Clinical decision making by experienced and inexperienced pediatric physical therapists for children with diplegic cerebral palsy.

Pediatric physical therapists regularly make clinical decisions that structure and guide their evaluations and treatments, yet surprisingly little is known about the processes used in making these decisions. Watts,[1] for example, reported that physical therapists make complex, costly, and important decisions based on an array of scientific principles but also may rely on judgments of which they are largely unaware.

Clinical decision making occurs under conditions of uncertainty.[2] In pediatric physical therapy, clinicians exercise judgments and implement interventions in hopes of achieving some positive future outcome. Outcomes often are difficult to quantify. Clinicians, therefore, out of necessity must develop clinical decision-making skills based on delayed and often ambiguous feedback. Undoubtedly, clinical decision making is an abstract topic to investigate. Yet, this does not diminish the importance of understanding decision making, as these processes represent the foundation on which physical therapy interventions are based.

To access these elusive processes in this qualitative study, we used three sources of information. First, existing literature provided conceptual examples of how clinical decision making has been applied in related disciplines. Second, in a pilot study we assessed preliminary characteristics of clinical decision making and field-tested research methods. Third, we used our study to clarify and describe four characteristics of clinical decision making using qualitative research methods. As customary in qualitative research, data analyses were combined with reassessment of literature to define, refine, and describe complex phenomena.

Four Characteristics of Clinical Decision Making

First, clinical decision making in physical therapy requires a clinical application of "domain-specific" knowledge, knowledge usually described as fitting into one or more of the following specialty areas: cardiopulmonary pathology,[3] cardiopulmonary rehabilitation,[4] stroke rehabilitation,[5] orthopedic spinal rehabilitation,[6] low back rehabilitation,[7] neonatal rehabilitation,[8] infant diagnosis,[9] and sports rehabilitation.[10] Knowledge, however, is only one component of clinical decision making. Payton[11] reported that physical therapists and physicians used similar "clinical reasoning processes" in applying their knowledge. Jensen et al,[12] reported that experienced clinicians demonstrated "effort less integration" of clinical information based on their ability to access and apply previous experiences to current cases. Jensen et al[13] also reported that master clinicians in orthopedic physical therapy presented a much more "elaborate cognitive framework" by accessing and applying easily accessible schemata. Schemata[14, 15] are the underlying cognitive structures that individuals use to encode, store, recall, and apply information based on previous experiences. Jensen and colleagues,[13] reported that these schemata appeared to be used clinically for evaluating and treating patients efficiently.

These authors suggest that applying domain-specific knowledge is needed to make decisions, but they do not suggest how clinicians might use cognitive processes to apply information in making clinical decisions. Glaser[16] suggested that underlying cognitive structures are applied based on "conditionalized knowledge," or knowledge that is connected with specific conditions. Further, Glaser contends that "knowledge structures" stored from past experiences serve as reference points for retrieval and comparison with current cases.[16] Perhaps schemata may be applied in pediatric physical therapy practice. Thus, cognitive processes used in clinical decision making were assessed in this inquiry.

Second, divergent views regarding "improvisation" versus "formal decision analyses" exist. Improvisation, defined here as practice without prior planning, is believed to be an important component in physical therapy clinical practice.[12,13,17,18] Jensen and colleagues[13] reported that therapists apply improvisation while responding to individual patient needs. Improvisation is believed to be applied quickly in practice and therefore may be in conflict with a more time-consuming formal decision analysis advocated by other authors.[19-24] Some authors have suggested that formal decision analysis, or a decision tree (ie, a diagram showing the relationships over time among the decisions made and the unfolding of events), should be applied in clinical decision making. Rothstein and Echternach[19,20] for example, have proposed a nine-step "hypothesis-oriented algorithm" for making clinical decisions. Palisano et al[24] have advocated this model for pediatric physical therapy. Yet, it is unclear whether time constraints permit formal analyses during everyday (within-session) decision making.

Third, authors have affirmed that physical therapists should be socially responsive to their patients during clinical practice. Authorities on pediatric physical therapy[25,26] contend that clinicians should engage their patients through play as a means of motivating children, directing therapy activities, building self-esteem, and nurturing assertiveness. Other authors[27-30] encourage a circular social reciprocity between adults and caregivers to foster positive social interaction. A "goodness of fit" is recommended between a child's coping resources and demands from the environment.[31] Thus, in our study we examined psychosocial interactions as a component of clinical decision making.

Fourth, authors have suggested that self-monitoring during practice is a key element in making clinical decisions. Chi et al[32] reported that experts appear to be more aware than are novices of their own mistakes, and they emphasized the need or individuals to self-monitor their practice. Dorner and Scholkopf[33] reported that in complex systems experts pay lose attention to the configuration of the facts and adapt their behavior to changing environments. Schon[34] and Shepard and Jensen[35] advocated that reflective practitioners should monitor their practice regularly. Therefore, self-monitoring appeared to be germane to our investigation. These reports in related areas of physical therapy and other disciplines provide helpful guidelines for examining clinical decision making in pediatric physical therapy. To develop and evaluate methods for investigation, and to initiate the study of clinical decision making in pediatric physical therapy, a preliminary study was conducted by the principal author (DGE).[36] The pilot study, as well as the study reported here, applied an "experienced versus inexperienced" methodology for two reasons. First, this approach provided a sharp contrast for comparison. For example, Jensen and colleagues,[12] reported data comparing the insights of clinical practice of experienced and inexperienced therapists but excluded data from clinicians with intermediate experience who "did not contrast sharply" with the other groups. Second, applying the "experienced versus inexperienced" research design provided clinically meaningful findings. For example, the clinical decision-making characteristics of experienced clinicians provided characteristics of exemplary practice. Equally important, understanding clinical decision making among inexperienced clinicians helped identify possible deficits in their education and training that may a potentially be addressed.

Findings in the pilot study detected two characteristics of clinical decision making for further investigation. Experienced clinicians, for example, appeared to be more "child oriented" in their intervention, whereas novices appeared to be "activity oriented." These findings suggested that experienced clinicians may have been able to be more responsive to the child's psychosocial needs compared with the novices, and further study seemed to be warranted. The pilot study also detected cognitive schemata (large meaningful patterns called "movement scripts") that were used to process clinical information.

