Abstract
Case notes of patients treated at the Maudsley Hospital during the interwar period provided data about diagnosis, symptoms and beliefs about mental illness. In the absence of effective treatments, patients were investigated in detail in the hope that connections between disease processes might be revealed. We analysed a randomly-selected sample of 700 patients taken equally from 1924, 1928, 1931 and 1935. Eight groups (three representing psychosis and five indicating psychological disorders) were identified on the basis of symptom clusters. Formal diagnosis did not correlate with clusters. Although there was a measure of agreement between patients and doctors about the cause of mental illness, stigma may have inhibited discussion of some themes. Psychiatric diagnosis was informed by symptoms but not determined by them. In an era before classification systems were tested for reliability, diagnosis was fluid, reflecting changing hypotheses about causation, pathology and treatment. Attributions were associated with diagnosis rather than symptoms.
Keywords
Introduction
Patient records have been identified as crucial indicators of ‘the interaction between physicians and their patients in which individual personality, cultural assumptions, social status, bureaucratic expediency, and the reality of power relationships are expressed’ (Risse and Warner, 1992: 189). Furthermore, Hurwitz (2006: 217) argued that the clinical case report ‘is a foundational text that enables clinicians to depict, reason and instruct others about a sick person’s medical situation’. Despite their potential value, they remain a problematic and under-used primary source. In 1992, Risse and Warner believed that they were ‘coming into vogue’ because they lent themselves to ‘large-scale statistical analyses’ by electronic means (Risse and Warner, 1992: 184, 195). Kemp and colleagues, in their historical study of birth weight and other obstetric data from the 1930s and early 1940s, concluded that ‘the richest sources of data were hospital clinical records, particularly case notes’ (Kemp, Gunnell, Davey Smith and Frankel, 1997: 312). Despite the availability of patient records for the twentieth century, the hopes of Risse and Warner have not been fulfilled. As they had acknowledged, the records are not without methodological problems. Primarily designed to serve the needs of doctors, case notes represent a partial view of the interaction between medical staff and patients. They are not necessarily an objective or inclusive record of illness and its treatment. Patients were not given a designated space to record their views, and their actual words were rarely transcribed verbatim.
Furthermore, new legislation and a concern for patient rights have also placed limitations on the use of case notes. A legal judgment delivered in May 1999 suggested that the storage and analysis of anonymized patient data could constitute a breach of confidentiality (Walton et al., 1999). The Data Protection Act, which came into force in March 2000, was ambiguous on the issue of informed consent about patient-identifiable data in large population studies, and many believed that new restrictions hindered legitimate study (Al-Shahi and Warlow, 2000). Perceiving an imbalance between the rights of patients to confidentiality and gains that might accrue to society from epidemiological research, doctors campaigned for alternative ethical approaches (Ward et al., 2004). These restrictions also applied to medical historians seeking to research patient records unless it could be shown that the subjects were no longer living. Thus most post-1945 investigations were caught within this legislation.
Using adult case notes created at the Maudsley Hospital between 1924 and 1935, this study explores the experience of being a patient. In particular, the symptom patterns of random samples of in- and out-patients are compared with the formal diagnosis on discharge. Statements about the cause of mental illness have also been gathered from both patients and doctors alike to assess the extent to which beliefs about the aetiology of psychological disorders were shared by clinicians and their patients.
Maudsley patient notes
The Maudsley Hospital was opened in February 1923 to treat Londoners with mental illness (Jones, Rahman and Woolven, 2007). Unlike the traditional county asylums, all patients were voluntary and no one, however ill, could be detained against their will. This policy was designed to restrict admission to those in whom illness was in its early stages, to improve the chance of reversing pathological processes (Jones and Rahman, 2009). With a higher staffing ratio than found in county asylums and the benefit of post-graduate trainees, each patient was subject to an exhaustive examination. This was not simply an academic exercise but driven by the hypothesis that mental illness was causally connected with recognized diseases (Carswell, 1924; Mayou, 1989). Before World War I, Frederick Mott (1899) had shown that the psychiatric symptoms of general paralysis of the insane (GPI) had an organic basis in the neurological effects of tertiary syphilis. In the 1920s, it was hypothesized that viral or bacterial infection might play a causal role in dementia praecox and other psychiatric disorders (Scull, 2005: 112–20) or that they were the consequence of a metabolic imbalance produced by a malfunctioning endocrine system. Pituitary and thyroid extract had been given to patients suffering from shell shock, and in the post-war period trials were extended to major mental illness (Jones, 2010). As a result, doctors at the Maudsley conducted wide-ranging physical investigations of new patients in the belief that large-sample data would reveal connections between what appeared to be discrete disease processes.
