2 Michael Bruch I have no data yet. Sherlock Holmes This chapter will provide a description of the current status of the model as developed at the University College London (UCL). The current design draws predominantly on the work of Ira Turkat and David Lane, who were both associated with the Cognitive–Behavioural Psychotherapy unit at UCL and have made strong contributions to the model in recent years. Unlike in psychiatry, where the interview is mainly guided by a categorical classification system (e.g. DSM; see previous chapter), case formulation is an experimental, hypothesis-driven procedure in pursuit of an explanatory model that subsequently assumes a guiding role for the ongoing therapeutic process. In addition to explaining acquisition and maintenance of a presented disorder, the problem formulation facilitates intervention hypotheses, leading to individually tailored treatment programmes. As the initial interview is of central importance in this procedure, special attention shall be devoted to this aspect. Case formulation is about developing a ‘clinical theory’ of an individual problem under investigation. This is done within the cognitive–behavioural framework. To achieve this, relevant information has to be obtained and suitable hypotheses developed and tested. As Vic Meyer would put it, the therapist acts like a Sherlock Holmes type of detective: Guided by cognitive–behavioural knowledge as well as personal life experience, he is observant and vigilant and picks up clues from appearance, behaviour, language and so on. Out of this process, hypotheses about the aetiology and nature of a problem can be generated, which in turn would inspire questions for the purpose of testing such hypotheses. Such subjective experimental procedure appears appropriate if we accept the reality of individual differences, especially with complex and deeply rooted problems. Moreover, important decisions and interventions have to be made that involve long-term consequences for the client. Furthermore, it is expected that the quality and outcome of the interview will be crucial for building motivation and the patient’s insight into the problem as well as being instrumental in establishing the beginnings of a constructive therapeutic relationship. Let us consider this model in some more detail. The case formulation procedure is conducted in a dynamic and deterministic manner. It rests on the assumption that disordered behaviours can be understood according to established cognitive–behavioural knowledge when assisted by experimental principles of investigation. In this approach, the initial interview is of pre-eminent importance as we set out to achieve a plausible framework, the problem formulation, to account for the behaviour in question in terms of its causal history and maintaining factors. Such clinical theory is also expected to enable us to make predictions for target behaviours in specified situations and is expected to facilitate intervention hypotheses. Another important aspect are individual differences in problem behaviours. Pioneering clinicians have always acknowledged this fact. Meyer’s (e.g. 1957) original groundbreaking work focused on the individualized analysis of complex cases in the psychiatric setting as it became obvious that psychiatric diagnosis was not facilitating psychological treatment of individual patients. On the level of labelling, presenting complaints might appear similar; however, individual analysis often reveals discrepancies in terms of causal histories and mechanism of the disorder, thus suggesting different problem formulations. The case of social phobia may serve as an example: The underlying mechanism may relate, e.g. to either lack of social skills or fear of negative evaluation: each would suggest different treatment priorities and sequencing of treatment strategies. Finally, the narrowness of classical and operant conditioning models merely focusing on functional analysis of presenting complaints has proved insufficient, especially for complex clinical cases. We consider it useful to keep an open mind concerning other knowledge concepts, both from within and outside the learning framework. Employing innovative and creative practice, we are keen to extend our scope and to improve our model for understanding, in order to match the complexity of human behaviour and abnormal behaviour in particular. In case formulation, we attempt to utilize any relevant knowledge from any discipline that can further the application of operationalized methods to effect behavioural changes. However, such knowledge concepts should be understood and applied according to the learning frame of reference involving experimental methodology. This is necessary to avoid conceptual confusion, which in our opinion would undermine the integration of ‘new’ ideas. As clinicians, we consequently intend to do this in view of clinical applications to the individual case. For example, Meyer (1970) has suggested such diverse fields as social and cognitive experimental psychology, physiology, anatomy, neurophysiology, sociology, pharmacology, and even electronics as suitable sources to broaden our methodology. The UCL case formulation model was originally pioneered by Meyer (1957). This was already outlined in the last chapter. Meyer preferred to apply established learning principles directly to clinical settings with psychiatric inpatients. He felt that learning theories were often too speculative and lacked empirical support. To facilitate this, the distinction between learning principles and learning theories seemed important to Meyer: Principles referred to observable phenomena that were experimentally verified but were not indicative of any theoretical explanation. Theories, on the other hand, were considered as highly speculative (especially in the 1950s) and thus neither valid nor useful for experimental assessments and the development of individually tailored treatment programmes. This issue has been critically discussed elsewhere (Meyer, 1975). More recently, Turkat and Maisto (1985) have proposed case formulation as a ‘scientific approach to the clinical case’. They emphasize both the process and the outcomes of science, the former referring to the experimental method requiring hypothesis generation and testing and the latter producing a method to modify the problem behaviour under investigation. This dual-aspect approach seems important, as it emphasizes the practical implications of a problem formulation, unlike psychiatric diagnosis that has rather limited relevance in guiding the therapeutic process. The outcome achieved by this type of formulation process has been clinically demonstrated without any doubt. The fact that this approach began as a clinical effort using single cases highlights individual differences regarding the mechanism of presented problems. In more detail, clinical–experimental analyses of seemingly similar complaints often reveal significant differences in structure, predisposing factors and learning histories. For example, one can usually detect complex interactions between presenting complaints and earlier learnt underlying maladaptive behaviours (e.g. schemas; Bruch, 1988). As the development of a problem in the biography of an individual can be highly idiographic, we suggest that assessment strategies should go beyond a description of the presenting complaint (as attempted in psychiatric classification) and subsequent treatment strategies should be designed accordingly. Such thinking was hardly fashionable in the early days of behaviour therapy, when clinical efforts were mostly directed at ‘symptoms’. The main purpose of the interview process is to collect relevant data, integrate these into an operationalized description for further expert analysis and eventually formulate a model for explanation and prediction covering all presented problems. This resulting ‘clinical theory’ will be provisional and subject to ongoing experimental scrutiny. Wolpe and Turkat (1985) have listed the critical questions to be answered as