Based on the findings of the pilot project and literature review, we designed and conducted the study reported here. Retrospective think-aloud procedures were selected to evaluate four characteristics of clinical decision making by pediatric physical therapists. Two questions were addressed: (1) What are the characteristics of clinical decision-making processes used by pediatric physical therapists? and (2) What are the differences and similarities in clinical decision making used by experienced and inexperienced pediatric therapists? An overview of this study is provided in Figure 1. As depicted in this illustration, an ancillary inquiry estimated the possible benefits of the clinical decisions made and subsequent interventions provided.

Method

Therapists

Therapists were recruited by contacting the directors of outpatient neuromuscular centers in the Pacific Northwest. Each director nominated therapists who were considered to be skilled, commensurate with their experience level, in the evaluation and treatment of children with cerebral palsy. The four centers in this study served primarily children with neuromuscular disorders.

The experienced therapists (E1, E2, and E3) had 10, 25, and 15 years of experience, and the inexperienced therapists (N1, N2, and N3) each had less than 2 years of experience. One inexperienced therapist (N3) possessed unique qualifications as a certified aquatics instructor and had 18 years of experience working with children, including children with cerebral palsy. This novice also had received a more intensive "apprenticeship" than therapists N1 and N2. Specifically, therapist N3 had received weekly 1-hour problem-solving sessions for 3 months with an experienced therapist (E3) as a routine part of her orientation and training. All three novices received routine consultation with other staff members, including their clinical supervisors, upon request.

Although therapist N3 possessed unique qualifications, she met the inclusion criteria and her data were included in the data analysis. Her participation in the study provided an interesting, albeit limited, contrast to estimate the effects of prior experience with children with cerebral palsy and of more intensive training on clinical decision making.

Children

Children selected for this study were diagnosed with diplegic cerebral palsy based on medical records at each neuromuscular center and were between 2 and 10 years of age (mean age=4 years 10 months). Children with this diagnosis were selected because they are seen frequently in clinical facilities and provide abundant opportunities for clinical decision making. The children were selected based on convenience and willingness to participate. All participants in the study signed a consent form.

Procedures

The pilot study assessed the feasibility and merit of three methods: retrospective and concurrent think-aloud procedures and semistructured interviews. Retrospective think-aloud procedures were conducted by an experienced therapist who was videotaped evaluating a child and later, while viewing the videotape, was asked to verbalize his thinking about this evaluation. Concurrent think-aloud procedures occurred when the other therapists described their thinking while viewing this videotape.

Comparing the transcripts derived from these methods, the researcher was able to determine that the type of verbalizations in these two think-aloud procedures were similar, with one exception: Concurrent methods did not allow the second two clinicians to have access to kinesthetic and tactile information (eg, resistance to passive movement, muscle strength), as they described their thinking only from videotaped information. Thus, retrospective think-aloud procedures were selected for the current study.

In the pilot study, semistructured interviews were conducted in which therapists were asked both structured questions and spontaneous questions that occurred to the researcher during the interview. Semistructured interviews occurred as therapists reviewed medical histories and planned treatment programs. Semistructured interviews were not selected for the current study because we believed the interviewer could consciously or subconsciously lead the clinicians in their dialogue and thus potentially bias the data.

Based on these findings, retrospective think-aloud procedures were selected as the best method for assessing clinical decision making in pediatric physical therapy. This method, however, required conditions under which accurate "think-aloud data" could be collected.

Some scholars[37-39] have suggested that certain think-aloud data collected using some procedures can lead to inaccurate findings. Nisbett and Wilson,[37] for example, reported that researchers have no direct access to higher-order mental processes such as those involved in evaluation, judgment, and problem solving. Consequently, it is likely that attempts to directly access these processes from verbal reports or introspection will produce inaccurate information. Therefore, "protocol analysis" as developed by Ericsson and Simon[40] was used in our study to elicit primarily level-1 verbalizations (those stored in short-term memory) and level-2 verbalizations (descriptions of recently acquired information). Therefore, instructions (Tab. 1) were given to the therapists prior to each think-aloud session. These instructions were designed to minimize level-3 verbalization (those requiring the individual to explain his or her though processes that required higher-level thinking or long-term memory). Table 1.

Instructions for Evaluation and Treatment Think-Aloud Sessions As you watch the (evaluation or treatment videotape of yourself with (child's name), we will be conducting a procedure called a "think aloud." During this process, I'd like you to watch the videotape and verbalize what you are thinking, using these guidelines:

1. Verbalize what you are thinking at this point in time, not what you may have been thinking during the treatment session.

2. Do not try to overanalyze. Just talk about whatever comes to your mind.

3. Speak as continuously as possible. Say something at least every S seconds, even if only "I'm drawing a blank."

4. You may stop the videotape at any time to share your thoughts by pressing on the right side of the foot pedal.

5. You may rewind the tape by pressing on the left side of the pedal.

Any questions?

Periodically, I may encourage you to share your thoughts.

Have fun: Let's get started.

Yinger[39] contended that data collection using videotapes produced "a very rich and detailed, yet somewhat foreign behavioral account" of the original event. He suggested subjects should focus on immediate thoughts when viewing themselves rather than trying to remember their thoughts at the time of the actual event. Yinger also argued that individuals can gain cues from videotape viewing that may not be available in the actual event. For example, therapists could concentrate on describing their thoughts rather than attending to the children and therapy procedures.

We used a five-step process for the retrospective think-aloud procedures. First, interventions were videotaped by the principal investigator in the typical clinical environment. Second, retrospective think-aloud sessions were conducted by asking therapists to watch the videotapes and describe what they were thinking (Tab. 1) at the time of viewing. Third, these verbalizations were audiotaped and videotaped. Audiotapes were transcribed by a medical transcriptionist. Fourth, transcripts were edited by the principal investigator while watching and listening to videotaped recordings of the think-aloud sessions. The principal investigator thereby corrected errors in transcripts and augmented verbal data with visual data. For example, if a therapist reported that a child demonstrated increased weight bearing on "that extremity," visual data from the videotape could be edited to denote "[left lower extremity]." Fifth, these edited and corrected transcripts were coded and used for analysis.

Coding Criteria and Coder Agreement

Coding criteria were developed for four characteristics of, clinical decision making: movement scripts, procedural changes, psychosocial sensitivity, and self-monitoring (Tab. 2). The coding criteria were developed and refined through repeated analysis of the transcripts.