Maudsley in-patient notes were exhaustive, commonly running to 20 pages. The notes were of two kinds: printed forms and free text. The standardized front sheets, defined by subheadings and boxes, required the doctor to gather basic medical and biographical data. Most of the information was recorded by the senior medical officer, though juniors and nurses made additions. Patient statements were transcribed, though rarely as direct quotations. For example, a factory worker diagnosed with schizophrenia was reported by Dr William Moodie as believing that ‘his fellow workers used to make noises with the idea of making him a man’. Moodie interpreted this statement as a persecutory delusion (Anon., 1928a). The man himself was not given space in the notes to explain why he believed this to be the case. The status of case notes was revealed by the rule governing their access: they were kept in locked cupboards, only the ward sister and doctors having access to the key (Trevor and Bond, 1923: 1).
The comprehensive nature of the files, William Sargant argued, reflected a sense of clinical impotence; detailed note taking gave ‘us a feeling that we were doing something for the patient by learning so much about him, even if we could not yet find any relief for his suffering’ (Sargant, 1967: 36). A full physical examination was undertaken together with a detailed family and personal history. Although some of these data were objective (blood pressure, height, weight, qualifications), psychiatrists were required to record ‘moral and volitional tendencies’, although in practice this section was seldom completed. Patients were regularly written up, initially on a daily basis but as time passed at greater intervals. As Hurwitz observed, notes compiled in the twentieth century increasingly focused on pathological findings relegating the patient’s account, his or her perceptions, beliefs and emotions, to a remnant (Hurwitz, 2006: 228, 234). This was no less true at the Maudsley where notes targeted medical investigations, symptoms, behaviour and changes to medication.
With a high staffing ratio, doctors had more time to study patients in detail. David Clark, a post-war trainee, recalled the difference in practice between an asylum and the Maudsley: ‘where before I had been responsible for a hundred patients, I now had no more than six but I was expected to study them with an intensity far greater’ (Clark, 2000: 6). As a result, psychiatrists had an opportunity to hold extended conversations with their patients. Eliot Slater recalled that Aubrey Lewis ‘spent hours, and hours and hours talking to this [Jesuit] priest’ who had been admitted ‘tormented by obsessions … They shared a common fund of arcane knowledge because Aubrey himself had been brought up in a Jesuit school’ (Wilkinson, 1993: 8).
A printed form for ‘nurses notes’ contained boxes for weight checks and was followed by daily observations of diet, sleep and behaviour. Value judgements were recorded, as the entries from a case admitted in 1928 demonstrated (Anon., 1928b):
19 March, patient is very quiet. Eats and sleeps well. 8 April, cheerful and willing to help on the ward. 22 April, nothing unusual to note. 18 June, patient is showing little improvement, has occasional attacks and lying dully in bed with full rapid pulse. 15 August, thinks she cannot swallow and cannot walk. Very difficult and resistive with her diet.
In addition, printed sheets recorded changes in medication, and the notes also contained copies of letters sent to the referrer on admission and discharge.
Patients
Women outnumbered men in UK mental hospitals. In 1907, females represented 58.5% of admissions in England and Wales, a proportion that rose to 69.4% by 1937 (Andrews et al., 1997). The Maudsley reflected the national bias towards the attribution of mental illness in women. A random sample of 400 in-patients drawn from 1924, 1928, 1931 and 1935 showed that 39.8% were male and 60.3% female, similar proportions being recorded for a sample of 300 out-patients: 41.3% and 58.7%, respectively (Jones and Rahman, 2008). Despite the hospital’s strategic aim of treating the young in whom mental illness was incipient, most patients were middle-aged. The mean age of in-patients in the sample was 37.3 years for females and 41.5 for males, while out-patients were 38.3 and 36.2, respectively.
