One-time physical therapist consultation in primary health care.

In the Netherlands, as in many other countries, primary care physicians (PCPs), also called "general practitioners" or "family doctors," occupy a pivotal position in the provision of care. (1-3) Their role is to focus on patient medical diagnosis and provide intervention, prevention, and continuity of care. (1-3) The current emphasis on cost containment in health care has, in some settings, led to an increase in the role of PCPs as "gatekeepers." Because most health care services in the Netherlands are accessed through written referral from the PCP, there appears to be an assumption that PCPs have a knowledge of the scope and breadth of practice of other health care professionals.

Patients with disorders of the musculoskeletal, cardiovascular, and pulmonary systems represent the most prevalent groups managed by PCPs in the Netherlands, (3) where approximately 80% of patients seen by physical therapists are referred by PCPs and 20% are referred by medical specialists. (4) Twenty-five percent of patients with disorders of the musculoskeletal system in the Netherlands are referred by PCPs to other health care professionals: 18% are referred to physical therapists, 6% are referred to medical specialists, and 1% are referred to other professionals. (1,4-6) Problems between PCPs and these disciplines, however, have been identified, including poor communication, (1,7) insufficient PCP knowledge about physical therapy, (1,8,9) unclear indications for referral,(1,8,9) and questions about the efficacy of physical therapy intervention. (1,10-13) Because of this referral rate, we believe communication between PCPs and physical therapists is crucial to address these problems.

Because in the Dutch health care system physical therapy can only be provided following a physician referral, a request by a PCP for a one-time physical therapist consultation prior to referral for treatment by a physical therapist or other health care professionals appears to provide one option for improving the appropriateness of patient referral. (1,14) If the PCP is uncertain as to whether physical therapy intervention will be beneficial for a patient, a consultation mechanism provides a valuable tool to enhance the communication between both professionals. This enhanced communication could improve the utilization of physical therapy services and help to develop a sense of shared care. (1,14)

Questions from PCPs about what is an appropriate referral for physical therapy intervention can prevent unnecessary referrals, and we believe this emphasizes the need to develop strategies to improve utilization of physical therapy services. (1,6-15) This need is illustrated by large variations in the kinds and numbers of patients referred by PCPs to physical therapists. (1,15-21) Factors that may contribute to the variation in PCP referrals include: the patient's health status, (1,15-21) the PCP's sociodemographic and personal characteristics, (1,18,19) the PCP's approach to patient care, and the PCP's attitude about physical therapy and decisions (beliefs) about referral to a physical therapist. (1,8,9,20,21) Furthermore, there is evidence to suggest that PCPs will refer more patients to physical therapists when they have more knowledge about physical therapy, recognize physical therapists' capabilities to diagnose, and believe in the benefit for patients and effectiveness of physical therapy intervention. (1,9-15,18-21)

By enabling PCPs to consult physical therapists tot care, a new dimension in the relationship between PCPs and physical therapists can be introduced. A referral for a one-time physical therapist consultation in the context of this study was defined as a written request by a PCP to a physical therapist to examine and evaluate a patient to generate information regarding (functional) diagnosis and prognosis to facilitate the management plan of the PCP, specifically when it concerned the possibilities for physical therapy interventions. (1,14,22) A one-time consultation was not intended to create access to the physical therapist without physician referral. Within the Dutch health care system, physical therapy is accessible only after referral by a physician for physical therapy intervention. The PCP retains final responsibility for the continuity of care of the patient. In our study, consultation was designed for PCPs who said they would like to make appropriate use of the expertise of the physical therapist to enable them to make better decisions with regard to the course of patient management.

Use of physical therapist consultation by PCPs does not appear to have been studied or described in the literature. (1,14,23) We investigated the frequency of use of a one-time physical therapist consultation, the kind of information requested by PCPs, their opinions on the benefit and process of consultation, and their management decisions or referral patterns following the consultation. In addition, we explored to what extent the demographic and personal characteristics of the PCP and the physical therapist and the demographic and clinical characteristics of the patient influenced PCP referral for patient management, the PCP's opinions on the outcome of consultation, and the number of consultation requests.

Method

Design

An observational study was conducted during a 7-month period in which PCPs were given the opportunity to use physical therapists as consultants. A random sample of 72 physical therapists in private practices in primary care was obtained from a list of all registered physical therapists in 4 different regions of the Netherlands (N = 1,533). In the Netherlands, primary care is structured by the PCPs, who select patients for referral to professionals in other disciplines in primary care (eg, physical therapists, podiatrists, psychologists, occupational therapists) or to medical specialists in hospitals (or secondary care). The physical therapists who participated in our study received patient referrals from the PCPs. The selected physical therapists were stratified by degree of urbanization (ie, rural, semiurbanized, or urban) because PCP referral patterns tend to be related to the geographical location of both PCPs and patients. (1,24) Once a physical therapist agreed to participate, a PCP was randomly selected from a list of PCPs who usually referred to that therapist and was paired with the therapist for the purposes of the study. If a selected PCP refused to participate, another PCP was randomly selected from the list until a match was made. The physical therapists received an additional fee for each report following a referral for a one-time consultation to cover the costs of the additional time spent for the consultation. The PCPs received a set fee for training prior to the study (eg, training in study procedures and the use of the registration and communication forms) and for participating in the study, irrespective of the number of consultation requests they made. All participants (patients, PCPs, and physical therapists) signed an informed consent statement.

