Development of the Professional Role Behaviors Survey (PROBES).

Organizational roles are never static because they are modified as those who occupy the roles (role incumbents) adapt to changes occurring within the organization.[1] The roles of health care practitioners provide numerous examples of this evolution, an evolution that has been accelerated by the advent of hospital restructuring.[2-4] In particular, practitioners' roles have been altered directly as hospitals have replaced traditional, hierarchical organizational structures with more product- or team-oriented structures.[5-7] Furthermore, practitioners' roles have changed as role incumbents respond to the expectations of their coworkers.[1] For example, some authors believe that, in hospitals, this interaction among practitioners has led physical therapists to become more flexible, interactive, and outcomes oriented.[8-11]

To understand how change in an organization can alter the roles of its members, it is useful to examine the nature of organizational roles, how they are developed, and how they are transmitted among the members. From a purely functional perspective, roles are composed of behaviors that reflect the responsibilities defined in job descriptions, such as evaluating patients, planning programs, and serving as a content specialist or consultant for a care team.[11] However, if role enactment were as simple as following a job description, it would be a relatively simple matter to see how changing the job description could change a practitioner's behavior. However, any clinical manager will tell us that it is not that simple to change behaviors.

From an organizational perspective, roles are defined by more than a set of behaviors written on paper. Roles link individuals to a work group (role set) by virtue of the tasks that need to be performed to get the work done. The role of each member of the group is shaped through a reoccurring exchange of expectations, which are sent by work group members and received by the role incumbent, and receiver behaviors. That is, work group members attempt to influence individuals to conform to group expectations about how roles should be enacted. In turn, the individual in a role (ie, the incumbent) perceives and interprets the role expectations sent by the work group based on her or his perceptions and beliefs. Therefore, the role of any individual member of the work group reflects that person's perceptions and beliefs as well as those held by the group. If the role expectations of the work group are perceived as congruent with the person's perceptions, beliefs, and experience, it will influence and motivate her or his behavior in a manner consistent with the work group's intent.[1] However, if the role expectations of the work group are perceived to be incongruent, illegitimate, or coercive, the individual may strongly resist meeting the work group's expectations. The response evoked is, in turn, fed back to the work group, and it reinforces or alters the group's expectations and subsequent role messages.[1] In essence, organizational structure is created by multiple cycles of exchanges, which define organizational tasks. Predictability in how the members of the work group enact the exchanges not only forms the basis for defining roles within the organization, it also serves as the basis for the effectiveness of the organization.[1]

However, the predictability of role behavior is complicated by the sheer complexity of role sets within organizations. Individuals working in organizations are members of several formal work groups as well as informal groups and are called upon to meet the expectations of each. Ostensibly, the formal organizational work group to which a member belongs, such as a department or a patient care team, should dictate the behavior of the member because this immediate work group typically controls the organization's formal, extrinsic rewards. For example, a physical therapist may be required to collaborate with other practitioners to provide patient care and may be rewarded with a status change. Role behavior, however, may be influenced by other groups, such as informal support systems, that are subordinate to the formal work groups within the organization or by other formal groups, such as professions, that are outside of the organization. In many instances, therefore, the roles that emerge can be complex, unclear, and often contradictory as members of the organization attempt to meet the role expectations of several groups. This complexity often leads to role conflict, role ambiguity, or role overload, which can create sizeable problems within the organization and lead to a diminution in individual as well as organizational performance.[1,12,13]

Organizational change often requires members to alter their shared conceptions of individual roles and the boundaries of these roles within a work group.[14] Major organizational changes often challenge assumptions about the core, distinctive, and enduring attributes that members may admire about their work, their organization, and perhaps even their profession. This challenge often requires them to change deeply ingrained beliefs and attitudes.[11,15-18]

Members of organizations often find it hard to embrace major organizational changes. Numerous theories have been developed to explain why individuals initially resist change but eventually accept change and alter their perceptions about their role within the organization.[19-21] Although a discussion of the mechanism of how individuals respond to organizational change is beyond the scope of this article (for such a discussion, see Gutek and Winter[19] and Lau and Woodman[21]), I believe it is important to note that the ability to embrace change or the speed with which one embraces change may depend on the people involved, the strategies used to carry out the change, the magnitude of the change, the degree of involvement of the people whose roles and responsibilities are changing, or the passage of time.[20]

A classification scheme can be used to help examine the effect of organizational change on the roles of professional practitioners. In the literature on so-called patient-focused care, hospital tasks are classified as either clinical tasks or administrative or operational support tasks.[22] This classification scheme fails to recognize the multidimensional nature of the clinical work performed by professionals.[22] To remedy this shortcoming, an alternative classification contains 2 categories of clinical tasks: care production and care management.[7] These 2 categories are highly interconnected and occasionally overlap.[7]

Care production refers to the processes through which all necessary elements are brought together in the delivery of care to the patient.[7] In a sense, care production represents the "hands-on" execution of the patient's care plan. Several studies in the literature on hospital restructuring identify changes in 'care production role behaviors for physical therapists and nurses. For example, Lopopolo[9] found that, following hospital restructuring, physical therapists were expected by management to be more flexible in assuming and carrying out work assignments. D'Aunno et al[3] and Shindul-Rothschild and Duffy[18] found that practitioners needed to be able to do a greater variety of tasks.

