3.1 Decision-making in clinical practice

Clinical decision-making is influenced by many factors related to the client, the clinical environment and the clinician. Examples of these factors include the attitudes, beliefs and priorities of both the client (together with his or her family and friends) and the clinician, professional experience and knowledge, the perspectives of colleagues and multidisciplinary team members, and organisational expectations and resources. Rogers and Masagatani [2] described these diverse factors as being either intrinsic (related to the client) or extrinsic (related to the clinical context or environment and to the clinician). A number of consistent influences on clinical decision-making have been identified across different settings and client diagnostic groups (Table 3.1).

What do we know about the decision-making and intervention choices of clinicians working with children and adults with stroke, traumatic brain injury or cerebral palsy? Several studies have found that it is the intrinsic, client-related factors related to the person’s condition (including neuromuscular impairment-level components, such as spasticity, muscle weakness or contracture and joint changes, atypical or stereotyped posturing, acuteness or chronicity of the condition, and sensory or cognitive performance problems) that are most likely to influence decision-making in neurological settings [9, 27].

A qualitative study with expert clinicians identified two additional types of client-related factors as influencing decision-making and intervention choice. These included the client’s personal attributes, such as age, attitudes to participating in therapy, and living situation (for example, in a metropolitan or non-metropolitan area, in one’s own home or a residential setting), as well as the client’s particular daily performance needs, such as being able to answer a telephone at work, or to produce written work at school. In the same study, an extrinsic factor was found to include clinicians’ embodied knowledge regarding the ‘feel’ of a person’s limbs and joints when affected by different severities of hypertonicity. In particular, clinicians indicated that extensive experience working with similar clients was essential for guiding their decision-making, given the limited availability of research evidence that supported the use of one treatment approach or technique above another. It was also evident that clinicians demonstrated personal preference for particular types of intervention options, based on confidence and personal convictions relating to the perceived efficacy of certain techniques [27]. Level of expertise is a further clinician-related factor identified as influencing decision-making, with clinicians who have a higher level of expertise in neurological practice demonstrating more consistency in decision-making and greater capacity for discrimination between client characteristics. Interestingly, in this and another study, level of expertise was not found to be related to length of practice [10, 28].

Work settings and organisational policies are also reported to influence clinical decision-making. Rassafiani et al. [9] found that treatment choice for children with cerebral palsy and upper limb hypertonia was strongly correlated with the setting in which the clinician worked. Using the same case vignettes, clinicians working in educational settings were most likely to choose splinting as the appropriate treatment choice, clinicians working in a hospital context were most likely to choose Botulinum toxin injections, while in community settings the preferred choice was upper limb casting. It is interesting to note that despite their different clinical decisions, there was no difference between these clinician groups in terms of the factors they considered in the case vignettes when making treatment decisions. The authors explained the findings in terms of the influence on decision-making of different organisational policies across settings, that is, a more disease- focused approach in hospital settings compared with a less medical approach in community and educational settings.

3.2.1 Practice-based knowledge: knowing how

Practice-based knowledge or ‘knowing how’ is a highly valued skill in therapy-based rehabilitation practice that develops out of familiarity with, and repeated performance in, particular clinical situations. Prolonged contact with similar client groups within those clinical situations leads to the construction of domain-specific, contextualised knowledge and, ultimately, to expertise [38, 47]. Experienced clinicians develop internal representations, or stereotypical images or ‘scripts’, of frequently encountered clinical problems, allowing recognition and naming of meaningful patterns or clusters of cues, signs or features. This personal, contextualised ‘body of knowledge’ enables the clinician to make fine discriminations between cues that are typical, atypical, expected, unexpected or missing in clients with particular conditions. Pattern recognition is then proposed to lead to the almost automatic generation of potentially useful assessments, interventions, goals, and predictions concerning possible client outcomes, based on past experiences with successful (or unsuccessful) interventions [48–50].

However, although EBP (and its medical counterpart, EBM) are typically described as the integration of high quality research evidence with clinical expertise [44, 51], reliance on experience and expertise is also described as resulting in biased and sub-optimal clinical decision-making [52]. In an attempt to clarify the place of expert clinical judgment in EBM, Howick [42] differentiates between different types and roles of clinical judgment and expertise. General clinical judgment is considered unreliable as a source of evidence and is therefore appropriately placed at the bottom of, or omitted from, evidence hierarchies. However, recommended roles for clinical experience within EBP include the integration of information about intervention effectiveness with client values and circumstances (described as integrating expertise), maximising the benefit of interventions through ‘therapeutic use of self’ (described as therapeutic expertise), and performing tasks that involve tacit knowledge, such as recognising signs and features (described as clinical expertise).