Table 2.
Coding Criteria for Movement Scripts, Procedural Changes,
Psychosocial
Sensitivity, and Self-Monitoring
Descriptor              Criteria
Movement scripts        Movement characteristics are reported in
                        clusters. Three specific body parts,
                        muscles, or movements must be identified
                        as descriptors. Examples include flexion,
                        extension, abduction, adduction, medial
                        (internal) rotation, lateral (external)
                        rotation, reciprocation, disassociation,
                        rotation, foot, knee, hip, trunk, spine,
                        shoulder, elbow, wrist, hand.
Active participant      The therapist must verbalize, using a
                        personal pronoun or related part of
                        speech (eg, I, me, my, we, let's) and
                        verbalize being an active participant with
                        the child by guiding or helping.
Procedural changes      Intervention is reported to be changed.
                        Intervention, activity, or position is
                        reported to be changed by the therapist
                        and should include personal pronouns or
                        related parts of speech that identify the
                        therapist (eg, "I," "I'm," "me," "my,"
"we")
                        and a verb denoting or describing a
                        procedural change in position, activity,
                        or environment.
Psychosocial            The therapist reports a collaborative
sensitivity             relationship with the child. Therapist
                        reports attending to or anticipating the
                        child's needs. Each descriptor should
                        address a need of the child, using a
                        phrase such as "play," "fun," "help,"
                        "choice," "entertain," "want," or identify
                        the therapist's response to the child's
                        need, such as reluctant, coaxed,
                        insecure, tired.
Positive psychosocial   The occurrence of psychosocial sensitivity
sensitivity             is judged to be positive when the activity
                        is beneficial for the child socially or
                        emotionally.
Self-monitoring         Therapist verbalizes monitoring his or her
                        therapeutic effectiveness. Judging phrases
                        (self-monitoring) are used, such as "I'm
                        not giving enough stability here," "I was
                        pleased," "So I gave that up," "I'm
                        struggling with this," "I'm trying to
                        emphasize more."
Positive                Self-monitoring is judged to be positive
self-monitoring         when the occurrence is reported to be
                        therapeutically beneficial for the child.

Movement characteristics of the children, for example, were often verbalized in clusters and seemed to represent common themes and patterns. Through data analysis, these patterns appeared to occur more frequently with experienced clinicians than with novices. In an attempt to quantify the differences, criteria were developed for movement scripts, as shown in Figure 2. Generally, three body parts, muscles, or movements(*) were selected, as two of these items did not seem to represent a cluster of clinical information and four items appeared to exclude many occurrences. The following dialogue, for example, represents a movement script that would have been excluded if four body parts, muscles, or movements were selected as criteria:

He's using his abdominals as I get more mobility into

an anterior pelvic tilt to get his trunk moving forward.

(E3-child 1 [C1])

This specific movement script illustrates a movement pattern common to children with diplegic cerebral palsy, as will be described later. Thus, at least three items were required to represent cognitive reference points to index a complete schema. For this study, movement scripts were defined as "cognitive schemata representing prototype clusters of at least three body parts, muscles, or movements." The term "movement" represents the best description of the data, and the term "script" was adopted based on parallel terminology in the clinical literature. The term "prototype" was selected because this denotes an "abstract representation of a category" rather than an exemplar (a cognitive schema representing an actual patient).[2] Therefore, movement scripts are believed to represent schemata or cognitive structures used by pediatric physical therapists to encode, organize, retrieve, and apply clinical information based on prototypes.

Procedural change provides another example of how the coding criteria were developed. Verbalizations that merely denoted a change in the therapy procedure did not meet the criteria. The dialogues required the therapists to use personal pronouns and related parts of speech (ie, "I," "I'm," "me," "my," "we") to denote that the procedural change directly involved the therapists and thereby excluded changes that may have been initiated or directed by the children .

Periodically, segments of the transcripts represented more than one characteristic. The following dialogue, for example, represented examples of both psychosocial sensitivity and procedural changes:

He's at an age where his attention span is going to be

fairly short, and I want to keep him happy, and I want to

keep him interested, so I sensed it was time for a change.

(E3-C1)

Intracoder agreement was conducted by the principal investigator, who coded the same transcripts twice, 3 days apart. This procedure required analyzing 623 pages of transcripts representing 23 therapy sessions. Interrater agreement was collected between the principal investigator and a pediatric physical therapist with 10 years of experience. These therapists rated 103 pages of transcripts representing two therapy sessions. A level of 90% point-by-point agreement and Cohen Kappa of .80 were accepted as the general criteria for coding data.[41-43] Intracoder and intercoder agreement results are provided in Table 3. These data suggested a moderate to high degree of coder agreement for each of the four characteristics of clinical decision making.

Table 3.
Intracoder and Intercoder Point-by-Point Agreement and Cohen Kappa
                             Point-by-Point
Characteristic of Clinical   Agreement          Cohen
Decision Making              (%)                Kappa
Movement scripts
Intracoder                   96                   .91
Intercoder                   99                   .97
Procedural changes
Intracoder                   93                   .85
Intercoder                   94                   .88
Psychosocial sensitivity
Intracoder                   92                   .84
Intercoder                   95                   .89
Self-monitoring
Intracoder                   89                   .79
Intercoder                   91                   .82

The four characteristics of clinical decision making described in this report are not exhaustive. Additional research may certainly refine and expand these characteristics to increase our understanding of clinical decision making in pediatric physical therapy.

Pretreatment and Posttreatment Videotaping

Due to the importance of clinical decision making in clinical practice and the uncertain nature of these processes in pediatric physical therapy, we believed it was important to estimate the possible outcome of children treated. Therefore, movement patterns and postures of six children (one child for each therapist) were examined.

A three-step process was used to collect and examine the outcome data. First, the movement patterns and postures for each child (Tab. 4) were videotaped by the principal investigator before and after 4 months of weekly physical therapy. These movements and postures were selected and videotaped for each of the children and included (1) the primary means of mobility, (2) the most functional posture, and (3) a multifaceted upper-extremity movement. Second, pretreatment and posttreatment videotape segments were randomized. A coin was flipped to determine which segment (pretreatment or posttreatment) would be first or second (A or B) in each vignette. Videotapes were further randomized by sequentially drawing all vignettes (pretreatment-posttreatment pairs) "from a hat" to determine the order in which pairs would be viewed. These procedures reduced the probability the vignettes of the same children would be rated sequentially and thereby reduced the order effect. Third, these edited and randomized vignettes were rated by 13 pediatric physical therapists and occupational therapists who worked in a university-affiliated pediatric hospital. These therapists, whose pediatric experience ranged from 2 to 18 years ([X var]=9 years 2 months), were not otherwise involved in the study. These therapists were asked to rate the vignettes and designate which segment, "A" or "B," appeared "better" based on their clinical judgment.