Care management refers to the planning and coordination of care delivery using a patient-focused care approach, which involves the integration of patient care across traditional practitioner role boundaries through communication and coordination among clinicians.[7] Care management generally involves relatively high levels of decision making, autonomy, and accountability. Three examples of changes in care management role behaviors following the implementation of patient-focused care from both nursing and physical therapy have been identified in the literature. First, many authors[7,11,15,18,23] have identified active participation in interdisciplinary care teams as a role behavior change. Second, it appears that practitioners have been expected by the patient care team to be more autonomous and accountable for decision making and care provision since the introduction of the patient-focused approach to care.[9,23] Finally, practitioners involved in patient-focused care have had to assume a more assertive role in interagency collaboration in an effort to ensure that continuity of care occurs beyond the acute care hospital setting.[9,23]

Other relevant role behaviors, such as those that fall into the administrative or support category of Lathrop et al,[22] are also important in today's hospital environment.[9,11] These administrative role behaviors reflect organizational responsibilities and are developed with the intention of improving work flow and service integration. As such, I believe they play an increasingly important role in today's restructured hospital environment. Examples of these tasks include performing more administrative tasks and being willing to work at multiple clinical sites such as inpatient or outpatient sites or skilled nursing facilities.[11]

Changes in role behaviors resulting from shifting clinical and organizational responsibilities have created a work environment that is much different than it was 10 years ago, and the advent of hospital restructuring has had an effect on this change.[8-11,24,25] Some of these role behavior changes, however, have been evolving since the implementation of the prospective payment system in the mid-1980s.

A general picture of how physical therapists' roles have changed following hospital restructuring has emerged.[9,11] Yet, the prevalence of these role behavior changes is not known. Furthermore, it is not known how changes affect practitioner views of the work experience, which can affect outcomes for both the individual and the organization.[26] To begin to understand these issues, an instrument is needed to measure the nature and magnitude of the changes.

The purpose of my study was to extend prior research on the changes in practitioner role behaviors through the development and validation of a survey instrument, the Professional Role Behaviors Survey (PROBES). Specifically, this study was undertaken to:

1. Determine whether the role behaviors previously identified in my qualitative role behavior survey[11] adequately reflect the changes that have occurred in hospital-based practice,

2. Examine the internal consistency of the survey instrument, which examines these role behaviors (the PROBES), and

3. Explore the congruence between the underlying dimensions of the role behaviors in this survey and those identified in previous research.[11]

Genesis of Survey Items for the PROBES

In 1999, a report on a role behavior study using qualitative research methods of 100 physical therapy clinical managers was published.[11] Managers were asked to identify the effect of hospital restructuring on the roles of hospital-based physical therapists working during the spring of 1998. Through this study, a list of role behaviors that changed and did not change in response to hospital restructuring was identified, and the frequency of occurrence of each of these role behaviors was measured. In addition, the role behaviors were grouped along 3 general dimensions reflecting direct patient care, professional interaction, and other work-related activities. Similar role behaviors were identified in a majority of the restructured hospitals participating in this study.

The qualitative role behavior survey has provided the only comprehensive compilation of professional role behaviors following hospital restructuring. Because this survey used the Delphi method--using input from a group of content experts to achieve consensus on a topic--one can assume that the role behaviors identified in this study reflect the domain of role behaviors for at least one group of practitioners (physical therapists) working in this setting.[27,28] However, because a Delphi survey uses a qualitative research approach, the psychometric properties of the list of role behaviors and their underlying dimensions were not examined. I believe, therefore, that a survey instrument that examines the role behaviors identified through this method needs to be examined for its content validity, internal consistency, and underlying dimensions before it can be used in future research.

Methods

Survey Instrument and Procedure

Twenty-six mutually exclusive physical therapist role behaviors (survey items) were extracted from the changed and unchanged role behaviors identified in my previous study.[11] In developing the PROBES, 3 major changes were made to the list of role behaviors to improve their content validity and reliability for the current study.

First, in those instances in which a changed role behavior was considered to be an extension of an unchanged role behavior, the 2 role behaviors were combined into one survey item. For example, the changed role behavior "an increase in educating and teaching of patients, family and other health care providers" and the unchanged role behavior "providing patient and family education" were combined into one item: "teaching of patients, family and other health care providers." Second, in those instances in which 2 behaviors were identified conjointly, the 2 behaviors were put into separate survey items for the current study. In this case, the role behavior "an increase in patient evaluation and program planning with a decrease in patient treatment" was divided into "time spent in patient evaluation and program planning" and "time spent in direct patient care (eg, treating patients)." These double-barreled items (ie, items offering 2 choices in one) were eliminated to avoid interpretation problems by the respondent.[29]

Third, all survey items were worded in neutral terms rather than using the words "increasing" or "decreasing." For example, "an increase in focus on functional needs of patients" was reworded to "the focus on the functional needs of the patient has...." This format was used for 2 reasons. First, I believe it avoided respondent agreement biasing by eliminating a presumption of an expected direction of role behavior change.[29] Second, it allowed the respondents to specify the direction and magnitude of change in each role behavior that reflected what had occurred. A 4-member panel of clinicians who were experienced in acute care practice reviewed the initial list of survey items in an effort to ensure item clarity and to eliminate redundancy before the final survey instrument was produced. The resulting 26 survey items are identified in Table 1.