The notion of reliance on explicit (external) knowledge as more reliable than tacit (internal) knowledge lies at the heart of the EBP movement’s assumptions about what constitutes knowledge (or epistemology). Thus, the EBP movement assumes that “the best available external clinical evidence from systematic research” [44, p. 70, italics added] provides more reliable knowledge on the basis of the methodological design of clinical research trials, than does internal, tacit knowledge learned through experience. This assumption that experimental research methodologies are more reliable and always provide greater ‘strength of evidence’ for clinical practice is increasingly being called into question, since clinical judgment and the knowledge gained through experience are acknowledged as being inextricably linked with clinical decision-making, even in research trials [42, 53, 54]. For example, if a clinical trial is comparing the use of resting hand splints to manual stretches to reduce spasticity, then clinical experience is still being relied upon to make decisions about how the stretches will be provided, how it will be implemented, how the splint will be made and when it is worn, how long after splint removal outcome measures will be taken and so on. The credibility of practice-based knowledge is recognised as having a valuable contribution to make in medical as well as therapy settings. For instance, expert knowledge and clinical know-how are judged to be reliable for decision-making and diagnosis in situations of medical uncertainty, and for teaching and mentoring of both medical and therapy students and novice clinicians [36, 55–58]. Practice-based knowledge is also recognised as having an essential role in translating statistical, population-based research findings to individual clients [42, 45, 53, 59, 60].

In addition to practice-based, embodied knowledge, expert clinicians in occupational therapy, physiotherapy, medicine and nursing have been shown to have access to specialised, domain-specific knowledge that includes a combination of theoretical/scientific knowledge and technical skill [50, 61–65]. The organisation of this diverse (but highly individual) body of knowledge helps differentiate between more and less experienced clinicians [21, 38, 66, 67].

3.2.2 Science- and theory-based knowledge: knowing that

Science- and theory-based knowledge is concerned with facts. Theory provides clinicians with principles and assumptions for defining function and dysfunction in relation to particular conditions or diseases [38]. Such knowledge assists clinicians to view (or ‘frame’) each client’s situation as an example of a general category. Relevant knowledge is used as a ‘conceptual lens’ through which to interpret presenting cues and to plan actions accordingly [68, 69]. However, within evidence hierarchies, reliance on theory and knowledge of underlying mechanisms is reduced in importance, and reliance on science takes precedence [70, 71].

In recent rehabilitation practice (during the growth of the EBP movement), science-based knowledge has been predominantly equated with that generated from research using empirical, quantitative methods. Thus, practice described as ‘evidence-based’ has placed greatest value on research methods that promote the highest internal validity (that is, identification of the ‘true’ causes of observed outcomes), primarily through blinding of researchers and/or participants, and randomisation to either a control or experimental participant group [72, 73]. However, research evidence that is generated by traditional experimental methodologies and that describes outcomes in terms of average, group-level statistical models is beginning to be viewed as having limitations in rehabilitation settings. In particular, such models are not designed to represent the complexity of clinical practice that is influenced by many factors (including heterogeneity or diversity of client presentation) or to evaluate the effectiveness of interventions for individual clients [74–77]. The assumption that higher ‘internal validity’ automatically leads to high generalisability (that is, to high external validity) has been criticised, particularly in regard to translating results to real-life clinical situations that are typically influenced by many uncontrollable factors or variables [73, 78]. Group-level research focuses on the average results of ‘typical’ groups of people. Therefore, another criticism of sole reliance on RCTs as evidence in rehabilitation is that the causes of within-group variability for both research and control groups cannot be clarified (this is sometimes described as a ‘black box’ research model because only the inputs and outputs are known, whereas what goes on inside the ‘box’ cannot be known). Once again, translating study results to different populations becomes problematic [70, 78]. In such situations, clinicians are encouraged by some academic/researchers to adopt alternative research designs to augment findings from controlled designs [75], to use reasoning based on knowledge about pathophysiological mechanisms and effects to translate study results to individual circumstances (mechanism-based reasoning), or to understand the possible reasons for study results [79].