follows: The style of interviewing is empathic but can also be directive when required. We do not regard direct questioning as ‘unethical manipulation’ as is sometimes suggested by dynamic or humanistic psychotherapists. With these approaches, to be directive is taboo and a scientific discussion of their ‘golden rule’ is not encouraged. Turkat (1986), on the other hand, has argued that, if logic prevails, the whole therapeutic procedure should follow directive principles: after all, it is the client who is looking for direction in order to cope better with his life problems. He further advances his case as follows: The directive style of interviewing as advocated here has often been the target of criticism. Such objections stem from theoretical notions that seem to have little scientific basis. For example, the author has seen many beginning therapists berated by their supervisors for being ‘too directive’. Often, they are accused of being non-empathic and ‘threatening the fragility’ of the patient. These assertions deserve some comment. First, one cannot be empathic if one does not understand the specifics of a problem. It would seem that the most efficient way to understand is to ask direct questions. Second, the notion of a patient to be ‘too fragile’ to handle direct inquiry has little scientific support. Descriptively speaking, the patient encounters daily the problems he or she is purported to be ‘too fragile’ to discuss. The author finds that directness and openness seem highly valued by most patients. (p. 127) At the start of the interview, a great amount of time is spent explaining the rationale and purpose to the patient. This involves clarification of his understanding and expectations of therapy, the procedure of cognitive–behavioural psychotherapy, the collaborative (team) approach and his participation in what is described as an ‘active therapy’. Unlike in the classical psychiatric ward round, the patient becomes very much an active member of the group of professionals, who is invited to comment and contribute with his own questions or request further clarification and so on. A large white board is used to record all relevant information for all participants to refer to as much as possible. When teaching the initial clinical interview, we prefer small group sessions with the patient and trainee therapist(s) present. In this way, every step is commented on and explained. For example, the supervisor might ask: What is the purpose of a particular question? Is there a relevant hypothesis being pursued? What conclusion can be drawn from any outcome? Does it make sense to the patient? and so on. Typically, the interview will be started by an experienced clinician. From time to time – given the consent of the patient – trainees will be given the opportunity to put forward questions, however, on the condition that they can provide a clear rationale including a supporting hypothesis for the purpose of any question. Meyer likened this procedure with ‘being a detective on a mission’. Above all, it is important to explain all processes and outcomes to the patient in a language that is as jargon free as possible. To facilitate and develop a problem formulation, it is paramount to generate hypotheses of cause and maintenance of the problem behaviour under investigation. These will drive the interview strategy in order to test their validity. This experimental procedure is mainly applied to phases 1–3. Whereas phases 1 and 2 are designed to identify, specify and describe problem behaviours, it is attempted in phase 3 to test and verify the problem formulation with clinical experiments. This shall be elaborated in more detail later on in this chapter. The role of hypothesis generation in the interview is twofold: (1) Any information starting with the first contact should be utilized to develop hypotheses (2) which in turn form the basis for further questions designed to verify or reject adopted hypotheses. However, in a clinical context, experimental rigour, as was originally proposed by Shapiro (see Chapter 1), is normally neither possible nor desirable. As a compromise, Turkat (1985) suggested confirming or eliminating untenable hypotheses by means of systematic questioning. In fact, a similar strategy had already been practised and taught by Victor Meyer since the early 1960s. As such a strategy can be influenced by the clinician’s own personal (life) experience as well as expert knowledge of cognitive–behavioural principles, it is important for the clinician to review the process of hypothesis generation continuously. For further validation, it seems desirable to communicate this information to colleagues, trainees and patients in order to confirm its logical basis and achieve professional consensus. Obviously, this procedure must be described as ‘pseudo-experimental’ as it involves a rational appraisal of information instead of controlled experiments. In consideration of the complexity of lifestyles and the idiosyncratic nature of problems, a high standard of clinical skills, (life) experience and therapist ingenuity are called for. In most cases, the clinician will have to perform a triple role: develop sensible hypotheses with the information provided, define suitable questions and evaluate the outcome of this process, all at the same time. It is important that all information provided by the patient is accounted for. Discrepancies and contradictions need resolving, even if this means discarding otherwise suitable hypotheses and restarting the process. Generating hypotheses begins at the very first contact with the client, i.e. the clinician may even use the initial impression to develop hypotheses as to what the problem may be. Turkat (1986) gives a fitting illustration: The clinician scrutinises the manner in which the patient speaks, such as tone, pitch, style, choice of words and phrases, intensity, latencies between words, sentences, questions, and replies, searching for a clue. The clinician…also scrutinises the patient’s physical presentation such as hairstyle, clothing, posture, motor activity, and so forth. …An example will help to illustrate this point. If a young man in the clinic lobby whose physical presentation includes poorly matched, ill-fitting clothes, unstyled hair, thick-rimmed eyeglasses, and uneasy movements and facial expressions when introduced to the therapist, then a preliminary general hypothesis of a social skills deficit is suggested. Depending on subsequent inquiry, one might hypothesise further certain consequences of this social skills deficit such as loneliness, depression, and so forth. (p. 121) Any item of information provided by the client is treated in this way, which provides guidance throughout the interview. As we progress towards the problem formulation (detailed below), we collect information in a systematic and logical manner. Therapists are encouraged to base questions on reasoned hypotheses and reject questions that do not fit these criteria. This strategy will prevent the collection of too much and, especially, unnecessary information, a problem one often finds with inexperienced therapists who might continue the interview over many sessions. Typically, numerous topics might be covered and in the end a therapist might find it difficult to make sense of the accumulated information. Eventually, the elegance, simplicity and purposefulness of the experimental approach are in danger of being lost as an inexperienced therapist begins to ‘drown’ in a vast amount of data. (A typical question to the supervisor is ‘what shall I do next?’). It is not recommended to jump from topic to topic in a random fashion as this tends to lead to disorientation or even confusion. Such interviewing styles are usually associated with inexperience, uncertainty and lack of knowledge. In these circumstances, close guidance and supervision is needed. By contrast, the discipline of the hypothesis-driven approach provides a sense of orientation and guidance throughout the interview. Thus, it is useful to conduct teaching sessions to