research workers (Ross et al., 1983; Clayer et al., 1984; Arrindell et al., 1983). Originally the items were grouped in 14 arbitrary subscales covering the follo wing aspects of rearing practices: abusive, depriving, puni tive, shaming, rejecting, overprotective, overinvolved, tolerant, affectionate, performance oriented, guilt-engende ring, stimulating, favouring siblings and favouring subject.
Two additional questions refer to a global judgement about the strictness of the parental rearing practices and to the extent to which they have been consistent. Subsequent factor analyses have led to the identification of the following three principal factors (Arrindell et al., 1983): ’rejec tion' (reflecting an assertive, punitive and rejecting rea ring practice), 'emotional warmth' (including items refer ring to an affectionate, tolerant and stimulating attitude) and 'overprotection' (reflecting practices characterized by overinvolvement and overprotection). These factors have a satisfactory internal consistency and have also proved to be transferable across samples (Arrindell et al., 1985). 23
6. Social-psychiatric questionnaire
This questionnaire comprised a variety of socio-demographic
variables such as: education, profession, profession of
parents, civil status, number of children, sibship size,
native place, changes of residence etc.
For the classification into social classes the profession-
based criteria drawn up in the 'Låginkomstutredning' (Low
Income Survey) (Johansson, 1970) have been used. According
to this survey, 7.8% of the Swedish population belong to
social group I, 34.7% to social group II and 57.5% to social
group III.
Procedure
Immediately after admission and before any treatment was
initiated the patients were rated by means of the Comprehen
sive Psychopathological Rating Scale (CPRS).
The other inventories and questionnaires were consistently
administered when the patients had recovered from their de
pressive episode, i.e. in most cases during the last days before discharge from hospital. When appropriate (e.g.
assessment of social network) the patients were explicitly
asked to report the conditions existing prior to onset of
their present depressive episode. 24
Statistical methods
The statistical analyses were carried out at the Umeå Uni versity Computer Centre (UMDAC) using SPSS standard program mes (Statistical Package for the Social Sciences, Nie et al., 1975, 1981). The significance of inter-group differen ces was tested by means of the Student's t-test, the Kru- skal-Wallis analysis of variance or the non-parametric
Mann-Whitney U-test, as appropriate.
Differences in distributions were tested by the Chi square test, with Yates' correction for continuity when the number of patients was small (Ferguson, 1959). Relationships bet ween variables have been calculated as Spearman's rank cor relation coefficients.
To statistically distinguish between samples on a set of variables a discriminant analysis was performed.
RESULTS
Paper I : 'Contact difficulties and experience of loneli
ness in depressed patients and non-psychiatric
controls'.
There is empirical evidence from the literature, that lone liness and depression often co-occur (Russel et al., 1980).
The purpose of the present study was to investigate possible differences in the experience of loneliness and contact dif ficulties between depressed patients and non-psychiatric controls. Furthermore, it was tried to elucidate the rela tionship between these variables and social network measures and severity of the depressive syndrome. The results show that depressed patients felt lonelier and suffered more from their loneliness than did controls. Among the diagnostic 25
subgroups the neurotic depressives scored highest. This group also reported most contact difficulties. For the dep ressives, the experienced loneliness and the number of regu lar contacts with family members were negatively correlated.
In the unipolar and bipolar group even the number of friends was negatively correlated. In all groups there was a positi ve relationship between contact difficulties and suffering from loneliness. There were significant correlations between severity of the depressive syndrome and the extent of lone liness suggesting that loneliness not only precedes but also produces symptoms. For a further elucidation of this issue it is finally suggested to study the relevance of personali ty characteristics in this context.
Paper II: 'Leisure activities of depressive patients'.
The objective of this study was to explore the leisure beha viour of depressed patients and their possible relationships to personality characteristics. Based on the observation that depressed patients have a restricted social network they were expected to engage in fewer leisure activities and to be associated with fewer people during these activities than controls. The results supported these hypotheses. The controls were engaged in significantly more leisure activi ties than were the depressives. The former conducted also more activities involving other people and more activities occurring outside home. As concerns personality, the 'achie vement' scale from the CMPS was correlated with activities which were active, involving others and done daily. The sca les measuring 'defense of status', 'succorance' and 'passive dependence' were negatively correlated with most types of leisure activities. Similar relationships were observed for the KSP scales 'detachment', 'psychic anxiety', 'somatic anxiety' and 'psychasthenia'. 26
Paper III: 'The availability of confiding persons in de
pressed patients'.
Social support has increasingly attracted the attention of research workers. The original finding by Brown et al.
(1978) that 'lack of an intimate confiding relationship' was a vulnerability factor for psychiatric morbidity was confir med by subsequent investigators (Miller & Ingham, 1976; Roy,
1978, 1981; Winefield, 1979). The aim of the present inves tigation was to re-examine these earlier findings and to extend these studies by classifying the patients into well- defined diagnostic subgroups of depression. Throughout, the depressed patients reported the presence of confidants both within and outside the household to a significantly lower degree than did controls. The only exception were the bipo lar patients who had about the same availability of confi dants outside the household as the control group.