Representation of the Participants

The willingness to participate in the study was 86% (62/72) for the physical therapists and 74% (62/84) for the PCPs. Of the 62 PCP-physical therapist pairs originally created, 59 completed the study and were included in data analysis. Three PCPs withdrew from the study for personal reasons. Table 1 lists the characteristics of the participants. A comparison of the data obtained for the PCPs and physical therapists in our study with data for all PCPs (N = 6,548) (25) and physical therapists (N = 10,187) (26) in primary care in the Netherlands indicated that the PCPs in our study appeared to be representative of classifications of sex, age, time (in years) in practice, total number of patients in a PCP's practice, number of professional partners or colleagues in the same private practice, and mean referral rate for physical therapy intervention. The mean referral rate for physical therapy intervention of the group of PCPs was 104 (SD = 20.7) referrals per PCP per 1,000 patients per year. The physical therapists in our study were representative of classifications of sex, age, and type of postgraduate education for the total population of physical therapists in primary care in the Netherlands, hut they were more experienced in terms of years in practice. Seven of the 10 physical therapists who declined to participate in the study had less than 5 years of experience in practice. Single-practitioner physical therapy practices were under-represented when compared with the total population of physical therapy practices (8.5% versus 34.9%).

Process of Consultation

The process of a one-time physical therapist consultation by written communication was as follows. When a PCP was undecided as to the possible benefit of physical therapy, the PCP then decided to initiate a written consultation request to the physical therapist with which he or she was matched. The PCP communicated the reason for the request for consultation and the patient's clinical status to the physical therapist. On receipt of the request for consultation, the physical therapist examined the patient and evaluated the patient's condition and functional status. The physical therapist communicated the examination findings, diagnosis, and management recommendations for a plan of care in writing to the PCP. The PCP evaluated the information from the physical therapist and the recommendations of the physical therapist and decided on the preferred patient management (eg, whether or not to refer the patient for physical therapy intervention). No attempt was made by the researchers or the physical therapist to influence the PCP's management decisions or referral decisions. All patients received information about the purpose and procedures of the study.

Data Collection

Data were collected regarding the PCPs, the physical therapists, and the referred patients. Self-administered questionnaires given at the start and completion of the study, consultation requests, standardized forms, and treatment referral reports from health insurance agencies were used to obtain data. The development of the questionnaires was based on the literature). (1,8,9) Items for the questionnaires used in this study were field tested and selected following a feasibility study. (14) We did not investigate the reliability and validity of data obtained for the questionnaire items.

Data on the characteristics of the PCPs and the physical therapists were obtained by use of self-administered questionnaires given at the beginning of the study. These questionnaires identified personal and practice characteristics of both groups, as well as the PCPs' perceptions of physical therapy. The personal and practice characteristics included sex, age group, years in practice, postgraduate education, type of certified specialization(s), number of colleagues in private practice, and total number of patients in a PCP's practice. The PCPs' perceptions of physical therapists' knowledge, level of cooperation, and diagnostic capabilities as well as their belief in the efficacy of physical therapy intervention were measured on a 5-point Likert scale. The Likert scale responses varied from "poor" (1) to "excellent" (5). The PCPs' behavior in referring patients for physical therapy, in the case of diagnostic uncertainty or doubt about the indication for physical therapy intervention, also was measured on a 5-point Likert scale, with the scale responses varying from "never" (1) to "very often" (5). (1)

To describe the patient and the consultation, information was collected during the process of consultation by use of consultation forms. (1,14) Characteristics of the patient consisted of the following: sex, age group, health insurance, level of education, employment, and marital status. The health-related characteristics were: the PCP's medical diagnosis, duration of the complaint(s), prior diagnostic and therapeutic medical interventions, and prior visits to other health care professionals for the same complaint in the past 12 months. Variables that describe the characteristics of the physical therapist consultations were: the reason for each consultation, the date of request, the date of patient appointment, the date on which the consultation was provided to the PCP, the recommendations of the physical therapist, and the PCP's evaluation of the consultation process and the extent to which the consultation question was answered and reported.

