Physical Therapists' Perceptions of Factors Influencing the Acquisition of Motor Abilities of Children With Cerebral Palsy: Implications for Clinical Reasoning

Selection of Participants

Physical therapist participants were recruited from 18 of the 19 centers affiliated with OACRS; one chief executive officer did not respond to an initial letter. The majority of children with cerebral palsy in Ontario receive services through these centers. Information packages were mailed to 75 physical therapists; 64 consents were returned after a second mailing, for a response rate of 85%. Of the 64 initial respondents, 57 physical therapists from 12 OACRS centers participated in the focus groups (phase 1). Group sizes ranged from 2 to 7 participants. Although 5 to 9 participants per group is recommended for the NGT,²⁰ groups were conducted with fewer participants because this represented the total number of eligible physical therapists at these centers. Four additional physical therapists consented to participate in the third phase of the study, which involved completion of a questionnaire. The questionnaire was mailed to the 68 physical therapists. After one reminder, 60 survey questionnaires were returned, for a response rate of 88%.

Fifty-six participants were women, and 4 participants were men; the majority worked full-time. As a group, these therapists had worked in pediatric rehabilitation for an average of 13.7 years (SD=7.4, range=1–35). Although the preferred interventions used by these participants cannot be stated with certainty, service providers in the OACRS centers usually provide family-centered services, particularly with respect to: enabling and partnership, providing specific information about the child, and providing respectful and supportive care.²⁴ There is, in our view, a trend among physical therapists working at centers similar to these to a move from a neurodevelopmental treatment approach to family-centered functional therapy²⁵ incorporating motor learning and cognitive strategies²⁶ when working with children with cerebral palsy.

Phase 1: Consensus Within Focus Groups

A package of information containing a description of the study, the agenda for the focus group, a 2-page introduction to the model and the definitions of the 4 constructs of interest, and a descriptive survey instrument was mailed to each physical therapist. Examples of determinants were described in very broad terms, so as not to constrain responses.²⁰ Participants were asked to think about what they believed influences the acquisition of basic motor abilities of young children with cerebral palsy relative to the 4 constructs. They were asked to record their ideas in the survey instrument provided in the information package in preparation for the meeting. The focus groups lasted 60 to 90 minutes, and the NGT was used.²⁰ Although the primary author (DJB) had not previously used this technique, the “Guidelines for Conducting NGT Meetings”²⁰ was used for the sessions. Each focus group began with an introduction of the facilitator (DJB), a review of the agenda, and a brief overview of the model. This introductory portion ended with the question “What are the important factors in the 4 constructs of child characteristics relating to the primary neurological impairments, child characteristics relating to the secondary impairments, child personality characteristics, and family ecology contributing to change in motor abilities of children with cerebral palsy?” Participants were assured that this was not a test. The facilitator reiterated the point that we were interested in learning from them and their experiences.

Step 1 was completed in 5 to 10 minutes. During this time, each participant independently generated a list of determinants of motor outcomes for the 4 constructs. Next, a round-robin format was used to generate a composite list of determinants of motor abilities. Each participant was asked to suggest one determinant at a time in an effort to ensure equal participation. Participants were encouraged not to repeat ideas, but to suggest variations that they thought were important. The group facilitator recorded ideas on a flip chart with a separate page for each construct. This process lasted 30 to 40 minutes. The round-robin format was conducted rapidly based the assumption that doing so would minimize ideas being attributed to individual participants.²⁰

After all potential determinants were recorded for the 4 constructs, each determinant was reviewed for clarity and distinctiveness from other determinants through discussion guided by the facilitator. Although many of the determinants might be partially correlated, the main consideration was whether each determinant could potentially affect the acquisition of motor abilities in a unique way. Ideas were grouped together only if all members of the group agreed. The average number of determinants for each construct generated by the focus groups were: primary impairments=13.9 (SD=2.8, range=9–17), secondary impairments=10.8 (SD=3.1, range=6–16), child personality characteristics=11.6 (SD=4.2, range=6–20), and family ecology=14.8 (SD=3.5, range=9–21).

Each focus group ended with the participants identifying and ranking the 5 most important determinants for each of the 4 constructs. As recommended by experts on the NGT,²⁰ each participant first recorded the 5 most important determinants within a construct on 5 separate cards and then identified the most important determinant from the 5 cards (rank=5), the least important determinant from the remaining 4 cards (rank=1), the most important determinant from the remaining 3 cards (rank=4), and the least important determinant from the remaining 2 cards (rank=2). The last remaining card received a rank of 3. Consensus within each focus group was established by determining the average rank for each determinant generated by the participants.