As important factors for the understanding of the mechanisms behind the capability of eliciting social support, 'ego functioning' (Flaherty et al., 1983) and the size of the social network are suggested.
Paper IV: 'The relationship of personality to social net
work aspects and loneliness in depressed pati
ents ' .
The role of personality in depression has been given consi derable attention in the psychiatric literature. Attempts have been made to find predisposing personality patterns.
Hirschfeld & Klerman (1979) found depressive patients sco ring higher on introversion, neuroticism, obsessionality, solidity and stability than normals. According to von Zers- sen (1976) the 'melancholic type' was, among others, charac terized by a dependency on close personal relationships. In 27
his review article Akiskal et al. (1983) regarded 'introver sion' to be the trait with the 'strongest evidence to pre dispose' to depression. In the past social disintegration has been associated with mental illness (Chen & Cobb, 1960).
More recently the social networks of psychiatric patients have been studied both quantitatively and qualitatively
(Tolsdorf, 1976; Miller & Ingham, 1976). In a previous article (Eisemann, 1984a) it was suggested to investigate possible relationships between personality characteristics and the experience of loneliness, contact difficulties and social network aspects. The present paper complies with this suggestion. The results show that patients with a personali ty characterized by somatic and psychic anxiety, detachment, psychasthenia, irritability and inhibition report a limited number of friends. This possibly suggests that making friends requires a more well-functioning personality than associating with relatives. Even regarding loneliness and contact difficulties the scales reflecting anxiety and psychasthenia correlate most strongly. In addition, there is a positive relationship between 'suspicion' and suffering from loneliness. Our results point to the importance of per sonality traits for the understanding of the social beha viour of depressive patients.
Paper V : 'Depressed patients and non-psychiatric controls:
discriminant analysis on social environment'.
The interest in the social environment as a possible factor in disease etiology dates back to the end of the last cen tury when Durkheim (1883) pointed to the importance of social relationships for the social development. In more recent studies associations were found between deficiencies in the social environment and psychiatric morbidity (Post,
1962; Blöschl, 1976; Brown et al., 1978). In the present study we tried to distinguish between depressed patients and 28 controls on a variety of social variables (N=ll) by means of discriminant analysis. From the eleven original variables nine were retained in the following order of their entrance: feeling of loneliness, confiding persons within household, number of family members (regular contact), number of lei sure activities involving others, number of relatives (re gular contact), confiding person outside household, size of household, civil status, number of friends (regular con tact) . By means of the discriminant function as much as 83% of the subjects could be correctly classified. The results indicate shortcomings in the social environment of depressi ves reported in earlier studies (Eisemann, 1984a,b) and underline the discriminating power of such factors.
Paper VI : 'Social class and social mobility of depressed
patients'.
Studies on the relationship between social class and depres sion have yielded divergent results. Some authors (Stern-Pi- per, 1925; Hollingshead & Redlich, 1958) found higher pro portions of depression in higher social classes whereas others (Birtchnell, 1971; Brown et al., 1975; Weissman &
Myers, 1978) established greater prevalences of depression in lower social classes. So far little attention has been paid to the possible variations of social class among diffe rent diagnostic subgroups of depression. To overcome this shortcoming the relationship of social class to distinct depressive subgroups was studied. To exclude the possibility that the results are influenced by a decline in social class as a consequence of the depressive illness the social class of the family of origin was also assessed. By this approach it was possible to scrutinize the social mobility of the subjects. The results show that social class III is overrep resented in the unipolar, bipolar and unspecified group com pared with figures from the general population. Within the 29
depressive subgroups the unipolars show an overrepresenta tion of social class III compared with the neurotic depres sives. As regards the parental social class there are no significant differences between the samples. The unipolar and the bipolar patients are significantly more downwardly mobile than the other groups. The latter result is in cont rast to the findings of other studies (d'Elia et al., 1969;
Weissman & Myers, 1978; Woodruff et al., 1971; Weiner et al., 1979).
Paper VII: ’Perceived parental rearing and depression’.
During the last years there has been a shift in the study of parental deprivation in depressed patients. Earlier studies had focused on quantitative aspects of deprivation (parental loss) during childhood (for a review see Granville-Grossman,
1968) whereas in recent investigations it is suggested that qualitative aspects should be taken into account (Raskin et al., 1971; Parker, 1979; Perris et al., 1980; Jacobson et al., 1975). Both Jacobson et al., (1975) and Parker (1979) found that depriving childrearing practices were related to adult depression. To test the hypothesis that depressed patients would report more negative and depriving rearing practices from their parents than would controls, the EMBU, a Swedish instrument aimed at assessing the experience of parental rearing attitudes was applied. The following three factors, obtained from factor analysis and suggested by
Arrindell et al., (1983), were investigated in the present study: 'rejection', 'emotional warmth', 'overprotection'.
