A Review of Treatment Decision-Making Models and Factors in Mental Health Practice

Rational Choice Decision Making

EBP assumes that practitioners will evaluate, sort, and prioritize relevant information (e.g., research evidence, assessment data) and then use this information to optimize their decisions and practices (Webb, 2001). As such, one’s understanding of EBP is grounded in presuppositions about decision making. Writing from a medical perspective, Spring (2008) has described decision making as the “lynchpin” of EBP. Given the centrality of decision making to the EBP process, decision-making theories can provide a useful framework for investigating the clinical decision-making processes of mental health professionals (Baker-Ericzen et al., 2015). Historically, EBP has been most directly associated with the classical or rational choice decision-making tradition (van de Luitgaarden, 2009; Webb, 2001; White & Stancombe, 2003). As envisioned by rational choice theory, a decision is rational if it meets four criteria: (a) based on the decision maker’s current assets (e.g., resources, physiological state, psychological capacities), (b) based on possible consequences of the choice, (c) based on evaluated likelihood (as governed by probability theory) when consequences are uncertain, and (d) adaptive within the constraints of probabilities and the values associated with each possible consequence of the choice (Hastie & Dawes, 2010). As implied by these criteria, rational decision-making theories are focused on decision events—the point at which a decision maker surveys a known and fixed set of alternatives, weighs the likely consequences of choosing each, and makes a choice (Orasanu & Connolly, 1993).

One can readily recognize the implicit connection of rational decision-making principles to the EBP process; van de Luitgaarden (2009) has gone so far as to describe EBP as “an operationalization of a rational choice approach to judgment and decision making” (p. 244). Conceptualized thusly, the clinician—facing Client Problem A—draws upon the best available evidence to consider the relative probability that Intervention B compared with Intervention C or Intervention D will lead to desired Outcome E. In effect, the clinician is assumed to operate in a quantitatively oriented, statistically minded mode of clinical decision making (van de Luitgaarden, 2009). The degree to which clinicians’ decision making actually reflects this notional process of rational choice has been challenged in the scholarly literature. For example, Webb (2001) has argued that the evidence-based model assumes that social workers act as “information processors” in closed systems for decision making but they actually live in constantly changing systems (e.g., legal, organizational) and often revert to their practice wisdom, values or ethics, and personal preferences/comforts in decision making. Proponents of the EBP model have countered such arguments by noting that EBP arose from the recognition that “professionals are not rational agents and that in spite of intentions of professionals to provide competent, ethical services informed by practice-related research, they do not do so” (Gibbs & Gambrill, 2002, p. 463). Because professionals are not guaranteed to make the rational choice, the EBP framework acknowledges the uncertainty and bias that might affect practitioner decision making and thus encourages rigorous evaluation criteria when appraising the available evidence (Gibbs & Gambrill, 2002).

Both sides of the debate acknowledge that, despite the similarities and shared intellectual lineage, EBP is not dependent upon a rational choice conceptualization of decision making. The question thus arises, “Can the basic propositions and processes of EBP be supported by an alternative understanding of practitioner decision making?” The theory of naturalistic decision making (NDM) has been proposed as a viable alternative to the rational choice theory for conceptualizing, implementing, and evaluating EBP (Baker-Ericzen et al., 2015; Falzer, 2004).

NDM

The NDM approach emerged in response to the realization that “researchers were not likely to find out how people actually made decisions by conducting experiments to test hypotheses derived from statistical and mathematical models of ideal choice strategies” (Klein, 2008, p. 456). In contrast to the classical rational decision-making researchers, NDM researchers were not only interested in the decision event itself but also in the human decision maker within the natural decision-making setting (Lipshitz et al., 2001; Orasanu & Connolly, 1993). They sought to explore the strategies that individuals used to make difficult decisions in complex situations.

Early NDM research provided compelling evidence that individuals, when faced with decisions in “real world” settings, were not identifying and comparing a set of options but rather were drawing on an experience-based schema to rapidly assess, categorize, and select a course of action (Klein, 2008). In 1997, Zsambok offered the following concise definition of NDM: “the way people use their experience to make decisions in field settings” (p. 7). As opposed to decisions made within the laboratory conditions traditionally favored by rational decision-making researchers, decisions made in a naturalistic setting are complicated by a number of factors including ill-structured problems, dynamic environments, shifting or competing goals, action/feedback loops, time stress, high stakes, multiple players, and organizational goals and norms (Orasanu & Connolly, 1993). As such, field settings “establish the eliciting conditions for making decisions and shape decisions through their constraints and affordances” (Lipshitz et al., 2001, p. 334). The human (and thus rationally bounded) decision maker is of central importance in the NDM framework. The ideal, omniscient, logically consistent decision maker posited by classical decision theory has little relevance in the real world (Beach & Lipshitz, 1993). By contrast, NDM presupposes that decision makers are shaped by prior experience and content knowledge and seeks to identify those who are proficient and demonstrate expertise (Lipshitz et al., 2001). Within the NDM framework, the quality of the decision-making process is judged not by its procedural rationality or logical consistency but rather by the quality of the decisions produced by the process (Bordley, 2001).