The respondents were asked to indicate the direction and magnitude of change for each of the 26 role behaviors for the physical therapists working in their respective facilities. Their responses were measured using a 7-point Likert-type scale with a neutral midpoint (1=greatly decreased, 2=somewhat decreased, 3=slightly decreased, 4=not changed, 5=slightly increased, 6=somewhat increased, 7=greatly increased).

Finally, a cover letter explaining the purpose of the survey, the survey instrument, and a stamped return envelope were mailed to each respondent. The return envelopes were coded to facilitate a second mailing to those who had not responded to the original mailing. The second mailing was sent 6 weeks later. Completion and return of the survey instrument indicated acknowledgment of informed consent. The data were analyzed using SPSS-PC.(*)

Survey Respondents

During the fall of 1999, all members of the American Physical Therapy Association's Section on Administration and the Acute Care/Hospital Clinical Practice Section received an invitation to participate in the study if they were clinical managers in hospitals that had undergone restructuring within the past 8 years. These members received an introductory letter explaining the purpose of the study and a form they could use to indicate their willingness to participate in the survey. Four hundred fifty-nine physical therapist clinical managers agreed to participate and were subsequently sent the survey instrument.

Because the term "restructuring" is used to describe a wide range of organizational changes within hospital environments, the cover letter included with the survey instrument further defined the term. The clinical managers were asked to complete and return the survey instrument if the restructuring within their hospitals met the following definition:

Hospital restructuring or reengineering involves major organizational changes, which alter the structure, reporting relationships or operation of the hospital departments including physical therapy and alter the delivery of patient care services provided by physical therapy. These organizational changes may include any or all of the following:

* Patient aggregation or the grouping of patients by product line or resource needs.

* Decentralization of services or the movement of services closer to the patients typically in the form of patient focused care teams.

* Cross-training of staff members to perform tasks formerly outside of their area of expertise.

* Simplification of administrative or service delivery processes such as documentation, transportation, etc.

* Staff reductions following the reorganization of hospital departments.

Based on this definition, the number of potential respondents was reduced to 447. The sample of respondents was further reduced to 412 when 35 survey questionnaires were returned because they could not be delivered.

Results

Of the 412 potentially valid survey questionnaires, 253 were completed and returned, yielding a return rate of 61.4%. Although the return rate is somewhat low for surveys, the number of responses was almost twice the amount needed to eliminate subject variance as a significant statistical concern for a 26-item survey instrument.[30] Based on the adequacy of the response rate, further analysis of the data, in my view, was warranted. All completed and returned survey instruments were included in the data analysis, even though a small number of them had missing data.

To ascertain how closely the perceptions of role behavior change for respondents of this survey matched the findings of the qualitative role behavior survey, the PROBES' data were recoded to reflect a trichotomous state similar to the response format used in the qualitative role behavior survey. To make this comparison, the response format for each PROBES role behavior was recoded to reflect either a decrease, an increase, or no change. That is, the respondents who indicated that the role behavior had decreased (choices 1 through 3) had their responses recoded into the category "decreased," whereas respondents who indicated that the role behavior had increased (choices 5 through 7) had their responses recoded into the category "increased." Respondents who indicated that there was no change in the role behavior (choice 4) had their response recoded as "not changed." The recoded data were then analyzed to determine mean responses for each role behavior. The results of this comparison are displayed in Table 1.

Table 1.
Comparison of Perceptions of Role Behavior Changes Following Hospital
Restructuring

                                               Perception of Role
                                               Behavior Changes

                                            Qualitative
                                            Role
                                            Behavior
Role Behavior                               Survey        PROBES(a)

 1 The focus on the functional needs of
   patients has                             Increased     Increased
 2 Delegating and supervising of physical
   therapy treatment has                    Increased     Increased
 3 Teaching of patients, families, and
   other health care providers has          Increased     Increased
 4 Using critical pathways/care paths to
   guide care has                           Increased     Increased
 5 The focus on efficiency/productivity
   in the performance of activities has     Increased     Increased
 6 The integration of physical therapists
   into multidisciplinary teams has         Increased     Increased
 7 The physical therapist's role as a
   consultant, specialist, or advanced
   clinician has                            Increased     Increased
 8 Administrative activities have           Increased     Increased
 9 Work in more than one clinical area
   of the hospital or site has              Increased     Increased
10 Work on weekends/holidays on a
   rotating schedule has                    Increased     Increased
11 The ability to plan and control how
   the work will be done has                Not changed   Not changed
12 A professional approach to patient
   care has                                 Not changed   Not changed
13 Interacting with and assisting other
   physical therapists with patient
   care has                                 Not changed   Not changed
14 Time spent doing care other than
   physical therapy has                     Not changed   Not changed
15 The professional identity of the
   physical therapist has                   Not changed   Not changed
16 Participating in the clinical
   education of students has                Not changed   Not changed
17 Participating in self-education has      Not changed   Not changed
18 Attending staff meetings has             Not changed   Not changed
19 Social interaction with other physical
   therapists in the facility has           Not changed   Not changed
20 Involvement in professional activities
   outside of work has                      Not changed   Not changed
21 Teaching of groups in the
   community has                            Not changed   Not changed
22 Assuming the formal responsibility of
   a case manager has                       Not changed   Not changed
23 Documenting the results of patient
   care has                                 Not changed   Increased
24 Communication/collaboration with other
   health care professionals has            Not changed   Increased
25 Time spent in patient evaluation and
   program planning has                     Increased     Not changed
26 Time spent in direct patient care
   (eg, treating patients) has              Decreased     Not changed

