Modeling physical therapy clinical research centers.

Until recently, no physical therapy clinical research center (PTCRC) had been established in the United States. (*) This is in spite of the fact that the American Physical Therapy Association (APTA) Plan to Foster Clinical Research in Physical Therapy, which was adopted in 1978, [1] included a goal regarding the development of PTCRCs. Further, since April 1980, physical therapists have had available to them the document Guidelines for Establishing Physical Therapy Clinical Research Centers. [2] Another document, Goals and Objectives for Physical Therapy Clinical Research Centers, approved in 1984 by the APTA Board of Directors, states

Solutions to important physical therapy clinical problems, improvements in the quality of physical therapy care, contributions to the science and theory of physical therapy, preparation of clinical investigators in physical therapy, acceleration of the development of clinical research in physical therapy, and enhancement of the credibility and visibility of physical therapy clinical research can best be served not by redirecting or influencing existing clinical research centers but by establishing physical therapy clinical research centers under the direction of physical therapists. [3]

Various non-physical therapy clinical research centers exist today in the United States and are funded through a variety of mechanisms. [4,5] These centers traditionally receive directives regarding research priorities (categorical funding), or they receive broad-based (noncategorical) funding and determine their own research agendas. Categorically funded centers exist in the National Institutes of Health (NIH) in specific program areas. For example, the National Cancer Institute (Cancer Control Centers), the National Heart, Lung, and Blood Institute (Lipid Research Centers), the National Institute of Child Health and Human Development (Maternal and Child Health Pregnancy Centers), and the National Institute of Diabetes and Digestive and Kidney Diseases (Diabetes Research and Training Centers) have a long history of development and monitoring of clinical research through agency-directed initiatives. Other agencies, such as the Arthritis Foundation, have invested in the future of investigators and clinical research centers by awarding broad-based undesignated funds to promote and further develop the current research agenda of the particular site. [6] In addition, each of these agencies provides new investigator awards and a number of other avenues for individuals to continue their research careers.

The use of systems analysis has been well established and can be used for a description of the components of clinical research centers. [7,8] The approach is used to determine the important components of the system under study. [9,10] This approach uses a basic systems framework from which each existing model of a clinical research center can be analyzed. Common components are then grouped and form the basis on which further data are collected. We examined NIH centers for common systems components, and, as the model for integrated research efforts, these components formed the basis of our Delphi survey. Delphi methodology was selected because the

...intended purpose [of the Delphi technique] is to make the best use of a group of experts in obtaining answers to questions requiring reliance, at least in part, on the informed intuitive opinions of specialists in the area of inquiry. The Delphi technique was designed to accomplish this with a minimum of interference from the kind of psychological distractions that usually attend open-forum discussions among panels of experts, and to achieve as close a consensus as possible compatible with individual divergences from the central tendency of the panel's opinion. [6]

Thus, we selected this approach as the most cost-effective and efficient method of achieving the goal of this project. Examples of applications of this process can also be found in the literature. [11,12]

We believe that, if PTCRCs are to develop, we must be familiar with and assess existing models and their applicability to physical therapy. Because multiple models exist for other disciplines, it is feasible that a number of models may be appropriate for physical therapy. To address the matter of developing PTCRCs, this project (1) assessed available resources for the formation of PTCRCs, (2) completed a systems analysis, and (3) applied the results of a Delphi survey to the resulting PTCRC models.

Methodology

A conceptual framework guided the methods used in this study. Generally, we sought to determine whether the mechanisms and personnel in physical therapy had features and characteristics similar to those of other disciplines within the health sciences. More specifically, we used the conceptual framework shown in Figure 1. Many of the same characteristics of medically oriented research centers will also influence the development of PTCRCs (eg, factors associated with personnel, facilities, philosophies, funding, and so forth) are depicted in Figure 1. Our methods consisted of (1) a questionnaire survey of physical therapy academic programs and rehabilitation facilities, (2) a systems analysis with a clinical component or focus within the NIH, and (3) a Delphi survey of a panel of physical therapy researchers. The selection and ordering of these techniques allowed us to gather data about existing resources and compare these resources with those required for the formulation of PTCRCs. Each of the techniques used is described in detail in the following paragraphs.

