Empirical approaches to family assessment.

Empirical approaches to family assessment Carr, A. (2000). Empirical approaches to family assessment. Journal of Family Therapy, 22, 121-127. 1 Empirical approaches to family assessment EDITORIAL EMPIRICAL APPROACHES TO FAMILY ASSESSMENT The aim of this special issue of the Journal of Family Therapy is to familiarise clinicians with important empirical approaches to family assessment. This volume contains five invited papers by originators of major empirically based models of family assessment from both sides of the Atlantic. These models are: • The Beavers Family Systems model (Beavers & Hampson) • The Circumplex model (Olson) • The McMaster Model (Miller et al ) • The Family Process model (Skinner et al ) • The Darlington Family Assessment System (Wilkinson). Authors of these papers were invited to outline the conceptual framework or model of family functioning on which their assessment instruments are based and then to describe these self-report questionnaires and observational rating scales. They were also asked to give a summary of the main research findings arising from studies in which their family assessment instruments have been used and to outline the implications of their work for clinical practice. By way of introduction to these papers, a number comments will be made below on the following issues with reference to the five models: • The number of dimensions for conceptualising family functioning • The degree to which different models focus on subsystems • The focus on family strengths • Characteristics of the empirically based assessment instruments • Findings from process and outcome research based on the models 2 Empirical approaches to family assessment • Clinical utility of the models • Technical challenges for further development of the models and • Structural models as social constructions. Number of dimensions The five models hold in common the view that for clinicians and researchers it is useful to conceptualise families or their subsystems as differing along a limited number of dimensions. However, they differ in the number of dimensions that they contain. On the one hand the Beavers and Circumplex models confine themselves to two or three dimensions while the McMaster, Family Process and Darlington models contain more than twice this number. There is a balance to be sought between parsimony and oversimplification in selecting the number of dimensions to include in a clinically useful model of family functioning. What is noteworthy is that the models with more dimensions separate out constructs that are lumped together in models with fewer dimensions. For example, roles is an aspect of the flexibility dimension of the Circumplex model but within the McMaster and Family process models, roles is an entire dimension unto itself. Focus on subsystems The models differ in the degree to which they allow clinicians and researchers to conceptualise the status of family subsystems. The Darlington Family Assessment System explicitly distinguishes between dimensions which apply to the child, parental, and parent-child subsystems. Within the other four models less clear provision is made for conceptualising the functioning of specific subsystems. However, for the Circumplex model and the Family Process Model special versions of the self-report assessment instruments have been developed for evaluating different subsystems. For example, for the Family Process Model, versions of the FAM III have been developed which permit 3 Empirical approaches to family assessment 4 the assessment of family members perception if the whole family, specific dyadic relationships within the family and individuals perception of their specific family roles. Focus on strengths A focus on strengths and competence is central to all five models. Some or all of the dimensions of family functioning in each of the models are conceptualised in terms of family resources and competencies rather than deficits and problems. For example, the central dimension of the Beavers model is labelled competence; those of the Circumplex model are flexibility, cohesion and communication; the McMaster model includes dimensions for problem-solving, communication, roles, affective responsiveness, affective expression and behaviour control. However, all five models are clear in highlighting that an absence of competencies and strengths as defined by each model's dimensions is indicative of dysfunction. The models are less influenced by a deficit discourse than traditional mainstream mental health conceptualisations of family problems. For example, research on highly functional families was conducted using the Beavers model as a conceptual framework and Olson's group used the Circumplex model as the basis for a major study of normal family functioning across the lifecycle. Characteristics of assessment instruments All 5 of the models have led to the development of assessment instruments. These have included self-report questionnaires for completion by family members or observational rating scales for completion by clinicians who have interviewed or observed a family. The psychometric characteristics, especially the reliability and validity, of the instruments described in this volume have been evaluated in numerous studies. These have led to a range of important findings, three of which deserve particular mention. First, the rating scales for four of the models (Beavers, Circumplex, McMaster, Empirical