Table 4
Movement Activities for Selected Pretreatment-Posttreatment
Videotaping Vignettes
Clinician-Child                         Upper-Extremity
Interaction(a)    Mobility   Posture    Function
E1-C1             Walking    Sitting    Five cubes in a cup
E2-C1             Running    Standing   Catching a 15.2-cm
                                        (6-in) ball
E3-C1             Commando   Sitting    Five cubes in a cup
                  crawling
N1-C1             Walking    Sitting    Five cubes in a cup
N2-C1             Running    Standing   on Catching a tennis ball
                                        one leg
N3-C1             Creeping   Sitting    Five cubes in a cup
"E1=experienced therapist with 10 years of experience,
E2=experienced therapist with 25 years of experience,
E3=experienced therapist with 15 years of experience;
N1, N2, and N3=inexperienced therapists; C1=child 1.

A child was considered to have made observable progress if at least 10 of 13 raters judged the child's posttreatment segment of the vignette as "better." (Note: The binomial probability of 10 of 13 raters making such judgments coincidentally is <.05.44)

Results

Movement Scripts

The numbers of movement scripts described by the experienced clinicians ([X var]=19.2 per session) and inexperienced therapists ([X var]=6.5 per session) are listed in Table 5. These schemata enabled clinicians to describe movement patterns of children which represent movement patterns typically demonstrated by children with diplegic cerebral palsy. Three movement patterns, for example, were frequently described in the transcripts. First, hip flexion, adduction, and medial (internal) rotation often were described in conjunction with knee flexion and plantar flexion in a variety of functional in an standing position appeared to be described in conjunction with weakness of the abdominal musculature and tightness of the hip flexor and hamstring muscles. Third, a posterior pelvic tilt in a sitting position was often described, with tight hamstring muscles, lumbar and thoracic kyphosis, an(1 rounded shoulders and hyperextended cervical spine. These three movement scripts accounted for 52% of all descriptions. The criteria (at least three body parts, muscles, or movements) seemed to serve as "reference points" to efficiently describe movement scripts without having to list the entire patterns characterized previously.

Table 5.
Movement Scripts Described by Experienced and Inexperienced
Therapists
                          Mean No. of
                          Descriptions   Active
Clinicians                Per Session    Verbalizations (%)
Experienced therapist 1    17.9           49
Experienced therapist 2    15.9           52
Experienced therapist 3    23.8           67
Total                      19.2           55
Novice 1                    4.9           23
Novice 2                    3.1           10
Novice 3                   11.4           34
Total                       6.5           28

Interestingly, about half of the movement scripts were not contained in these patterns, which demonstrates that other movement patterns (eg, upper-extremity patterns) existed. In our study, however, other patterns did not occur frequently enough to warrant coding. Some of the movement scripts not represented in the three common patterns in our study also may represent highly idiosyncratic schemata because they are developed by personal experiences, rather than by textbook cases.[45]

Two of the 18 children demonstrated movement patterns that apparently did not match the therapists, expectations f`or children with diplegia, as indicated by the following remarks:

The first time I saw him, I was thinking "this is a child with

diplegia?" You know like right here just as he's sitting

there, he looks like a real, regular little kid. He's rotated

out, his knees are out, and he's keeping them flexed, and a

little bit of a tush tuck, maybe a tiny bit of sacral sitting.

(N3)

She's primarily in steep equinus, relatively symmetrical,

and not a whole lot of internal or external [lateral] rotational,

torsion kinds of elements at the hip. I'm pleased about that.

(E1)

These observations document that the experienced and inexperienced clinicians in our study were able to encode, organize, and retrieve clinical information that enabled them to recognize movement patterns as atypical patterns. These clinicians apparently recognized the children's atypical pattern as different and, therefore, could not assimilate (accept into their existing schemata) this clinical information as examples of typical movement scripts for children with diplegic cerebral palsy. These clinicians apparently needed to accommodate or adapt their existing schemata based on the atypical information. Experienced therapists seemed to use movement scripts automatically to encode, organize, access, and apply clinical information such as kinesthetic and tactile information([dagger]):

There is an awareness of what the child's body is doing,

but also about what your own body is doing, too. You have

to be aware of where your hands are, what kind of

pressure they are providing, what direction the pressure is

going, and how that affects the movement pattern. You

also have to anticipate their movement, and it's more of a

feed forward. And that's something that comes with time.

Experience. A variety of kids, a variety of experiences.

Being acutely aware of what you're doing. (E3)

Transcripts of experienced therapists illustrated that movement scripts occurred with greater mean frequency compared with those of the inexperienced clinicians (group mean of 19.2 versus 6.5 per session). Furthermore, movement scripts described by experienced therapists appeared to be associated with intervention procedures, which enabled them to easily adapt the therapy activities to assist the children as needed:

I'm seeing some compensatory extension through his cervical

spine to help maintain his balance, and I'm giving a

little bit of anterior and downward pressure across his

sternum to help facilitate a little bit more flexion and chin

tuck here. (E3)

Thus, transcripts were further coded (Fig. 2) to include active participation in which therapists reported "guiding or helping" the children. This variable appears to represent a marked difference in the clinical practice of experienced and inexperienced clinicians. Specifically, active participation was coded in the transcripts of experienced therapists with a group mean of 55% as compared with 28% for the novices. Apparently, experienced clinicians could encode, organize, retrieve, and apply clinical information in the form of movement scripts to help or guide the children in therapy more frequently than novices could. Transcripts of inexperienced clinicians, however, showed that movement scripts primarily described the characteristics of the children:

He has that typical standing pattern that children with

diplegia have, where he uses adductors. His knees come

together and internally rotate. (N1)

Alternatively, experienced therapists appeared to have a larger repertoire of movement scripts, which allowed them to help and guide the children more spontaneously (based on the relative number of active participations described per session).

