Assessing the conduct of juveniles: diagnosis and delinquency, 1900–2013

Abstract

American child psychiatrists have long been interested in the problems of delinquent behaviour by juveniles. With the rise of specific psychiatric diagnoses in the 1960s and 1970s, delinquent behaviour was defined within the diagnosis of conduct disorder. Like all psychiatric diagnoses, this concept was shaped by particular historical actors in context and has been highly contingent on assumptions related to race, class and gender. The history of conduct disorder illustrates the tensions in child psychiatry between the expansive goals of the field and the often limited uses of its professional authority, as well as individual children as the target of intervention and their interactions in groups.

In 1975, psychiatrist Dorothy Otnow Lewis and psychologist David Balla published a critique of the concept of sociopathy as applied to children in the American Journal of Psychiatry (AJP). They complained that the various terms that the American Psychiatric Association (APA) had enshrined in the successive editions of its Diagnostic and Statistical Manual (DSM) were negative and used to dismiss patients, especially children, without treatment. The sociopath was described in DSM-I as ‘always in trouble, immature, disloyal, hedonistic, irresponsible and unable to benefit from experience or punishment’. Individuals with antisocial personality disorder (the newer term from the 1968 DSM-II) were characterised as ‘grossly selfish, callous, irresponsible and impulsive’, as well as lacking loyalty to other people or groups. The problem, Lewis and Balla explained, was that many – perhaps all – of the children who were diagnosed with sociopathy had treatable psychiatric conditions that were being written off. ¹

Iowa child psychiatrist Richard Jenkins responded to defend the APA’s nomenclature, for which he claimed authorship. He said that just because there were some antisocial reactions that were misdiagnosed did not mean that the entire diagnosis should be thrown out. Instead,

we need to recognize a deficit in the capacity of some persons to relate with empathy to others. This deficit commonly results in a failure of socialization…Since this is a fundamental defect in character structure and has far-reaching and, at times, catastrophic consequences, we cannot ignore it.

Jenkins cited his own 1973 book that described behaviour disorders in general and insisted that he had a better understanding of the sources cited by Lewis and Balla than they did. ² Lewis technically had the last word in this published dispute, although she only briefly commented that she was working on a book on this topic and that she had corresponded privately with Jenkins, so that he would know she was not ignoring him. ³

Less than a decade later, they were at it again in the pages of the AJP. By this time, DSM-III had been released, and the conflict centred on a new diagnosis used to describe young people with problematic or illegal behaviours: conduct disorder. Lewis published an article in 1984 with a group of colleagues (including her child psychiatrist husband, Melvin Lewis) on the problems with the diagnosis, arguing that it was neither valid nor useful. Their major complaint was that the diagnosis focused on behavioural symptoms rather than causes, especially issues that arose in the brain, and missed other treatable conditions: ‘The types of disorders most frequently masked include psychosis, neurological impairment, learning disabilities, borderline retardation and even manic states.’ ⁴ Once again, Jenkins weighed in to critique the group’s research findings. He noted that they had dismissed an important distinction in DSM-III conduct disorder, that between socialised and unsocialised individuals. ⁵ This time, when the Drs. Lewis had the last word, they expressed respect for Jenkins’s work in this area but disagreed with his conclusions. ⁶

The conflicts between Lewis and Jenkins are a good example of the instability and contingency at the heart of child psychiatric diagnoses. As we know from numerous scholars, psychiatric diagnosis in the United States was (and is) highly subjective and periodically contested. ⁷ But the squabbles over how to classify child misbehaviour are particularly revealing, even as they connect to stories about adult diagnoses. Martyn Pickersgill has pointed out, for example, that Jenkins also attempted to influence the formation of the adult version of conduct disorder, antisocial personality disorder. ⁸ Child psychiatrists as a group were wary of the potential effect of diagnoses on children, fearful that labels could have long-term, negative consequences. There was no agreement on what caused problems – were they due to parenting, unconscious conflicts, physical problems, peer interactions, economic crises or neighbourhood issues? Furthermore, the Lewis and Jenkins disagreement illustrated one of the most difficult issues for child psychiatry as it evolved as a medical specialty in the twentieth century: what to do with juvenile delinquency.

As many historians have described, the child guidance movement that took shape in the first few decades of the twentieth century in the United States emerged as an effort to combat the pressing social problem of juvenile delinquency, that is, behaviour by young people that did not conform to the social and legal norms of society. Conversations about conduct problems took place against a backdrop of changing ideas and expectations of children, methods to raise and educate them and assumptions about how they should relate to each other. ⁹ As we might expect, then, the concept of juvenile delinquency has been fluid, contingent and used in wildly different ways depending on the expectations for children of any particular generation, as well as the race and class of the child in question. ¹⁰ Child psychiatry as a field began as an effort to bring medical expertise and attention to what had been seen as a social and economic problem. As Kathleen Jones pointed out, though, by the 1930s and 1940s, mental health professionals were distancing themselves from the seriously delinquent and instead focusing attention on middle-class children with less severe problems. ¹¹

And yet, the crises of poorly behaved children continued to haunt child psychiatry, even as the specialty was formally organised beginning in the 1950s. ¹² The professional discussions over the diagnosis of conduct disorder illuminate the tensions in the field over bad behaviour. Child psychiatrists debated whether they should treat every child with issues or just give advice? Furthermore, which children were they looking at? What were psychiatrists expecting to see? What were they proposing to do? Throughout the discussions of the conduct of badly behaved children and adolescents, psychiatrists came back again and again to questions about social and individual perspectives and interventions, the role of aetiology in description of behaviour and the question of how to balance statistical views with the unique perspective of an individual patient. While the leadership of child psychiatry had a strong bent towards psychoanalysis from the beginning of the specialty through the 1980s, the topic of bad behaviour in children seemed to attract clinicians and researchers who had a broader and more eclectic approach. Some used general Freudian concepts of unconscious conflict, and most agreed that families had a critical role in creating problems for kids. The conversations about conduct disorder were seldom monolithic, though, and the varieties of approaches that characterised early writings on juvenile delinquency carried through in later child psychiatrist writings about conduct disorder. ¹³

This paper explores the evolution of the diagnosis of conduct disorder, from its origins with child psychiatry in the child guidance movement, to its specific criteria in the DSM, to the conversations about the use of the diagnosis into the most recent edition of the APA’s diagnostic manual. In particular, I focus on the role of Richard Jenkins and his assumptions about boys that were embedded into the criteria he helped to write. I also use Dorothy Otnow Lewis’s objections to the diagnosis as a window on the contrasts between the populations of patients seen by different types of psychiatric experts. As with other psychiatric diagnoses, conduct disorder was developed at a particular time and place by individuals who were embedded in their own historical contexts. For the historian or cultural critic, psychiatric diagnoses can serve as a kind of Rorschach, a projection of the concerns of the framers about the individual and the world they perceived. Conduct disorder, in particular, highlights the tensions between the expansive goals of child psychiatry and the often limited uses of its professional authority, as well as individual children as the target of intervention and their interactions in groups.