demonstrate interviewing in a transparent manner with both trainee therapist and patient present. Hypotheses are put forward and are commented upon and discussed until consensus can be achieved. It seems particularly important to encourage trainees to formulate hypotheses that can be supported by the cognitive–behavioural psychotherapy framework. Furthermore, hypotheses may be derived from sources other than information provided by the patient. All aspects of the immediate environment may be relevant, including a patient’s key relationships with others. Finally, it is important to recognize that hypotheses at all stages of the interview may be wrong and prove to be unsupported by data. To remedy wrong hypotheses, open mindedness and flexibility are called for. For example, highly experienced therapists may develop biases based on preferred models of explanation leading to matching hypotheses. This bias may subsequently lead to selective perception and direction in the interview and may even include shaping of the patient’s verbal behaviour. Clinical illustrations for hypothesis generation and testing can be found with Turkat (1985, 1987/see appendix). A more in-depth discussion regarding generation and testing of hypotheses in the context of the interview is given by Richard Hallam in the next chapter. To enhance the model’s practicality, Lane (1990) has proposed five basic phases (see Chapter 1). This involved more precise definitions and clearer operationalization of the procedure, originally developed for the assessment and treatment of children and adolescents. This design allows for correcting feedback and continuous verification of the process by means of hypothesis generation and testing. An updated version of this stepwise procedure is detailed below in Table 2.1. The first three phases are covered during the initial interview. Table 2.1 Phases in the Case Formulation Process. These steps shall be explained in more detail below. This phase is focussed on the individual narrative of the patient. The purpose is to explore the client’s main problem(s), why he is seeking help at this point in time and what changes he may envisage? A detailed and personally meaningful description of present difficulties is sought to facilitate generation of reasoned hypotheses. We actively encourage clients to use their own ‘language’ and refrain from jargon or interpretations of other professionals acquired during the course of previous assessments or treatments. The patient’s statement normally inspires a host of hypotheses that can be investigated further. Already at this stage, the patient is encouraged to express his views and expectations regarding therapeutic change and outcome (to be reviewed later in the light of the problem formulation). In cases that involve relationship problems (e.g. marital or family issues), it may be appropriate to obtain statements and opinions of other individuals involved. Discrepancies and conflicting information might have important implications regarding both motivation and outcome of therapy. Consider the case of a person with obsessions and compulsions who disrupts his marital life through repeated checking and cleaning rituals. As the situation becomes unbearable, his wife might threaten divorce if he refuses to be treated for his problems. In cases like this, a consensus about treatment motivation and goals has to be found before a modification programme can be envisaged. In other words, we support a process of increasing awareness to achieve a consensus for all individuals who may get involved in the therapeutic process. How do we get started in the initial interview? Routinely, we begin by explaining to the client the purpose of the interview and may discuss briefly various options of approaching behavioural problems (e.g. emphasizing a psychological versus a psychiatric approach; or more specifically, emphasize and explain differences between various psychological approaches). Depending on a client’s expectations and previous knowledge, we may also explain the rationale and practicalities of cognitive–behavioural psychotherapy: stressing active participation of the patient, frequency of sessions, a goal-oriented focus, continuous engagement and learning experience, ‘homework’ assignments, behavioural contracts and so on. If this is a teaching session involving trainee therapists, one introduces each individual and facilitates adaptation to the situation. The patient is assured that he is a volunteer and is only expected to offer information that he feels comfortable with and that he may be able to talk to a therapist in private later on. Finally, the patient is actively encouraged to participate in the interview by asking questions and by making comments or suggestions, etc. Next, we may take some biographical details, e.g. age, sex, marital status, and profession. An account of the way we start focusing on the presenting problem has been provided by Meyer and Turkat (1979): We begin typically by generating a list of all the behavioural difficulties the client is currently experiencing. Each problem is listed in general terms with the aim of generating an exhaustive list. The list of behavioural difficulties serves a variety of purposes such as structuring the clinical interview, specifying the range of problems the individual is experiencing, and, most importantly, providing the therapist with information for generating hypotheses. Preferably, the list of problems and subsequent information is recorded on a blackboard or some other medium which the client and therapist can visually refer to (as the wealth of information to be elicited is usually beyond memory capabilities). Visual inspection of the behaviour problem list often provides clues as to how the presenting complaints may be related and account for one another. If such relationships are discovered, then clinical efficiency is facilitated. For example, with a particular client it may be hypothesised that this person is depressed because he is sexually impotent. Consequently, for clinical expediency, sexual impotence will be examined first. In certain cases, the list of behaviour problems does not facilitate the formulation of such an hypothesis. Therefore, the most incapacitating behaviour difficulty is examined first. In either case, the next step in conducting the initial interview involves a developmental behaviour analysis of each individual problem the client is experiencing (p. 262). In cases where it seems impossible to elicit a clear description of the main presenting complaint, it can be appropriate to ask the client to list all problems in session or as a homework task. Such list should be used to facilitate hypotheses regarding relationships between single complaints and might thus be helpful in detecting the underlying problem mechanism. Turkat (1986) provides an example for this: …assume the following list of problems is generated: (1) Depression, (2) Lack of friends, (3) Excessive hand washing, (4) Inability to leave the house, (5) Difficulty sleeping, (6) Excessive cleaning. The therapist attempts to find an explanatory hypothesis for all of these complaints. A striking hypothesis from the problems listed in this case is a fear of contamination. Such a hypothesis is derived from the following type of thinking: The patient probably washes her hands and cleans her house excessively to prevent possible contamination by dirt, germs, and so forth. Further, she avoids leaving her home in order to prevent exposure to more contaminating stimuli. This results in social isolation, rumination about her predicament at night (which produces sleep onset insomnia) and thus, depression. Other problems predicted from the general mechanism of ‘fear of contamination’ might include: avoidance of touching others, sexual problems, hosing down or vacuuming others when they enter her house(this is not as uncommon as it might sound), preventing others from entering her house, and so forth. (p. 123) This style of reasoning is at variance with some ‘case formulators’ who