Procedural Changes

The transcribed data presented in Table 6 showed that experienced therapists changed their treatment procedures (procedural changes) about every 46 seconds (78 times per 60-minute session) and inexperienced therapists changed their treatment procedures about every 86 seconds (42 times per 60-minute session). Changes in the therapy procedures probably occurred more frequently than was observed in the data, simply because during each think-aloud session substantial time was obviously used to verbalize other information, such as descriptions of movement patterns. Thus, these procedural changes suggested that the clinicians altered their interventions frequently. Clinicians also could have decided not to change therapy procedures, which is certainly a decision, but it would not have met the criteria. Thus, it is likely that the data presented in Table 6 represent the minimum frequency of procedural change.

Table 6.
Procedural Changes Described by Experienced and Inexperienced
Therapists
                           Mean No. of Descriptions
Clinicians                 Per Session
Experienced therapist 1      83.2
Experienced therapist 2      83.1
Experienced therapist 3      68.5
Total                        78.3
Novice 1                     36.2
Novice 2                     33.0
Novice 3                     56.9
Total                        42.0

Verbal data from therapists N1 and N2 illustrated the least frequent procedural changes. Transcripts suggested an explanation for this low frequency. These two inexperienced therapists reported Using either formal or informal lists of activities, which guided their therapy sessions:

I kind of made a list of different things I wanted to do with

him today, and I felt like a lot of the things I chose were

maybe a little bit higher, or required more than he could

do. (N1)

Similarly, therapist N2 reported providing a "report card" of activities:

I believe we had six activities. One was riding a trike,

another was balancing on a Swedish ball, another involved

climbing stairs, the next was a combination of activities of

stepping over blocks and climbing up a ramp. One activity

was free time for 5 minutes, kind of as a reward. I believe

that was all of them. (N2)

This "report card" was used consistently during the 4 months of intervention for this child. Stickers were used to denote the completion of each activity, and positive reinforcements were provided if all activities were successfully accomplished. This procedure appeared to the investigator to provide both a motivation for the child and a specified list of treatment activities for this therapist.

In contrast, the three experienced therapists relied on readily available intervention strategies, as noted by this report:

I initially felt a lot of adduction and internal rotation

through his lower extremities, so I'm trying to get my arms

positioned to inhibit that a little bit and keep the alignment

[anterior pelvic tilt] on the ball while we're working.

(E1)

In addition, procedural changes often did not appear to require advanced planning:

I just had to reposition him. He was kind of sagging down,

so I'm going to do a little bouncing to get him up. (E1)

These procedural changes, without advanced planning, may

have represented "improvisation." The data in our study,

however, could not quantitatively assess improvisation; they

could only estimate the frequency of procedural changes.

Psychosocial Sensitivity

Psychosocial sensitivity appeared to be a key factor, as shown in Table 7. These verbal reports suggested that the frequency of psychosocial sensitivity was variable, based on the level of experience of the therapists. Nevertheless, the frequency of psychosocial sensitivity from both experienced (33.5 per session) and inexperienced therapists (22.5 per session) illustrates the importance of meeting the emotional and social needs of the children.

Table 7.
Psychosocial Sensitivity Described by Experienced and Inexperienced
Therapists
                          Mean No. of
                          Descriptions   Positive
Clinicians                Per Session    Verbalizations (%)
Experienced therapist 1     31.8            76
Experienced therapist 2     39.4            83
Experienced therapist 3     29.5            83
Total                       33.5            81
Novice 1                    19.9            33
Novice 2                    20.5            39
Novice 3                    27.2            46
Total                       22.5            40

Experienced therapists reported that the children frequently participated in directing the therapy activities, whereas novices indicated that children infrequently helped direct the sessions. Apparently, experienced clinicians could follow the child's lead with spontaneous interventions, as suggested by positive psychosocial sensitivity interactions 81% of the time compared with 40% of the time for the novices. Novices appeared to be more "activity oriented." These data suggest that inexperienced therapists are less able to attend to the psychosocial needs of the children, perhaps because their attention is focused on the therapy procedures. Dialogue from this novice supports this viewpoint:

Yeah, she wasn't ready to leave the toy, but I wanted her to

get some of the active climbing up and down the bolsters.

Working on the lower-extremity dissociation and the

eccentric control coming down. (N1)

In many instances, the novices, attention appeared to be focused primarily, and sometimes exclusively, on the therapy activities, to the exclusion of the needs of the children. The apparent challenge for the novice was to provide a therapeutic activity and simultaneously be socially engaged with the children.

Self-Monitoring

Self-monitoring was coded from the transcripts of the experienced and inexperienced therapists about 18 times per session, suggesting that these clinicians regularly self-assessed their interventions with the children (Tab. 8). Thus, selfmonitoring appeared to be an important mechanism for evaluating and directing interventions. Furthermore, selfmonitoring served as a mechanism for developing and refining the repertoire of intervention strategies used by pediatric physical therapists, as noted here:

I'm trying two different ways and seeing for my own

learning. Exploring with different handling methods and

seeing what might possibly work differently. (N3-C1)

Table 8
Self-monitoring Described by Experienced and Inexperienced
Therapists
                         Mean No. of
                         Descriptions   Positive
Clinicians               Per Session    Verbalizations (%)
Experienced therapist 1   13.9          90
Experienced therapist 2   25.1          73
Experienced therapist 3   16.4          86
Total                     18.5          81
Novice 1                  19.2          35
Novice 2                  14.8          26
Novice 3                  20.4          44
Total                     18.1          36

A notable difference between the data of the experienced and inexperienced therapists was the percentage of positive self-monitoring. Generally, self-monitoring comments of experienced therapists were coded positive 81% of the time. In contrast, the self-monitoring comments of the inexperienced therapists were coded positive only 36% of the time, as illustrated in these verbal reports:

I wanted to get the ball back so that he is using more of his

abdominals and stuff again. I'm not sure really what I'm

doing here. It doesn't look very good. (N1)

I just don't have enough experience to know. What are

some pretty common points that I can control. How much

of it is variable from child to child? I'm trying to get her

to blow bubbles and she is totally absorbed in her balloon,

so give that up. (N3)

These data suggest that novices are aware of what they do not know and are frustrated by this dilemma, but are unable to rectify the deficits.

Pretreatment and Posttreatment Videotape Ratings

Data indicated posttreatment videotapes were rated as "better" (showing clear progress, P>.05, based on binomial probability of 10 of 13 raters[44]) for six of nine activities involving experienced therapists and for three of nine activities involving inexperienced therapists. These findings suggest that the impact of decisions made by experienced therapists may be more frequently positive than the decisions made by inexperienced therapists may be. These are tentative data, however, and it should be noted that one of the children treated by inexperienced therapists achieved ratings as high as those of the highest rated child treated by experienced therapists. Interestingly, the child treated by the advanced novice was not the child who had these high ratings.