During the initial stage of the project, we surveyed the 118 academic institutions that housed physical therapy education programs and 22 other major rehabilitation facilities that were identified from listings and registries of units offering rehabilitation services. No follow-up was done of nonrespondents. The purpose of this phase of the study was to determine the current state of academic and clinical research centers. The survey questionnaire did not specifically define PTCRCs or their contents, but addressed topics such as the status of PTCRCs and elements of organization as related to such centers at each facility. Scholarly productivity, the likelihood of developing a PTCRC, and the identification of impediments to the development of a PTCRC were among the items included in the questionnaire (Appendix 1).

During the spring and summer of 1988, we completed a systems analysis of research centers with a clinical component or focus within the NIH for the purpose of determining the center models used in medical research. We then developed a description of the components of clinical research centers within each specific context (ie, categorical, noncategorical, or nonfunded). Within each description, the context within which the system existed was analyzed and system models were developed for each type of center encountered.

In order to prepare the Delphi survey, the data derived from the questionnaires received from the respondents in physical therapy settings were subjectively correlated by one investigator (GLS) with the data from the systems models. In this stage of the project, we assessed how the models may apply to physical therapy. Some of these prospective models were further developed and discussed at a presentation during the Combined Sections Meeting of the APTA in 1989. Attendees were asked to further identify categories of PTCRCs that were to be studied by a group of "experts."

The Delphi survey we developed was used to systematically query a panel of physical therapy researchers regarding the components of a PTCRC from a systems perspective and to achieve consensus regarding the degree of importance of specific characteristics. [13] Nomination forms (N=137) were sent to a random Delphi methodology sample from the APTA Section on Research, to all present and past officers of the APTA Section on Research of the last 5 years, to all past members of the Research Committee of the APTA, to physical therapy researchers

Table 1. Administrative Units Identified by Respondents (a)
                                                 No. of
Administrative Unit                              Respondents
Division/College/School of Allied Health
  Professions/Health-Related Professions         52
Department of Physical Therapy                   13
School of Medicine                                6
College of Arts and Sciences/Liberal Arts         2
Department of Rehabilitation Services             2
Health Affairs                                    1
School of Health and Social Work                  1
State university                                  1
Department of Orthopedic Surgery and
  Rehabilitation                                  1
  (a) Of 105 respondents, only 79 provided a specific response.

who have published in Physical Therapy, and to three members of the Foundation for Physical Therapy Inc's Board of Trustees. Seventy responses were received, which yielded 233 nominations. Eighty-eight of these nominations were single nominations, and 65 were self-nominations. Thirty-two responses included nominations for more than one person. Helmer [13] states that once the criteria for selection have been established and a defined population from which nominations are to be selected has been established, those

[TABULAR DATA OMITTED]

who nominate themselves as experts are most likely to be experts. The next group most likely to be experts are those receiving multiple nominations. [13] Ninety-seven individuals either nominated themselves or were nominated by more than one peer.

The Delphi survey consisted of open-ended questions regarding the components of a PTCRC. The components were derived from the systems analysis of NIH clinical research centers completed in the second phase of the project. A cover letter and the first round of the Delphi survey were sent to the 97 individuals identified as "experts," of which 82 responded and became the study sample. Round 1 responses were collated and returned to the sample for review and additions as round 2. Like items were collapsed, round 2 suggestions regarding component items and current lists were considered, and new items were recorded. In the final round (round 3), the participants were asked to rank items according to appropriateness and importance. Round 3 also included a demographic survey, which was completed and returned by 60 of the participants.

Data analysis consisted of tallies of responses to questionnaire items. The systems analysis required subjective assessments by one investigator (JMW). The Delphi survey used data classification and interpretation of responses. Where appropriate, tallies were made and descriptive statistics were calculated.