(a) PROBES=Professional Role Behaviors Survey.

As shown in Table 1, a majority of the respondents from the 2 surveys had similar perceptions for 22 of the 26 physical therapist role behaviors. For the remaining 4 role behaviors ("Documenting the results of patient care," "Communication/collaboration with other health care professionals," "Time spent in patient evaluation and program planning," "Time spent in direct patient care [eg, treating patients]"), the respondents' perceptions were not substantially different; that is, no role behavior differed by more than one category ("increased" to "not changed" or "decreased" to "not changed"). The perception of the nature and direction of change in role behaviors, therefore, appears to be quite similar for the 2 studies.

Examination of Survey Items

The first step in examining individual survey items involved the inspection of the distributions. The descriptive statistics for the 26 role behaviors measured in this study using the full 7-point response format scale are displayed in Table 2. The mean response for individual survey items ranged from 3.6 to 6.1, with an average standard deviation of 1.2. These data give an overall impression that all of the role behaviors either had not changed or had increased following hospital restructuring. This finding is consistent with the data presented in Table 1. Although the responses to most of the survey items were not normally distributed, individual items were not found to be sufficiently skewed to prevent their inclusion in further data analysis.[31]

Table 2.
Descriptive Statistics for the Professional Role Behaviors Survey
Role Behaviors

Role Behavior                            N     [bar]X   SD    Median

 1 The focus on the functional needs
   of patients has                       251   5.4      1.5   6
 2 Delegating and supervising of
   physical therapy treatment has        252   5.0      1.3   5
 3 Teaching of patients, families,
   and other health care providers has   253   5.2      1.1   5
 4 Using critical pathways/care paths
   to guide care has                     253   5.5      1.1   6
 5 The focus on efficiency/
   productivity in the performance
   of activities has                     252   6.1      1.0   6
 6 The integration of physical
   therapists into multidisciplinary
   teams has                             253   5.2      1.2   5
 7 The physical therapist's role as a
   consultant, specialist, or
   advanced clinician has                248   5.0      1.3   5
 8 Administrative activities have        248   5.0      1.3   5
 9 Work in more than one clinical area
   of the hospital or site has           247   5.1      1.3   5
10 Work on weekends/holidays on a
   rotating schedule has                 247   5.3      1.2   5
11 The ability to plan and control how
   the work will be done has             246   3.9      1.4   4
12 A professional approach to patient
   care has                              252   4.2      1.1   4
13 Interacting with and assisting
   other physical therapists with
   patient care has                      253   4.2      1.3   4
14 Time spend doing care other than
   physical therapy has                  251   4.4      1.3   4
15 The professional identity of the
   physical therapist has                248   4.3      1.2   4
16 Participating in the clinical
   education of students has             253   3.9      1.1   4
17 Participating in self-education has   252   4.0      1.2   4
18 Attending staff meetings has          246   4.0      0.9   4
19 Social interaction with other
   physical therapists in the
   facility has                          248   3.6      1.2   4
20 Involvement in professional
   activities outside of work has        246   3.6      1.0   4
21 Teaching of groups in the
   community has                         248   4.2      1.2   4
22 Assuming the formal responsibility
   of a case manager has                 244   4.3      1.1   4
23 Documenting the results of patient
   care has                              253   5.1      1.4   5
24 Communication/collaboration with
   other health care professionals has   248   5.0      1.2   5
25 Time spent in patient evaluation
   and program planning has              253   4.1      1.4   4
26 Time spent in direct patient care
   (eg, treating patients) has           251   3.9      1.3   4

Next, the intercorrelation among survey items was examined using Pearson product moment correlation coefficients to assess the internal consistency (ie, how well the items reflect the true score of the underlying variable).[29] As displayed in Table 3, 54.2% of the 325 pairs of inter-item correlations were found to be significant at the .05 level and 44.3% were significant at the .01 level. Although none of the inter-item correlations were greater than .50, 32.9% of the them were greater than .20.