prefer to focus on psychiatric diagnosis using standardized assessment procedures (e.g. Persons, 2008). In hypothesis-guided case formulation, the use of tests or questionnaire batteries and such like (as performed in search of a diagnosis) is not considered helpful as it would undermine the experimental–investigative rationale; however, it might be useful, when normal interviewing is not possible, to use screening or other hypothesis-inspired measures (e.g. behavioural observation) to facilitate the generation of appropriate hypotheses for further investigation. The impetus for such investigations should always be on the individual! Finally, additional interviewing of partners, family members or even friends might provide useful information for further investigation; however, full consent of the client would be of crucial importance. Phase one concludes with a list of concretely described problem behaviours to be subjected to further analyses. It needs to be stressed that such analyses should not be a typical diagnostic or psychometric procedure but hypothesis-driven individualized assessments for all identified problem(s) answering all questions relevant toward a valid problem formulation. In our experience, most cases will require a functional analysis including triple response system analysis and a problem-focused developmental analysis. Further optional analyses (e.g. schema analysis) as suggested by hypotheses arising from clinical observations and initial interviewing may also be conducted. This analysis was originally suggested by Kanfer and co-workers as a core part of the learning equation model of behavioural analysis (e.g. Kanfer & Phillips, 1970). The principal goal is to investigate antecedent and consequent conditions of the problem behaviour. This analysis is still being recommended as a salient component in cognitive–behavioural investigations. More recently, tripartite response system analysis has been incorporated into this design to allow the systematic study of individual response modalities. This seems appropriate as response systems have been shown to be highly interactive (e.g. Lang, 1979; Turkat, 1979). A graphic outline of the complete model is provided in Figure 2.1. Figure 2.1 Functional analysis. The clinician scrutinizes each presenting complaint according to these criteria. One attempts to identify relevant triggers regarding high (S+) and low (S−) probability. It is also important to determine whether stimulus generalization has occurred, i.e. is the behaviour triggered by a single stimulus or a cluster of related stimuli (which may be hierarchically organized). Regarding organismic variables, it is useful to explore whether biological factors or related predispositions contribute to the expression of the investigated response. For example, an elevated level of habitual arousal has been shown to be of predictive value for anxiety patterns following stress (Lader & Wing, 1966). Most salient in the functional analysis is the examination of response systems as illustrated in Figure 2.1. This follows Lang’s (1971) influential model, which has proposed three related components of behaviour: verbal–cognitive, autonomic–physiological and behavioural–motoric. This conceptualization allows us to study the contents as well as the complex interactions between these responses. For example, it is useful to identify the primary and dominant response mode that may have causal impact on other response systems. Or, as in the case of anticipatory anxiety, we may detect a mutually enhancing interaction between cognitive and autonomic variables (Meyer & Reich, 1978). The study of such interactions can provide guidance for the design of suitable treatment method. Obviously, the importance of therapeutic aims would suggest tackling dominant and primary response systems first. Further, it is of interest to know whether response systems are in a state of synchrony (high correlation) or desynchrony (low correlation) (Rachman & Hodgson, 1974). Such patterns can be highly individual suggesting different treatment options. For example, decoupling synchronous systems can be a successful strategy to reduce anxiety by (cognitive) relabelling of autonomic cues. Also, the three systems should be assessed to determine whether there is a predisposed response mode in individuals as one can frequently detect primary ‘cognitive’ or ‘autonomic’ responders (e.g. Bandura, 1977; Bruch, 1988). Part of this investigation should focus on the sequence of events, i.e. in what way can response systems influence each other? The mode of assessing response systems requires further consideration. We have found that self-report may differ significantly from direct measures in response to stress. For example, we have detected significant discrepancies regarding autonomic–somatic responses. To complicate things further, individuals tend to respond to stress in different modalities. For example, we found in one experiment using two physiological measures that some individuals showed an increase in ‘heart rate’ whereas others rather responded with increases in ‘skin conductance’. This suggests that direct measures may vary independently. At the same time, it was found that subjective ratings of the same parameters showed significant correlations and increases in the expected direction. For clinical purposes, it seems of significance how physiological arousal is perceived, evaluated and labelled by the individual. For example, those who perceive an increase as positive activation would find that this facilitates a stressful task whereas quite the opposite would be true when the perception is negative (Bruch, 1988). To conclude, an analysis investigating response primary and sequences should provide important clues for understanding the problem mechanism. For example, an anxiety response that is initiated and dominated by the verbal–cognitive system may require a different therapeutic focus (e.g. cognitive restructuring) as compared with a strong autonomic reaction (e.g. biofeedback). The arrows in Figure 2.1 indicate possible interactions. Additional relevant parameters of each response system such as intensity, frequency and duration may also be studied to gain deeper understanding of the presenting complaint. This process is assisted by the behaviour analysis matrix (Table 2.2), which recommends additional assessment of cognitive, autonomic and motoric components under antecedent and consequent conditions. These can be either behaviours or environmental events. A fuller discussion can be found in Turkat (1979). Table 2.2 The Behaviour Analysis Matrix. The assessment of consequences in the functional analysis is designed to clarify the operant maintaining factors of the problem behaviour under investigation. It is important to identify whether there are conflicting short-term versus long-term consequences. For example, a social phobic might be able to reduce social anxiety by means of withdrawal and avoidance, which in the long run may to lead to complete isolation and subsequent depression. The balance between short-term and long-term consequences can also shed light on the level of self-control. (Maladaptive self-regulation is operating when immediate reinforcement is preferred to long-term gratification, i.e. reduction of anxiety by means of substance abuse.) Lane (1990) has suggested a variety of additional strategies that can be employed in the exploration phase if suggested by appropriate hypotheses or in the event that the functional analysis is inadequate. Any assessment method outside the traditional cognitive–behavioural methodology should be supported by a hypothesis with a clear rationale and only in pursuit of a clearly defined goal. For example, it may be appropriate to activate a deeply seated and avoided schema by means of emotive techniques (Young, 1990). It is also important to observe that any analysis of behaviours is not treated as an isolated ‘snapshot’. As behaviour is continuous, a functional analysis should be conceptualized as part of a wider