Central Research Question

In 1997, Witteman and Kunst observed that the process of treatment planning had not yet been adequately analyzed and no formal models of treatment decision making had been developed. The slow growth of academic knowledge in this area is evidenced by the declaration 10 years later that clinician decision-making research was “in its infancy” (Schottenbauer et al., 2007, p. 225) and again more than a decade after that (Gutierrez et al., 2018, p. 95). Nevertheless, some progress has been realized. Although a comprehensive treatment decision-making model is still lacking, a body of research that spans national and disciplinary borders has identified a number of factors that influence decision making. The following narrative review will explore studies relevant to decision making and will be separated based on characteristics of the therapist, and characteristics of the client. The question explored in this narrative review was as follows:

Decision Making Based on Characteristics of the Therapist

Cook and colleagues (2009) conducted an internet-based survey of 2,607 psychotherapists to identify the factors with the greatest influence on practice behaviors. Respondents indicated that decisions to adopt a new treatment approach were most heavily influenced by ease of integration with existing practice, endorsement by respected therapists, and accessible training opportunities. By contrast, endorsement by a professional organization as being evidence-based, and client testimonials regarding effectiveness were the least influential factors. Witteman and Kunst (1997) conducted a process-tracing study to explore how psychotherapists make treatment decisions for a depressed patient. Eleven clinical psychologists of different therapeutic orientations were presented with a case study of a depressed client and asked to think aloud how they would construct a treatment plan. All participants offered an interpretation of the case, proposed a treatment, and then selectively identified information to confirm this treatment option without considering alternatives, or potentially disconfirming information. Researchers posited that participants were driven by applying methods they already employ instead of developing a new approach. When researchers shared their findings with study participants and other practicing clinicians, therapists agreed that their treatment decisions were often quite unstructured and subjective. Witteman and Koele (1999) expanded on Witteman and Kunst’s (1997) exploratory findings by investigating the explicit and implicit explanations of clinical treatment recommendations among a larger sample of 56 registered psychotherapists. The participating psychotherapists were asked to read four case descriptions and answer questions regarding their proposed course of treatment. Findings indicated the best predictor of treatment decisions was the therapist’s theoretical background (i.e., psychodynamic) suggesting that treatment decisions were driven by the therapist’s theoretical orientation and practice experience.

Nelson and Steele (2008) examined the relative importance of various factors influencing treatment selection. Via an online survey, 206 mental health practitioners—including psychologists and social workers—were asked to rate 29 potential considerations in terms of their likelihood to influence treatment selection. They were also asked to rank (in order of relative importance) 10 broader characteristics that might influence treatment planning. Practitioners reported their decisions to employ a particular treatment were most heavily influenced by empirical support, flexibility, colleague recommendation, and appeal to clients. In terms of relative rankings, flexibility emerged as the most important characteristic in considering the use of a treatment. The second most important characteristic was that the treatment was supported by evidence. Baker-Ericzen and colleagues (2015) used case vignettes to investigate assessment and treatment formulations among a sample of pediatric clinicians, including psychiatrists, psychologists, social workers, and marriage and family therapists. Participating clinicians were grouped based on prior training in a specific evidence-based treatment (EBT). Using a “think aloud” technique, clinicians verbalized their case conceptualization and offered a treatment decision in response to a randomly presented vignette. Researchers coded responses by applying five primary decision-making processes identified in the NDM literature: (a) type of reasoning, (b) organization of information, (c) attention to information and level of abstraction, (d) finding solutions, and (e) incorporating actuarial information and flexibility in application. Results indicated clinicians with prior training in one or more EBTs demonstrated clinical decision-making skills consistent with NDM’s conceptualization of expert performance (e.g., prior experience and content knowledge). Importantly, this study demonstrates the viability of using the NDM framework to investigate treatment decision making within an EBT/EBP context.