Table 3.
Distribution of the Inter-item Correlations for the Professional Role
Behaviors Survey Role Behaviors

                                       No. of         Percentage of
r                                      Correlations   Correlations

[is greater than or equal to] .13(a)   176            54.2
[is greater than or equal to] .16      144            44.3
[is greater than or equal to] .20      107            32.9
[is greater than or equal to] .25       64            19.7
[is greater than or equal to] .30       35            10.8
[is greater than or equal to] .35       17             5.2
[is greater than or equal to] .40        8             2.5
[is greater than or equal to] .45        2             0.6
[is greater than or equal to] .50        0             0

Total no. of correlations performed: 325

(a) Correlation is significant at the .05 level (2-tailed). Remaining
correlations are significant at the .01 level (2-tailed).

In addition to inter-item correlations, corrected item-to-scale correlations were calculated to examine the correlation of individual items with the overall score. As depicted in Table 4, all survey items except items 10 ("Work on weekends/holidays on a rotating schedule"), 14 ("Time spent doing care other than physical therapy"), and 26 ("Time spent in direct patient care [eg, treating patients]") were shown to have good corrected item-scale correlations (ie, the corrected item-scale correlations were as large as the bivariate correlations at the .05 level of significance or [is greater than or equal to] .13).[11] The findings of low item-scale correlations for the 3 survey items suggested to me that they should be eliminated. However, several authors[29,30,32] have suggested examining the effect of eliminating items on the scale's reliability coefficient first because scale reliability depends on both the extent of the covariation between survey items and the number of items in the survey. Before eliminating any items, therefore, I examined the overall quality of the survey using the Cronbach alpha statistic.

Table 4.
Reliability Analysis for the Professional Role Behaviors Survey

                                            Corrected       Alpha
                                            Item-Total     if Item
Role Behavior                               Correlation   Deleted(a)

 1 The focus on the functional needs of
   patients has                             .41           .79
 2 Delegating and supervising of
   physical therapy treatment has           .16           .81
 3 Teaching of patients, families, and
   other health care providers has          .43           .79
 4 Using critical pathways/care paths to
   guide care has                           .22           .80
 5 The focus on efficiency/productivity
   in the performance of activities has     .14           .81
 6 The integration of physical therapists
   into multidisciplinary teams has         .41           .79
 7 The physical therapist's role as a
   consultant, specialist, or advanced
   clinician has                            .60           .78
 8 Administrative activities here           .18           .80
 9 Work in more than one clinical area of
   the hospital or site has                 .30           .80
10 Work on weekends/holidays on a
   rotating schedule has                    .01           .81
11 The ability to plan and control how
   the work will be done has                .54           .79
12 A professional approach to patient
   care has                                 .51           .79
13 Interacting with and assisting other
   physical therapists with
   patient care has                         .46           .79
14 Time spent doing care other than
   physical therapy has                     .09           .81
15 The professional identity of the
   physical therapist has                   .50           .79
16 Participating in the clinical
   education of students has                .35           .80
17 Participating in self-education has      .43           .79
18 Attending staff meetings has             .36           .80
19 Social interaction with other physical
   therapists in the facility has           .42           .79
20 Involvement in professional activities
   outside of work has                      .28           .80
21 Teaching of groups in the
   community has                            .45           .79
22 Assuming the formal responsibility of
   a case manager has                       .30           .80
23 Documenting the results of patient
   care has                                 .26           .80
24 Communication/collaboration with other
   health care professionals has            .57           .79
25 Time spent in patient evaluation and
   program planning has                     .25           .80
26 Time spent in direct patient care
   (eg, treating patients) has              .12           .81

(a) Cronbach alpha for the scale=.80.

The Cronbach alpha was calculated to determine the internal consistency of the survey elements (ie, the proportion of variance in the survey scores that is attributable to the true score for the underlying variable).[29] For the 26 survey items, the alpha coefficient was found to be .80, which I would consider good overall internal consistency.[29] As a second part of this analysis, I examined the effect of eliminating any one of the 3 items found to have low corrected item-scale correlations on the alpha coefficient. As displayed in Table 4, the results of eliminating any one of these items improved the alpha coefficient of the scale by less than .01. Given this negligible improvement in the reliability coefficient and the identification of these survey items as important role behaviors in the qualitative survey, I decided to retain all 26 survey items.

Examination of Underlying Dimensions

Given the number of inter-item correlations, I performed an exploratory factor analysis to examine whether there was a more elaborate underlying factor structure. In addition, I believed the factor analysis would help to determine whether the current dimensions (ie, factor structure) were congruous with the role behavior dimensions identified in the qualitative role behavior survey. For this analysis, a principal component factor analysis was used to examine potential relationships among the 26 survey items. The results of the initial factor analysis suggested that a 5-factor solution would have the best potential for accounting for a large portion of overall variance and for producing interpretable factors.[29,30] Based on eigenvalues and a Cattell's scree plot, the 5-factor solution accounted for 46.8% of the overall variance.[29,33]

Although the 5 factors produced what I consider a satisfactory representation of the data, the initial factor matrix did not provide a clear factor loading. That is, the factors were not easily interpreted because they did not provide clear relationships among survey items. Consequently, the data were subjected to a Promax, or oblique, rotation that was chosen because I believed that there was some collinearity between the factors.[30] Table 5 presents the results of the loading of the survey items on each factor following the factor rotation and the percentage of variance explained for each factor.