loop. Responses may become stimuli for subsequent behaviours and so on. The Behaviour Analysis Matrix (Turkat, 1979) provides a useful conceptualization for such assessments. As already discussed above, we do not recommend standardized, often very detailed and seemingly over-inclusive schemes of analysis (e.g. Kanfer & Phillips, 1970). Consistent with the case formulation model, the selection and focus on problem areas should be determined by hypotheses, as the potential number of investigations can be otherwise limitless. Apart from being time-consuming, this may create an unmanageable and sometimes confusing array of data that may become a distraction when attempting a problem formulation. In the worst case scenario, the purposeful, dynamic flow of the interview may get substituted by an over-concentration on ‘correct’ categorization. It has proved to be more useful to collect additional data at a later stage during assessment and treatment if and when relevant hypotheses suggest so and point to apparent information gaps. The usefulness of a developmental assessment has been the subject of controversy among behaviour therapists and was treated with suspicion by early learning theorists (who preferred a ‘here’ and ‘now’ focus, e.g. Stuart, 1970). They sensed a backsliding towards psychodynamic positions. It has to be emphasized that in case formulation such analysis should not be general but precisely focused on individual complaints. Turkat (1986) has summarized the main arguments for the usefulness of a developmental aetiologic analysis: First, there is sufficient scientific and clinical evidence that one of the best overall predictors of future behaviour is previous behaviour. Second, in order to change a behaviour in a meaningful way, one must know the potential causal and maintaining variables. Finally, one cannot prevent future behavioural problems unless one knows what the etiologic determinants are. In the initial interview, etiologic enquiry usually serves to either identify antecedents and consequences of relevance or to validate predictions from the hypothesised mechanism of disorder. In the former case, etiologic enquiry is used to develop an hypothesis about the mechanism of disorder. In the latter case, the clinician predicts what the history of the presenting problems are. In either case, every behavioural problem is examined from its very first occurrence through all changes in its development to the present. (p. 124) Historical information is usually also the main source for generating hypotheses about predisposing factors like biological vulnerabilities and earlier learned, deeply seated behaviours, etc. Turkat (1985) illustrates this for the case of a ‘socially inept individual’: …he is likely to have had parental models for such behaviour (i.e. vicarious conditioning), promoting a lack of opportunity to acquire appropriate skills (operant conditioning) and, perhaps, traumatic consequences such as social rejection resulting in the present social anxiety (classical conditioning) problem. (p. 30) In case formulation, the developmental analysis attempts to construct a focused timeline for each identified problem behaviour. We start with an assessment of the onset circumstances involving predisposing and precipitating factors that may have contributed to vulnerability for onset of these complaints. Thereafter, all significant further occurrences until the present date are carefully recorded, including any changes in manifestation of identified problems and their maintenance mechanism (antecedent and consequent conditions). Such changes can have implications for the intervention hypothesis (as suggested by the problem formulation; see below). For example, a phobia may be established through classical conditioning but over time increasingly maintained by operant factors such as a protective partner. Possible interactions between different problem areas (if in evidence) are also carefully investigated. This might provide clues as to what the main and central problem might be and thus be useful information for the construction of the problem formulation. Particularly when working with complex cases, we have often found evidence for early established maladaptive schemas being closely related to cognitive–behavioural problems. Schemas were hypothesized to promote behavioural disorders in general. We have conceptualized this mechanism as deficient self-regulation (Bruch, 1988). Hypotheses arising from such observations have inspired an optional schema analysis, which shall be addressed in the next section. Case formulations and related clinical research with complex and difficult problems have evidenced early learned maladaptive schemas that are strongly related to dysfunctional cognitive processing styles and interpersonal behaviours (Bruch, 1988). In more detail, we have found a cluster of excessive negative and pessimistic thinking, distorted attribution, low self-efficacy, social isolation and depression. This cluster of behaviours was shown to be strongly related to self-ideal discrepancy, negative self-image and low self-esteem. We have described such a scenario as deficient self-regulation, usually resulting in deeply seated self-perpetuating vicious cycles creating a complex pattern of multiple, often interrelated, problems (as typically found with personality disorders). In therapy, it is mainly the deficient self-regulation that can undermine all previous therapeutic efforts and achievements. The main underlying mechanisms appear to be maladaptive cognitive styles that prevent the correct processing and appraisal of therapeutic progress and outcomes. As a consequence, the building of positive self-schema structures is prevented. In clinical practice, such complications are not always obvious and are sometimes difficult to explore, as schemas may operate in a non-conscious mode. As well as promoting schema-consistent cognitive biases and behavioural complaints, clients are often unaware of their automatic and continuously operating nature. In the long term, negative schemas may lead to maladaptive ‘personal theories’ that are often responsible for maintenance of these problems and also resistant to treatment. It is also common that sufferers deny or suppress negative schemas because of the painful evidence they provide as well as the feelings of hopelessness and helplessness they may cause (Bruch, 1988; Young, 1990). In the long term, this leads to an avoidant, insecure, anxiety-prone and increasingly depressive lifestyle. Also, clients suffering from this condition tend to lack motivation and adherence in treatment. It is for these reasons that relevant schemas should be investigated very carefully in order to obtain full cooperation of the client. It also helps to explain the rationale of schema focus therapy and pointing out of possible treatment options. According to the self-schema model (Bruch, 1988), we have proposed that the pervasive maladaptive cognitive and behavioural patterns underlying presenting complaints should receive more attention. Obviously, the need for this arises in the context of evidence gained in the initial interview. For example, in analyzing client’s early history, hypotheses regarding basic anxieties and conflicts can be generated. Common basic anxieties include those associated with having to achieve in response to high standards or those associated with becoming independent from overprotective parents, etc. In terms of conflicts, it may be hypothesized that the client is in an approach–avoidance conflict when basic anxieties associated with taking responsibility conflict with the consequences of avoiding responsibility (e.g. parental, social or self-disapproval). A dysfunctional lifestyle often develops in response to such basic anxieties and conflicts. A schema analysis usually aids the explanation of the onset and development of problem behaviours. This is especially true for complex cases, where a lack of traumatic conditioning events renders explanations based on a principle of conditioning