The results show that patients in all diagnostic groups sco red lower than the controls on 'emotional warmth' and this applied to both parents. However, statistical significance was only obtained for the unipolar and unspecified group.
The same tendency became apparent for the factor 'maternal overprotection'. The depressives tended also to rate their 30
parents as less consistent in their rearing attitudes. In a
discriminant analysis the factors maternal emotional warmth
and maternal overprotection allowed 64% of the total patient
sample and 72% of the unipolar depressives to be correctly
classified. These results, like those of previous studies,
support the hypothesis that deprivation of love during
childhood represents an important psychological background
factor for depressive disorders.
Paper VIII: ’Perceived parental rearing practices in de
pressed patients in relation to social class'.
One of the background determinants of child-rearing practi
ces reported in the literature is the social status of the
parents. A number of authors (Sears et al., 1957; Jonsson &
Kälvesten, 1964; Nettelblad et al., 1981) regarded the
parental style of lower social classes as more authorita
rian, more punitive, and less stimulating than that of the
middle class. Other authors (Kuhn, 1959; Perrez et al.,
1981) could not verify these differences. However, none of
these previous studies addressed the issue as to whether
differences in 'emotional warmth' do occur in the parental
rearing practices of different social classes. Since possib
le differences in rearing practices between social classes
could bias the interpretation of our previous findings in
depressed patients the present study was carried out to in vestigate such differences in our sample. The results reveal no differences in the perception of 'emotional warmth' expe rienced by patients grown up in different'social milieus. On
the other hand subjects from higher social classes scored
their parents as more rejecting and overprotecting. These aspects have not been considered separately in earlier stu dies, but might be of relevance as a risk factor for depres sion (Perris et al., 1984; Arrindell et al., 1984). 31
GENERAL DISCUSSION AND CONCLUSIONS
There has been a number of studies, reported in the medical literature, on the relationship between psychosocial factors and health in general (Cassel, 1976) and mental health in particular (Greenblatt et al., 1982). In most of the studies the 'social support1 derived from social relationships (net works) has been regarded as the crucial variable in this context. However, the lack of agreement about operational and conceptual definitions complicates the comparison of studies in which the empirical effects of 'social support' on health has been investigated. There seems to be some evi dence suggesting that individuals lacking an adequate pro portion of 'social support' are at higher risk of morbidity.
A number of studies have shown that depression was associa ted with impoverished social networks and lack of 'social support' (Bloschl, 1976; Roy 1978, 1981; Winefield, 1979).
However, the mechanisms behind these observations are, so far, poorly understood.
The different parts of the present investigation have been conducted in an attempt to shed some light on this issue.
There are a few indications from the literature on how dep ressives look at the members of their social network e.g. what demands they have on them. Bibring (1968) has pointed to the specific dependency of depressives on their social relationships, having exceptionally high demands on their attention and affection. The study by Libet and Lewinsohn
(1973) suggests that depressed patients concentrate their social life on a single relationship. Under such circumstan ces any neglect from the side of this person leads to inc reasing frustration.
In his theory of depression Becker (1964) focuses on the ego and on the social determinants of self-esteem. According to 32
Becker self-esteem is derived and nourished by a meaningful life which in turn depends on satisfactory social interac tions ('games’) with others ('objects'). He regards loss of self-esteem, game-loss and object-loss as inseparably part of meaning-loss. However, Becker gives primacy to game-loss since he considers object-loss as not crucial per se when there is a possibility of obtaining esteem and affirmation from a variety of objects. In this way he expanded the
'battlefield' of dynamic forces from the intra-psychic to the social area.
Some findings from the present study support Becker's view.
If leisure activities can be seen as one of the potential areas where 'games' can be played, depressives seem to be handicapped. The results of the present study point both to a reduced overall activity among depressives and to a prefe rence of leisure activities not -involving other people.
The significant correlations between the experience of lone liness and aspects of social network size indicate that even quantitative measures have implications for the quality of social support available.
The application of Engel's biopsychosocial model (Engel,
1980) based on systems theory gives the possiblity to see depression as a product of a dynamic process of forces from different structures from the living system. Starting in childhood the results show a lack of perceived emotional warmth (corresponding to Bowlby's 'emotional attachment') amongst the depressed patients. That these results cannot have beeen biased by possible differences in the social class of their parents is confirmed by the findings from the study on perceived parental rearing and social class.
Results from one of our earlier studies (Perris et al., 33
1983) pointed to the relation between perceived emotional warmth during infancy and personality characteristics (e.g.
socialization, psychasthenia, suspicion, anxiety) in adult
hood. The numerous relationships which have been found in
the investigation on personality and loneliness and social network features stress once more the importance of persona
lity traits for a better understanding of the social beha
viour of the depressed patients.
One of the obstacles of being able to compare studies of
other authors on this subject has been the utilization of
badly-defined diagnostic groups. The numerous differences on
a variety of variables between the well-defined diagnostic
subgroups in the present investigations point to the neces
sity and usefulness of such an approach.