Child guidance and juvenile delinquency

As numerous historians have pointed out, the child guidance movement was inspired by the goal to understand, and hopefully treat, children who behaved badly. At a time when many in the United States held a sentimental view of childhood, the prospect of children committing crimes seemed to indict society, not just the individual children. ¹⁴ In Chicago, at the beginning of the twentieth century, reformist efforts coming out of the settlement house movement prompted the creation of a clinic to assess children brought before the courts. Psychiatrist William Healy and psychologist Augusta Bronner, who later included a social worker into what became a classical treatment triad, created the framework for decades of mental hygiene efforts. The work of Healy and Bronner took place within the broader increase in the authority of social sciences. ¹⁵ Although child guidance teams acknowledged that social issues were important, historian Kathleen Jones has emphasised that Healy and Bronner shifted the focus towards the individual child. ¹⁶

Child guidance programmes became prevalent by the 1920s. At the Bureau of Child Guidance in New York in the late 1910s and into the 1920s, for example, psychiatrists collaborated with social work colleagues to gather data about children referred to the bureau. As Bernard Glueck explained to his social work students, there was not enough known about children and their mental states to be able to effectively create a classification. He created a broad list to group patients that included arrests of intellectual development, constitutional problems, psychoses, organic or functional nervous disorders and conduct disorders. ¹⁷ Lawson Lowrey complained, though, that the concept of ‘conduct disorders of childhood’ seemed to just be a way of stating the population of patients they were being asked to assess. ¹⁸ By gathering more extensive data on their patients, including development, relationship with parents and siblings, sexual issues, school performance and recreational pastimes, the psychiatrists hoped to translate intervention with kids into prevention of adult problems in the future. ¹⁹

In their 1936 book on delinquency, for which they presented research conducted at the Yale Institute of Human Relations, Healy and Bronner divided delinquents into three main groups: those with personality deviations or unstable neurosis, those with overwhelming external issues and those who were not grossly neurotic or with grossly socially pathological circumstances. On a continuum of problems that they explained were entirely within the child and those that were entirely external to the child, Healy and Bronner stressed that the children who were in the middle were the easiest to help. Healy and Bronner also emphasised that the family relations were among the most important to address when trying to get a handle on delinquency. ²⁰ The goal was not to diagnose but rather to understand. Healy and Bronner created categories in which to lump delinquents, but the groupings were meant as a kind of shorthand to indicate which delinquents would be more amenable to change.

Within the broad topic of juvenile delinquency, psychiatrists joined with other kinds of professionals in advancing the importance of expert opinion directed towards the problem. An excellent example was the Michigan Child Guidance Institute (MCGI), a data-gathering and research organisation that formed as a partnership between the state and the University of Michigan (UM) in 1937. The university created a Delinquency News Letter beginning in 1934 with contributions by many of the schools and institutes within the university, including perspectives from social work, the justice system, sociology and mental health. The MCGI continued the newsletter, and its UM sociologist director, Juilliard Carr Lowell, worked with a team of associate editors from the state welfare department, the judiciary, the police and UM faculty from religious education, psychology, law, child development and psychiatry. ²¹ The MCGI assessed cases of children and adolescents referred from all over the state, and offered an array of expert opinions on the problems and recommendations to the state. In 1940, the MCGI emphasised the need for more guidance clinics, community education and prevention programmes, creation of a juvenile probation system as well as a new twenty-bed children’s unit as part of the University Hospital’s Neuropsychiatric Institute. ²² By the time the United States entered World War II, the MCGI literature contained a sense of urgency and an overt connection between domestic safety and juvenile delinquency. The MCGI’s advisory role was valued by the state legislature, although by 1943 the competing priorities of wartime led to the dismantling of the programme.

This expansive approach to the issues of juvenile delinquency was echoed nationally by the number of multidisciplinary conferences and conversations about the problem. Although the experts could agree that professionals had important roles to play, it was less clear what the ultimate outcome could or should be. In 1949, for example, William Healy led a roundtable for the American Orthopsychiatric Association (Ortho) on the topic of psychiatry and delinquency, which included comments by psychiatrists, leaders in juvenile justice and educators. Psychiatrist George Gardner from the Judge Baker Guidance Clinic insisted that a psychiatric diagnosis – which would be informed by perspectives from sociology and anthropology – would explain the individual delinquent. But how useful was the psychiatrist explanation and diagnosis? A Toledo judge complained that psychiatrists used excessive jargon and that their diagnoses were not helpful. The president of the New York City Welfare Council agreed that psychiatrists were too much enamoured of jargon and that they did not actually conduct the treatment they recommended in general during child guidance assessments. Meanwhile, Healy insisted that the ‘core of the problem of delinquency is the child’s inner life of emotions, ideas and attitudes’, regardless of environment or circumstance. ²³

In the 1950s and 1960s, many psychiatrists focused on individual problems with a particular emphasis on internal conflicts that could be analysed and explained with Freudian theories. Kurt Eissler published an edited volume in 1949 on psychoanalytic studies of delinquency, while in 1952, Ortho conducted another roundtable on delinquency on the topic of psychodynamics. ²⁴ As US psychiatry as a whole became focused on the unconscious conflicts of individuals, the role of diagnosis in general was vague. ²⁵ The APA created its first DSM in order to provide an accounting mechanism for individuals in psychiatric treatment. The original categories in what would be known as DSM-I (1952) sounded explicitly Freudian. ²⁶ However, there were members of the APA’s nomenclature committee who were ready to take a significant step away from a narrow, Freudian interpretation with regard to understanding and defining juvenile delinquency. Perhaps the most influential of these was Richard Jenkins, who used his own clinical experiences to define the emerging concept of conduct disorder.