Given the stigma of mental illness and the requirement that the Maudsley treat only voluntary patients, the question arises why did people agree to an admission? The notes show that many patients were in distress, as revealed by their behaviour and symptoms, but others may have been persuaded by family and friends. Some could no longer carry out their jobs properly and this had come to the attention of their employers. Although the Maudsley was designed in appearance and operation to stand apart from the asylum system, the local population knew that it was a mental institution. While doctors at the Maudsley were denied the power of certification and could not compel a patient to stay, they were not without leverage. Patients who did not recover were referred to outer London institutions, notably Bexley, Horton and the Bethlem. A doctor could legitimately argue that if a patient refused admission to the Maudsley or decided to leave prematurely, then they would find themselves committed to one of the larger suburban asylums where their freedoms would be curtailed.
Symptom patterns
Patient records were archived alphabetically by surname according to year of first appointment. Using a random number generator, a sample of 175 subjects was selected for each of the years 1924, 1928, 1931 and 1935 to capture temporal changes in admissions policy. A standard form was used to gather data in a uniform manner. No names were recorded, and each case was identified by a unique number. Basic biographical details were collected together with 94 possible symptoms. To be consistent with cultural assumptions, occupations were rated according to the ‘social class grading’ of the 1931 Census (Anon., 1931). This had five groups: (1) professionals and senior managers, (2) intermediate professionals and managers, (3) skilled workers, (4) semi-skilled, and (5) unskilled and unemployed. Married women who were not employed were classified according to their husband’s occupation. The largest single group that could not be categorized were single women for whom no job was listed. The random sample of 700 adults treated at the Maudsley Hospital was analysed by frequency of symptoms and the 25 most common were identified for further study (Table 1).
Symptoms in order of frequency
Statistical analysis
Cluster analysis is a generic term for a wide range of numerical techniques that try to discover homogeneous groups or clusters of individuals within a sample of data. There are many such techniques available, and most are described in Everitt, Landau and Leese (2001). In the present study, a model-based method, proposed by Banfield and Raftery (1993) and extended by Fraley and Raftery (1999, 2002), was used. This statistically respectable method has the advantage of an associated test for number of clusters known as the ‘Bayesian Information Criterion’ (BIC); see Fraley and Raftery (1999). A K-means cluster analysis was applied to the first 10 principal component scores of the data set. These components accounted for 52% of the variation in the data, and use of the Bayesian information criterion indicated an eight-group solution.
Results
Cluster solutions
The eight-cluster solution (Table 2) divides itself into two groups: clusters 1, 7 and 8, which suggest psychosis, and clusters 2, 3, 4, 5 and 6, which reflect a range of psychological disorders. Cluster 1 has the characteristics of bipolar disorder, while cluster 7 represents severe mental illness, being composed of delusions, auditory and other forms of hallucination together with disturbed behaviour. Cluster 8, a predominantly male group, includes delusion, problems of speech and violent behaviour that suggest a severe psychotic disorder. Cluster 5 is the largest single group, comprising 38% of the total sample. These were depressed and anxious patients, some so severe as to have psychotic features. In addition, cluster 2 is indicative of a depressive group with symptoms of sleep difficulties, poor concentration, memory impairment and suicidal ideation. Cluster 4 has a psychosomatic quality (headache, weakness, dizziness, tremor and exhaustion) and includes many of the cases diagnosed as neurasthenic.
Percentage of subjects who report a symptom within each cluster
Gender
Reflecting the general trend, groups were predominantly female with the exception of cluster 8 (75% male), and cluster 3 (52% male). These findings imply that gender was not a crucial variable determining cluster membership.
Diagnosis
Diagnoses given by Maudsley doctors at discharge did not correlate with cluster groups (Table 3). Cluster 5, for example, contained 35 cases of dementia praecox or schizophrenia, but these represented only 39.8% of all such cases in the sample. In other words, no single group based on symptoms alone encapsulated a formal diagnosis. The implication of this finding is that doctors did not make diagnostic judgements solely on symptoms but relied on other data such as family history, behaviour and reports from other health-care professionals.