Decision Making Based on Characteristics of the Client

Lucock et al. (2006) surveyed qualified psychotherapists and psychologists in clinical training to identify the primary influences on their clinical practices. Data were structured into four categories: training, literature, practice, and personal factors. Responses indicated that qualified psychotherapists were most influenced by psychological formulation, current supervision, postqualification training, and client characteristics. Clinical psychologist trainees were most influenced by current supervision, professional training, psychological formulation, and client characteristics. When responses were analyzed by therapeutic orientation, practitioners with a cognitive-behavioral background reported being more influenced by EBP guidelines and research-based journal articles than other orientations. By contrast, therapists with a psychodynamic, person-centered, integrative, or elective orientation were most influenced by intuition or judgment when making treatment determinations. Altogether, findings suggest treatment decision making is heavily influenced by client-level factors (i.e., psychological formulation, client characteristics, client feedback). As part of a larger publication discussing the need for improved theory and research regarding the decision-making processes of integrative psychotherapists, Schottenbauer et al. (2007) conducted a preliminary study exploring one dimension of treatment decision making among 171 practicing psychotherapists (including psychologists, social workers, psychiatrists, and counselors). Using an online questionnaire to collect data, the researchers asked participating clinicians how they would adjust their treatment planning in the case of a client who was not experiencing noticeable improvement or treatment gains. The most common responses among participants included reassessment of the client, reassessment of the client’s environment or motivation, or reconceptualization of the client’s problems.

Eells and Lombart (2003) explored case conceptualization and treatment planning among a sample of cognitive-behavioral and psychodynamic therapists. The researchers recruited clinicians of three types: novice, expert, and experienced. Each therapist was presented with six vignettes representing three common psychiatric disorders. The researchers explored differences in case formulation and treatment decision prediction among the three experience categories as well as between the two therapeutic orientations. Results for case formulation suggested cognitive-behavioral therapists focused most heavily on symptoms and problems, while those with a psychodynamic orientation placed more emphasis on childhood history, coping mechanisms, strengths, and treatment obstacles. When asked about the effectiveness of treatment, cognitive-behavioral therapists anticipated greater improvement than did psychodynamic therapists. Falvey et al. (2005) investigated the process of treatment planning among a small sample of mental health professionals including psychologists, social workers, and counselors. Participants were administered a notional case representing a client with ADHD (attention-deficit hyperactivity disorder). After reviewing the psychosocial history, client interview, and parent interview, clinicians were instructed to write a case conceptualization and develop a treatment plan, followed by a structured interview with the research team. Results were then compiled into four treatment-planning clusters. In the first cluster, clinicians relied on direct diagnostic matching to inform their treatment planning. In the second cluster, clinicians relied on heuristic strategies to plan treatment. In the third cluster, clinicians provided thorough case reviews in which supporting and disconfirming evidence was carefully weighed against initial hypotheses. In the fourth cluster, clinicians completed a rapid assessment and sought direction from others involved in the case (e.g., parents, teachers).

de Kwaadsteniet and colleagues (2010) used a cognitive mapping approach to examine intervention selection among a sample of Dutch clinical psychologists. Study participants were presented with case studies and outcomes from psychological assessments for two children with symptoms of depression. Clinicians were asked to construct a causal map and select five interventions they would employ, ranked in order of effectiveness. Results indicated low levels of agreement on causal factors for the two case studies. However, individual clinician’s ratings of effective interventions were significantly predicted from his or her own causal model. Findings suggest causal models may lead to differential views of intervention effectiveness and selection, which may then lead to clinicians choosing different interventions for the same client based upon these views of effectiveness. Groenier et al. (2014) investigated the degree to which problem complexity influences case formulation and subsequent treatment decision making. A sample of Dutch psychologists was presented with two case vignettes—one with a simple problem presentation and one of greater clinical complexity. After reviewing each vignette, the participants were asked to select the most likely Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) classification, describe in their own words the etiology of the client’s problems, and select one or two interventions from a list of 18 specific treatment methods. Results indicated study participants formulated higher quality case formulations for the less complex case than for the more complex case. The treatment methods selected by the psychologists were neither associated with the DSM-IV classifications they identified nor related to the hypothesized cause of the clients’ problems. Rather, treatment decisions were most highly associated with clients’ pattern of complaints as described in the case materials.

Gutierrez et al. (2018) explored the treatment-planning process of a small sample of experienced counselors. Participant were provided a 10-min video vignette of a client discussing her clinical concern (family conflict and stress). While watching the vignette, participants were asked to think aloud concerning their initial impressions and treatment decision strategy. Upon completion of the video, researchers conducted a semistructured interview with each counselor to elicit additional information. Of primary interest to the researchers was the process whereby treatment decisions were reached. Results were divided into four domains through which the treatment-planning process progressed. The first domain, assessment steps, involved identifying unhealthy behaviors. The second domain, clinical impressions, consisted identifying client strengths. The third domain, treatment factors, entailed identifying family dynamics, and client readiness. The final domain, treatment strategies, including identification and selection of specific treatment techniques deemed most appropriate for use with the client. Zuber (2000) investigated the relationship between a client’s own problem formulation and subsequent psychotherapeutic treatment decision recommendations. The sample consisted of 159 Swedish patients receiving public financial support for psychotherapy; all had undergone an extensive clinical interview and assessment process which included an opportunity for the client to describe clinical concerns in his or her own words. Upon completion of the clinical interview and review of the client’s diagnoses and problem formulation, a psychologist would recommend the patient to a specific psychotherapist. The orientation of the psychotherapist to which the client was referred was used to classify treatment decisions. Results indicated clients who formulated their problem as relational in nature were most frequently recommended to insight-oriented psychotherapies while those whose problem formulations focused on symptoms were more likely to be recommended to directive-oriented psychotherapies. Furthermore, with the sole exception of anxiety disorders, clients’ problem formulations were found to have more influence on treatment recommendation than did formal psychiatric diagnoses. These findings suggest that a client’s own problem formulation may have a substantial effect on the clinician’s treatment-planning process.