Richard Jenkins and the construction of conduct disorder

Richard Jenkins was born in 1903 and attended Stanford University before going to medical school at the University of Chicago. ²⁷ He started his career as the medical director for the MCGI in 1941, two years before it closed, and then transitioned to the Institute for Juvenile Research in Chicago. He subsequently moved to Washington DC, and from 1949 to 1961, he was the research chief for the Veteran Administration’s (VA) psychiatric evaluation unit. While at the VA, Jenkins embraced a critical stance towards psychoanalysis and the perceived truths of psychiatry in general. ²⁸ He finally landed as head of the child psychiatry division at the University of Iowa, where he also worked with the boys from the local training school. Throughout his career, Jenkins insisted that data were forming his assessment of the conduct of juveniles and that he was at the vanguard of scientific research.

One of Jenkins’s first projects was an extensive analysis of data from the MCGI. In this project, Jenkins collaborated with research fellow Lester Hewitt to review charts from 500 delinquents who had been referred for evaluation to the institute. Of note, the MCGI had data for girls as well as boys, although Jenkins and his collaborators who used this type of approach only discussed boys. Hewitt and Jenkins (with the assistance of Sara Lou Mann Hewitt, a clinical psychologist who was married to Hewitt but who was not named or given formal credit for the work) described a large number of behavioural traits from review of the boys’ charts. ²⁹ They then transcribed the traits onto punch cards for statistical analysis and published the results in 1946. Hewitt and Jenkins clustered the traits into three types: the overinhibited child (characterised as shy, apathetic, sensitive and prone to neurosis), the unsocialised aggressive child and the socialised delinquent. ³⁰

Jenkins used this general methodology and delinquent types in numerous publications over subsequent decades. Around the same time that he was finishing the project with Hewitt, he also collaborated with another research associate from Ann Arbor, Sylvia Glickman, to analyse data from the Institute for Juvenile Research in Chicago. For this paper, as with the publication with Hewitt, Jenkins used a combination of statistical data and narrative case-type explanations to describe a typical boy within the category. The unsocialised aggressive boy ‘is characterised by assaultive tendencies, initiating fights, cruelty, defiance of authority, malicious mischief and inadequate guilt feelings. He is selfish, jealous, vengeful, deceitful and prone to place upon others the responsibility for his own misconduct. He is suspicious of others, profane and obscene in language, and precociously interested in sex’. The socialised delinquent, on the other hand, ‘is characterised by stealing in the company of others, furtive stealing, habitual truancy from school, staying out late at night, desertion of home, bad companions and gang activities’. ³¹ This latter separation between boys who acted badly alone and those who violated the law in the company of others was something that Jenkins emphasised throughout his career.

In fact, Jenkins argued that the ability to relate to peers was a key distinction for both description and prognosis for boys. The difference between boys who could relate to others and those who could not was based, Jenkins said, on the boys’ relationships with their mothers. In an analysis of another set of data, this time from the New York State Training School at Warwick, Jenkins and Glickman looked at the boys’ home lives. Their article explained that the unsocialised aggressive child was a product of maternal rejection, while the socialised delinquent was wanted but neglected. By this publication, Jenkins was becoming more favourably disposed towards the socialised delinquents. As Jenkins and Glickman described, the socialised delinquent ‘exhibits some loyalty towards the members of his group, his companions in delinquency, and he is likely to show courage and fortitude in adhering to the code of his group – for example, in resisting pressure towards informing upon fellow delinquents’. The bigger problem for the socialised boys was that their fathers were absent or ineffective. ³²

In his separation between unsocialised, loner boys and those who moved in gangs, Jenkins reflected hopes, fears and expectations of how boys should relate to one another, particularly with regard to race and class. ³³ Many mental health professionals recognised the value of boys’ friendships. Harry Stack Sullivan, for example, said in the 1930s that a boy needed a ‘chum’ and that this friendship was protective against mental illness. ³⁴ Reformers had been arguing for decades that boys could be influenced by positive peer relationships. ³⁵ Groups such as the Boy Scouts had long been used to cultivate certain kinds of White, middle-class boy habits and interactions. ³⁶ In the post-war years, prevention for juvenile delinquency often meant intervention at the level of boys’ interactions, based of course on White, middle-class expectations. ³⁷

It is obvious that the way that experts viewed boys and their socialisation depended on social and cultural factors in the United States at the time, as well as common understandings of how youth and adults should interact. Although Jenkins insisted that his conclusions about types of delinquent boys were based in statistics and science, it became clear over the decades that he had some degree of admiration for the socialised delinquent boys. By the 1950s, Jenkins was characterising their behaviour as adaptive – meaning that it served some function to help the boys survive. This was in contrast to his assessment that the unsocialised boys were engaged in maladaptive, self-destructive behaviours. ³⁸ For the socialised delinquent,

Its typical antecedents include physical vigor, an aggressive attitude toward life, a good musculature, an experience of having received maternal care adequate to have developed a social responsiveness to others, and either the absence of paternal guidance or conflict with the father in pre-adolescence and adolescence.