Phase 2: Compilation of Determinants Identified by Focus Groups

Typically, the NGT includes steps 5 and 6 in which the results of the preliminary vote are discussed and a second vote is taken. In our study, the second vote was conducted by having all participants complete a questionnaire. In order to develop a questionnaire for completion by all participants (phase 3 of the study), the rank-ordered lists generated by the 12 focus groups were integrated into a single list. This task was accomplished using a strategy referred to as “the induction of categories using content analysis”²⁰ and was conducted first by the primary author and then independently by another pediatric physical therapist who had 10 years of clinical experience and a Master of Science degree (ie, research training beyond the entry-to-practice degree).

First, the rankings for each potential determinant generated by each focus group were written on paper, and the paper was cut into strips, with one determinant per strip. The primary author then sorted the determinants into groups, with each group capturing one set of ideas. For example, the determinants of “muscle tone” and “spasticity” were included in one group. Participants in some groups elected to combine some issues (eg, muscle contracture and skeletal malalignment); when this occurred, these ideas were pooled for all similar items. Through this process, redefinitions of categories and the development of new categories occurred. Thus, all potential determinants were categorized by new themes, and an average rank for each new theme was produced, reflecting input from the 12 groups. This average rank was determined by dividing the sum of the rankings provided by each individual focus group by 12 (ie, the number of groups).

Only those items with average ranks above 0.7 were selected for inclusion in the questionnaire. We arbitrarily selected this cut-point of 0.7 in order to get about 5 potential determinants for each construct. Although this cut-point seems low in the context of the scale of 1 through 5, many groups did not suggest some of the potential determinants. That effectively contributed zero points to the ranking for that determinant. Accordingly, the value cannot be considered an estimate of importance of the determinant; instead, this was simply the strategy that was used to identify variables ranked most highly for the next phase. Our strategy resulted in 5 potential determinants for primary impairments, 5 potential determinants for secondary impairments, 4 potential determinants for child personality characteristics, and 6 potential determinants for family ecology. The results of the average ranked values of these top 4 to 6 determinants for each construct are shown in Table 2.

To examine interrater reliability, a second pediatric physical therapist independently categorized the determinants generated from the focus groups (ie, she grouped the determinants into categories without knowledge of the primary author's categorization). Independent classification resulted in agreement on 17 of 20 determinants (ie, 85% item agreement). The 2 physical therapists were in agreement on 4 of the top 5 determinants for the construct of primary impairments, on all 5 determinants for the construct of secondary impairments, on 3 of the top 4 determinants for the construct of child personality characteristics, and on all except the fifth ranked of 6 determinants for the construct of family ecology. The questionnaire was constructed using the categorization by the primary author.

Phase 3: Consensus Among Participants

In the final phase, all participants completed a questionnaire that was developed for this study. The Appendix contains an example of one of the items under construct of family ecology. The potential determinants for each of the 4 constructs were presented in random order on the questionnaire. Questions were focused on the importance of each determinant in influencing the acquisition of motor abilities, an estimate of the feasibility of collecting data in a clinical setting, and an indication of what resources would be required before the data could be collected (eg, training, time, “other” [for example, financial reimbursement]). Participants were asked to rate the question about the importance of each item on a 10-point scale, with 10 being “very important” and 0 being “unimportant.” Feasibility was rated on a 3-point scale, with 3 being “feasible” and 1 being “not at all feasible.”

In order to identify a set of variables for future research to test the model, we selected a cut-point of greater than 7. The intent was to identify a relatively small number of variables that therapists perceived to be important in the acquisition of basic motor abilities by children with cerebral palsy. We did not aim to identify all possible variables. In the context of testing the model in the future with structural equation modeling, we believe in the principle of parsimony.^27–29 In addition, we believe that when doing research in a clinical setting, it is important not to burden families and service providers with data collection.

Implications for Clinical Reasoning

This consensus exercise was conducted with 60 physical therapists with an average of 13.7 years of experience in pediatrics. Although the participants neither developed the conceptual model illustrated in the Figure nor explicitly stated their assumptions about factors influencing the acquisition of motor abilities outside of this study, we contend that the results shown in Table 3 provide a rare glimpse of factors that physical therapists might use implicitly in practice when making clinical decisions. These determinants can be used in one of 2 ways. Determinants perceived as not likely to change have implications for prognosis, including establishment of realistic and attainable goals. In contrast, determinants perceived as amenable to change are potential areas for intervention. Although the therapists in this study were not asked to make this distinction, the results provide a foundation for further inquiry on how knowledge of determinants is applied in decision making.