Results

Of the 140 questionnaires distributed during the first survey of institutions, 105 were returned, a response rate of 75%. The distribution for those responding to the question about administrative unit is shown in Table 1. Several questions were answered relative to the status of the development of PTCRCs (Tab. 2). Regarding funding, respondents identified sources from which funding has been received (Tab. 3). Table 4 summarizes the responses associated with the questions about the degree of scholarly productivity of the personnel associated with the facilities. When asked about the likelihood of developing a PTCRC at their institution or facility, 38% of the respondents indicated a high likelihood of developing a PTCRC. Thirty-eight percent and 24% of the respondents thought that there were medium and low likelihoods, respectively, of developing a PTCRC. The responses to the question about the existence of impediments to the development of PTCRCs are summarized in Table 5. Finally, in response to the question about whether a specific model would need to be followed in the respondent's institution or facility, 41 of the 44 respondents indicated that no specific model would need to be followed.

The systems analysis demonstrated the following common components of the system: (1) the environment in which the model exists, (2) the inputs (ie, time, money, resources), (3) the throughputs (ie, the processes used to accomplish the goal, such as interdisciplinary, cooperative, and other arrangements), (4) the outputs (ie, type and volume of data the system generates), and (5) the evaluation system (ie, how the funding agency and the clinical research center measure effectiveness). These components provided the basis for the structure of the Delphi survey.

Results from the meeting on PTCRCs held at the APTA Combined Sections Meeting in February 1989 also influenced the formulation of the Delphi survey. Thirty attendees discussed various issues related to PTCRCs, including whether there should be a disease, systems, or other focus. General opinion was that the focus could be any topic relevant to physical therapy and

[TABULAR DATA OMITTED]

Table 3. Sources of Funding
Identified by Respondents (a)
                              No. of
Source                        Responses
Private agencies, including
  foundations                 20
Intrainstitutional            15
No funding, volunteered       15
Federal agencies              11
Other                          7
Interinstitutional             6
  (a) Answered only by those respondents whose
institution or facility is operating, formally or
informally, as a physical therapy clinical research
center.

that the selection of the focus should be left to the applicant. Cooperative constituency was also discussed, as were the criteria for review. An important final point was that these models should include patient access.

Rounds 1, 2, and 3 of the Delphi survey experienced 61%, 66%, and 74% return rates, respectively. The respondents were academic administrators, faculty members, researchers in private practice, and institutionally based researchers. Of the respondents who participated in round 3, 46 held doctoral degrees and 10 held master's degrees. An analysis of the number of publications in peer-reviewed journals (N=16) showed that the mode was tied at 0 and 1 (tab. 6).

The Delphi survey queried respondents regarding characteristics of 10 center components and subcomponents. Consensus was achieved regarding some of the characteristics of 6 of these components. That is, all 61 respondents who completed round 3 of the Delphi survey indicated that certain characteristics were appropriate or most appropriate (Appendix 2). By also looking at near-consensus (59 of 60 respondents), an additional component was characterized and the consensus elements further defined (Appendix 3). The only component elements that were determined by consensus to be inappropriate were two potential groups of collaborators--speech pathologists and chiropractors.

Delphi survey results indicated that there was no consensus or near-consensus among the expert panel for the following components: environment, principal investigator (PI), other

[TABULAR DATA OMITTED]

investigators, and nature of collaboration. Sixty-two percent of the respondents indicated that the environment should be an academic medical center; 97% felt that a doctoral degree was appropriate for the PI, and 52% indicated that a master's or baccalaurate degree was inappropriate. In addition, 67% felt the PI must have potential for future work, and 66% felt that the PI must show evidence of creativity, perseverance, and focus. In describing other collaborators, 51% indicated that they should have academic degrees commensurate with their responsibilities.

Discussion

We believe that the questionnaire response rate of 75% provides a representative view of the state of potential formation of PTCRs. Most of the non-respondents were from facilities in which a physical therapy entry-level education program did not exist. Most of the administrative units are located in the allied health professions and may thus be less likely to be recognized or identified as autonomous units.