The only difference between socialised delinquents and normal boys, he said, was that they were striving towards activities that society had deemed illegal. For the unsocialised delinquent, ‘The picture is one of rather gross poverty of the personality and poor personality organisation coupled with a pattern of vengefulness, inability to get along with other children, sullenness, negativism and suspicion.’ For those with ‘a good musculature and a high level of native vigour’, there was risk for aggressive action. ³⁹

It is perhaps not surprising, then, that Jenkins assumed a racial correlation with the two types. In his study with Glickman on data from the New York State Training School at Warwick, he included one sentence on race: ‘The socialised delinquent is more often White, and the unsocialised aggressive child more frequently is Negro.’ ⁴⁰ In 1960, Jenkins co-authored a paper with psychologist Eva Blodgett on the evaluation of delinquent boys with a sentence-completion task, a projective test in which the boys were given the beginning of a sentence. Their conclusion of the sentence was coded for concern for others and a willingness to conform to society. Jenkins and Blodgett had data from before and after treatment at a couple of different training facilities, but they struggled, because they said that the Negro boys had a higher rate of recidivism to criminal activities, and it was hard to balance the data. They decided to focus on lessons learned from the White boys about evolving a sense of personal responsibility and moral success. ⁴¹ The same year, Jenkins also published an article about antisocial personality disorder and its antecedents in adolescence. He stressed that humans were social animals, but that the psychopath had no capacity for loyalty and was ‘a basically unsocialised animal who nevertheless lives in society’. He did acknowledge that social discrimination could worsen antisocial behaviours, as those who were being discriminated against would be more likely to act out against society. ⁴²

Jenkins’s dislike of unsocialised delinquents became more and more palpable over the decades. He was more positive and optimistic about what could be done for socialised delinquents, who he said could be treated with strong, supportive father figures in a training school environment. As he explained,

The good comrade of the delinquent gang is a lad of strong loyalties, and this loyalty can usually be captured by a socialized adult who is willing to work at it. The gang boy easily admires and often will spontaneously pattern himself after a strong masculine figure, particularly if this person seems to like and understand him. ⁴³

The unsocialised boy, however, required constant limits, redirection and an adult who was capable of withstanding the boys’ rejection of the authority figure without becoming hostile in return. In the 1970s, Jenkins admiringly cited the fictionalised account of a typical juvenile delinquent created by Clifford Shaw, The Jack-Roller. Although Shaw, a sociologist, had told the story to illustrate the problems with poverty and deprived neighbourhoods, Jenkins took the protagonist of the story as a model of a delinquent who (although his behaviour was maladaptive) could be redeemed through alternative socialisation. ⁴⁴

Jenkins had strong opinions about his own work and his value to the field, and he also had an outsized influence on the diagnostic upheaval that occurred in psychiatry in the 1970s. Jenkins was one of a small number of psychiatrists who participated in the process to revise DSM-I. DSM-II, which was published in 1968, was the first edition to include explicit diagnoses applicable to children and adolescents. As the introduction to DSM-II explained, the revisions from the older edition were intended in part to harmonise international diagnoses (in the International Classification of Disease) with those used by the APA. ⁴⁵ The DSM-II categories for children and adolescents, which Jenkins later claimed to have written, included hyperkinetic reaction, withdrawing reaction, overanxious reaction, runaway reaction, unsocialised aggressive reaction, group delinquent reaction and other reaction. ⁴⁶ Jenkins particularly laid claim to the reactions that correlated to his research samples, which he identified as the runaway reaction, the unsocialised aggressive reaction and the group delinquent reaction. ⁴⁷

New York State Psychiatric Institute biometrician Robert Spitzer joined the process to finish DSM-II as a consultant in the last year or two of the APA nomenclature committee work. By the early 1970s, he had taken over a much more ambitious project: to completely revamp the diagnostic system to create discrete criteria. As a result of this effort, Spitzer helped to usher in an era of major professional, social and economic focus on diagnosis within psychiatry. ⁴⁸ Spitzer worked with small groups of experts in each area of the proposed DSM-III, including a cross section of child psychiatrists who could help with the categories applicable to children. Although Jenkins was not initially one of the members of the committee, he became so involved with the work of describing the diagnosis of conduct disorder that he was given credit when DSM-III was published in 1980. ⁴⁹ As committee member Stella Chess remarked to a colleague about the process, Spitzer wanted everyone to agree on every point. He shared all of the information given to him by each consultant and committee member, and he continued discussions until everyone could agree. ⁵⁰

The name for the DSM-III diagnosis of conduct disorder was probably related to the ICD-9 classification of disturbance of conduct, which in turn was also likely shaped by the older usage of the concept of conduct disorders from the early dates of child guidance. ⁵¹ The committee also created a new – and somewhat contested – diagnosis of oppositional defiant disorder, although no one was ever clear on whether it really existed. Jenkins eagerly jumped into the committee discussions about conduct disorder and provided Spitzer and other members of the committee with numerous drafts of the criteria, copies of his published work and frank opinions about what needed to happen. ⁵² As Spitzer passed around Jenkins’s drafts and comments, he did not editorialise about Jenkins’s passion or his persistence regarding his own research. ⁵³ Spitzer noted to committee member Rachel Gittelman that, although they were not obligated to use the language that Jenkins proposed, it was an efficient place to start. ⁵⁴

Spitzer attempted to reach a compromise among his consultants. He included elements suggested by Jenkins with rewrites conducted by Gittelman, along with clarifications offered by California psychiatrist Dennis Cantwell. Jenkins, however, was not interested in compromise. The group decided to try to split off a part of the conduct disorder spectrum into a nonmedical term of antisocial behaviour. Spitzer attempted to circulate the new criteria. But Jenkins wrote back with line-by-line feedback, objecting to the changes. Jenkins wanted the diagnosis to reflect the patient populations he treated at training schools, and thought that the criteria he wrote best expressed this. As other members of the committee expressed varying degrees of frustration, Jenkins locked in and increasingly cited his own work as the scientific proof of his statements. ⁵⁵

The resulting diagnosis reflected Jenkins’s influence, although he never stopped complaining about it, because it did not precisely match his research categories from the 1940s. The DSM-III conduct disorder included two main types, socialised and undersocialised. Each type also included aggressive and nonaggressive. ⁵⁶ Jenkins did not like the subgroup of socialised, aggressive conduct disorder – he said it did not exist. His emphasis in the conversations about the specific criteria for the types of conduct disorder was on the ties to outcome. At the same time that he was participating in the DSM-III committee conversations, he was also completing a study on boys from the Eldora Training School in Iowa. Jenkins insisted that these data demonstrated the major differences in outcome between socialised and unsocialised delinquents and also illustrated that aggression was only an issue for unsocialised boys. ⁵⁷ Jenkins complained to Spitzer in 1979 that the diagnostic categories from DSM-II could be demonstrated statistically but that the division of socialised delinquency into aggressive and nonaggressive types did not make sense, because it was all about circumstance:

The undersocialized individual is more stereotyped in his behavior, less adaptive, and more typically has an automatic pattern of striking out or running. The socialized individual has a wider range of responses and in this regard is more adaptable and more healthy. ⁵⁸

Even as Spitzer tried to ease the committee into agreement when the nosology needed to go to press, Jenkins continued to complain. ⁵⁹

Jenkins emphasised the difference between the socialised and unsocialised conduct disordered boys, because it fit his experience of which boys seemed to respond to the structure and interventions at his training centre. When the categories were applied, though, the racial bias that was built into the diagnosis could be seen. As Rutgers psychologist Harriet Hollander explained to Spitzer when he asked her to be on the task force to revise DSM-III, the undersocialised conduct disorder diagnosis was being used by the family court in New Jersey as a reason to waive an accused teenager from the juvenile system to the adult courts. While this was validation of the usefulness of the APA’s system of diagnosis, ‘An unintended adverse consequence of using DSM-III to support waivers could be to transfer disproportionate numbers of Black and/or low IQ youths to adult prisons. Such a practice would undo years of work by child advocates to separate juvenile from adult offenders.’ ⁶⁰ University of Pittsburgh psychiatrist Anthony Costello also pointed out that trying to differentiate delinquents based on their capacity to form friendships could lead to troubling issues: the socialisation issue ‘slides very easily into class and ethnic biases about expected behaviour, and also raises the spectre of psychiatrists offering a diagnosis which “excuses” behaviour when the diagnosis merely states the social context in which the behaviour occurred’. ⁶¹ Jenkins believed that his diagnostic distinctions were based in science, but the applications revealed the broader social and cultural implications of the diagnostic label.

Sociologist Lee Robins, a leading authority on the long-term patterns of antisocial behaviour with its antecedents in childhood who acted as a consultant for Spitzer’s group, recommended that the task force remove the conduct disorder types for what became DSM-IIIR in 1987. Her 1966 book Deviant Children Grown Up had examined the cases of hundreds of children who had been referred for child guidance interventions as children or teens. The vast majority of them grew up to have criminal histories. Robins had not differentiated among types of delinquents, but did support the idea that these behaviours constituted a psychiatric condition. ⁶² Spitzer took up Robins’s suggestion with the task force. Although there was hesitation and a great deal of discussion, the group ultimately decided to take the subtypes out of the diagnostic system but leave in the text a discussion about what might contribute to a more favourable prognosis. ⁶³ When Jenkins got wind of this, he wrote a series of strongly worded letters to Spitzer and the other members of the task force. He concluded, ‘If this proposed move is successful, I will regard it as the longest step in child psychiatry diagnosis in the twentieth century. Unfortunately, it is a step backwards.’ ⁶⁴ His colleagues did not agree. The questions they were starting to ask about the potential consequences of a diagnosis had already been taken up by Jenkins’s AJP adversary, Dorothy Otnow Lewis.

Dorothy Otnow Lewis and the consequences of diagnosis

Although Lewis was by no means the only child psychiatrist who wrestled with the concept of conduct disorder as it was used in practice, she developed her observations and arguments over decades and articulated important issues. Lewis spent much of her career trying to understand why people behaved badly, and began her work with the assumption that something must have happened to an individual to make him commit crimes. ⁶⁵ In her studies, in which she collaborated with colleagues from the Yale Study Center and New York University, she worked to overcome both the psychiatric assumption that delinquents just behaved badly and the expectation by sociologists that social circumstances were what drove children’s behaviour. In 1976, her research group noted that parental issues could lead whole families into difficulties, and argued that both family environment and genetics could contribute to a child’s delinquency. ⁶⁶ In 1977, Lewis and Shelley Shanok pointed out that children who had been referred to juvenile court had a greater than expected number of encounters with health care systems due to accidents, injuries and illnesses. They argued that these hospital encounters could be important places to try to address issues, including uncovering medical problems that may have led to delinquent behaviours. ⁶⁷

In their efforts to uncover factors that might lead children to behave badly, Lewis and her team found consistent – though perhaps not surprising – differences in outcomes based on the children’s race, socio-economic status and family structure. They discovered that children referred to child guidance centres were more likely to have a biological mother in the home, perhaps taking the role of protecting the child from the legal system. ⁶⁸ It became clear that race was a major factor. As Lewis and her colleagues explained in a paper published in 1979, they explored ‘the possibility that black delinquent children and their families are being denied access to psychiatric treatment because of a reluctance or inability of mental health professionals to diagnose severe psychopathology in his population’. ⁶⁹ While Jenkins had asserted that the traits of the boys he saw determined whether or not they were sympathetic and could benefit from treatment, Lewis and her group called out the fact that prejudice shaped how psychiatrists and others interpreted the child in the first place. As they explained in 1980,

Our clinical and epidemiological findings indicate clearly that many seriously psychiatrically disturbed, aggressive black adolescents are being channeled to correctional facilities while their equally aggressive white counterparts are directed toward psychiatric treatment facilities. As a result of this practice, correctional facilities in the United States are being asked to function as the mental hospitals of the lower socioeconomic class black population. ⁷⁰

At a time when sociologists and psychologists in the United States were still grappling with contentious arguments about families and social environments creating Black delinquent behaviours, Lewis and her colleagues argued that bias on the part of White professionals kept them from diagnosing and treating real medical problems. ⁷¹

Lewis and her group followed in the spirit of William Healy in their efforts to identify psychiatric problems in kids who had been pushed into the juvenile justice system, although their ideas were at odds with much of the established child psychiatry of their time. ⁷² Lewis edited a special issue of the Journal of the American Academy of Child Psychiatry in which she criticised existing frameworks for juvenile delinquency and the psychiatric response. ⁷³ One major issue that Lewis and her colleagues emphasised was that children became violent, because they had experienced violence themselves. In the early 1960s, paediatricians had begun to call attention to patterns of physical injury that suggested child abuse, while Congress passed legislation intended to protect children from abuse in the mid-1970s. Most of the discussion on this issue centred on children as fragile and dependent. ⁷⁴ What Lewis did was to try to connect concerns about vulnerable children with the problems of children who went on to commit violent acts. ⁷⁵ Her group compared violent juvenile delinquents with their nonviolent correctional school peers and found a higher degree of neuropsychiatric symptoms in the violent group. They concluded that

enlightened psychological, educational, and medical programs can and should be devised to meet the needs of these multiply damaged children. Programs designed to diminish violence which focus primarily on socioeconomic and psychological factors are likely to be unsuccessful if they ignore the medical problems (e.g., psychotic symptoms, neurological impairment) that contribute so strongly to the expression of violence. ⁷⁶

Lewis’s group emphasised that it was necessary to look at what might be causing delinquent behaviour rather than assume that delinquency was what needed to be managed.