Implications for treatment
If the importance of social relationships as a mediator of
social support is accepted, some re-orientation in the theo
ry and practice of mental health professionals is required.
The emphasis on intrapsychic explanations and therapies on
the one hand and on drug therapy on the other hand, must be
curtailed in favour of an interpersonal orientation aimed at
the development of social skills. Admittedly, antidepressive
drug treatment is effective in the acute phase of depres sion. In addition, the prophylactic effects of lithium in the treatment of recurrent affective disorders have been demonstrated (Schou et al., 1970; Smigan, 1984). However, there is still a significant number of patients responding
incompletely to drugs or who are even resistant to therapy
(de Francisco, 1979; Freyham, 1979). There is growing evi dence (Kovacs et al., 1981; Blackburn & Bishop, 1983; Teas- dale et al., 1984) that a combination of drug and psychothe rapy is very likely the most effective treatment of depres 34
sion. Results from the present study point to the suffering
from loneliness among depressives. Young (1982) outlined a
systematic framework for the understanding of loneliness and
for the treatment of socially isolated people. He pointed to
the importance of cognition, behaviour and emotion in defi ning and treating loneliness. Young regarded the concomitant
behaviours and emotions of loneliness mainly as a function
of the individual's thoughts, attributions and assumptions.
After having understood why lonely people act and feel as
they •do it could be possible to help them to overcome their
loneliness by applying cognitive behavioural techniques. A
prerequisite for this treatment is a thorough assessment of
the patient's social network by means of a standardized
instrument which should become a routine procedure.
A future objective for mental health workers could be to
identify individuals and groups at high risk by virtue of
their lack of social relationships and to determine the necessary and suitable form of social supports that can and should be strengthened in order to protect such persons from disease manifestations or relapses. For example, in the
Sovietunion where life, to a much greater extent than in
Western societies, is built up on the collective, the mental health workers often use members of the collective for the rapeutic interventions such as for visiting depressed patients to break or prevent isolation and detachment (Zi — ferstein, 1976).
Suggestions for future research
Since there is some evidence (Sokolovsky, 1978) that the impact of social relationships may vary between different types of mental illnesses, in future studies a group with another psychiatric disorder should be included. Without the inclusion of such a control group, it is difficult to iden 35
tify those factors that render the person at risk for any psychiatric disorder from those for depression specifically.
So far, most of the research is retrospective, which makes it difficult to sort out consequences from causes. Prospec tive, longitudinal studies should be conducted to confirm an etiological role of social relationships. In this context the following issues could be examined: the predictive value of social network characteristics, association of changes in social network with changes in psychopathology between assessment periods.
Another important area of investigation would be the study of the interaction between factors from the psychosocial and the biological (organismic) systems. E.g. how changes in the social environment are accompanied by neuroendocrinal chang es which in turn, affect among others, the adrenocortical pituitary system. According to Schoenheimer (1942) these endocrine substances are responsible for the maintenance of the ’dynamic steady state' of the organism, which can be threatened by disease agents of biological or social origin. 36
REFERENCES
AKISKAL HS, HIRSCHFELD RM, YEREVANIAN BI. The relationship
of personality to affective disorders. Arch Gen Psychiat
ry 1983:40:801-810.
ALBRIGHT MJ. Psychosocial aspects of development. In: Ballis
GU, Wurmser L, Me Daniel E, Grenell RG (eds) Dimensions
of Behaviour. 1. The psychiatric foundations of medicine.
Butterworth, London 1978.
AMERICAN PSYCHIATRIC ASSOCIATION. Committee on Nomenclature
and Statistics. Diagnostic and Statistical Manual of Men
tal Disorders, ed 3, American Psychiatric Association,
Washington DC 1980.
ARIETI S, BEMPORAD I. Severe and Mild Depression. Basic
Books, New York 1978.
ARRINDELL WA, BRILMAN E. The role of perceived parental rea
ring practices in the aetiology of phobic disorders: a
controlled study. Brit J Psychiatry 1983:143:183-187.
ARRINDELL WA, EMMELKAMP PMG, BRILMAN E, MONSMA A. Psychomet
ric evaluation of an inventory for assessment of parental
rearing practices. Acta Psychiatr Scand 1983:67:163-177.
ARRINDELL WA, PERRIS C, EISEMANN M, PERRIS H, VAN DER ENDE
J, ROSS M, BENJAMINSEN S, GASZNER P, DEL VECCHIO M.
Cross-national study of generalizability of patterns of
parental rearing behaviour and attitudes: invariance of
EMBU dimensional representations of healthy subjects from
Australia, Denmark, Hungary, Italy and the Netherlands.
Acta Psychiatr Scand 1985, submitted.
ARRINDELL WA, PERRIS C, EISEMANN M, VAN DER ENDE J. Cross-
cultural invariance of dimensions of parental rearing
behaviour: comparison of self-ratings of Swedish normals
and depressives with Dutch target ratings on the EMBU.