When considering the answers to questions about the state of development of PTCRSs (Tab. 2), note should

Table 5. Impediments to Development of Physical Therapy Clinical Research Center Perceived to Exist

                          No. of
Impediment                Responses
Funding                   35
Administrative
  structure/philosophy    18
Space                     18
Research expertise        13
Personnel/staffing        11
Patient availability      10
Time allocation            9
Equipment                  7
Faculty                    7

be made that the responses are from physical therapists; therefore, there is no assurance that "administration" agrees with the development or endorsement of a PTCRC. This is perhaps reflected when considering all the responses, which show that fewer than 50% of the facilities have discussed the concept. Further, only 18% have established committees to pursue the concept. Even fewer (12%) have developed structural organization/operations charts. Interestingly, 21 respondents answered that their facility was operating in principle, but without recognition, as a PTCRC. This response may reflect a parochial view or that the mechanisms associated with recognition are encumbered by the facility's administration.

The responses regarding funding shown in Table 3 are consistent with general knowledge about the level of funding associated with research in physical therapy. The respondents identified only 11 federal agencies and 20 private agencies and foundations that have funded physical theraphy-related research. Either additional numbers of applications for funding or higher-quality applications need to be generated in order to beter use the available funds. This may also be a consideration when establishing criteria for the success of any funded PTCRC, because sites designated as PTCRCs may not be able to generate additional funds to support projects or to guarantee ongoing research activities.

Responses to questions about scholarly activity (Tab. 4) were extremely varied, but probably were proportionally related to the mission and goals of the respective institutions or facilities. Fewer than 10 sites have established records in communicating the results of scholarly endeavors. The finding that only 4 site each had more than 20 currently active research projects at the time of the survey does not provide much encouragement that there will be a fast rate of development of new knowledge. Therefore, although an argument may be made for allocation of funds to more than one facility for the purpose of fostering PTCRCs, only a select number appear to have sufficient publication records--and credibility--to justify awards. Further emphasis on research must be forth-coming if institutions are to become competitive for the larger fiscal awards and recognition usually associated with clinical research centers.

The belief of 38% of the responents that there is a high likelihood of developing a PTCRC was based entirely on conjecture. The perspective of respondents regarding the requirements for a PTCRC may have been quite broad, considering that no definition of a PTCRC may have been quite broad, considering that no definition of a PTCRC was provided within the cover letter or the survey questionnaire. In response to survey questions about the status of PTCRCs (Tab. 2), 21 respondents indicated that their institution or facility was operating in principle as a PTCRC. An additional 19 respondents indicated that the development of a PTCRC at their institution or facility is highly likely given their local circumstances. Considered in relation to responses regarding scholarly activity (Tab. 4), showing fewer than four facilities with an established publication record, the number of 40 PTCRC facilities would appear to be excessive. The few CRCs in other disciplines throughout the country would support the position that 40 facilities is too large a number. Given the state of faculty preparedness, resources, commitments, and philosophies, as indicated by the survey responses in this study, this number seems unreasonably high.

Question 10 of the survey asked the respondent to indicate the impediments to the development of a PTCRC that wefre perceived to exist (Tab. 5). Responses were formulated in many different ways, for example, "funding," "budget," and "resources." Common responses were placed into general categories for purposes of compilation. Using this method, funding was considered to be the greatest constraint. Unfortunately, the questionnaire did not ask a follow-up question about the sources that the respondent perceived to be available to solve this problem. Personnel, in terms of faculty number, expertise in research, and staffing, also appeared to be a significant impediment. The difficulty of recruiting and retaining personnel in health science centers, and in academia in particular, is a well-known problem. The fact that 18 respondents identified administrative structure and philosophy as an impediment to the establishment of a PTCRC was correlated with responses to the question regarding the likelihood of establishing a PTCRC. We found that this structure and philosophy impediment was associated with responses indicating a low likelihood of developinga PTCRC.

That 41 of the total of 44 respondents indicated that no specific model of PTCRC would need to be followed seems to provide a maximum amount of flexibility in creating structures. This finding should be taken into account when funding and establishing PTCRCs. Our questionnaire, revealing that 64% of the respondents had an interest in having a vist to their institution for the purpose of consultative assistance, identified a high level of interest in fostering research.