The PCPs' consultation request form was developed according to the guidelines for referral of the Dutch College of General Practitioners (NHG). (27) The physical therapists' consultation report form was developed according to the guidelines for documentation of the Royal Dutch Physical Therapy Association (KNGF). (28) The medical diagnosis, referral diagnosis, referral data, and types of complaints were classified according to the International Classification of Primary Care (ICPC). (29) The patients' functional status was classified in terms of the International Classification of Impairments, Disabilities (Activities), and Handicaps (Participation) (ICIDH) (30) as proposed by Heerkens et al. (31) In a reliability study, (32) the intraobserver and interobserver reliability for identifying selected impairments and disabilities, in our view, were satisfactory for survey research. Kappa values varied between .40 and .91. (32)

At the conclusion of the 7-month study, questionnaires were given to the PCPs and the physical therapists to examine their opinions related to the appropriateness and feasibility of a referral for one-time consultation and consultation forms and the willingness of PCPs and physical therapists to continue with this process of consultation.

To evaluate the patient referral and treatment patterns following a one-time consultation, the health status and treatment characteristics of the patients in our study were compared with data of 17,201 patients included in the Netherlands' national database for physical therapy intervention. (33,34) Our comparisons primarily concerned the characteristics of the patients.

To investigate the PCPs' patient management decisions after each consultation, their management decisions were compared with their intended patient management prior to the physical therapist consultation. The intended patient management (referral decision) was defined as the hypothetical management decision the PCPs reported they might have adopted if a referral for a one-time consultation had not been available. The referral rate was defined as the numerator representing the actual number of referrals per year and the denominator representing the total number of patients in a PCP's practice. (1)

Data Reflecting One-Time Consultations

Dependent variables were: (1) the number and nature of consultation requests by the PCPs, (2) differences in patient profiles between patients referred for consultation compared with the national database of patients referred for physical therapy intervention, (3) PCPs' opinions on the benefit and process of consultation, (4) whether PCPs accepted the physical therapists' recommendations, and (5) PCPs' management decisions.

The dependent variables that we studied using linear or logistic regression analyses were: (1) the number of consultation requests; (2) the extent to which the PCPs thought the consultation question was answered satisfactorily (for the purpose of the analysis, the satisfaction measure was dichotomized into high ["excellent"] satisfaction versus low ["poor," "slight," "fair," or "substantial"] satisfaction); and (3) whether the intended (hypothetical) management decision as compared with the actual management decision and referral behavior after consultation was changed or not.

Data Analysis

Statistical methods to compare the characteristics and number of patients referred for a one-time consultation with the patients referred for physical therapy intervention (reference data) included chi-square tests for independent proportions for the categorical variables or Student t tests for comparison of means for the continuous variables. In all cases, a 2-tailed level of P [less than or equal to] .05 was set as significant.

Multiple linear regression was used to explore the relationships between the predictor variables (ie, characteristics of the patients, the consultation, the PCPs, and the physical therapists) and the number of referrals for a one-time consultation. Separate hierarchical logistic regression analyses were used to explore the relationships of the characteristics of the patients, the characteristics of the PCPs, and the characteristics of the physical therapists with: (1) PCPs' opinions of the benefit of consultation (high versus low satisfaction) and (2) whether PCPs' intended patient management decision was changed compared with the actual management decision (changed versus not changed). The predictor variables used for further analysis are listed in Tables 2 and 3. Table 2 also presents the relevant summary data for the classes of each predictor variable. All analyses were completed using SPSS-PC for Windows, version 6.1.2.*

Predictor variables were consecutively entered in 3 blocks: (1) characteristics of the patients and characteristics of the consultation, (2) characteristics of the PCPs, and (3) characteristics of the physical therapists. We included all predictor variables to examine and control for their effect.

Multi-level analysis was used because the data had an intrinsically hierarchical nature in which patients (lower level) are nested within PCPs (higher level). Because referred patients of one PCP are more alike than patients of different PCPs, the patients cannot, a priori, represent completely independent observations. By applying the statistical linear regression tool of hierarchical linear modeling, (35-37) clustering of data was taken into account. This permitted analysis of data without aggregation of patient data to the PCP level or distribution of patient characteristics to the PCP level (see, for example, Kerssens et al (38)). The strength of the associations was assessed by nonstandardized regression coefficients [beta] or odds ratios (ORs) and by probability values.

Results

Frequency of Use and Reason for Consultation

During the 7-month study, 352 patients were referred for a one-time physical therapist consultation by 59 PCPs (mean per PCP = 5.97, median = 5, range = 0-20). Six of the 59 PCPs did not refer any patients for consultation. None of the patients refused the referral for physical therapist consultation. Ten patients, however, preferred to see a physical therapist of their own choice and not one of the therapists participating in the study.

The mean referral rate for physical therapist consultations made was 4.9 per 1,000 registered patients per PCP per year (SD = 4.6, range = 0-21.4). The referral rate for consultation by PCPs located in urban areas (X = 6.3, SD = 5.6) was higher than in rural areas (X = 3.9, SD = 2.8) (P = .05). The main reasons for referral by PCPs for a one-time physical therapist consultation were for information on diagnosis and health status (92/352 [26%]), for information about the indication for physical therapy intervention (102/352 [29%]), or for both reasons (158/352 [45%]). Frequently mentioned additional reasons were: to obtain a second opinion from a physical therapist's perspective (96/352 [27%]), to support the intended management decision of the PCP (55/352 [16%]), and to prevent referral to a medical specialist (44/352 [13%]).