inadequate. Furthermore, a conceptualization of problem behaviours that involves self-schemas and lifestyle can have predictive value with regards to ‘high-risk’ situations where exacerbation of problem behaviours or relapse might occur. This analysis is highlighted here as a typical creation arising from the case formulation approach with complex problems – clearly a far cry from symptom–technique matching technology. Finally, we are keen to establish knowledge about the assets and potentials of patients. This prevents us from falling into a one-dimensional understanding of problems and may also provide important clues for involvement of the patient in treatment. For this purpose, we assess areas of positive adjustment with particular emphasis on self-control skills. It may also be of interest to find out whether clients can cope with their problems under certain circumstances or have done so previously and for whatever reasons. This enumeration of procedures for the initial interview cannot be considered complete. There are no limits for inventiveness as long as the general experimental strategy is followed. In principle, we try to keep assessment routines to a minimum and adopt a flexible, hypothesis-led approach. This avoids the collection of unnecessary information, which can muddy the waters when attempting a plausible problem formulation. Although the client is regarded as the principal source of information, we might also consider it appropriate to use other sources, such as direct observation in the natural environment or interviewing partners, friends or family members. Information that is not consistent with prevailing hypotheses during the interview must always be accounted for. This may be due to a wrong hypothesis, insufficient interview technique or even manipulative behaviour of the client. It is recommended not to cling to hypotheses that cannot be supported. Also, the interview procedure should carefully avoid the confirmation of favoured hypotheses. Should there be justifiable doubts about the client’s account, one might consider interviewing other individuals who know the client well, provided full consent has been given. The problem formulation assumes a central role as the locus of data integration from which all further therapeutic steps should logically evolve. Meyer and Turkat (1979) have defined the problem formulation as a clinical theory which… …(1) relates all the client’s complaints to one another, (2) explains why the individual developed these difficulties, and (3) provides predictions concerning the client’s behaviour given any stimulus conditions. (p. 261) Meyer has argued against reductionism in behaviour therapy (Meyer & Liddell, 1977) and has pointed out that functional analysis or other singular assessment tools cannot be regarded as sufficient for a full understanding of behavioural problems – instead a conceptual system is preferred that approaches the whole person in cognitive–behavioural terms. The ultimate questions in cognitive–behavioural analysis that ideally have to be answered are as follows: Why has this individual at a particular point of time acquired this specific presenting complaint? How did the problem develop? Which (underlying) conditions (if any) were instrumental? And what is the functional value of the problem for his life in general? Such an analysis should also clarify any predisposing factors that may explain why a patient tends to behave in a particular way in a given situation or why a specific situation is more influential than another one and so on. By integrating all relevant data gathered in the initial interview, we hope to arrive at such a comprehensive formulation to explain aetiology and maintenance of problem behaviours that should enable us to make predictions for specified situations. Subsequently, it is important to present and discuss the formulation with the client. In so doing, we are not just seeking approval but are trying to make the assessment and conclusions as transparent as possible, thus hoping to motivate the client for active participation in the subsequent treatment programme. However, any criticism or even rejection by the client is also considered as helpful as this may point to lack of information, faulty conclusions or other inconsistencies that might require clarification. Turkat (1986) has recommended a checklist for this procedure: The other important purpose of a problem formulation is to provide guidance for all further therapeutic steps, i.e. to enable the therapist to decide on appropriate treatment strategies, determine priorities and sequencing, develop or select suitable techniques and so on. For example, this should clarify whether the main complaint is isolated or is supported by other underlying problem behaviours acting as independent variables. We should also be able to explain why the individual has been vulnerable to stressful conditions leading to a variety of complaints whereas another might appear immune. We believe that problem formulations based on such multi-level analysis guided by individual clinical hypotheses are best suited for understanding the underlying mechanism of individual problems. Problem formulations should allow predictions of target behaviours and enable intervention hypotheses (see below). It is difficult to see how a diagnosis or psychometric procedure could provide equally relevant information. To conclude, factors that contribute to a valid problem formulation include (a) recognition of individual specificity, (b) multi-level analysis guided by cognitive–behavioural principles and (c) the relation of complaints to underlying conditions (e.g. schemas). The experimental method applied throughout this analysis serves as a guiding and integrating factor. In rare cases where a formulation cannot be achieved (when information is insufficient or unobtainable), one adopts a more pragmatic approach. Obviously, this can be expected to reduce therapeutic effectiveness greatly, especially when complaints are treated in an isolated fashion. To validate the problem formulation, i.e. its explaining and predictive power, suitable behavioural tests and measurements should be employed in relevant clinical settings. Precise hypotheses should be delineated and tested with in vivo experiments set under relevant stimulus conditions. The selection of procedures will depend on client variables, nature of hypotheses as suggested by the formulation and inventiveness of the clinician. In practical terms, the clinician should provide stimulation across all response modalities and select appropriate measures covering all response systems. As outlined earlier, this can be useful for investigating response patterns, e.g. to determine discrepancies between verbal reports and motoric behaviours or to determine the dominant response mode that may react first and have impact on the others. Distorted perception and dysfunctional labelling of physiological arousal is also a common aspect of response system interactions. For example, an actress who perceives physiological activation before or during a performance as negative and debilitating is likely to label the outcome as stage fright. At other times, physiological cues may get exaggerated, which is likely to promote the vicious cycle of anticipatory anxiety (operationalized as enhancing interaction of cognitive and autonomic variables). Another issue that may need to be investigated is a possible discrepancy between verbal reports and actual behaviours as assessed in vivo. For example, this can occur when phobic situations have been avoided for long periods of time. Some other relevant measures might also be appropriate. These could pertain to information provided by other individuals involved, in vivo observations, self-monitoring data or even questionnaire measures, which should be carefully selected to be relevant to the hypothesized problem. Naturally, converging evidence of multiple measures covering several dimensions will provide strong support for the problem formulation. The clinician continues to