Brit J Psychiatry 1985, in press.
BARNES IA. Class and committees in a Norwegian island
parish. Human Relations 1954:7:39-58. 37
BECKER E. Toward a comprehensive theory of depression: A
cross-disciplinary appraisal of objects, games and mea
ning. J Nerv Ment Dis 1962:135:26-35.
BECKER WC. Consequences of different kinds of parental
discipline. In: Hoffman YL, Hoffman LW (eds) Review of
Child Development Research, voi 1. Russel Sage Foundation
New York, 1964.
BERKMAN LF, SYME SL. Social networks, host resistance, and
mortality: a nine-year follow-up study of Alameda County
Residents. Am J Epid 1979:109:186-204.
VON BERTALANFFY L. General systems theory. Braziller, New
York 1968.
BIBRING E. The mechanisms of depression. In: Greenacre PH
(ed) Affective Disorders. Intern Univ Press, New York
1953.
BIRTCHNELL J. Social class, parental class and social mobi
lity in psychiatric patients and general population cont
rols. Psychol Med 1971:1:209-221.
BLACKBURN IM, BISHOP S. Changes in cognition with pharmaco
therapy and cognitive therapy. Brit J Psychiatry
1983:143:609-617.
BLÖSCHL L. Zur intra- und extrafamiliären Kontaktstruktur
depressiver Patientinnen. Psychol Beitr 1976:18:465-480.
BOYD JH, WEISSMAN MM. Epidemiology of affective disorders.
Arch Gen Psychiatry 1981:38:1039-1046.
BOWLBY J. The making and breaking of affectional bonds. I.
Aetiology and psychopathology in the light of attachment
theory. Brit J Psychiatry 1977:130:201-210.
BOWLBY J. Attachment and loss. Voi 1, Attachment. Hogarth
Press, London 1969.
BROWN GW, BHROLCHAIN MN, HARRIS T. Social class and psyc
hiatric disturbance among women in an urban population.
Sociology 1975:9:225-254.
BROWN GW, HARRIS T. Social origins of depression. Tavistock,
London 1978. 38
CAPLAN G. Support systems and community mental health. Beha
vioral Publications, New York 1974.
CASSEL J. An epidemiological perspective of psychosocial
factors in disease etiology. Am J Pubi Health 1974:64:
1040-1043.
CASSEL J. The contribution of the social environment to host
resistance. Am J Epid 1976:104:107-123.
CESAREE Z, MARKE S. Mätning av psykogena behov med frågefor-
mulärsteknik. Skand Testförlaget, Stockholm 1968.
CHEN E, COBB S. Family structure in relation to health and
disease: a review of the literature. J Chron Dis
1960:12:544-567.
CLAYER JR, CAMPBELL RL, ROSS MW. Parental rearing and tran
sient thought disorder in young adults. Psychopathology
1984:17:9-16.
COBB S. Social support as a moderator of life stress. Psyc-
hosom Med 1976:38:300-314.
COHEN Cl, SOKOLOVSKY I. Schizophrenia and social networks:
Ex-patients in the inner city. Schizophrenia Bull
1978:4:546-560.
COOLEY CH. Social Organization: A study of the larger mind.
Scribner's, New York 1909.
DURKHEIM E (transi Simpson G). Division of Labor in Society.
The Free Press, Glencoe 111 1947. (Originally published
1883).
EISEMANN M. Leisure activities of depressive patients. Acta
Psychiatr scand 1984a : 69:45-51.
EISEMANN M. The availability of confiding persons in depres
sed patients. Acta Psychiatr Scand 1984b : 70:166-169.
EISEMANN M, JACOBSSON L, VON KN0RRING L, PERRIS C, PERRIS H.
Choice of symptoms in a multi-aspects classification
model applied to depressed patients. Psychiat clin
1980:13:84-89.
D'ELIA G, VON KNORRING L, PERRIS C. Non-psychotic depressive
disorders: A ten-year follow up. Acta Psychiatr Scand 39
1974:Suppl 255:173-186.
D'ELIA G, LAURELL B, PERRIS C. Comparison of sociological
background factors in different kinds of depressive sta
tes. Proc of the 2nd Intern Congr Social Psychiatry, Lon
don 1969.
ENGEL GL. The clinical application of the biopsychosocial
model. Am J Psychiatry 1980:137:535-544.
ERIKSON EH. Childhood and Society. Norton & Co, New York
1950.
EYSENCK SBG, EYSENCK HJ. Impulsiveness and venturesomeness:
their position in a dimensional system of personality
description. Psychol Report 1978:43:1247-1255.
FEIGHNER IP, ROBINS E, GUZE B, W00DRFUFF RA, WINOKUR G,
MUNOZ R. Diagnostic criteria for use in psychiatric
research. Arch Gen Psychiatry 1972:26:57-63.
FERGUSON GA. Statistical Analysis in Psychology and Edu
cation. McGraw-Hill, New York 1959.