Patient Profiles

In Table 4, an overview is given of the demographic characteristics of the patients in this study compared with those from the Netherlands' national database for referrals by PCPs for physical therapy intervention. (33,34) The groups differed in age, level of education, employment, and marital status. The patient group undergoing a one-time consultation in this study had a greater percentage of patients under 25 years of age, a lower percentage of patients older than 65 years of age, and a greater percentage of patients with a high-level education compared with the national data.

In both this sample of patients referred for consultation and the national sample of patients referred for physical therapy intervention, (33,34) almost all patients were referred for consultation or intervention due to complaints of the musculoskeletal system (97.5% versus 87.5%). Very few patients were referred with disorders of the neurological system (2.5% versus 7.8%). In Table 5, an overview is given of the patients' health status-related characteristics. When compared with the national data-base, a greater percentage of patients in the consultation group were referred with knee pain, back pain, and "other pain in the leg." In 57% (196/352) of the patients referred for consultation, the duration of complaints was longer than 12 weeks, and very few patients had acute complaints (of less than 1 week in duration). Patients referred for consultation had less contact with physical therapists prior to consultation and had undergone fewer medical diagnostic evaluations or therapeutic interventions compared with the national database of patients referred for physical therapy intervention. Patient complaints of less than 13 weeks' duration and larger PCP patient base were associated with fewer referrals for a one-time consultation (Tab. 2). Nineteen percent of the variance in number of referrals for a one-time consultation was explained by these 2 variables.

Level of Satisfaction Regarding the Outcome

The level of satisfaction of the PCP regarding a one-time consultation was evaluated for each referral and via a questionnaire completed by the PCPs. Because 10 patients who were referred by a PCP did not receive a one-time consultation by a physical therapist in the study, the results of 342 of the 352 referrals could be evaluated. In 94% (321/342) of the consultation requests, the PCPs were of the opinion that the physical therapy reports had answered their questions "excellently" (219/342 [64%]) or "substantially" (102/342 [30%]). In the remaining cases, their opinions were reported as "fair" (11/342 [3%]) or "slight" (10/342 [3%]).

Predictor variables associated with PCPs' satisfaction with the consultation are given in Table 3. The PCPs were less likely to be satisfied with the referral for a one-time consultation when the request sought additional diagnostic information only. The PCPs were more likely to be satisfied with a one-time consultation when they had a poor knowledge about the possibilities of and indications for physical therapy intervention, when they were more positive about the diagnostic skills of physical therapists, and when they were initially less positive about their cooperation with physical therapists. The PCPs were more likely to be satisfied when the physical therapists had more experience in advising the PCPs (ie, physical therapists who had some experience in informal consultation), when the physical therapists had taken more postgraduate courses, and when the physical therapists had a certification in manual therapy. The final model explained 35% of the variance in PCP satisfaction. A stepwise analysis revealed that a larger percentage of explained variance was derived from characteristics of the physical therapists ([R.sup.2] = .22) than from the PCPs' characteristics ([R.sup.2] = .11).

Primary Care Physicians' Opinions on the Process of a One-Time Consultation

Data from the questionnaire administered at the conclusion of the study were based on the responses of 59 PCPs, including the 6 PCPs who did not use the consultation option, and indicated that all PCPs received all physical therapist consultation report forms within an appropriate amount of time. Forty-four percent (26/59) of the PCPs indicated that the consultation had been "often or very often" useful for their decision making, 30% (18/ 59) indicated that this had been the case "sometimes," and 26% (14/59) indicated that this had never been the case (including the 6 PCPs who have never used the opportunity of a one-time consultation). Almost all PCPs who made use of the consultation option (48/53 [90%]) indicated that their consultation requests were "substantially" or "excellently" answered by the physical therapists including their recommendations for patient management. Eighty-six percent (51/59) of the PCPs believed that the process of a one-time consultation could easily be incorporated into their daily practice. Eight percent (5/59) of the PCPs indicated that they would find this difficult, and 5% (3/59) had no opinion. Fifty-three of the 59 PCPs who made use of the opportunity of a one-time consultation were satisfied with the consultation request form for allowing the physical therapists insight into the patient's condition. The average amount of time taken to complete the form (including the research questions) was 3 minutes (range = l-6). Seventy percent of the PCPs indicated that they would like to continue having the opportunity of a one-time physical therapist consultation.