approaches to family assessment 5 Darlington) have been shown to be reliable insofar as pairs of raters tend to rate the same family similarly. Second, the self-report instruments for four of the models (Beavers, Circumplex, McMaster and Family Process) are reliable, insofar as there tends to be a high level of consistency in the way that the same family members respond to the differing items in the same scale. Third, for each model either the rating scales or self-report instruments or both have been shown to be valid, insofar as extremely deviant scores are associated with poorer adjustment of family members. Process and outcome research From a research perspective, a number of the models, notably the Beavers and McMaster models have spawned research programmes into process and outcome in marital and family therapy. The incorporation of measures of dimensions of family functioning into these studies has thrown light on aspects of family functioning that change during treatment and also on aspects of family functioning that should be taken into account when planning effective treatment, since not all families respond well to the same type of therapeutic style or approach. For example Beavers and Hampson found that families rated as more competent and families who were characterised by a centripetal style fared best when their therapists were more open about their therapeutic strategy, more egalitarian in the power differential they established with their clients, and more joined in partnership with families within the therapeutic alliance. Families rated as more dysfunctional and more centrifugal in their style made greater therapeutic progress when their therapists were less open about their therapeutic strategy, and established a more hierarchical therapeutic relationship characterised by interpersonal distance and directiveness. These results suggest that different therapeutic styles are appropriate for differing types of families as defined by the Beavers Model of Family Functioning. Empirical approaches to family assessment 6 The work of the McMaster group provides another example of clinically useful findings from treatment related research based their model of family functioning. The McMaster group have shown that families of depressed patients report severe levels of family dysfunction on the Family Assessment Device during the depressive episode; that the level and type of family dysfunction is associated with an increased risk for suicide; that impairments in family functioning, while improving somewhat, still remain even after remission of the acute depressive episode; and that impaired family functioning is associated with a longer time to recovery and increased risk for relapse. Clinical utility From a clinical perspective, assessment instruments from all five of the models have been used by clinicians in routine practice to evaluate family strengths and weaknesses, to plan treatment, to monitor progress and to assess outcome. A particularly interesting example of changes occurring over the course of treatment on dimensions of the Circumplex model for a family in which the father had a heart attack is given in David Olson's paper. In the current climate where increasing emphasis is being placed on auditing clinical services, demonstrating treatment effectiveness and evidence based practice, it may be useful for clinicians to incorporate assessment instruments described in this volume into routine clinical practice to monitor the impact of family therapy in their clinics. Technical challenges for further development of the models Each of the models reviewed has its shortcomings and technical challenges which need to be addressed in systematic programmes of empirical research. What follows are some illustrative examples of such challenges, not an exhaustive inventory. With the Beaver's Family Systems Model, it is clear that the family competence dimension is strongly related to family members adjustment but further validation of the family style Empirical approaches to family assessment 7 dimension is required. With the Circumplex model, there is a need to integrate the original 'flat' two dimensional model (which is assessed by the curvilinear subscales of the Clinical Rating Scale) with the new three dimensional model which is assessed by the linear subscales of the Circumplex model's self-report instrument, the FACES. For the McMaster Model, it may be useful to develop instruments for assessing specific family subsystems such as the parenting or the marital subsystems. For the Process Model of Family Functioning, validation studies of the clinical rating scale have not yet been conducted and are required to complement the self-report instrument for this model, the FAM III. For the Darlington Family Assessment System, the initial validation study requires replication and extension across a number of different clinical and nonclinical populations. For all of the models described in this volume, studies which examine the relationship between scores on assessment instruments and progress through different styles of family therapy are required to furnish clinicians with useful information on the clinical implications of assessment instrument profiles. Structural models as social constructions All five of the models described in this volume fall within the structural-functionalist tradition of social and psychological studies of the family. Within the family therapy field, all