FLAHERTY IA, GAVIRIA FM, BLACK EM, ALTMAN E, MITCHELL T. The
role of social support in the functioning of patients
with unipolar depression. Am J Psychiatry 1983:140:473-
476.
DE FRANCISCO-PEREZ C. Therapy resistant depression. Int
Pharmacopsychiat 1979:14:71-78.
FREYHAN FA. Treatment resistant or untreatable? Int Pharma
copsychiat 1979:114-119.
FROST WH. Introduction to 'Snow on Cholera'. Common Wealth
Foundation, New York 1936.
GRANVILLE-GROSSMAN KL. The early environment in affective
disorders. In: Coppen A, Walk A (eds) Recent Developments
in Affective Disorders. Brit J Psychiatry, Spec pubi 2,
1968.
GREENBLATT M, BECERRA RM, SERAFETINIDES EA. Social networks
and mental health: An overview. Am J Psychiatry 1982:139:
977-984.
GRINSPOON L. The threat of nuclear war. Am J Psychiatry 40
1982:139:1313-1314.
HAGNELL 0, LANKE J, RORSMAN B, ÖJESJÖ L. Are we entering an
age of melancholy? Depressive illnesses in a prospective
epidemiological study over 25 years: the Lundby study,
Sweden. Psychol Med 1982:12:279-289.
HARLOW HF. Primary affectional patterns in primates. Am J
Orthopsychiat 1960:30:676-684.
HARLOW MK, HARLOW HF. Affection in primates. Discovery
1966:27:11-17.
HARLOW HF, SUOMI SJ. Induced depression in monkeys. Behav
Biol 1974:12:273-296.
HENDERSON S, DUNCAN-JONES P, MC AULEY H, RITCHIE K. The
patient's primary group. Brit J Psychiatry 1978:132:74- 8 6 .
HENDERSON S, DUNCAN-JONES P, BYRNE DG, SCOTT R, ADCOCK S.
Social relationships, adversity and neurosis: a study of
associations in a general population sample. Brit J Psy
chiatry 1980:136:574-583.
HIRSCHFELD RM, KLERMAN GL. Personality attributes and affec
tive disorders. Am J Psychiatry 1979:136:67-70.
HOLLINGSHEAD AB, REDLICH FC. Social Class and Mental Ill
ness. Wiley, New York 1958.
HULL CH, NIE NH. SPSS Update 7-9. McGraw-Hill, New York
1981.
JACOBSON S, FASMAN I, DI MASCIO A. Deprivation in the child
hood of depressed women. J Nerv Ment Dis 1975:160:5-14.
JOHANSSON S. Låginkomstutredningen (Low Income Survey). All
männa förlaget, Stockholm 1970.
JONSSON G, KÄLVESTEN A-L. 222 Stockholmspojkar. Almkvist &
Wiksell, Stockholm 1964.
KLERMAN GL. Affective disorders. In: Nicholi A (ed) The Har
vard Guide to Modern Psychiatry. Harvard Univ Press,
Cambridge Mass 1978.
KOVACS M, RUSH I, BECK AT, HOLLON SD. Depressed outpatients
treated with cognitive therapy or pharmacotherapy. A one 41
year follow-up. Arch Gen Psychiatry 1981:38:33-39.
KRAUS A, LILIENFELD A. Some epidemiologic aspects of • the
high mortality rate in the young widowed group. J Chronic
Dis 1959:10:207-217.
KUHN ML. Social class and the exercise of parental authori
ty. Am Sociol Rev 1959:24:352-366.
LIBET JM, LEWINSOHN PM. Concept of social skill with special
reference to the behaviour of depressed persons. J con
sult clin Psychol 1973:40:304-312.
LORENZ K. Die angeborenen Formen möglicher Erfahrung. Z
Tierpsychol 1945:5:235-409.
MASON JW, BRADY IV. The sensitivity of the psychoendocrine
system to social and physical environment. In: Shapiro D
(ed) Psychobiological Approaches to Social Behavior.
Stanford University Press, Stanford 1964.
MC KINLAY JB. Social networks, lay consultation and helpsee-
king behaviour. Soc Forces 1973:51:275-280.
MILLER P McC, INGHAM JG. Friends, confidants and symptoms.
Soc Psychiatry 1976:11:51-58.
MURPHY LB, M0RIARTY AE. Vulnerability, Coping and Growth.
Yale University Press, New Haven 1976.
MURRAY HA. Exploration in personality. Oxford University
Press, Oxford 1938.
NETTELBLADT P, UDDENBERG N, ENGLESSON I. Sex-role patterns,
paternal rearing attitudes and child development in dif
ferent social classes. Acta Psychiatr Scand 1981:64:12-
24.
NEWMAN BM, NEWMAN PR. Development Through Life: A psycho
social approach. Dorsey Press, Homewood IU 1975.
NIE NH, HULL CH, JENKINS IG, STEINBRENNER K, BENT DH. Sta
tistical Package for the Social Sciences. Me Graw-Hill,
New York 1975.