Physical Therapists' Opinions on the Process of a One-Time Consultation

Data from the questionnaire administered at the conclusion of the study were based on 53 of the 59 physical therapists, because 6 physical therapists did not receive a consultation request by their matched PCP. The results indicated that 94% (50/53) of the physical therapists believed that the process of consultation could easily be incorporated into their daily practice. Nineteen percent (10/53) of the physical therapists indicated that it was sometimes difficult to complete the consultation within the requested time period. Fifty-three of the 59 physical therapists who received a consultation request were satisfied that the standard consultation request and report forms allowed the physical therapists insight into the patients' health problems and the PCPs' management decisions and included a clearly stated consultation request and structured written communication.

The average amount of time needed to complete the history and physical examination (including the research questions and use of standard forms) was 55 minutes (range = 50-70). The time needed for writing the reports varied between 10 and 30 minutes. All physical therapists indicated that a one-time consultation offered them an important communication tool for helping them to tailor treatment for a patient and to engage in joint (shared) patient care.

Recommendations and Patient Management

Of the 342 consultations included in the analysis, in l0 reports the physical therapists recommended multiple options for the PCPs' management plan. The data analysis, therefore, was restricted to the 332 consultations that presented a clear choice for the PCPs. The PCPs accepted and implemented the recommendations of the physical therapists in 93% (310/332) of the cases, and there was no difference between the recommendations of the physical therapists and the actual management decisions of the PCPs ([chi square] = 3.16, df = 3, P = .368). In 7% (22/332) of the cases, the PCPs did not implement the physical therapists' recommendations. The physical therapists were of the opinion that physical therapy intervention was indicated in 70% (232/332) of the consultations. Table 6 compares the management decisions that the PCPs would have made if the physical therapists had not been consulted, with the actual decision made following the onetime physical therapist consultation. The PCPs indicated that they made different decisions after a consultation than they would have made prior to the consultation. After the consultation, they stated that they referred fewer patients to medical specialists (55 versus 101) and more patients to physical therapists (232 versus 181).

Table 6 shows, for example, that of the 101 patients whom PCPs would have referred to a medical specialist before consultation, 21 (21%) were referred to a medical specialist, 67 (66%) were instead referred for physical therapy intervention, and 11 (11%) were treated by the PCPs. For 49% (164/332) of the patients, the intended management decisions of the PCPs presumably changed due to the information and recommendations obtained through the physical therapist consultation.

Predictor variables associated with whether or not the intended management decision was changed are shown in Table 3. The PCPs were less likely to change the intended management decision if the age of the patient was low (younger than 25 years) and the level of education of the patient was low. The PCPs were more likely to change their intended management policy if they were satisfied with the outcome of the consultation and they had a higher frequency of patient communication per month with the physical therapist.

Discussion

The results of the study demonstrate that the PCPs who participated in the study were satisfied with the opportunity for a referral for one-time physical therapist consultation by written communication and expressed the view that it was potentially beneficial to their patient management. Both the PCPs and the physical therapists found that the procedure was easy to incorporate into their daily practice. The PCPs reported that they changed their management or referral strategy based on the physical therapists' recommendations in nearly 50% of the cases. The PCPs also reported that they referred less often to medical specialists than they would have without the opportunity of a one-time physical therapist consultation. Reported changes in management decisions were especially true of the PCPs who reported prior to the study that they were not completely knowledgeable about the role of physical therapy and the diagnostic capabilities of physical therapists.

The PCPs who indicated that their cooperation with physical therapists was not ideal prior to the study more often stated that they were satisfied with the physical therapist consultation and the relevance of information for directing treatment than PCPs who had previously had a good working relationship with physical therapists. In general, the PCPs were more satisfied when they had consulted more experienced and educated physical therapists. In this respect, the physical therapists with relatively little experience (less then 5 years in practice) more often declined to participate than those in the other "time in practice" categories. This differential participation rate may have influenced the generalization of the results. For the purpose of the analysis, satisfaction with the consultation was dichotomized into high ("excellent") satisfaction and low ("poor," "slight," "fair," or "substantial") satisfaction. This dichotomy was chosen because of the possibility that respondents would opt for the "substantial" option as a way to avoid expressing very positive (rating of "excellent") or very negative (rating of "slight" or "poor") opinions about the consultation. There is evidence that people are reluctant to express negative perceptions and to give socially desirable answers. (39-41)

Profiles of patients referred for a one-time consultation differed from those described in records of patients previously referred by PCPs for physical therapy intervention. Patients with acute or subacute complaints (less than 13 weeks in duration) and those from PCP practices having a larger number of patients were less often referred for a one-time consultation. Most frequently, patients with disorders of time musculoskeletal system were referred for a one-time consultation. The lower referral rate for consultation for patients with subacute or acute complaints may be related to the wait-and-see policy we believe many PCPs exhibit and the expected course for normal recovery. When time patient's recovery was delayed, it appears the PCPs believed the potential for physical therapy intervention following the consultation might prevent chronicity or disablement. The relevance of the number of patients in a PCP's practice may indicate the influence of increased workload affecting the PCP's management decisions). (1) The referral rate for a one-time physical therapist consultation by PCPs was substantially lower in rural areas than in urban areas, and this finding is in accordance with the literature about the influence of referral rates to medical specialists based on degree of urbanization of the practice location. (1,24)

In time 2-stage selection procedure for the participants, time PCPs having low referral rates for physical therapy intervention were under-represented. The study's sample of PCPs was chosen from the lists of PCPs who usually referred patients to the physical therapists who were selected for the study.