Some of the primary impairments might be the factors that are least amenable to change through physical therapy intervention. Physical therapy intervention is not likely to change either the distribution of involvement or the nature of the sensory impairment. These factors, therefore, are unlikely to be the focus of intervention, but they could be taken into consideration when selecting intervention strategies and procedures for children with different attributes. Similarly, although muscle tone (ie, “the force with which a muscle resists being lengthened”^31(p577)) has been a focus of physical therapy interventions in the past,³² current rehabilitation practice favors medical or surgical interventions such as use of intrathecal baclofen,³³ use of botulinum toxin A,³⁴ or selective posterior rhizotomy³⁵ for the management of spasticity. From this study, it is difficult to speculate how physical therapists would use the factor they identified as “movement patterns” in decision making. Therapists using a neurodevelopmental framework based on the work of the Bobaths would likely target the quality of movement as an area for intervention, albeit in what they consider a functional context.^36–38 Conversely, therapists using what they call a functional approach²⁵ might use information from the assessment of movement patterns to assist in establishing realistic functional goals. Finally, little information was obtained with respect to balance or postural control; however, the consensus statements of a conference on the management of cerebral palsy concluded that physical therapy is effective in enhancing postural control,³⁹ supporting therapists' focus on enhancing balance and postural control during intervention.

In contrast to the mixed picture under the construct of primary impairments, physical therapists try to prevent secondary impairments through interventions.⁷ The secondary impairments identified in our study as determinants of motor change are currently recommended as areas for intervention by health care practitioners working with children with cerebral palsy. These areas for intervention include prevention or minimization of muscle and joint contractures and skeletal malalignment and promotion of general health, muscle force, fitness, and endurance.^40,41

Motivation was the only child personality characteristic unrelated to the diagnosis of cerebral palsy that was identified as a determinant of change in motor abilities. Based on their experiences, participants might perceive that children who are motivated to achieve motor abilities are more likely to actively participate in intervention than children who are not motivated. That is, therapists might regard a child's “motivation” as important during the goal-setting process. In addition, some therapists might use strategies during intervention that are intended to enhance motivation. Some people believe that motivation can be increased if the child perceives the activity to have some type of intrinsic reward.⁴² Campbell⁴³ related the concept of “flow” (defined as an optimal psychological experience) to a pediatric physical therapy session by suggesting that if a therapist is able to engage a child deeply in an activity, outcomes of therapy might be enhanced.

Finally, therapists identified a family's support to the child, expectations of the child, and the support that the family has as being important determinants of motor change for children with cerebral palsy. We believe that these ideas reflect a philosophical approach related to family-centered service.⁴⁴ Each of these determinants is a potential target for intervention or supportive care. Based on the assumption that optimal child functioning occurs within a supportive family and community context, service providers should support families and encourage the use of community supports, as necessary.⁴⁴

In our study, therapists were not asked to describe how they perceived these factors to influence acquisition of motor abilities of children with cerebral palsy; however, we believe that this is a useful area for future inquiry. Campbell⁴³ recently emphasized the importance of using explicit decision-making paradigms in the context of a conceptual framework to improve physical therapy practice and optimize outcomes of children with neurological impairments. Conceptual models have been identified that influence daily practice by giving guidance on what to do, how to do it, and—most importantly—why to do it.⁴⁵ Thus, although it has been argued that clinical reasoning involves more than the ability to offer reasons justifying clinical decisions,⁴⁶ we believe that explicit acknowledgement of the content and process of clinical decision making will improve both the outcomes among children with cerebral palsy and the learning experiences of inexperienced physical therapists.

Comparison of Consensus Results With Current Knowledge

Informing Future Research

The determinants identified by the participants will be used, along with future research data, to test a multivariate model of determinants of motor change for children with cerebral palsy. Although practicing clinicians reached consensus about potential determinants, this does not mean that the identified factors are the most influential. We believe that their perceptions must be tested. Before this research can be conducted, measures of some of the variables will need to be operationally defined and developed, and the validity and reliability of data obtained with these measures will need to be examined.

Limitations

Several limitations of this work exist. First, participating clinicians might not have identified some critical factors that might influence the acquisition of basic motor abilities of young children with cerebral palsy. Some practicing clinicians do not keep up with the research literature, and they are not necessarily aware of the broader concepts being developed. However, many of the participants in this study are, in our view, reflective practitioners,⁷¹ and make the most of their practice experiences.

Second, because the therapists generated lists of determinants and a common set of operational definitions was not used, the participants may actually have had differing views of the determinants. Muscle tone, for example, may have been tested by multiple therapists, but they may not all have had the same idea as to what the term means.

Third, the model and this consensus exercise focus on the acquisition of basic motor abilities such as rolling, crawling, and walking, and not motor skill acquisition, which relies on cognition to a greater extent.¹⁵ Among children with cerebral palsy, these basic motor abilities are acquired by 7 years of age.¹⁶ The combined effect of these 2 limitations might explain the lack of consensus about child characteristics. Variables such as cognition, attention, memory, anticipation, reasoning, and judgment are clearly important in acquiring new motor behaviors that are described as “skills” versus “basic abilities.”