Table 5.
Factor Analysis of the Professional Role Behaviors Survey Role
Behaviors

                                       Factor

                                           3
                                           Eva-    4
                                  2        luat-   Admi-
                                  Infor-   ing     nistra-   5
                         1        mation   and     tion/     Pro-
                         Inter-   Sha-     Plan-   Clini-    ducti-
Role Behavior            acting   ring     ning    cal       vity

19 Social interaction
   with other physcial
   therapists in
   the facility          .800
16 Participate in the    .698
   clinical education
   of students
11 Ability to plan and
   control how the
   work will be done     .579
13 Interact with and
   assist other
   physical therapists
   with patient care     .558
 9 Work in more than
   one clinical area
   or more than
   one site              .549
18 Attending staff
   meetings              .463
20 Involvement in
   professional
   activities outside
   of work               .403
 6 Integration of
   physical therapists
   into multidis-
   ciplinary teams                .643
 4 Use of critical
   pathways/care paths
   to guide care                  .638
 7 Role as a
   consultant,
   specialist, or
   advanced clinician             .617
22 Assume formal
   responsibility of
   a case manager                 .612
 3 Teaching of
   patients, families,
   and other health
   care providers                 .584
21 Teach groups in the
   community                      .566
17 Participate in
   self-education                 .493
15 Professional
   identity of the
   physical therapist             .389
24 Communication/
   collaboration with
   other health care
   professionals                  .312
25 Time spent in
   patient evaluation
   and program
   planning                                .731
 1 Focus on the
   functional needs
   of patients                             .686
23 Documentation of
   the results of
   patient care                            .598
 8 Administrative
   activities                                       .714
14 Spend time with
   patients doing care
   other than physical
   therapy                                          .568
26 Time spent in
   direct patient care
   (eg, treating
   patients)                                       -.478
12 Professional
   approach to patient
   care                                            -.445
 2 Delegation and
   supervision of
   physical therapy
   treatment                                                 .652
 5 Focus on
   efficiency/
   productivity in the
   performance of
   activities                                                .629
10 Work on weekends/
   holidays on a
   rotating schedule                                         .506

Percentage of variance
  explained              20.40    8.09     5.85    5.47      5.32

The factor rotation yielded 5 fairly distinct factors with factor loadings greater than .40, which is considered a substantial loading for a survey with 26 items.[33] Four factors had survey items with like signs indicating that the items were positively related. However, factor 4 had 2 survey items with positive signs and 2 with negative signs, indicating that the 2 groups of items were negatively related and may represent opposite orientations toward a dimension of the physical therapist's role.[33] When the survey items loading on this factor were viewed using the classification scheme proposed by Lathrop et al,[22] this factor appeared to represent the dichotomy between administrative and clinical tasks. That is, increases in administrative and non-care-related activities appear to represent tasks that take time away from patient care, which is represented by both "direct patient care" and "maintaining a professional approach to care."[11]

Generally, the intent of examining the underlying structure of a survey instrument is to create a smaller subset of items to use in further research or to determine whether multiple constructs are being measured.[29] For my study, Cronbach alphas were calculated to examine the homogeneity of the survey items within each factor, to assess the underlying dimensionality of the data, and to aid in factor interpretation. In addition, the content of the role behaviors loading on each factor and their signs were examined to understand the nature of the variables underlying the factors. Factors 1 and 2 had alpha coefficients of .70 and .76, respectively, which indicated to me acceptable levels of internal consistency.[29,30] When the survey items loading on each factor were examined, factor 1 appeared to represent behaviors involving practitioner interactions both within and outside of the hospital, whereas factor 2 appeared to represent behaviors involving the sharing of information. Thus, I labeled these factors "interaction" and "information sharing," respectively.

The remaining 3 factors had alpha coefficients that indicate that the survey items within each factor were not very homogeneous. For factors 3 and 5 (alpha coefficients of .53 and .41, respectively), this apparent lack of homogeneity most likely had 2 causes that combined to diminish the potential shared variance of the survey items and thus the internal consistency of the factors.[32] The combination of a relatively small number of survey items (4 or fewer) loading on a factor and relatively low inter-item correlations ([is less than] .50) can have a considerable depressing effect on the reliability coefficient.[29,30] The low alpha coefficient (.09) for factor 4, in my view, was due to the 2 groups of survey items that loaded in opposite directions.[29]

The effect of removing these survey items on the survey's alpha coefficient was examined. Through this analysis, I found that, although these 3 factors were not as homogeneous as the others, their removal from the survey had only a minor effect ([is less than] .01) on improving the internal consistency of the survey. Given this finding, their removal from the survey instrument, in my view, was not necessary from a statistical standpoint or desirable from a conceptual perspective.[29,30]

Despite the low factor alpha coefficients, in my view, the interpretation of factors 3 and 5 was fairly clear. That is, factor 3 appeared to represent behaviors concerned with patient evaluation and program planning, and I labeled it "evaluating and planning." The role behaviors in factor 5 appeared to be concerned with being productive, and I labeled the factor "productivity." Finally, if the role behaviors that loaded on factor 4 are viewed together, this factor appeared to represent a rivalry for the clinician's time between administrative and clinical role behaviors, and I, therefore, labeled it "administration/clinical."