In clinical practice, mechanism-based reasoning is defined as making an inference “from (supposed) knowledge of the relevant mechanisms to claims that treatments have patient-relevant effects” [42, p. 126, original italics]. Thus, in rehabilitation practice, for example, mechanism-based reasoning might link an intervention (such as serial casting) with a clinical outcome (such as reduced muscle contracture) and/or an individual client goal (such as improved ability to push one’s arm through a sleeve), based on knowledge about relevant mechanisms (the sliding filament model of actin and myosin muscle components, Section 2.1.2.1). Mechanism-based reasoning is placed on the lowest level of the OCEBM 2011 Levels of Evidence hierarchy [45]. However, in addition to using mechanism-based reasoning for generalising the results of clinical trials to individual clients, it has recently been recommended for use as evidence for defending clinical diagnoses and intervention efficacy [42, 70].

Based on the clinician’s knowledge and understanding of relevant pathophysiology, mechanism-based reasoning is described as having a role in diagnosis, where the clinician is able to determine a problem based on the features or cues that are present, and also in treatment, where the clinician predicts that the effects of an intervention will address the identified mechanism-based problem. Similarly, mechanism-based knowledge is suggested as useful for applying population-based statistical results or EBP guidelines that are derived from such results, in circumstances of multiple co-morbidities affecting a person’s functioning, or where the individual’s circumstances are distinctly different from the participant and/or contextual circumstances in the relevant clinical trial [42, 70, 79, 80]. This type of knowledge, about the potential effects of an intervention in individual circumstances on the basis of pathophysiological, mechanism-based reasoning, is described by Andersen [70] as expertise, built up over time through repeated clinical experience and reflection, and therefore valuable in the provision of EBM. As noted previously, Howick [42] described this capacity for generalisation as integrated (medical) expertise. In contrast to the medical literature, the use of pathophysiological or mechanism-based knowledge by therapists would typically be described as scientific/theoretical reasoning, that is, reliance on knowledge drawn from known or generally assumed facts about how a part of the body operates (Section 3.1.1.1). This discrepancy between how ‘mechanism-based’ reasoning is perceived in medical, compared with therapy practice, suggests that the typical EBM hierarchies are not as easily extrapolated to therapy-based rehabilitation as is typically supposed.

3.2.3 Research questions and study designs

The way in which knowledge is understood or believed to be created shapes the types of questions that researchers seek to answer, as well as the study design that is chosen as appropriate for investigating those questions [81, 82]. The EBP movement is grounded in the logical-positivist paradigm where knowledge is assumed to develop out of logical deductions that are grounded in observable facts. This approach includes a belief in empiricism which emphasises evidence as discovered in experiments. Thus, the EBP movement in healthcare (initiated within the biomedical, rather than therapy/rehabilitation, sciences) is concerned with questions related to intervention effectiveness and causal relationships, and advocates quantitative, experimental research designs (clinical trials) as producing the highest level of evidence (see Figure 3.1). In contrast, the interpretive-constructivist paradigm assumes that knowledge develops out of (or is constructed out of) the interpretations, values and experiences that occur during social interactions. Knowledge is therefore viewed as diverse, shaped by perceptual, social and environmental circumstances [82]. The aim of research in this approach is to describe and explain experiences within social contexts. Thus, research designs focus on idiographic (contextualised) research, using qualitative methodologies [83].

These different approaches to scientific inquiry, and the focus within the EBP movement on traditional empirical research designs, have caused tension among researchers as well as clinicians whose focus is on delivering services to individual clients [36, 75, 84]. Grimmer and colleagues [85, p. 189] describe the situation faced by allied health clinicians as they try to integrate ‘best evidence’ from both logical-positivist (empirical/quantitative knowledge and evidence) and interpretive-constructivist (experiential/practical/qualitative knowledge and evidence) approaches into their clinical practice as “the ‘therapies’ dilemma”. There is now a growing recognition of the need to accept a diversity of study designs that complement findings from traditional research designs while promoting the investigation of intervention effectiveness for individual clients [73, 75, 80, 86–91]. For example, although RCTs have the greatest capacity for minimising risk to internal validity through random allocation of participants, well-conducted non-RCTs that use control strategies such as paired observations and stratified sampling reduce sampling error satisfactorily to produce credible evidence [92]. Further, single case experimental designs (SCED) are designed to answer causal questions of effectiveness, and are regarded as especially appropriate for use in clinical settings [93]. They are now recognised as producing Level 1 evidence for patient treatment decisions, equivalent to the evidence provided by the systematic review of multiple RCTs [45] (see Box 3.1). This shift reflects the understanding that, for many conditions (including brain injury), some problems occur too infrequently to enable large sample sizes for research, and also that client heterogeneity is a barrier both to undertaking RCTs and transferring their findings to clinical practice [94].