Even as Lewis and her colleagues reached different conclusions about delinquents in training schools from Jenkins, legal reformers were trying to overhaul the juvenile justice system itself. In 1974, Congress passed the Juvenile Delinquency and Youth Offenses Control Act that was intended, among other things, to offer alternatives to incarceration for youth who engaged in certain kinds of crimes (such as status offenses, defined as things that would not be criminal if committed by adults such as truancy or running away). ⁷⁷ As scholar Barry Feld has pointed out, the juvenile justice system in the early twentieth century was originally intended to serve more as a social service organisation, with discretion for judges and courts prioritised over due process for juveniles. Juvenile justice reforms in the 1960s and 1970s, especially in concert with increased migration of Blacks to urban areas, made the risk for minority boys much higher to be incarcerated with less process and more punishment. ⁷⁸ As Elizabeth Hinton has pointed out, the social service programmes embedded within urban centres in the 1970s and 1980s combined resources with increased surveillance and punishment, and led to dramatic increases in incarceration for youth, especially Black children and adolescents. ⁷⁹ At the same time, as social and juvenile justice reforms were leading to expanded incarceration for some populations, other well-intentioned efforts to reduce institutionalisation of all kinds for kids had unintended consequences. As Deborah Doroshow has shown, the 1970s move away from institutions resulted in the closure of a number of residential treatment centres that had been intended for children and adolescents who had not been manageable in their communities. ⁸⁰

The confluence of changes in juvenile justice, institutional or residential options for emotionally disturbed children, and the new diagnostic categories of DSM-III made for a great deal of argument among psychiatrists in the 1980s. It became increasingly obvious that providers who advocated for a particular classification or treatment intervention were not necessarily talking about the same patient populations as others who advocated just as strongly for something entirely different, nor were they necessarily using the same diagnostic labels. As Michael Rutter, from the Institute of Psychiatry in London, and David Shaffer, from Columbia University, complained in 1980, ‘a child presenting with socially disruptive behaviour might be included under one of the three attention deficit disorder (ADD) categories, under one of the five varieties of conduct disorder, oppositional disorder, identity disorder, personality disorder, adjustment disorder or under the V code for childhood antisocial behaviour’. ⁸¹ The model of ADD, which as a diagnosis continued to expand into broader clinical populations who were increasingly treated with medications, offered a suggestion of how to manage poorly behaved children in clinic settings. ⁸² Although medication was not a common modality for many child psychiatrists of this era, the severity of the kids within the treatment populations seemed to suggest that the addition of a biochemical agent might help. ⁸³ It made a difference whether researchers were looking at behaviour in clinic settings or took a longitudinal perspective, as in the work of Lee Robins. ⁸⁴

For Lewis and her colleagues, though, the kids of their studies were much more disturbed and their treatment raised more challenging issues. In comparing neuropsychiatric problems for the poorly understood population of delinquent girls against what they had learned about delinquent boys, Lewis’s group found that the girls were more likely to have been abused but not more likely to be more psychiatrically ill. In a comparison of psychiatrically hospitalised boys with boys who had been in detention facilities, they found that both groups had major issues. Lewis’s group cautioned that society was moving towards seeing psychiatrically ill children as adults who were criminally responsible, and ‘Our findings clarify the need for society to decide whether the psychiatric system or the correctional system will take responsibility for providing appropriate care for these multiply handicapped, often extremely aggressive youngsters.’ ⁸⁵ Lewis and her group acknowledged the problems but insisted that these children needed and deserved some kind of help.

A few psychiatric researchers replicated the work of Lewis and her colleagues in terms of assessing populations of kids in detention facilities. ⁸⁶ For the most part, though, others publishing in the psychiatry literature used the DSM-III (and the 1987 revised version, DSM-IIIR) diagnosis within clinic populations, not for children already within the juvenile justice system. Lewis did not shy away from the tough cases. She and collaborators from New York University looked at boys who had been evaluated when they were younger and compared a group that had gone on to commit murder with those who did not. They found

that severe CNS dysfunction, coupled with a vulnerability to paranoid psychotic thinking, created a tendency for the nine homicidal subjects to act quickly and brutally when they felt threatened. Living within psychotic households, they were frequently the victims of and witnesses to psychotic parental rages, experiences that undoubtedly further exacerbated their tendencies toward the physical expression of violence.

They suggested that early intervention might be a possibility, especially for those children who already had a tendency towards violence. ⁸⁷

Lewis and her group wanted to help. They articulated the lofty aim – one that had been repeated over decades of child guidance and child psychiatry – to try to prevent serious delinquency by identifying at-risk children early on. But the move towards diagnosis with DSM did not seem to add clarity or direction. As University of Pittsburgh psychologist Edward Mulvaney and John LaRosa from the Middletown Youth Services Commission in Connecticut pointed out in 1986, delinquency treatment and prevention programmes did not work as well as everyone wanted. They suggested that the problem might be in how it was framed: ‘In the case of adolescent antisocial behaviour, the formulation of this problem as a “disorder” to be “treated” may have limited our ideas about possible solutions.’ ⁸⁸ Lewis went farther than that to say that the particular diagnosis of conduct disorder was used to dismiss. Instead, she and her group wanted a new diagnosis of Limbic-Psychotic-Aggressive Syndrome to capture the population of violently behaving children and adults. ⁸⁹

In 1989, Lewis and her group conducted a study to follow up on the criminal careers of ninety-seven boys who had been incarcerated in Connecticut in the 1970s. The boys’ classification as being more or less violent as juveniles did not predict adult violence. Instead, both intrinsic vulnerabilities (such as psychotic symptoms, neurological impairment or cognitive problems) and violent homes seemed to predispose individuals to more violence as adults, with particular danger from those who had both multiple vulnerabilities and an abusive home. ⁹⁰ Lewis’s group did not make use of the conduct disorder diagnosis, a fact that Richard Jenkins seized on in his letter commentary on the study in 1990. While he agreed with Lewis’s group on their data, he said that there was a public health reason to include conduct disorder:

Conduct disorder, and particularly, solitary aggressive type, adds a cautionary note of the risk of violence, including violence to infants, whether there is an organic or other diagnosis or not. There is a measure of public protection in making such a risk explicit when it is present. ⁹¹

The role of a diagnosis of conduct disorder was itself a point of contention within child psychiatry.