The congruity between the underlying dimensions of the role behaviors from this study and the qualitative role behavior survey was examined and related to the classification scheme previously described. I compared the 2 surveys by inspecting how the role behaviors were distributed among the dimensions or factors of each survey. The results of this comparison are displayed in the Figure. Overall, the role behaviors included under each qualitative role behavior survey dimension loaded onto 3 or more factors of the PROBES; the majority of the role behaviors loaded onto a maximum of 2 factors. Individual role behaviors from each dimension, however, loaded onto additional factors, as indicated by the use of broken lines in the Figure. Although the research methods used for the 2 studies were different, the PROBES data generally appear to be consistent with that of the qualitative role behavior survey and fit reasonably well within the classification scheme.

[ILLUSTRATION OMITTED]

Discussion and Conclusions

Despite the methodological differences between the current study and the qualitative role behavior survey and the fact that approximately 1 year elapsed between the 2 studies, the findings are quite similar. Of the 26 role behaviors, only 4 demonstrated a different direction of change between the studies. This similarity in findings over time could indicate that the nature of the role behavior changes for hospital-based physical therapists might now be more a matter of degree rather than substance. This notion, however, would need to be examined at a later date to determine whether it is valid.

Of the remaining 4 role behaviors, 2 may be attributed to differences in methods between studies. The finding of no change in "time spent in patient evaluation and program planning" and "time spent in direct patient care" in the PROBES study could be considered inconsistent with the literature on clinical practice. Specifically, patient evaluation and program planning are generally believed to have increased, whereas time spent in patient treatment is believed to have decreased since the advent of hospital restructuring.[8,9,11,25] The inconsistency of the findings related to these behaviors between the current study and the previous research may be the result of the use of the words "time spent" in the PROBES instrument compared with the use of "importance of" or "focus on" in previous work.[11] Patient evaluation and program planning, in my view, is considered to be an important part of physical therapy practice in hospitals, and although the actual time spent may not have increased, the importance in patient care of this role behavior appears to have increased as the length of stay for patients in hospitals has decreased.[8,9,11] However, the remaining 2 role behaviors that were dissimilar between studies--"documenting the results of care" and "communication/collaboration with other health care professionals"--were perceived to have increased, which is consistent with what I view as a current clinical belief and may reflect the evolution of these role behaviors toward greater interaction with other professionals.[8,9,25]

The analysis of the data from the current study indicates that the survey instrument made up of the 26 role behaviors has good overall internal consistency and content validity in relation to the role change that has occurred for physical therapists working in restructured acute care hospitals. Thus, I believe that the PROBES provides a practical tool for use in future research concerning the effect of role behavior change and outcomes that are important and relevant to the organization. Given the paucity of research in this area, however, the ability to confirm the construct validity of this survey instrument is currently limited. Furthermore, I believe that care must be taken in using this survey instrument for practitioners in other disciplines and in other areas of clinical practice, because the PROBES was designed to measure role behaviors of physical therapists in a hospital setting. At a minimum, researchers who are interested in using this instrument to measure practitioner role behaviors in these other professions will need to validate the survey items for their particular sample.

Finally, the data suggested a factor structure with similarities to the role behavior dimensions previously reported by Lopopolo.[11] Moreover, the factor model in the current study provided for clearer differentiation of role behaviors by underlying dimensions than previous work. Beyond providing a general sense of the relationships among the survey items, I contend that generalizing the data on factors to a presumption of the existence of multiple underlying constructs is not warranted. Three reasons lead me to caution against making this assumption. First, the factor model accounted for less than 50% of the total variance in the data. Second, a clear separation of role behaviors by factors was not achieved. Third, there were weak relationships among the role behaviors for several of the factors. Thus, these findings favor viewing the survey as a single construct rather than as multiple constructs.

In summary, the 26-item PROBES appears to provide a useful measure of the role behaviors of physical therapists in today's hospital environment. It demonstrates both consistency with previous work describing the nature of role change following hospital restructuring and internal consistency. The next phase in this line of research would be to use the PROBES to ascertain whether the perceptions of role behaviors held by practicing clinicians are the same as those held by the clinical managers surveyed in this study. Certainly, the perceptions of the actual role incumbents regarding the nature of their roles are important. This is especially true if we are interested in determining whether or how organizational change is affecting practitioners' roles and their feelings about organizationally relevant outcomes such as satisfaction with their jobs, commitment to their organizations, or even commitment to their occupations.

(*) SPSS Inc, 444 N Michigan Ave, Chicago, IL 60640.

References

[1] Katz D, Kahn R. The Social Psychology of Organizations. 2nd ed. New York, NY: John Wiley & Sons Inc; 1978:chap 7.

[2] Aiken LH, Fagin CM. Evaluating the consequences of hospital restructuring. Med Care. 1997;35(suppl 10):OS1-OS4.

[3] D'Aunno T, Alexander JA, Laughlin C. Business as usual? Changes in health care's workforce and organization of work. Hospital and Health Service Administration. 1996;41(1):3-18.