We believe it is plausible that PCPs with high referral rates refer patients to several physical therapy practices, potentially contributing to a selection bias of PCPs with higher referral rates for physical therapy intervention. However, additional analysis examining PCPs with low (lower quartile) and high (upper quartile) referral rates for physical therapy intervention did not show a difference in time frequency of use of one-time physical therapist consultation (4.4 [SD = 4.7] per 1,000 patients per PCP per year).

The referral rate for a one-time physical therapist consultation was low compared with the documented rate of referrals for intervention by physical therapists (4.9 versus 104 referrals per PCP per 1,000 patients per year) (1) or compared with a previous study (12 referrals per PCP per 1,000 patients per year), (14) but seems to be in line with referral rates for a one-time consultation of medical specialists. The referral rates by PCPs for a one-time consultation to medical specialists in the Netherlands (referrals per PCP per 1,000 patients per year) are: 9.7 for neurologists, 9.4 for orthopedists and orthopedic surgeons, 2.3 for rheumatologists, and 1.4 for physiatrists. (42,43)

To facilitate understanding of the actual referral rate to physical therapists and medical specialists following a one-time physical therapist consultation, documenting all PCP contacts with their patients during the 7-month study would have been desirable. However, we believe the documentation of such a volume of data would have had an adverse effect on the PCPs' willingness to participate in the study. For this reason, health insurance figures (PCPs' referral data and number of patients in their practice) and PCPs' management data following the consultation were used in order to place the least possible burden on the PCPs.

The number of patients in a PCP's practice may indicate how workload can influence a PCPs' behavior. The literature on the influence of the number of patients in a PCP's practice on their management decisions is scarce. Evidence suggests that there are fewer referrals per PCP contact in smaller practices. (1) In larger practices, however, the number of contacts per patient may be lower, but the number of referrals per contact increases. We conclude that the influence of the number of patients in a PCP's practice on referral rates is more likely to be determined by the number of contacts patients have with the PCP.

Our results indicate that patient management decisions of the PCPs were often changed or adjusted after the additional information was obtained through the physical therapist consultation. We found that PCPs who changed their intended management decision (Tab. 3) were more likely than not to be satisfied with the outcome of consultation. Although the opinions of the PCPs on the benefit of consultation do not provide complete insight into variations in demand for physical therapy services, in our view it is possibly the most important variable in relation to their management decisions.

The PCPs indicated they changed their management decisions after consultation in 49% of the cases. From the physical therapists' perspectives, there was no indication for physical therapy intervention in 30% (100/ 332) of the cases, and PCPs' decisions to refer to medical specialists prior to consultation were changed after consultation in 46% (46/101) of the cases. Our findings suggest that referrals to physical therapists rather than referrals to medical specialists could be beneficial in primary health care and that this could result from physical therapist consultations. In general, the PCPs were satisfied with the way in which the physical therapists answered their queries and how relevant that information was for directing referrals for intervention. Confirmation of a substitution effect on rates of referral, however, should be viewed with caution because of the nature of the study.

Based on the results of the study, we believe that improving access to physical therapy services through physical therapist consultation might reduce unnecessary referrals for physical therapy intervention and referrals to medical specialists. Primary care physicians also could have a better idea of services physical therapists can offer, and they therefore might change their referral patterns. A one-time physical therapist consultation may be beneficial when the PCP is uncertain about indications for physical therapy or the services physical therapists can provided in a changing health care environment. (10-13,44,45) In theory, however, a one-time physical therapist consultation also could lead to an increase in referrals for physical therapy intervention and overutilization of physical therapy. Based on the literature and the results of our study, there is some evidence to suggest that the implementation of a well-constructed process of consultation could improve the outcome of care. Accountability and quality assurance, we contend, are essential aspects of the consultation process, and the process should be subjected to peer review and system analysis by all participants at both local and national levels.

Further research is needed to determine if there are benefits of physical therapist consultation in primary health care--benefits for the patients and for other health care professionals. If evidence of the benefits can be documented, it could be used to convince policymakers and those who finance care of the value of PCPs' consultation with physical therapists. In the Netherlands, PCPs determine the need for physical therapy based on the biomedical or medical diagnosis of the patient. (1,44) A medical diagnosis alone, however, may not be a sufficient guide for intervention. (11,31-34,44-53) In recent years, physical therapists and other health care professionals have been encouraged to adopt a "biopsychosocial" model of health care, encompassing the physical, psychological, and social characteristics of the patient (54,55) and a functional approach of physical therapy. (44,48-55) Physical therapist consultation, therefore, could be educational for PCPs because it may improve and update their knowledge about the indications for physical therapy intervention.