conduct ‘experiments’ to verify or eliminate untenable hypotheses as new information arises. Finally, the most plausible explanatory formulation is tested for its generality with further clinical experiments. In some cases, it may not be possible to verify a problem formulation. This can be caused by several factors including faulty information, a ‘weak’ formulation, inappropriate hypotheses or unsuitable measures. In some circumstances, reformulation of problem behaviours becomes necessary and clinical experimentation has to be repeated. If this does not solve the problem, additional information sources may be consulted or otherwise a pragmatic approach to treatment is adopted, e.g. to focus on the most salient complaint. Such an approach may also provide new information for a revised formulation. Observations arising from clinical experimentation might also suggest prototypic therapeutic techniques (e.g. in vivo exposure) including measurements for baseline and change during intervention and follow-up. Obviously, rigorous clinical experimentation might not be required as a routine measure, especially with simple problems that are formulated unambiguously and are generally agreed. Unexpected surprises may also occur when clinical experiments assume a treatment effect. For example, a client with a reported travel phobia (resulting from traumatic conditioning) had avoided travel for many years. She reported mild anticipatory anxiety before a clinical approach/exposure experiment. However, once confronted with the fearful situation, spontaneous immediate habituation took place. No further treatment was required. Finally, we like to emphasize that this experimental and hypothesis-guided procedure should not be confused with standardized psychometric test batteries, which we consider as inappropriate in a psychotherapeutic context where the focus should be on individual change processes. Any selected measure should serve a purpose within this rationale. Typical examples for clinical experimentation have been provided by Turkat and Carlson (1984) and Turkat and Maisto (1985). A more detailed discussion of the method can be found in Carey, Flasher, Maisto, and Turkat (1984). A plausible and verified problem formulation explaining the underlying mechanism of the problem(s) should enable the clinician to make reasoned proposals for treatment goals, strategies and prototypical methods of intervention. Regarding goals, we carefully review the patient’s original statement (as obtained in phase one) in the light of the formulation. Any major discrepancies should be fully explored and discussed with the client so that acceptable agreement can be found jointly. Obviously, this process is greatly facilitated when the patient comprehensively understands and accepts the problem formulation. Hopefully, full consent for any modifications, refinements or revisions of the previously stated goals can subsequently be achieved. Only in rare cases of disagreement must the therapist decide whether he can adopt a pragmatic compromise or rather disengage from the treatment process. Understanding the underlying mechanism of a disorder in terms of learning, cognitive or additional theoretical principles is the key source for intervention hypotheses and proposals. Obviously, the therapist is expected to have sound knowledge of cognitive–behavioural methodology based on single case–based and group-based research. Handbooks on techniques and quality manuals might also be useful in providing technical expertise about methods but should be adapted to the individual case. Finally, we find that clinical experiments in problem-relevant settings can be highly instrumental for designing specific treatment methods. For example, the knowledge that a phobic reaction is started with catastrophic thoughts would point to a matching treatment strategy. In addition to clinical experiments, the predictive power of the problem formulation can be further tested using self-report measures that can be subsequently used to establish baselines and to evaluate change during therapy as well as for outcome and follow-up. Suitable measures may range from questionnaire-type to rating scale–type assessments. All measures should be individualized and as meaningful as possible. In case formulation, we reject standardized psychometric test approaches based on trait personality conceptualization as inappropriate as clinicians should predominantly be concerned about evaluating the therapeutic change process rather than a comparison with normative groups. In training and supervision, we do encourage therapists to develop personal questionnaires or rating scales using statements in the personal language of patients. For example, such work led to a “dog phobia questionnaire” (Tuomisto, 1994), which was subsequently used as a prototype for similar problems but required adjustment for each new case. For comprehensive evaluation, we recommend a multidimensional approach: client/therapist (or team); global/specific; short-term/long-term. To state this in more detail, any changes or outcomes ought to be evaluated independently by both patients and therapists; measurements should be directed to both specific problems as well as overall life adjustment and, finally, short-term changes (e.g. response to an exposure session) as well as long-term changes (e.g. rating of severity of problems) should be considered. Furthermore, it is desirable to conduct a long-term follow-up for realistic assessment of therapeutic achievements. Obviously, multidimensional assessments may yield contradictory results that need to be addressed in an overall clinical evaluation. For example, our own data suggest strong pre-treatment agreement between the client and therapist, whereas strong discrepancies tend to be more typical for the follow-up stages (Bruch, 1988). This appears especially true for global long-term measures like life adjustment, indicating the need for greater processing and generalization of treatment achievements. These issues ought to be discussed fully and disagreements need to be reconciled. Lane (1990) has provided some examples for suitable measurement techniques. Phases one to three pertain to the initial interview and are considered to be the most salient and instrumental elements in the case formulation procedure. The intervention and evaluation phases in case formulation are in principle not distinct from best standard cognitive–behavioural psychotherapy practice. However, apart from formulation-guided design or selection of a treatment model, there will be a strong emphasis on the individual tailoring of all applied methods, be it adjustment of established techniques or the creation of innovative methods. Another important element in case formulation is close monitoring of the efficacy of any applied method in order to be able to provide corrective feedback to the formulation phase in case the programme does not work as expected. This may require revision of earlier phases where mistakes were possibly made or important information was overlooked. Formulation-guided intervention is designed to produce general options for the modification of problem behaviours. This would include all response systems as outlined earlier. Any chosen procedures should be specified in detail and explained and discussed with the client. For example, is it clear which (and how) behaviours are to be increased, decreased or instilled? Have the techniques to be used been clearly explained to the participants? (in brief: Who, What, When, Where and How?). Does the treatment procedure follow the logic of the formulation and subsequent discussion with the client and is it consistent with treatment objectives? This discussion can be followed by an intervention contract that provides an agreed description of the programme including the therapist’s role and the tasks to be carried by the client, etc. Finally, the agreed programme is enacted and monitored in a structured way. This should involve a suitable environment and active participation of the client