OTTOSSON J-0, PERRIS C. Multidimensional classification of
mental disorders. Psychol Med 1973:3:238-243.
PARKER G. Parental characteristics in relation to depressive 42
disorders. Brit J Psychiatry 1979:134:138-147.
PERREZ M, PATRY J-L, ISCHI N. Verhaltensstörungen bei
Schulkindern im Zusammenhang mit Erziehungsstil, ökolo
gischen und sozialstrukturellen Variablen. Z Klin Psychol
Psychother 1981:29:141-157.
PERRIS C. New approaches to the classification of the affec
tive disorders. In: Lader M, Garcia R (eds) Aspectos de
la Depresion. Pubi by the authority of the WPA, 1973.
PERRIS C, ARRINDELL WA, PERRIS H, VAN DER ENDE J, MAJ M,
BENJAMINSEN S, ROSS MW, EISEMANN M, DEL VECCHIO M. Cross
national comparison of perceived parental rearing beha
viour in healthy subjects from Australia, Denmark, Italy,
the Netherlands and Sweden. Acta Psychiatr Scand 1985, in
press.
PERRIS C, EISEMANN M, ERIKSSON U, JACOBSSON L, VON KNORRING
L, PERRIS H. Variations in self-assessment of personality
characteristics in depressed patients with special refe
rence to aspects of aggression. Psychiatr Clin 1979:12:
209-215.
PERRIS C, EISEMANN M, ERIKSSON U, VON KNORRING L, PERRIS H.
Parental rearing behaviour and personality characteris
tics of depressed patients. Arch Psychiatr Nervenkrh
1983a:233:77-88.
PERRIS C, EISEMANN M, ERIKSSON U, VON KNORRING L, PERRIS H.
Personality characteristics of depressed patients classi
fied according to family history. Neuropsychobiology
1983b:9:99-102.
PERRIS C, EISEMANN M, VON KNORRING L, PERRIS H. Presentation
of a subscale for the rating of depression and some addi
tional items to the Comprehensive Psychopathological
Rating Scale. Acta Psychiatr Scand 1984:70:261-274.
PERRIS C, JACOBSSON L, LINDSTRÖM H, VON KNORRING L, PERRIS
H. Development of a new inventory for assessing memories
of parental rearing behaviour. Acta Psychiatr Scand
1980:61:265-274. 43
PERRIS C, PERRIS H. A biological, psychological and social
approach to the study of depression and its implications
for treatment. Neurologia Psichiatria Scienze Umane 1984,
in press.
POST F. The social orbit of psychiatric patients. J Ment Sci
1962:108:759-771.
PRESIDENT'S COMMISSION ON MENTAL HEALTH TASK PANEL ON COM
MUNITY SUPPORT SYSTEMS: Report of the Task Panel on Com
munity Support Systems. US Government Printing Office,
Washington DC 1978.
RASKIN A, BOOTHE HH, REATIG NA, SCHULTENBRANDT IG, COLE D.
Factor analysis of normal and depressed patients' memo
ries of parental behaviour. Psychol Rep 1971:29:871-879.
ROSS MW, CLAYER JR, CAMPBELL RC. Dimensions of childrearing
practices. Acta Psychiatr Scand 1983:68:476-483.
ROY A. Vulnerability factors and depression in women. Br J
Psychiatry 1978:133:106-110.
ROY A. Vulnerability factors and depression in men. Br J
Psychiatry 1981:138:75-77.
RUSSELL D, PEPLAU LA, CUTRONA CE. The revised UCLA loneli
ness scale, concurrent and discriminant validity eviden
ce. J Pers Social Psychol 1980:39:472-480.
SCHALLING D. Psychopathy-related personality variables and
the psychophysiology of socialization. In: Hare J, Schal-
ling D (eds) Psychopathic Behaviour. Approaches to
Research. Wiley, Chichester 1978.
SCHALLING D, ÅSBERG M. Biological and psychological correla
tes of impulsiveness and monotony avoidance. In: Strelau
I, Farlay FM, Gale A (eds) The Biological Foundations of
Personality and Behaviour. Hemisphere Press. New York
1981.
SCHOENHEIMER R. Dynamic State of Body Constitution. Harvard
University Press, Cambridge 1942.
SCHOU M, BAASTRUP P, GROF P, WEIS P, ANGST J. Pharmacologi
cal and clinical problems of lithium prophylaxis. Brit J 44
Psychiatry 1970:116:615-619.
SEARS RP, MACCOBY EE, LEVIN H. Patterns of child rearing.
Row Petersen, Evanston, 111 1957.
SHULMAN N. Life-cycle variations in patterns of close rela
tionships. J Marriage Fam 1975:37:813-821.
SMIGAN L. Some clinical, biological and psychological
aspects of long-term lithium therapy. Umeå University
Medical Dissertations, New Series No 130. Umeå, Sweden
1984.
SOKOLOVSKY J. COHEN J, BERGER D, GEIGER J. Personal networks
of ex-mental patients in a Manhattan SRO hotel. Hum Org
1978:37:5-10.