The primary purpose of consultation is generally believed to be to improve the quality of care by making the expertise of physical therapists available to PCPs and patients. The actual decision to refer a patient for a one-time consultation depends on factors such as expertise, the medical environment, and the patient's wishes and needs.

Conclusion

The PCPs' management decisions changed following a one-time consultation with physical therapists about the indications for physical therapy intervention. Future studies are needed to assess the effects of a one-time physical therapist consultation over longer periods of time. Although a one-time consultation also, in theory, could result in unnecessary referrals and overutilization of physical therapy services, we expect that a revised referral pattern following physical therapist consultation may lead to more efficient utilization of physical therapy services.

Table 1.
Characteristics of the Primary Care Physicians (PCPs) and the Physical
Therapists

                                          Primary Care     Physical
                                          Physicians       Therapists
                                          (n = 59)         (n = 59)

Characteristic                            No.      %       No.   %

Sex
   Male                                   55       93      37    63
   Female                                  4        7      22    37

Age (y)
   <35                                     3        5      24    41
   35-45                                  31       53      28    47
   46-55                                  19       32       6    10
   >55                                     6       10       1     2

Time in practice (y)
   <5                                      6       10       4     7
   5-10                                   18       31      20    34
   >10                                    35       59      35    59

Postgraduate education (a) (yes)          46       78      55    93
   <3 courses                                              21    36
   Sports physical therapy                                 15    25
   Manual therapy                                          17    29

PCPs' perception (b) of:                  Median   Range
   Their knowledge about the               3       1-4
      possibilities of and indication
      for physical therapy intervention
   Diagnostic skills of physical           4       3-5
      therapists
   Efficacy of physical therapy            4       2-5
   Cooperation with physical therapist     4       1-5

Frequency of patient communication
   between physical therapist-PCP pairs
   (<2 per month)                         39       66%

Referral behavior toward physical         Median   Range
   therapy (c)

   Referral with uncertainty               2       1-5
      indication
   Referral for appropriateness            2       1-5
      indication
   Advise by telephone prior to
      decision of referral                 2       1-5

No. of partners
   Single-handed practice                 28       47%      7    12%
   1 or more partners                     31       53%     52    88%

No. of patients in PCP's practice
   <2,000                                  7       12%
   2,000-2,350                            13       22%
   2,351-2,700                            21       35%
   2,701-3,050                            13       22%
   >3,050                                  5        9%

Level of urbanization (urban              17       29%     17    29%
   areas)

(a) Concerning the musculoskeletal system.

(b) 5-point Likert scale ("poor" [1]-"excellent" [5]).

(c) 5-point Likert scale ("never" [1]-"very often" [5]).

Table 2.
Variables Used in the Analysis (Percentages and Medians) and the Linear
Regression Results (Nonstandardized Regression Coefficients [B]) for
Patient, Primary Care Physician (PCP) (and Practice), and Physical
Therapist Characteristics Associated With the Number of Consultation
Request (n = 352)

                                                           No. of
                                          Percentage       Consultation
Predictor Variable                        or Median        Requests (B)

Level 1

Characteristics of patients (n = 342)

   Sex (female vs male)                   53%              -0.72
   Age (<25 y vs [greater than or equal   27%               2.1
      to] 25 y)
   Health insurance (public vs private    75%               2.11
      insurance)
   Education (low vs middle-high)         43%               0.35
   Marital status (single vs living       09%              -2.81
      together)
   Duration of complaints (<13 wk vs      43%              -4.31 (a)
      [greater than or equal to] 13 wk)

Characteristics of consultation

   Request for diagnostic                 25%               3.99
      information only (yes vs no)
   Request for indication for physical    29%               2.49
      therapy only (yes vs no)
   Opinion of the report (low vs high     36%
      satisfaction) (b)

Level 2

Characteristics of PCPs (n = 59)

   Knowledge about the possibilities of    3                1.24
      and indication for physical
      therapy (c)
   Perception of diagnostic skills of      4                0.15
      physical therapists (c)
   Perception of efficacy of physical      4                0.80
      therapy (c)
   Perception of the cooperation with      4               -1.07
      physical therapists (c)
   Frequency of patient communication     66%               0.76
      between physical therapist-PCP
      pairs (<2 per month vs [greater
      than or equal to] 2 per month)
   Experience (time in practice            3                0.04
      [y]) (d)
   Postgraduate education (yes vs no)     78%               0.75
   Referral with uncertainty about         2               -0.68
      indication (e)
   Referral for appropriateness about      2                0.31
      indication (e)
   Advise by telephone prior to            2                0.04
      decision of referral (e)

Characteristics of practices

   No. of partners (single vs 1 or more   47%              -0.53
      colleagues)
   Mean (range) number of patients        2,516            -0.006 (a)
      in PCP's practice                      (450-3,150)
   Level of urbanization (urban vs        29%               0.32
      rural, semiurbanized areas)

Characteristics of physical therapists (n = 59)

   Experience (time in practice            3               -0.54
      [y]) (d)
   Experience with informal                2                2.08
      consultation (e)
   Postgraduate education (<3 vs          36%               0.19
      [greater than or equal to] 3
      courses)
   Sports physical therapy (yes vs no)    25%               0.56
   Manual therapy (yes vs no)             29%               1.53

(a) Significance level set at P [less than or equal to] .05.