These pretreatment and posttreatment data cannot support or imply a cause-and-effect relationship because of uncontrolled variables, but the data provide insight into the possible benefits of clinical decision making and subsequent interventions. Furthermore, due to the small sample, these results should not be generalized to other therapists, children or facilities.

Discussion

Movement Scripts

Movement scripts frequently occurred in the transcripts of experienced and inexperienced clinicians in our study. These cognitive processes appeared similar to schemata reported in the orthopedic physical therapy and cognitive psychology literature. Schmidt and colleagues,[45] for example. reported cognitive schemata called "illness scripts" were used by physicians to "index" clinical information from past experiences in the form of prototypes or actual patients (exemplars). Similarly, Irby[46] reported curriculum scripts were used by clinical teachers in medicine to respond to unknown cases without advanced preparation. Benner[47] suggested that nurses apply "critical incidents" to obtain an intuitive grasp of each situation and respond accordingly. occupational therapists apply the term "clinical reasoning" to discuss the decision-making processes during professional practice.[48-51] Based on these analogous conceptual frameworks and the data from our study, movement scripts evidently represent cognitive schemata used by pediatric physical therapists in making clinical decisions.

Furthermore, movement scripts provided by the pediatric physical therapists in our study appeared to demonstrate three phenomena. First, both experienced and inexperienced therapists used movement scripts to describe movements and postures of children with diplegia. Second, experienced clinicians indicated that these schemata were internalized, perhaps in the form of kinesthetic and tactile awareness. Thus, clinical decision-making processes may require sensorimotor skills in pediatric physical therapy clinical practice. Allard and Starkes52 reported that motor skills and cognitive skills are quite different and therefore warrant further investigation in pediatric physical therapy. Third, these cognitive prototypes appear to be associated with intervention strategies, based on a repertoire of past experiences.

Movement scripts may provide an interesting tool for additional research. Research of movement scripts may enable investigators to focus on more explicit intervention strategies for movement deficits, such as the common movement characteristics demonstrated by children with diplegic cerebral palsy. Eventually, an understanding of the clinical application of interventions guided by movement scripts may assist in efficacy research for children with cerebral palsy.

Procedural Change

Based on group means from the transcripts, procedural changes were frequently noted for both experienced and, inexperienced therapists, suggesting that clinical practice (at least during within-treatment sessions) requires rapid decision making. These findings suggest that time-consuming formal decision analyses for clinical decision making may be impractical for within-sessions decision analysis.

When clinical information did not match the typical movement scripts, however, both experienced and inexperienced clinicians noted this discrepancy. Perhaps when clinical information is atypical, clinicians may need to apply more formal decision analyses, such as those suggested by Rothstein and Echternach.[19,20] Although this application seems reasonable, it is only speculative and warrants further investigation.

Procedural changes noted in the transcripts of the true novices (therapists N1 and N2) often were characterized by formal and informal lists that appeared to guide their therapy sessions. In contrast, experienced therapists frequently changed their treatment procedures, probably based on a large repertoire of past experiences. Although "lists" may be perceived as elementary strategies, they enabled the novices to organize their treatment sessions and may provide a practical approach for inexperienced clinicians. Further investigation is warranted to determine whether these strategies are effective.

Psychosocial Sensitivity

Transcripts revealed that both experienced and novice clinicians acknowledged the importance of attending to the social and emotional needs of the children. Furthermore, psychosocial sensitivity findings of this study were consistent with social reciprocity recommended by other authors. Social reciprocity obviously involves circular relationships between therapists and children, and the methods used in this study did not enable the investigators to assess the circular nature of the interactions. Nevertheless, the findings of our study support the concept that social reciprocity is an important component of pediatric physical therapy.

Differences occurred in reporting positive psychosocial sensitivity based on the level of experience. The data from the experienced therapists revealed positive responses to the psychosocial needs of the children 81% of the time, compared with 40% of the time f`or the novices. Experienced therapists evidently had a large repertoire of intervention skills and possessed an array of previous experiences that enabled them to respond positively to the therapeutic needs of the children and simultaneously attend to the emotional and social needs of the children. Inexperienced therapists described being aware of the need to respond positively to the psychosocial needs of the children, but seemed overwhelmed by the technical nature of the therapeutic environment. Based on these findings, the importance of positive psychosocial sensitivity should be emphasized during physical therapy professional education and early clinical practice. Perhaps, if inexperienced clinicians focused on the psychosocial needs of the children, the technical components of interventions could be directed in a more positive manner.

All experienced therapists ended each therapy session with a positive activity. Frequently, these activities were chosen by the children and appeared to be emotionally bonding for the children and the therapists. Typically, this positive closure enabled the children to end the therapy sessions with pleasant memories and served as a motivating factor for subsequent sessions. Although therapists did not describe "emotional bonding" and "motivation," it appeared evident to the principal investigator based on observing each therapy session and watching the videotapes at least twice. Periodically, the inexperienced therapists described ending the session with similar activities, but not as consistently as the experienced therapists. Emphasizing positive psychosocial sensitivity in training inexperienced clinicians may eventually lead to more effective clinical practice.

Self-Monitoring

Think-aloud dialogue of the experienced and inexperienced therapists documented frequent self-monitoring. Interestingly, the findings of our study suggested that experienced and inexperienced pediatric physical therapists reported self-monitoring with similar frequency (about 18 instances of self-monitoring per session). In other disciplines, experts have been reported to self-monitor more frequently than nonexperts. These conflicting findings may be due to the methods used in our study or the differences in selfmonitoring reported by pediatric physical therapists compared with those reported by individuals in other fields, or the "experienced" therapists in our study may not have been true "experts." Consequently, self-monitoring, as described by pediatric physical therapists, appears to warrant further investigation.

The self-monitoring reported in our study appeared to support the work of Schon[34] and Shepard and Jensen,[35] who advocate that clinicians should be "reflective practitioners." Based on the findings of our study, the frequent reporting of self-monitoring appeared to provide evidence of this type of ongoing self-reflection in pediatric clinical practice.

Comparatively, experienced therapists described positive self-monitoring more than twice as often as inexperienced clinicians described. These data suggested that experienced therapists were generally affirming of their clinical procedures. In contrast, inexperienced therapists frequently described an awareness of the limitations of their clinical practice, but often were unable to provide different intervention strategies. These data suggest that novices may benefit from mentor programs early in clinical practice.