five models hold more in common with structural family therapy than with other more popular second order and narrative approaches to clinical practice. In structural family therapy clinicians map out dysfunctional family structures based on data gathered in clinical interviews and then focus on reorganising the structure of the family so that it conforms to a normative conceptualisation of a functional or well structured healthy family. The five models considered in the volume implicitly or explicitly entail a similar view of therapy. Family structures may be mapped and classified as functional or dysfunctional and then modified through therapy. Furthermore, the positioning of the therapist within such structures is not considered. The therapist is assumed to be Empirical approaches to family assessment 8 outside of the family structure. That is the models of family assessment in this volume are more consistent with a first-order rather than a second order approach to family therapy. A question raised by these observations is the compatibility of the models described in this volume with a social-constructionist therapeutic positioning. This central question may be broken down into two sub-questions. The first concerns the knowledge claims or truth status of the general models themselves. The second concerns how the results of family assessments in specific cases may be incorporated into third-order practice: practice that reaps the benefits of narratives from multiple sources including social science and concurrently does not use such knowledge to marginalize family members experiences in the way that dominant discourses and global knowledges have been shown to marginalize subjugated discourses and local knowledges (Carr, In Press). In response to the first question, I think the most coherent position to adopt is to view the models of family functioning presented in this volume as social constructions. One or other of these socially constructed models may be more or less useful for solving particular problems. The usefulness of these different socially constructed models may be tested through rigorous research. For example, in the sixth paper presented in this volume we compared the capacity of the rating scales from three of the models to distinguish between families with children who had various types of problems and families with children without clinically significant problems. We found that the Beavers and McMaster models showed particularly high levels of sensitivity in detecting clinical cases, whereas the Circumplex rating scale was particularly good at classifying non-clinical cases accurately. In short, different models were good for solving different problems In response to the second question, it may be useful to consider the scores from assessment instruments and the implications of these, not as global knowledges but Empirical approaches to family assessment 9 specialised local knowledges arising from conversations among rigorous researchers. This knowledge may be introduced into therapeutic conversations not as "the truth" but as insights that have been found to hold up in many similar situations and so may be useful for the clients to take into account. For example, families who obtain deviant scores on the problem-solving and communication subscales of the Family Assessment Device, may find it useful to know that coaching in problem-solving and communication skills may reduce their sense of distress and help them to solve their presenting problems more efficiently. However, these deviant scores, do not mean that "the truth is they are essentially poor communicators or problem solvers". Rather these scores mean that compared to other families who participated in the research project in which the scales' norms were established, the clients are currently describing themselves as having more difficulties with communication and problem solving and so may find it useful to focus their energies on refining their skills in these areas. A commitment to social-constructionism as an overarching framework for practice does not preclude a commitment to quantitative research grounded in empirical models of family functioning such as those described in this volume. These models have been developed by communities of scientists and clinicians in conversation and have been found to fit with the needs and concerns of these scientists and clinicians. That is, they have been found to be useful for solving problems, the hallmark of a valid social construction (Carr, 1995, 1997, 1999). These models and the results of research based upon them are found to be useful because they let us know how to bring the collective wisdom of a community of scientists and clinicians to bear on the problems of any specific family which is referred to a specific clinician for treatment. There is an ethical imperative for clinicians not to ignore and where appropriate to use this collective wisdom to guide their treatment of clients. Alan Carr Empirical approaches to family assessment Editor REFERENCES Carr, A. (1995). Positive Practice: A Step-by-Step Approach to Family Therapy. Reading, UK: Harwood. Carr, A. (1997). Family Therapy and Systemic Consultation. Lanham, MD: University Press of America. Carr, A. (1999). Handbook of Child and Adolescent Clinical Psychology: A Contextual Approach. London: Routledge. Carr, A. (In Press). Family Therapy: Concepts, Process and Practice. Chichester: Wiley. ______________ 10