DSM-IV and diagnostic certainty

Even as its use was debated, conduct disorder remained within the APA’s diagnostic manual, which was taking on more and more authority as an authoritative system of psychiatric expertise. ⁹² Diagnostic categories continued to be adjusted within small workgroups where the influence of key individuals could still have an outsized effect. ⁹³ As DSM-IV editor Allen Frances later reflected, the pharmaceutical industry began to invest in the concept of psychiatric diagnosis that could help market particular pharmaceutical products. ⁹⁴ The biggest issue with conduct disorder seemed to be that psychiatrists increasingly expected that DSM diagnoses would reflect patient populations they could help, rather than descriptions of kids with behaviours outside their scope.

In the workgroup conversations around how – or whether – to revise disruptive behaviour disorder categories for DSM-IV in the early 1990s, a group of child psychiatrists considered what data they had and whether they might be able to obtain more. In both the process of revising criteria for DSM-IIIR and DSM-IV, participants were told that there was going to be field trials to gather information that could be used in their discussions. The workgroup members had different expectations of the role of the field trials than the directors of the DSM revision process, though. In 1985, University of Massachusetts neuropsychologist Russell Barkley had prepared an extensive packet of information for multiple sites to allow them to use extensive batteries of tests to assess children and determine the appropriateness of disruptive diagnoses, including ADD, oppositional defiant disorder and conduct disorder. ⁹⁵ Barkley discovered, though, that Spitzer intended the field trials to only include a mechanism for clinicians to check their assessments of patients against the existing DSM criteria. ⁹⁶ By the time it came to revise the criteria for DSM-IV, the mechanism of the field trial was established on Spitzer’s model. For the group that investigated disruptive behaviour and ADDs, they looked at 440 clinic subjects, 336 boys and 104 girls, who had been given a standardised interview questionnaire. ⁹⁷

The members of the DSM-IV workgroup, which overlapped with the previous workgroup by only one or two members, discussed some of the clinical issues of the diagnosis. Harvard’s Joseph Biederman, who later achieved notoriety for his use of pharmaceutical industry funds to promote (and treat) the diagnosis of bipolar disorder in preschoolers, argued that it would be helpful for research to have both oppositional defiant disorder and conduct disorder possible in the same patient. ⁹⁸ Other members of the workgroup discussed whether it would make sense to add a subtype for bullying. Columbia University psychiatrist David Shaffer pointed out that there were different data for boys than for girls, but there were not enough data to suggest sex-specific criteria, especially since bias in clinic intake might affect whether girls were diagnosed with conduct disorder at all. ⁹⁹

There had been some discussion of delinquent behaviour in girls, although most of it focused on the differences between boys and girls. One team from the University of Missouri pointed out that girls who were placed in juvenile detention facilities were more likely to have committed status offenses, while boys were more likely to have committed offenses against others and/or property. ¹⁰⁰ The older literature on the role of sex in delinquency, especially among girls, was generally not incorporated into conversations about conduct disorder. ¹⁰¹ Dorothy Lewis and psychologist Shelley Shanok looked at the medical histories of female delinquents and commented that the sparse literature that focused on girls tended to look at gynaecological problems. ¹⁰² Most of the literature on kids in detention facilities, though, discussed only boys (even when not being explicit about this in the title or abstract of their papers).

This began to shift by the conduct disorder discussions for DSM-IV. University of Chicago psychologist Benjamin Lahey commented that there was a paradox in trying to use symptom criteria to define a diagnosis – ‘one cannot examine the diagnostic utility of a given symptom without relating it to the full diagnosis, but one cannot define the diagnosis without knowing the full list of symptoms.’ One of the factors that the group needed to weigh, Lahey pointed out, was that changing the age threshold for some of the criteria would change how many girls could be diagnosed with conduct disorder. ¹⁰³ Lahey did propose some changes in the criteria based on what he said was a careful review of the literature:

Although there is much left to be learned, it appears that there is a persistent, male-dominated, aggressive childhood-onset type of CD, and a less persistent and less aggressive adolescent-onset type of CD for which there is a less marked gender difference (or that is more common among females). ¹⁰⁴

Toronto psychiatrist Susan Bradley commented that this conceptualisation was helpful, because it would make it possible to include girls with the conduct disorder diagnosis, a move she thought would correct problems with the older criteria. ¹⁰⁵ The DSM workgroup, in any event, was pleased with their work and suggested that the modified diagnosis was more reflective of the types of problems commonly seen by child psychiatrists in their practices.

Since conduct disorder had its origins in the juvenile delinquency literature and the detention system, the purpose of the diagnosis was never particularly clear. As British child psychiatrist Michael Rutter commented in 1996 in response to an effort to create practice parameters through the American Academy of Child and Adolescent Psychiatry (AACAP),

My major concern is that the document starts with an assumption that there is a diagnostic entity called conduct disorder and that the diagnostic criteria and boundaries of the syndrome are known. I presume that this has been required by the need to start with a DSM-IV category. The problem, however, is that antisocial behavior is much commoner than the diagnosis of conduct disorder and the controversies when and how this should be regarded as a psychopathological disorder are not really addressed here.

Rutter complained that the practice guidelines included an absurdly large number of treatment options, including psychodynamic psychotherapy. ¹⁰⁶ Rutter’s comments illustrated the increasing gaps among what individual practitioners and clinics might have to offer, the patients who might present for treatment and the broader problems of severe child behaviours.