Conceptual Framework

A conceptual framework of clinical decision making developed from our study is illustrated in Figure 2. Because transcripts could be coded as more than one characteristic of clinical decision making, the following conclusions seem to be warranted.

The pivotal characteristic of this conceptual framework is self-monitoring. Self-monitoring, for example, requires an individual to cognitively process domain-specific knowledge and assess current clinical information in comparison with past experiences. Logically, self-monitoring would access cognitive schemata, such as movement scripts, to encode, organize, retrieve, and apply clinical information. As clinicians become more experienced and have greater access to movement scripts, self-monitoring may become more positive and intervention may become more effective. Preliminary data from our study support this view (Tabs. 5 and 8). Similarly, procedural changes were at times based on selfmonitoring the children's responses to the interventions. Transcripts of experienced clinicians demonstrated a high frequency of positive self-monitoring (81 %), with rapid changes in intervention (about every 46 seconds). The novices reported a lower frequency of procedural changes (about every 86 seconds) concurrently with the relative low frequency of positive self-monitoring (36%). Psychosocial sensitivity also often required ongoing self-monitoring, namely of the children's emotional and social needs. Thus, self-monitoring was a key element in the clinical decisionmaking processes of these pediatric physical therapists.

Limitations

A limitation of this study was the attempt to study clinical decision-making processes using retrospective think-aloud procedures, rather than during actual clinical practice. These procedures were selected to access think-aloud dialogue without interfering with the clinical decision-making processes and simultaneously collect accurate data. Findings therefore did not necessarily reflect the decisions made during actual clinical practice, but provided insights into the processes used by pediatric physical therapists as they viewed their practice from videotapes.

A drawback of this study was the relatively small number of therapists and children studied, making the data less generalizable. The findings regarding clinical decision-making processes however, generally appear to support the theoretical framework suggested by other literature.([double dagger]) The results of our study add to the body of clinical decision-making literature, and findings may be generalized to these theories.

Another limitation of this study was the abbreviated assessment of outcome. A comprehensive assessment of efficacy was beyond the scope of this investigation. Additional studies are needed to rigorously assess the efficacy of clinical decision making in pediatric physical therapy.

This study attempted to address an inherent limitation of qualitative research. On one hand, qualitative investigators typically become immersed in the data to carefully examine the research topic. Alternatively, qualitative investigators are an integral part of developing the research design, collecting the data, and analyzing the data. Thus, examiner bias is a concern. Our study addressed this dilemma by applying precise methods designed to add numerical credibility to the verbal dialogue, noted to be relatively ambiguous. These methods did not ensure freedom from examiner bias, but because the data have been precisely collected, analyzed, and presented, the findings appear to be trustworthy.

Conclusions

Our study provided insight into four characteristics of clinical decision making in pediatric physical therapy: movement scripts, procedural changes, psychosocial sensitivity, and self-monitoring. These characteristics define and describe valuable attributes of pediatric clinical decision making in pediatric physical therapy. These characteristics do not represent an exhaustive list of such characteristics, but they provide dialogue for further investigations.

The four characteristics of clinical decision making identified in our study generally support existing literature and provide directions for future research. The abbreviated pretreatment-posttreatment findings suggested benefit for some of the children, following 4 months of interventions directed by the clinical decisions made by the experienced and inexperienced pediatric physical therapists.

Acknowledgments

We express our sincere appreciation to the dedicated therapists, families, and children who served as inspirations and participants in this investigation. Special accolades go to Eugene B Edgar, PhD, and Samuel S Wineburg, PhD, for their scholarly leadership in guiding this study.

[Figures 1 nad 2 ILLUSTRATION OMITTED]

[dagger] This dialogue does not represent a movement script based on Table 2 criteria, but depicts an excellent example of the automatic way experienced clinicians may process and apply movement information.

[double dagger] References 1, 3-13, 16-24, 28-35, 39, 45-52.

References

[1] Watts NT. Decision analysis: a tool for improving physical therapy practice and education. In: Wolf SL, ed. Clinical Decision Making in Physical Theraphy, Philedelphia, Pa: FA Davis Co; 1985:7-24.

[2] Colman AE. Companion Encyclopedia of Psychology. New York, NY: Routledge; 1994.

[3] Blessey RL. Clinical decision making: assessment and planning for treatment of patients with primary cardiopulmonary pathology. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy, Philadelphia, Pa: FA Davis Co; 1985: 115-132.

[4] Irwin S. Clinical decision making: cardiopulmonary rehabilitation. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy. Philadelphia, Pa: FA Davis Co; 1985:133-148.

[5] Holden MK. Clinical decision making among neurological patients: stroke. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy. Philadelphia, Pa: FA Davis Co; 1985:150-171.

[6] Paris SV. Clinical decision making: orthopedic physical therapy. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy. Philadelphia, Pa: FA Davis Co; 1985:215-254.

[7] Nyberg R. Clinical decision making in orthopaedic physical therapy: the low back. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy. Philadelphia, Pa: FA Davis Co; 1985:255-294.

[8] Campbell SK. Clinical decision making: management of the neonate with movement dysfunction. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy. Philadelphia, Pa: FA Davis Co; 1985:295-324.

[9] Young JE. Clinical decision making: neurological dysfunction in infancy. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy. Philadelphia, Pa; FA Davis Co; 1985:325-334.

[10] Donley PB, Peyton RC. Clinical decision making: sports physical therapy. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy Philadelphia, Pa: FA Davis Co; 1985:335-358.

[11] Payton OD. Clinical reasoning process in physical therapy. Phys Ther. 1985;65:924-928.

[12] Jensen GM, Shepard KF, Hack LM. The novice versus the experienced clinician: insights into the work of the physical therapist. Phys Ther. 1990;70:314-323.

[13] Jensen CM, Shepard KF, Gwyer J, Hack LM. Attribute dimensions that distinguish master and novice physical therapy clinicians in orthopedic settings. Phys Ther. 1992;72:711-722.

[14] Corsini RJ. Ozakli BD. Encyclopedia of Psycholoyy. New York, NY John Wiley & Sons Inc; 1984.

[15] Bartlett FC. Remembering: A Study in Experimental Social Psychology. Cambridge, Mass: Cambridge University Press; 1932.