In reviewing the results of the Delphi survey, we noted the inclusiveness associted with the distribution of the categories of respondents (Tab. 6). The results demonstrate that there is a population of therapists who consider themselves to be researchers in a private-practice environment. Although this area was perhaps not as thoroughly considered in our models as it could have been, these individuals may be employed in such a capacity as a guest lecturer or their institution or facility may be a clinical affiliation site.

Only 12 respondents stated that their institution afforded postdoctoral research opportunities. This may be one of the most significant barriers for physical therapists in accessing funding from NIH and other governmental agencies. The review process at the federal level is bound by the philosophy that the only way one can truly appreciate the state of knowledge in a particular area is by studying with an expert and then extending that work beyond what is currently known. Although the availability of postdoctoral research opportunities is perhaps a minor issue, it is a factor that should be considered in the establishment of a PTCRC.

Further review of the data in Table 6 showed that scholarly productivity, as measured by the number of peer-reviewed publications, was rather low, the generation of publications being no greater than 2.9 for any single category and the mode being tied at 0 and 1. This statistic was biased by the finding that several productive people reported more than 10 publications per year. Note should also be made that less than 50% of scholarly effort was found to be committed to research when the respondents were considered across categories. This is probably a critical factor in limiting the productivity.

Data revealed an equal mix of respondents who wanted to maintain the word "clinical" in the name of the center and those who wanted to emphasize the word "research" and deemphasize "clinical." The range of suggestions made is important because individual centers, investigators, or foundations/institutes may find that some of these items are more representative of a center when applying their criteria to a particular environment or focus. The list suggested an almost endless series of possibilities for models and for further research.

Data did not give a clear idea of the model of the "ideal" PTCRC. Whole components such as environment have not been included in the results, because there was no clear consensus. This led us to look for mean rankings to clarify the models. In addition, no single component received consensus under any one ranking. In order to consider consensus, the rankings had to be combined as the two highest and two lowest parts of the scale. The low ranking of chiropractic can probably be explained by the general view that physical therapists hold of chiropractic. The ranking of speech pathology as a least appropriate inclusion is perhaps due to the limited number of respondents working with patients who have had head trauma or stroke or holding the view that swallowing and speech formation is less of an integral part of physical therapy. The primary purpose of the Delphi technique is to try to gain consensus from a group of experts. Perhaps our group of experts was too large or they were not all experts, or perhaps physical theraphy research is so diverse that there is no true consensus. Our opinion is that the latter is the most logical explanation. The suggestion is that diversity should be valued in physical theraphy research and that PTCRC funding should support diversity.

In order to understand how the respondents ranked the individual components of the Delphi survey, the responses were tallied and means were calculated. Appendix 4 shows those components that received 29% or more of the highest ranking (ie, "must have," "most important," and so on). Reporting these means allowed us to evaluate the highest ranking, something we could not do with the consensus rankings, because no single item received consensus within any one ranking. Our analysis of the data in Appendix 4 led us to develop a variety of models that are dependent on whether the major factor is setting (first model), configuration (second and third models), focus (fourth model), or initiative (fifth model) (Fig. 2). In view of the responses, any of these options would be possible and could lead to different constructs and components within the construct. Formulation of a structure may be considerably dependent on the instituion in which PTCRCs are formulated. Whether a granting foundation would like to follow a model more in concert with NIH, the National Institute for Disability Rehabilitation and Research, or other federal agencies or adopt an independent system would appear to have significant implications. Perhaps a major consideration for any foundation or the appliant institution or facility will be the requirements of the funding agency.

To establish applicability of the models and to measure institutional readiness to become a PTCRC, visits were made to three sites that had demonstrated that at least initial steps had been taken to establish a PTCRC. This evidence was manifested via organization, personnel, and research productivity. Location and fiscal resources were not factors in the selection of sites to be visited. Based on responses to the initial questionnaire, one program in physical therapy, one physical therapy department, and one rehabilitation institute were selected for visitation. The project director communicated with facility personnel and disseminated a copy of the objectives for the visit. In all cases, on-site discussions included the concept of PTCRCs with facility and physical theraphy administrators, physical therapy staff and faculty, and personnel from other disciplines within the facility. Assessed was the applicability of the models, the facility's philosophy, and ability of the personnel to participate in a PTCRC.