Eight-cluster solution by diagnosis: numbers, with percentages in brackets
Social class
Most adult patients admitted to the Maudsley were from the skilled artisan or middle class; only 13% of male admissions were labourers or unemployed (Jones and Rahman, 2008). Cluster membership was not defined by differences in income or class. Social class 5 (unskilled or unemployed) were found almost equally in all clusters apart from cluster 8 (Table 4). In addition, there are no significant differences in other classes: professionals were distributed though all groups at low levels (3.3–8.7%) apart from cluster 8, which had only 20 subjects.
Cluster solution by social class: numbers, with percentages in brackets
Changes over time
The numbers in clusters 1 and 7, both groups characterized by symptoms of psychosis, decreased over time: cluster 1 fell from 24 subjects in 1924 to 11 in 1935, while those in cluster 7 fell from 7 in 1924 to one in 1935. This trend did not reflect a fall in the incidence of psychotic disorders, but was a consequence of a change in the admissions policy of the Maudsley Hospital. The optimism of the immediate post-war period, when it was thought that advances in biomedical science and treatments developed for shell shock and other psychological disorders would effectively address psychosis, had passed. It reflected the growing realization that schizophrenia and hypomania were intractable disorders, even in cases that appeared to have a good prognosis. More realistic expectations of what could be achieved and the need for a throughput of new patients for the growing number of postgraduate trainees led Edward Mapother, the hospital’s first medical superintendent, to operate tougher admission criteria during the 1930s. Indeed, Desmond Curran recalled that on seeing a patient with long-standing schizophrenia ‘Mapother would just say, smacking his palm “a stiff praecox”’ (Curran, 1960: 5), the terse phrase ‘providing an epigrammatic summary of his opinion on aetiology, course and prognosis’ (Slater, 1960: 8).
‘An atmosphere of cure’
When the Maudsley Hospital opened in 1923, it included an out-patient department that operated on two afternoons a week, though in response to rapidly increasing demand, this was increased to four (Jones et al., 2007). In part, this service was provided to mitigate the stigma of psychiatric treatment, but also operated as a means of finding interesting patients for study. Not surprisingly, a high percentage of the clusters characterized by psychotic symptoms were in-patients. Clusters 1, 7 and 8 were composed of 91%, 65% and 80% of in-patients, respectively, while cluster 5, depressive disorders, was predominantly an out-patient group (54%).
Although not opposed to programmes of ‘re-education’ and occupational therapy, Frederick Mott sought physical treatments: medicines or other interventions designed to address pathological processes (Jones, 2010). Despite the promise held by malarial treatments for GPI, this strategy had yielded little in the way of cures. In the interim, both Mott and Mapother attempted to create an ‘atmosphere of cure’ based on diet, fresh air, control of infection and graduated exercise (Mott, 1919: 275). Once recovered from the severest effects of mental illness, patients were encouraged to take part in games (two tennis courts were laid out), social activity (dances and coach trips) and occupational therapy (raffia, rug-making, weaving, knitting, sewing and carpentry).
During the 1920s, psychiatrists had little to offer in the way of treatment, apart from various forms of restraint and sedation. Eliot Slater, who came to the Maudsley in 1931, recalled: ‘the most appalling thing was the chronic melancholics, often people of most excellent personality, sunk deep in a depression which nothing could move. People have no idea now of what that illness could be … But then there was absolutely nothing.’ (Wilkinson, 1993: 4). Those at risk of suicide had to be watched continuously. One male patient who took his life was found on post-mortem to have consumed yew leaves, which he had carefully ground into a digestible form. Confined to bed, he had been wheeled into the hospital gardens for fresh air and sunlight and he had been left under a large yew. Concerned that the death would be reported in the newspapers, Mapother immediately ordered the felling of the tree (Wilkinson, 1993: 4). In the first nine months of the Maudsley’s operation, two of the 10 deaths were suicides (Trevor and Bond, 1923). Reporting by the Board of Control was not consistent, although in 1929 it was recorded that one of the 41 deaths was from self-inflicted wounds; in the same year two patients had cut their throats but had not died from their injuries (Mapother, 1929). Suicides were also reported in 1926 (2), 1927 (1), 1931 (1) and 1937 (2), while for 1930 and 1932 it was stated that there were no deaths from self-harm.