[16] Glaser R. Education and thinking: the role of knowledge. Am Psychol. 1984;39:93-104.

[17] Wolf SL. Introduction. In: Wolf SL, ed. Clinical Decision Making in Physical Therapy. Philadelphia, Pa: FA Davis Co; 1985:1-5.

[18] Magistro CM. Clinical decision making in physical therapy: a practitioner's perspective. Phys Ther. 1989;69:525-534.

[19] Rothstein JM, Echternach JL. Hypothesis-oriented algorithm for clinicians: a method for evaluation and treatment planning. Phys Ther. 1986;66: 1388-1394.

[20] Echternach JL, Rothstein JM. Hypothesis-oriented algorithm. Phys Ther. 1989;69:559-568.

[21] Raiffa H. Decision Analysis: Introductory Lectures on Choices Under Uncertainty. Reading, Mass: Addison-Wesley Publishing Co Inc; 1968.

[22] Schenkman M, Butler RB. A model for multisystem evaluation, interpretation, and treatment of individuals with neurologic dysfunction. Phys Ther. 1989;69:538-547.

[23] Harris BA, Dyrek DA. A model of orthopaedic dysfunction for clinical decision making in physical therapy practice. Phys Ther. 1989;69:548-553.

[24] Palisano RJ, Campbell SK, Harris SR. Clinical decision-making in pediatric physical therapy. In: Campbell SK, Palisano RJ, Vander Linden DW, eds. Physical Therapy for Children. Philadelphia, Pa: WB Saunders Co; 1994: 183-204.

[25] Wilson JM. Cerebral palsy. In: Campbell SK, ed. Pediatric Neurologic Physical Therapy: Clinics in Physical Therapy. New York, NY: Churchill Livingstone Inc; 1984:353-408.

[26] Olney SJ, Wright MJ. Cerebral palsy. In: Campbell SK, Palisano RJ, Vander Linden DW, eds. Physical Therapy for Children. Philadelphia, Pa: WB Saunders Co; 1994:489-523.

[27] Brazelton TB, Koslowski B, Main M. The origins of reciprocity: the early mother-infant interaction. In: Lewis M, Rosenblum L, eds. The Effect of the Infant on Its Caregiver. New York, NY: John Wiley & Sons Inc; 1987:49-76.

[28] Larin HM. Motor learning: theories and strategies for the practitioner. In: Campbell SK, Palisano RJ, Vander Linden DW, eds. Physical Therapy for Children. Philadelphia, Pa: WB Saunders Co; 1994:157-181.

[29] Calhoun ML, Rose TL, Hanft B, Sturkey C. Social reciprocity interventions: implications for developmental therapists. Physical and Occupational Therapy in Pediatrics. 1991;11:45-56.

[30] Calhoun ML, Rose RL. Early social reciprocity interventions for infants with severe retardation: current findings and implications for the future. Education and Training in Mental Retardation. December 1988:340-343.

[31] Zeitlin S, Williamson GG. Developing family resources for adaptive coping. Journal of the Division of Early Childhood. 1988;12:137-146.

[32] Chi MTH, Glaser R, Farr MJ, eds. The Nature of Expertise. Hillsdale, NJ: Lawrence Erlbaum Associates Inc; 1988.

[33] Dorner D, Scholkopf J. Controlling complex systems; or, expertise as "grandmother's know-how." In: Ericsson KA, Smith J, eds. Toward a General Theory of Expertise. New York, NY: Cambridge University Press; 1991:218-239.

[34] Schon DA. The Reflective Practitioner: How Professionals Think in Action. New York, NY: Basic Books Inc; 1983.

[35] Shepard KF, Jensen GM. Physical therapy curricula for the 1990s: educating the reflective practitioner. Phys Ther. 1990;70:566-577.

[36] Embrey DG. Clinical decision making in novice and experienced pediatric physical therapists. Pediatric Physical Therapy. 1993;5(4);193. Abstract.

[37] Nisbett RE, Wilson TD. Telling more than we can know: verbal reports on mental processes. Psychol Rev. 1977;84:231-259.

[38] Smith ER, Miller FD. Limits on perception of cognitive processes: a reply to Nisbett and Wilson. Psychol Rev. 1978;85:355-362.

[39] Ymger RJ. Examining thought in action: a theoretical and methodological critique of research on interactive teaching. Teacher and Teacher Education. 1986;2:263-282.

[40] Ericsson KA, Simon HA. Protocol Analysis: Verbal Reports as Data. Rev ed. Cambridge, Mass: The MIT Press; 1993.

[41] Miles MB, Huberman AM. Qualitative Data Analysis: A Source Book of New Methods. London, England: Sage Publications Ltd; 1984.

[42] Payton OD. Research: The Validation of Clinical Practice. Philadelphia, Pa: FA Davis Co; 1988:115-131.

[43] Kazdin AE. .Single-Case Research Designs: Methods for Clinical and Applied Settings. New York, NY: Oxford University Press; 1982.

[44] Siegel S. Nonparametric Statistics for the Behavioral Sciences. New York, NY: McGraw-Hill Inc; 1956.

[45] Schmidt HG, Norman GR, Boshuizen HPA. A cognitive perspective on medical expertise: theory and implications. Acad Med. 1990);65:611-621.

[46] Irby DM. How attending physicians make instructional decisions when conducting teaching rounds. Acad MPd. 1992;67:630-638.

[47] Benner P. From Novice to Expert: Excellence and Power in Clinical Nursing Practice. Menlo Park, Calif: Addison-Wesley Publishing Co Inc; 1984.

[48] Rogers JC. Eleanor Clarke Slage Lectureship 1983: clinical reasoning--the ethics science, and art. Am J Occup Ther. 1983;37:601-616.

[49] Rogers JC, Masagatani G. Clinical reasoning of occupational therapists during the initial assessment of physically disabled patients. Occupational Therapy Journal of Research. 1982;2:195-219.

[50] Cohn ES. Field education: shaping a foundation of clinical reasoning. Am J Occup Ther. 1989;43:240-244.

[51] Neistadt ME. Classroom as clinic: a model for teaching clinical reasoning in occupational therapy education. Am J Occup Ther. 1987;41:631-637.

[52] Allard F, Starkes JL. Motor-skill experts in sports, dance, and other domains. In: Ericsson KA, Smith J, eds. Toward a General Theory of Expertise: Prospects and Limits. New York, NY: Cambridge University Press; 1991:126-152.