Synthesis

There were five studies that framed decision making based on characteristics of the therapist, with findings from four studies (Baker-Ericzen et al., 2015; Cook et al., 2009; Witteman & Koele, 1999; Witteman & Kunst, 1997) suggesting an intervention was chosen based on prior experience and orientation, and ease of integration into existing practices. Two studies (Cook et al., 2009; Nelson & Steele, 2008) indicated endorsement by colleagues was a driver in choosing an intervention. One study suggested an intervention was chosen based on opportunities for training (Cook et al., 2009), while Nelson and Steele (2008) posited decisions for interventions were chosen based on empirical support for the intervention or flexibility of the intervention. Of the eight studies framing decision making based on characteristics of the client, six studies suggested decisions in choosing an intervention were driven by the client’s stated problem or concern (de Kwaadsteniet et al., 2010; Eells & Lombart, 2003; Groenier et al., 2014; Lucock et al., 2006; Schottenbauer et al., 2007; Zuber, 2000). Three studies posited interventions were chosen based on the client assessment and diagnosis (de Kwaadsteniet et al., 2010; Falvey et al., 2005; Lucock et al., 2006), or characteristics and motivation of the client (Gutierrez et al., 2018; Lucock et al., 2006; Schottenbauer et al., 2007). Two studies indicated decisions on an intervention strategy were driven by whether or not the client has a support system of family and friends (Falvey et al., 2005; Gutierrez et al., 2018).

Two studies explored the crosswalk of a client’s stated problem and the decision to assign a specific modality. Zuber (2000), the only study to include actual client interviews, found clients who presented with relational concerns were assigned an insight-oriented approach, and clients who presented with symptom-based concerns were assigned a directive approach. Later, Eells and Lombart (2003) explored similarities and differences between therapists with a relational or directive approach and found cognitive-behavioral therapists focused most heavily on symptoms and problems, while psychodynamic therapists placed more emphasis on childhood history, coping mechanisms, strengths, and treatment obstacles.

Limitations

This narrative review incorporated literature that elucidated drivers for decision making when choosing an intervention strategy. Inherent in narrative reviews are bias with regard to the literature chosen. Because the body of literature supporting this area of inquiry is so limited, the authors were unable to incorporate a more rigorous approach to the review. This is due, in part, to the vast nature of how the term “decision making” is defined. We sought to include literature that answered the question of “What factors influence a therapist’s decision to choose a treatment approach?” in its most broad terms. There is also literature that suggests psychotherapy interventions are noninferior to each other, which posits choosing an intervention may be of less importance than other treatment factors such as therapeutic alliance (Ahn & Wampold, 2001; Messer & Wampold, 2002). Given these limitations, one must exercise great caution in drawing any overarching conclusions about treatment decision making other than those of the most general nature.

Implications for Practice

The EBP model implies that treatment decisions should be guided by a number of factors, including the client’s state and circumstances (i.e., assessment findings), the client’s preferences, the person of the practitioner, theory and research evidence, and professional expertise (Gilgun, 2005; Regehr et al., 2007). A number of theoretical models and conceptual frameworks have been developed to describe the process of decision making. Historically, EBP has been most closely aligned with the classical (i.e., rational choice) decision-making tradition (van de Luitgaarden, 2009). However, the treatment decision-making research literature provides little empirical support for this perspective. As such, NDM has been suggested as an alternative framework for conceptualizing treatment decision making within EBP (e.g., Baker-Ericzen et al., 2015). The pattern of modality prescription within this service system is suggestive of an underlying situation–action matching decision rule as conceptualized in NDM theory. The NDM approach recognizes that just because a source of information is available to the decision maker does not mean that this information is in fact used in the decision-making process. As such, NDM-based models seek to identify what types of information actually influence decision making in the field setting. Understanding the processes by which these treatment decisions are made is critical to promoting best practice and client outcomes across a range of practice settings in social work.