The finding that there is little difference between the treatment effects and effect sizes found in RCTs and observational studies has also led to the use of practice-based evidence methodology (PBE) as a reliable alternative to RCTs in rehabilitation practice. The strengths of the PBE methodology (which uses an observational cohort design) include careful attention to patient characteristics to reduce bias, detailed standardised documentation for measuring the intervention process, patient diversity to improve generalisation or external validity, and the use of multivariate statistics to evaluate associations between factors and the effectiveness of intervention outcomes [79]. In relation to clinical evaluation research, occupational therapists working with children similarly described valuing the use of ‘practice-generated evidence’, which they used to refer to information generated from the documentation and evaluation of client programmes. The therapists identified that practice evidence could be generated at both individual and whole service levels. Individual level evidence referred to clinicians evaluating client outcomes on a case-by-case basis, while the whole service level involved evaluation of service processes and outcomes [96]. The importance of generating practice-based, clinical research using designs appropriate to answering genuine clinical questions is progressively being acknowledged as necessary in therapy-based rehabilitation [39].

In summary, then, rather than strict adherence to the notion of a single-hierarchy approach to developing and appraising research-based evidence in rehabilitation practice, a diversity of study designs suitable for answering clinical questions is becoming recognised as appropriate in therapy-based rehabilitation [39, 73]. In addition, rather than current approaches, where practice is driven by evidence generated by means of a traditional experimental research model, a pragmatic approach where practice is informed by research-based evidence generated through diverse models and combined with domain-specific clinical knowledge and client preferences, is increasingly advocated as having the best capacity for promoting client-centred rehabilitation outcomes [59,73,76,88,97–99]. While such an approach echoes the most commonly used definition of EBP (Section 3.2), it is different in that it values equally diverse ways of knowing (research-based evidence, clinical knowledge). In this approach, EBP does not involve a standard protocol applied to all clients and research knowledge is not prioritised above other forms of evidence. Although there may be protocol-driven elements in the procedures used once intervention decisions are made, for example, using a standard casting process, the application of the intervention is still likely to be individualised based on unique client characteristics, context and goals. It follows then, that EBP can best be described as the logical, systematic and transparent documentation/explanation of one’s clinical decisions. This includes the ability to cite the sources of information used to reach the decision, as well as the individual client factors taken into consideration to ensure that the decision is client-centred.

3.4.1 Description

The layout of the HIPM aims to assist the clinician to view the clinical situation with each client as a ‘whole’, while directing attention to salient upper limb-related information as well as critical decision points within the therapy process. It has nine sections, namely, assessment, group, impairment, functional abilities, clinical aims, personal and environmental factors that may affect the choice and outcome of intervention, negotiated (client-centred) goals, primary intervention options and supplementary intervention options.

The purpose of the first phase of the rehabilitation process, and therefore the top half of the model (encompassing assessment through to personal and environmental factors), is to develop a comprehensive picture of the individual’s functional performance status and to understand the changes that may be desirable and attainable. The aim of the assessment section is to facilitate the development of a clinical picture, and this is achieved by using assessment data to describe upper limb impairments and performance (or function). Identified upper limb problems (impairments and performance limitations) are grouped along a continuum, forming four groups, or clusters, of characteristics, grouped according to the severity of the impairments and their effect on upper limb performance. The aim of grouping upper limb characteristics in this way is to identify general patterns of upper limb performance against which individual clients can be compared [23].

Clinical (or clinician-identified) aims are developed on the basis of the particular upper limb impairments and functional abilities within each group. The purpose of developing clinical aims (see Box 3.2) is to facilitate links between identified upper limb problems and intervention objectives, as well as between problems and the most appropriate theoretical intervention approach or practice model for addressing the identified problems. Personal and environmental factors play a significant role in determining whether outcomes from intervention are sustained over time. While these variables are unique to each person, the purpose of this section is to prompt the clinician to consider all aspects of the person’s life that are relevant to achieving and sustaining a successful intervention outcome (keeping in mind that hypertonicity management is a long-term endeavour).