[4] Leatt P, Baker GR, Halverson PK, Aird C. Downsizing, reengineering, and restructuring: long-term implications for healthcare organizations. Frontiers in Health Services Management. 1997;13(4):3-37, 52-54.

[5] MacStravic RS. Product-line administration in hospitals. Health Care Management Review. 1986;11:35-43.

[6] Tammy-Harrington S. Changing with the times. PT Magazine. 1999; 7(9):40-44.

[7] Schwiekhart SB, Smith-Daniels V. Reengineering the work of caregivers: role redefinition, team structures, and organizational redesign. Hospitals and Health Services Administration. 1996;41(1):19-35.

[8] Arthur RA. Patient focused care: acute orthopedic services. PT Magazine. 1994;2(6):34-47.

[9] Lopopolo RB. The effect of hospital restructuring on the roles of physical therapists in acute care. Phys Ther. 1997;77:918-936.

[10] Wynn KE. Embracing change: hospital restructuring revisited. PT Magazine. 1997;5(1):38-43.

[11] Lopopolo RB. Hospital restructuring and the changing nature of the physical therapist's role. Phys Ther. 1999;79:171-185.

[12] Kim S, Price J, Mueller C, Watson T. The determinants of career intent among physicians at a US Air Force hospital. Human Relations. 1996;49:947-976.

[13] Rizzo JR, House RJ, Lirtzman SI. Role conflict and ambiguity in complex organizations. Administrative Science Quarterly. 1970;15(2): 150-163.

[14] Turner RH. Role change. Annual Review of Sociology. 1990;16:87-110.

[15] Brannon RL. Restructuring hospital nursing: reversing the trend toward a professional work force. International Journal of Health Services. 1996;26(4):643-654.

[16] Jones KR, DeBaca V, Yarbrough M. Organizational culture assessment before and after implementing patient-focused care. Nurs Econ. 1997;15(2):73-80.

[17] Davidson H, Folcarelle PH, Crawford S, et al. The effects of health care reforms on job satisfaction and voluntary turnover among hospital-based nurses. Med Care. 1997;35:634-645.

[18] Shindul-Rothschild J, Duffy M. The impact of restructuring and work design on nursing practice and patient care. Best Practices and Benchmarking in Healthcare, 1996;1(6):271-282.

[19] Gutek BA, Winter SJ. Consistency of job satisfaction across situations: fact or framing artifact? Journal of Vocational Behavior. 1992;41: 61-78.

[20] Reger R, Gustafson L, Demarie S, Mullane J. Reframing the organization: why implementing total quality is easier said than done. Academy of Management Review. 1994;19(3):565-584.

[21] Lau C, Woodman R. Understanding organizational change: a schematic perspective. Academy of Management Journal. 1995;38(2): 537-553.

[22] Lathrop JP, Seufert RJ, MacDonald RJ, Martin S. The patient-focused hospital: a patient care concept. Journal of the Society for Health Systems. 1991;3(2):33-50.

[23] Lepine S, Ahola-Sidaway J. An exploratory study of trends in hospital nursing and their impact on professional development initiatives [preliminary draft]. Paper presented at: Third International Interdisciplinary Qualitative Health Research Conference; October 30-November 1, 1996; Bournemouth, England.

[24] Arthur PR, Brumfield J, Bullock K in Wynn KE. Restructuring update. PT Magazine. 1997;5(1):44-50.

[25] Bullock K. Patient focused care: hospital wide. PT Magazine. 1994; 2(11):51-60.

[26] Meyer JP, Allen NJ, Smith CA. Commitment to organizations and occupations: extension and test of a three-component conceptualization. J Appl Psychol. 1993;78:538-551.

[27] Levine A. A model for health projections using knowledgeable informants. World Health Statistics Quarterly. 1984;37:306-317.

[28] Couper MR. The Delphi technique: characteristics and sequence model. Annals of Advanced Nursing Science. 1984;7:72-77.

[29] DeVellis RF. Scale Development: Theory and Application. Newbury Park, Calif: Sage Publications; 1991.

[30] Nunnally JC, Bernstein IH. Psychometric Theory. 3rd ed. New York, NY: McGraw-Hill Inc; 1994.

[31] Cryer JD, Miller RB. Statistics for Business: Data Analysis and Modeling. 2nd ed. Belmont, Calif: Wadsworth Publishing Co; 1994.

[32] Robinson JP, Shaver PR, Wrightsman LS. Measures of Personality and Social Psychological Attitudes. Vol 1. San Diego, Calif: Academy Press Inc; 1991.

[33] Hair JF, Anderson RE, Tatham RL, Black WC. Multivariate Analysis. 3rd ed. New York, NY: Macmillan Publishing Co; 1992.

RB Lopopolo, PT, PhD, MBA, is Associate Professor, Department of Physical Therapy, Arcadia University, 450 S East0n Rd, Glenside, PA 19038 (USA) ([e-mail]).

This study was approved by the Arcadia University Committee on the Protection of Research Subjects. The rights of human subjects were protected.

This article was submitted January 10, 2000, and was accepted January 14, 2001.