While child psychiatrists in the early twentieth century argued that juvenile delinquency and violence committed by children was part of their professional purview, late twentieth-century child psychiatrists were much more cautious about how to approach the bigger social issues. In 1994, AACAP president William Ayers asked Stanford psychiatrist Hans Steiner to chair a new committee on violence and mental health. Steiner eagerly wrote to him about a number of topics he thought would relate to this topic, including public policy on violence and trauma. ¹⁰⁷ Ayers wrote back to say that while the committee could certainly look at children in the juvenile justice system, racism in juvenile justice and society as a whole or violence as an aspect of wider problems in society, Ayers wanted Steiner to focus instead on conduct disorders in kids inside and outside the juvenile justice system. Ayers suggested that the name of the committee should be the Committee on Children and Adolescents with Conduct Disorders, Research and Clinical Issues. ¹⁰⁸

Steiner did work with his committee on conduct disorder practice issues, but also tried to broaden the discussion to mental health and juvenile justice. In 1998, he suggested that his committee survey private practitioners to assess their opinion of the usefulness of the conduct disorder diagnosis. ¹⁰⁹ He remained acutely aware of the bigger issues, though. In an article for a newsletter for the National Alliance for the Mentally Ill, Steiner brought the field of child psychiatry full circle by proposing that some of the kids in the juvenile justice system needed psychiatric help, not just punishment. He noted that the prospect of getting assessment by a psychiatrist was not necessarily welcome by kids in the justice system: ‘Many of the boys and girls on probation and in parole appear to be in need of our services, yet such a prospect seems threatening and scary to them. To be bad is better than to be crazy. Our job is to convince them they are neither.’ ¹¹⁰ In 1999, the president of the AACAP recognised the usefulness of Steiner’s perspective, and suggested that he create talking points for the media. ¹¹¹ Whether or not they were going to try to tackle the big social issues, child psychiatrists wanted their expertise to be disseminated.

Conclusion

Conduct disorder as a diagnostic entity has perhaps been even more arbitrary and unstable than many other psychiatric diagnoses. While psychiatrists’ increasing claims about the pathology of entities, such as sadness or shyness, served to buttress professional stature and help pharmaceutical sales of drugs to treat these conditions, the utility of conduct disorder has been less obvious. ¹¹² The current literature on juvenile delinquency has little or no connection to conversations in child psychiatry. Sociologist Clemens Bartollas, for example, has published recurring editions of a textbook on juvenile delinquency in which there is no discussion of the psychiatric diagnosis of conduct disorder and very little (if any) conversation about what kinds of psychiatric treatment modalities might be relevant for juvenile offenders. ¹¹³ Meanwhile, child psychiatry descriptions of the management of conduct disorder understandably assume that children and adolescents who meet criteria will be brought to the attention of treatment providers in typical clinic settings. ¹¹⁴ However, the existence of a psychiatric diagnosis for bad behaviour in kids does not necessarily indicate professional enthusiasm for the issue. As with its adult diagnostic counterpart, antisocial personality disorder, most psychiatrists in practice are uninterested in working with individuals who carry the diagnosis of conduct disorder. ¹¹⁵

The mental health professionals involved with child guidance and child psychiatry in the twentieth century in the United States had expectations about normal childhood behaviours, and their diagnostic categories for pathology embedded those assumptions. As Estelle Freedman has pointed out about the sexual psychopath laws from mid-century in the United States, the net effect of the conduct disorder was less about diagnosis and treatment and more about defining the boundaries of normal and abnormal. ¹¹⁶ This power to define normality – and its converse – of course, has been one of the major critiques of psychiatry over the last century. ¹¹⁷ For child psychiatry, these opinions reached far beyond the number of children who were treated within clinic settings. Child psychiatrists have debated for decades how to reconcile the relatively low numbers of specialists in their profession with the high demand for child mental health services. In 2014, the year after the latest edition of the DSM was published, the AACAP published a vision statement and a road map for the future of their field and of child mental health in general. Among other issues, the report called out the shortage of child psychiatrists, and suggested that leaders in the field extend themselves through technology and advising other providers, as well as advocating for child welfare. ¹¹⁸

As child psychiatrists continued to grapple with defining what is normal and what is not, they have also repeatedly wrestled with the question of what to expect for children and adolescents in terms of peer interactions. The criteria that Richard Jenkins insisted on, the presence or absence of ability to conform to some kind of group norms (even if those norms involved law breaking), also illustrated the tensions about how to see children as individuals or as members of a group. Jenkins’s colleagues rejected his observations about social ability for DSM-IIIR and DSM-IV, although socialisation made a comeback for the latest edition of the diagnostic manual, DSM-5, in 2013. The new definition of conduct disorder includes the same descriptive criteria as for DSM-IV, but with the specifier, ‘with limited prosocial emotions’ including lack of remorse or guilt, lack of empathy, unconcern about performance and/or shallow or deficient affect. ¹¹⁹ As child psychiatrists recently noted, perhaps unintentionally echoing much older literature, the degree of a child’s ability to socialise or empathise with others was strongly predictive of future criminal behaviour and might represent a place to intervene to prevent future problems. ¹²⁰

Juvenile delinquency and conduct disorder represent attempts to grapple with the tensions of viewing children as victims or perpetrators. The presence or absence of a diagnosis does not make the determination. Richard Jenkins used diagnosis to classify delinquent boys in a training centre and to divide them into those who could be helped and those who could not. Dorothy Lewis objected to the diagnosis, in part because she said that it was used to dismiss other treatable causes for bad behaviour in kids. The child and adolescent psychiatry organisations and the APA nomenclature committees have tried to have it both ways – to create broad enough criteria to diagnose anyone for any reason, but also to insist that the diagnosis is agnostic with regard to the cause of behaviours.

The history of the construction and use of conduct disorder, though, illustrates that the tension between victim and perpetrator is about more than how psychiatrists see the kids and more how and where children and adolescents are seen or defined by mental health experts within the broader social, cultural and economic landscape of the United States. As has been the case for the entirety of the history of juvenile delinquency, race, class and gender are strong determinants of what happens to kids who behave badly, and even whether their behaviour is identified as a psychiatric illness or as a problem to be managed by the courts. The diagnosis of conduct disorder provided a language for psychiatrists to describe children and adolescents for whom they claimed expert opinion. With regard to conduct disorder, child psychiatry as a profession may or may not have anything to offer, but it does have something to say.