STERN-PIPER L. Der psychopathologische Index der Kultur.
Arch Psychiatr Nervenkrh 1925:74:514-525.
STRANDMAN E. 'Psychogenic needs' in patients with affective
disorders. Acta Psychiatr Scand 1978:58:16-29.
TEASDALE ID, FENNELL MJV, HIBBERT GA, AMIES PL. Cognitive
therapy for major depressive disorder in primary care.
Brit J Psychiatry 1984:144:400-406.
WEISS P. Animal behaviour as system reaction: orientation
toward light and gravity in the resting postures of but
terflies (Vanessa) (1925). In: General Systems. Yearbook
of the Society for General Systems Research 1959:4:1-44.
WEISS P. The system of nature and the nature of systems:
empirical holism and practical reductionism harmonized.
In: Schaefer IKE, Hensel H, Brody R (eds) Toward a Man-
centered Medical Science. Brody R Mt Kisco, New York
1977.
WEISSMAN MM, MYERS JK. Affective disorders in a US urban
community: the use of research diagnostic criteria in an
epidemiological survey. Arch Gen Psychiatry 1978:35:1304-
1311.
WEISSMAN MM, MYERS JK. Rates and risks of depressive symp
toms in a United States urban community. Acta Psychiatr
Scand 1978:57:219-231. 45
WELNER A, MARTEN S, WOCHNICK E. Psychiatric disorders among
professional women. Arch Gen Psychiatry 1979:36:169-173.
WINEFIELD HR. Social support and the social environment of
depressed and normal women. Aust NZ J Psychiatry
1979:13:335-339.
WINOKUR G. Familial v(genetic) subtypes of pure depressive
disease. Am J Psychiatry 1979:136:911-913.
WINOKUR G, BEHAR D, VAN VALKENBURG C, LOWRY M. Is a familial
definition of depression both feasible and valid? J Nerv
ment Dis 1978:166:764-768.
WOODRUFF RA, ROBINS LN, WINOKUR G. Manic depressive illness
and social achievement. Acta Psychiatr Scand 1971:47:237-
249.
WORLD HEALTH ORGANIZATION. Manual of the International Sta
tistical Classification of Diseases, Injuries and Causes
of Death. WHO, Geneva 1977.
YOUNG JE. Loneliness, depression and cognitive therapy:
theory and application. In: Peplau LA, Perlman D (eds)
Loneliness: A Sourcebook of Current Theory, Research and
Therapy. Wiley, New York 1982.
VON ZERSSEN D. Der 'Typus melancholicus' in psychometrischer
Sicht (Teil 1). Z klin Psychol Psychother 1976:24:200-
226.
ZIFERSTEIN I. The roots of Russian psychiatry. In: Corson
SA, Corson EO (eds) Psychiatry and Psychology in the
USSR. Plenum Press, New York 1976.
ZUCKERMAN M. Dimensions of sensation seeking. J Consult Clin
Psychol 1971:36:45-52.
ÅSBERG M, PERRIS C, SCHALLING D, SEDVALL G. The CPRS - App
lications of a psychiatric rating scale. Acta Psychiatr
Scand 1978: Suppl 271. 46
ACKNOWLEDGEMENTS
I wish to express my sincere gratitude to all those who made this work possible and in particular to:
The patients who participated in the study, for their co operation, interest and patience in taking part in the interviews and investigations;
Professor Carlo Perris, my teacher and supervisor, for fur nishing me with his eminent knowledge, for stimulating discussions and for all the support he has given me;
Assoc Prof Hjördis Perris for co-authorship, for help with interviewing the patients and for continous support;
Assoc Prof Lars von Knorring for co-authorship and fruitful discussions ;
Prof Lars Jacobsson for pleasant collaboration in the ini tial phase of the project;
Bengt Hållberg, PhD, for indispensable assistance with com puter work and statistics;
Doris Cedergren, MA, for professional and generous help throughout the years ;
Ludovit Smigan, Dr med Sc, for fruitful discussions and warm support ;
Ulla Eriksson, BA, for help with interviewing the patients;
Tove Bengts, Ann-Sofie Forsgren and Margaretha Lindh for skilful secretarial assistance and for cheering up workaday life ;
The staff from the Umeå University Library for providing me with literature;
Ass Prof Bengt Mattsson and Mrs Ulla Larsson from the Pri mary Care Unit, , for help with collecting the control sample ;
Willem Arrindell, PhD, from the University of Groningen,
Holland for co-operation in the EMBU study;
Sonja Holmgren, laboratory assistant, for generous help and 47
friendly support;
Birgitta Hammarström, BSc and Anders Byström, BSc, from the
Umeå University Computer Centre for computer work.
Last but not least my mother and my sister with her family for 'long-distance' support.
This study was supported by a grant from the Swedish Medical
Research Council (grant no 5244) and by scholarships from the Faculty of Medicine, Umeå University.