Without a range of effective medicines or clinical procedures, emphasis was placed on the environment and in particular ‘open-air treatment’ (Trevor and Bond, 1923: 1). The Commissioner of the Board of Control in his annual inspection report for 1925 recorded that the ‘full use’ of veranda accommodation was a ‘pleasing feature’ of the Maudsley Hospital’s operation (Anon., 1925: 1). The concept of ‘rest and open air’ for the treatment of acute psychosis had been explored before World War I. Having observed ‘the improvement in tuberculous patients, sane and insane, undergoing the open-air treatment’, C.C. Easterbrook, medical superintendent of Ayr District Asylum, argued that ‘fresh air has an undoubted soothing and soporific influence on the nervous centres, and the cooler outdoor atmosphere stimulates general bodily metabolism and appetite, both of which effects render the open air of special value in treatment of active insanity’ (Easterbrook, 1907: 733, 743). Open-air therapy was conceptualized as an adjunct to continuous baths which were also designed to have a sedative and tranquillizing effect. It was also recommended that patients undergoing such treatment should be shielded from external stimuli, so screens were often placed between beds moved to verandas (Devine, 1929).
The ‘garden villa’, an 18-bed unit opened in 1931 for those in ‘a temporary phase of restlessness or excitement’ was a controlled therapeutic environment (Macleod and Rotherham, 1931: 1). The building was orientated to face due south so that patients could enjoy the benefit of sunlight. Each single room opened directly on to a veranda so that beds could be wheeled into the fresh air and screened from each other by ‘movable partitions’ (Anon., 1932: 1). As an experiment, only the ‘hinder half of the roof of the veranda was glazed’ with the idea that direct exposure to the elements would be therapeutic. The central corridor was lit by two rows of clerestory windows to provide the maximum amount of natural light.
Indeed, several rotating, wood-framed shelters had been installed in the hospital grounds so that patients could enjoy the benefits of fresh air while following the passage of the sun. These structures were sufficiently large to accommodate two beds, a locker and commode (Randl, 2008: 61). The front side of the hut was entirely open and the structure turned on a circular iron track. Originally designed for tuberculosis sanatoria, they reflected a popular belief that sunlight and fresh air had broad therapeutic properties. Sunlight, effective for the treatment of rickets and other vitamin-deficiency diseases, was also thought to destroy the tubercle bacilli (Campbell, 2005). In July 1932, when the Board of Control commissioners made their annual visit to the Maudsley, they were pleased to discover that 125 patients (60%) were ‘under continuous treatment in bed’, which in practice meant exposed to fresh air, in shelters, under verandas or on the bridges constructed between the two ward blocks. In a county asylum containing a higher proportion of long-stay patients, the proportion was commonly as low as 10% (Anon., 1932: 8).
Buildings constructed at the Maudsley during the late 1930s incorporated these therapeutic principles: the Children’s Department featured a flat roof which could be used as a play area, while the three-storey private patients’ block had a roof garden protected by high glazed walls and broad balconies so that beds could be moved directly from single rooms into the fresh air (Jones et al., 2007: 370). Indeed, it was hypothesized that a causal connection existed between tuberculosis and schizophrenia. In 1933 the Rockefeller Foundation paid £500 towards the cost of a bacteriologist at the Maudsley laboratories to research the pathological links between the two illnesses (Mapother, 1934). Popular culture emphasized purity, hygiene, fresh air and sunlight as generators of health (Campbell, 2005), though scientific experiment was only beginning to establish objective criteria.
While there were no treatments for melancholia before the introduction of cardiazol fits and later electroconvulsive therapy (ECT), patients suffering from schizophrenia or hypomania could at least be calmed by sedatives and in more mild cases by continuous baths. However, without short-acting barbiturates, sedation tended to be heavy-dosage with the associated risk of irreversible coma. Phenobarbitone, medinal, bromides and paraldehyde were given orally, while morphone and hyoscine came as injectables (Wilkinson, 1993: 5). Insulin coma therapy for the treatment of acute schizophrenia was introduced just before World War II. In an interesting reversal of priorities, the garden villa became an insulin-coma unit in 1947 run by Russell Fraser and William Sargant (Darwin and Maclay, 1947: 1) In the following year, however, the villa reverted to a place of calm and isolation for ‘disturbed and difficult patients’, while deep insulin therapy transferred to ward 2.