These site visits reinforced the concept of diversity. Each facility presented unique considerations in the areas of milieu, philosophy, personnel, articulation with other disciplines, and available resources. Although diversity is an important concept, our discipline perhaps needs to identify specific points upon which to focus. This would prevent a broad, very nonspecific approach to the solution of problems commonly encountered by physical therapists in the clinical community and allow for the development of a specific focus within any given institution or facility developing a PTCRC.

A number of other points were also discovered as a result of the site visits, including how to incorporate personnel from other disciplines in the PTCRC is their work is relevant and if they want to be included. Another point was the educative role of PTCRCs and how the local and nationwide communities of physical therapists would look to the PTCRC for advice and counsel.

Summary

This study evaluated the resources and models associated with the development of PTCRCs. Facilities responded to a questionnaire, and individuals participating in a Delphi survey reached a consensus or near-consensus regarding the essential components of PTCRCs. In view of the finding that no single model would be appropriate, several possible models were presented. Although selected facilities may be able to establish such centers, deficiencies in number and quality of personnel, resources, and productivity may limit the development of many PTCRCs.

(*) The first PTCRC in the United States was established at The University of Iowa in April 1991 by a grant from the Foundation for Physical Theraphy Inc.

References

[1] Plan to Foster Clinical Reserch in Physical Therapy. Alexandria, Va: American Physical Therapy Association; 1978.

[2] Guidelines for Establishing Physical Therapy Clinical Research Centers. Alexandria, Va: American Physical therapy Association: 1980.

[3] Goals and Objectives for Physical Therapy Clinical Research Centers. Alexandria, Va: American Physical Therapy Association; 1984.

[4] Annual Report of the Artbritis Foundation. Atlanta, Ga: Arthritis Foundation; 1986.

[5] Annual Report of NCI, HLBI, NICHD, NIDDK Bethesda, Md: National Institutes of Health; 1986.

[6] Grant Awards Program Atlanta, Ga: Arthritis Foundation.

[7] Miser HJ, Quade ES. Handbook of Systems Analysis: What Is Systems Analysis? New York, NY: Elsevier Science Publishing Co Inc; 1976;1: chap 1.

[8] Immegart GI, Pilecki FJ. An Introduction to Systems for the Educational Administrator. Menlo Park, Calif: Addison-Wesley Publishing Co Inc; 1983:45-47.

[9] Churchman CW. The Systems Approach. New York, NY: Dell Publishing Co Inc: 1968.

[10] Bertalanffy L. General System Theory. New York, NY: George Braziller Inc; 1968.

[11] Braddom CL, Braddom RL. A Delphi survey of traits of effective physiatric leaders. Arch Phys Med Rehabil. 1986;67:793-795.

[12] Miles-Tapping C, Dyck A, Brunham S, et al. Canadian therapists' priorities for clinical research: a Delphi Study. Phys Ther. 1990;70:448-454.

[13] Helmer O. Looking Forward. Beverly Hills, Calif: Sage Publications Inc; 1983.

GL Soderberg, PhD, PT, FAPTA, is Professor and Director, Physical Therapy Graduate Program, College of Medicine, The University of Iowa, 2600 Steindler Bldg, Iowa City, IA 52242 (USA). Address all correspondence to Dr Soberberg.

JM Walter, EdD, PT, was Associate Professor, Physical Therapy and Division of Clinical Immunology and Rheumatology. The University of Alabama at Birmingham, Birmingham, AL 35294, at the time of he study. The is now Director and Associate Professor, Department of Physical Therapy, Oakland University, Rochester, MI 48309.

This research was supported by a grant from the Foundation for Physical Therapy Inc.

This article was submitted August 23, 1990, and was accepted June 4, 1991.