according to the treatment contract. The course of treatment should be designed as a continuous learning process. There is no good reason why therapy should take place once a week in a consulting room. This means that, after the initial interview and formulation, scheduling of sessions should be designed to foster continuous learning with increasing engagement of the client and therapist-led session should increasingly focus on review and corrective feedback. Generally, from the start of therapy, we encourage self-regulation processes in the application of treatment methods. The intervals between sessions should be flexibly adjusted to therapeutic progress. Typically, the frequency of sessions should be reduced gradually when self-control of the client and subsequent consolidation as well as generalization of treatment effects improve. How and why should accomplished outcomes be monitored and evaluated? We have already suggested some suitable dimensions for measurement. It seems equally important to employ a broad range of modalities, e.g. involving clinical judgement, behavioural observation, diaries (thoughts, feelings, behaviours), self-ratings and questionnaires. Such a range of measures might be helpful in detecting inconsistencies in outcome evaluation, whereas strongly converging measures across all modalities might provide convincing evidence for accomplished outcomes. If the programme has worked well, the ultimate confirmation of the problem formulation can be concluded and one can proceed to the follow-up phase. If success is only partial, we recommend a review of the formulation to detect possible errors, omissions, etc. If the programme proves a complete failure, the initial definition of problem(s) and treatment targets should be looked at. Apart from the academic and clinical interest of the ‘scientist–practitioner’, we also advocate substantial involvement of the client. The interface between active treatment and follow-up should seek to refine and consolidate achievements, enhance and generalize treatment gains and guide/supervise a client’s self-regulation processes. A number of questions might help the therapist to optimize the intervention process further: Have new objectives arisen that require clarification and action? Have the implications of the ‘new behaviours’ for all involved been considered? Has a programme to maintain gains and/or meet the new objectives been introduced? Has a post-intervention contract been established so that each participant is aware of his/her role in maintaining gains and taking necessary action in the event of difficulties or further new objectives arising? For the follow-up, we recommend continuing review of both therapy goals and measurement. To provide continuous feedback and enhance processing of therapeutic experiences, patients should be encouraged to keep behavioural diaries to strengthen mindfulness and resourcefulness of newly acquired behaviours, such as coping with anxiety or activation of alternative behaviours when depressed. Finally, regarding follow-up, we recommend an open end that should be determined by the patient as judged by his progress and help needed. Short and unsupported follow-up has shown in clinical practice to be rather counterproductive for enhancement and consolidation processes, whereas open-ended arrangements appear to instil confidence and paradoxically are rarely needed by a client over a prolonged period. At the end of follow-up, a number of further questions may be asked: Curiously, in the early 60s, when behaviour therapy became more technique orientated, little attention was paid to relationship issues in the therapeutic process. This was perhaps appropriate when dealing with rather simple monosymptomatic problems where the focus was mostly directed at the diagnosis and a prescribed technique. Another explanation might point to an overreaction to a perceived indulgence of psychodynamic therapies with relationship issues. However, with growing interest in complex and difficult cases in the last two decades, often involving low motivation and poor adherence as well as deficient self-regulation, it became increasingly obvious that a formulation-guided therapeutic relationship should be given more attention. Clearly, the virtues of a positive therapeutic relationship hardly need to be recommended; however, we would like to emphasize some issues that appear particularly important in the context of case formulation. In principal, we conceptualize the therapeutic relationship as an integral part of the whole treatment process and we recommend that this aspect should not be treated in an isolated manner. Thus, emphasis and direction of such relationships should be guided by the individual problem formulation and subsequent treatment requirements. Turkat and co-workers (e.g. Turkat & Brantley, 1981; Turkat & Meyer, 1982) have argued against standard relationship enhancement techniques according to Rogerian principles (i.e. to indicate understanding and to demonstrate empathy), which appear to be endorsed by most clinicians (e.g. Goldfried, 1982). They may be incompatible with individual problem formulations as Wolpe and Turkat (1985) have illustrated previously: “Unquestionably, one must be able to empathise with the patient if one is to be able to formulate the case. However, the question as to what demonstrable empathy is remains the basis of difference. We would argue that accurate empathy is demonstrated when the therapist can accurately predict the patient’s behaviour. This difference in demonstrating empathy can be seen in the following example: patient:
The UCL Case Formulation Model
Clinical Process and Procedures
It is a capital mistake to theorize before one has data.
Insensibly one begins to twist facts to suit theories,
instead of theories to suit facts.
The Clinical Purpose
Foundations and Assumptions
The Initial Interview
Developing Hypotheses
Practical Steps
Phase One: Definition of Problems
1 A narrative of the problem(s) from those involved is obtained.
2 Problems are described and specified on the basis of initial information provided by the client.
3 Initial objectives of the client are clarified.
Theme: A process of growing awareness aimed at a therapeutic consensus.
Phase Two: Exploration
4 Defined problems are conceptualized in cognitive–behavioural terms.
5 Multi-level cognitive–behavioural assessment is conducted.
6 Data are collected to test hypotheses.
Theme: The process is one of increasingly refined expert assessments.
Phase Three: Formulation
7 A problem formulation and intervention hypotheses are established.
8 Explanation and discussion with participants and redefinition of therapeutic objectives.
9 Clinical experiments to check and verify the problem formulation.
Theme: The process is one of summarizing and integrating gathered and tested information until an adequate explanation is available.
Phase Four: Intervention
10 The procedures to be used are selected and specified.
11 An intervention contract is established.
12 The agreed programme is enacted and monitored.
Theme: The process is one of structured practice.
Phase Five: Evaluation
13 Accomplished outcomes are evaluated.
14 Any gains made are supported and enhanced, the programme is optimized and further objectives, if suggested, are pursued.
15 Continuing evaluation and review. Generation of further ideas to consolidate progress.
Theme: The process is one of monitored achievement and support.
Defining problems (Phase 1)
Exploring problems (Phase 2)
Functional analysis
Antecedent
Behaviour
Consequence
Cognitive
X
X
X
Autonomic
X
X
X
Motoric
X
X
X
Environmental
X
X
Developmental analysis
Schema analysis
Additional issues
The problem formulation (Phase 3)
Clinical experimentation
Intervention hypothesis
Baselines and measures of change
Intervention (Phase 4)
Evaluation (Phase 5)
The therapeutic relationship
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
The UCL Case Formulation Model
I get very nervous when I leave the house by myself, I just feel as if I were going to pass out.