Attributions
Statements made by patients and recorded by clinical staff about the nature of their illness, together with the observations of medical officers, were collected as free text. This was read to identify common and recurrent themes (Pope, Ziebland and Mays, 2000). Items of data were repeatedly compared across the dataset to ensure that themes, differences and relationships between categories were re-examined and confirmed or modified (Green, 1998).
In total, 14 themes were identified (Table 5). The most common attribution made by both doctor and patient was that their mental illness was a consequence of a recognized disease, physical illness or operation. Indeed, throughout the interwar period it was believed that mental disorder was associated with a common physical illness. The focal sepsis hypothesis, for example, attracted considerable support during the early 1920s (Scull, 2005). Furthermore, in 1925 Humphry Rolleston, professor of physic at Cambridge and president of the Royal College of Physicians, proposed that hereditary factors, hitherto assumed present at birth, might be acquired in early youth as a result of family influences. Alternatively, they could remain dormant until awakened by exciting factors such as physical trauma, infection, toxins, unhealthy environment, diet or psychological triggers such as ‘worry, emotional strain, overwork’ (Rolleston, 1925: 782). Indeed, some of the attributions expressed by both patients and doctors fall into this category, rather than statements about the fundamental cause of mental illness.
Attributions: numbers, with percentages in brackets
The proposition that severe mental illness was a form of organic brain disease was increasingly questioned during the interwar period when new hypotheses about social processes and psychodynamic mechanisms were explored. The latter were reflected in statements about family and relationship difficulties and personality traits. However, the scant attention given to the theme of family history of mental illness is curious. Perhaps it was a belief so widely held that it required little iteration, or mention was censored to avoid stigma.
Conclusions
It is not known who designed the printed front sheets to the Maudsley case notes, but it is likely that both Mott and Mapother had a hand in their production. These standardized records were designed to gather information in a uniform manner to assist research and provide a method of training junior psychiatrists. Aubrey Lewis, who became clinical director at the hospital in 1936, repeatedly urged the careful collection of statistics so that clinical work could be properly evaluated. In February 1944, for example, Lewis raised the issue of the inadequate filing and retrieval systems for patient records, which limited the ability to undertake representative research (Lewis, 1944).
Analysis of case notes has shown that clinicians at the Maudsley were informed by symptom clusters, although they did not serve as an infallible guide to diagnosis and treatment. More recent studies have shown that psychiatrists base their decisions on other factors, including behaviour, family history and clinical intuition about the course of a longitudinal disorder (Stephens et al., 1986; Vaillant, 1984). This was an era before formal classification systems were tested statistically for reliability (Klerman, 1984). Not until 1952 was the Diagnostic and Statistical Manual of Mental Disorders, (DSM-I) published by the American Psychiatric Association. The first version of the International Classification of Diseases, Injuries and Causes of Death (ICD-6) to contain a section on mental disorders appeared only in 1949. As a result, diagnosis in psychiatry was fluid and subject to cultural forces (Young, 1995). For example, Kraepelin’s term ‘dementia praecox’ (indicative of an organic brain disorder) was routinely employed during the early 1920s but by the end of the decade had been supplanted at the Maudsley by Bleuler’s label ‘schizophrenia’ as ideas of causation turned towards psychobiological processes (Jones and Rahman, 2008). Senior and junior medical officers, regularly sent to the USA on Rockefeller Fellowships, were responsible for importing new diagnostic terms such as ‘behavioural disorder’ (Evans, Rahman and Jones, 2008). Although a popular diagnosis in 1920, neurasthenia fell from favour by the 1930s, being replaced by specific terms such as anxiety disorder, obsessive compulsive disorder and psychoneurosis (Taylor, 2001). During the interwar period, diagnostic terms were driven by causal explanations rather than observations of psychopathology (Klerman, 1989), a fact which explains the modest relationship between symptoms and cluster groups. Not until the 1960s did studies of symptomatology raise the status of psychopathology, while the introduction of DSM-III in 1980 had been preceded by three years of clinical trials by a designated ‘task force’ to establish diagnostic reliability (American Psychiatric Association, 1980: 5).
Footnotes
Acknowledgements
This research was supported by a grant from the Wellcome Trust (086071/Z/08/Z). The authors are also grateful to the staff of the Bethlem Royal Hospital Archive and Museum for their assistance in preparing case notes for digitization.