(b) Measured on a 5-point ordinal scale, but dichotomized in low
("poor" [1], "slight" [2], "fair" [3], or "substantial" [4])
satisfaction versus high ("excellent" [5]) satisfaction.

(c) Measured on a 5-point ordinal scale ("poor" [1]-"excellent" [5]).

(d) Measured on 3 categories (<5 y [1], 5-10 y [2], >10 y [3]).

(e) Measured on a 5-point ordinal scale ("never" [1]-"very often" [5]).

Table 3.
Hierarchical (Multi-Level) Logistic Regression Results (Odds Ratios
[ORs]) of Patient, Primary Care Physician (PCP)(and Practice), and
Physical Therapist Characteristics Associated With: (1) Whether or Not
the PCPs Were Satisfied With the Outcome of the Consultation and (2)
Whether or Not Their Intended Management Decision Was Changed (a)

                                                           Intended
                                            High           Management
                                            Satisfaction   Decision
Predictor Variable                          (OR)           Changed (OR)

Level 1

Characteristics of patients (n = 342)

   Sex (female vs male)                     0.78           0.86
   Age (<25 y vs [greater than or equal     0.89           0.52 (b)
      to] 25 y)
   Health insurance (public vs private)     0.90           0.86
   Education (low vs middle-high)           1.12           0.61 (b)
   Marital status (single vs living         1.00           1.20
      together)
   Duration of complaints (<13 wk vs        0.89           1.12
      [greater than or equal to] 13 wk)

Characteristics of consultation

   Request for diagnostic information       0.34 (b)       0.94
      only (yes vs no)
   Request for indication for physical      0.58           1.03
      therapy only (yes vs no)
   Opinion of the report (low, vs high                     2.07 (b)
      satisfaction) (c)

Level 2

Characteristics of PCPs before the study (n = 53)

   Knowledge about possibilities of and     2.78 (b)       0.88
      indications for physical
      therapy (d) (>3 vs [less than or
      equal to] 3)
   Perception of diagnostic skills of       3.35 (b)       1.07
      physical therapists (d) (>3 vs
      [less than or equal to] 3)
   Perception of efficacy of physical       1.82           1.04
      therapy (d) (>3 vs [less than or
      equal to] 3)
   Perception of the cooperation with       0.52 (b)       1.02
      physical therapists (d) (>3 vs
      [less than or equal to] 3)
   Frequency of patient communication       0.84           1.24 (b)
      between physical therapist-PCP
      pairs (<2 vs [greater than or equal
      to] 2 per month)
   Experience (time in practice [y]) (e)    1.11           1.22
   Postgraduate education (yes vs no)       1.80           1.62
   Referral with uncertainty about          0.86           0.76
      indication (f) (>3 vs [less than or
      equal to] 3)
   Referral for appropriateness about       1.22           0.90
      indication (f) (>3 vs [less than or
      equal to] 3)
   Advise by telephone prior to decision    0.84           1.19
      of referral (f) (>3 vs [less than
      or equal to] 3)

Characteristics of physical therapists (n = 53)

   Experience (time in practice             1.08           0.81
      [y]) (e) (<5 vs [greater than or
      equal to] 5)
   Experience with informal                 2.16 (b)       0.97
      consultation (f) (>3 vs [less than
      or equal to] 3)
   Postgraduate education (<3 vs [greater   0.07 (b)       0.98
      than or equal to] 3 courses)
   Sports physical therapy (yes vs no)      1.28           0.70
   Manual therapy (yes vs no)               5.21 (b)       0.76

(a) 342 level-1 variables (patients/consultations) and 53 level-2
variables (PCPs and physical therapists).

(b) Significance level set at P [less than or equal to] .05.

(c) Measured on a 5-point ordinal scale, but dichotomized in low
("poor" [1], "slight" [2], "fair" [3], or "substantial" [4],
satisfaction versus high ("excellent" [5]) satisfaction.

(d) Measured on a 5-point ordinal scale ("poor" [1] "excellent" [5]).

(e) Measured on 3 categories (<5 y [1], 5-10 y [2], >10 y [3]).

(f) Measured on a 5-point ordinal scale ("never" [1]-"very often" [5]).

Table 4.
Comparison of Demographic Characteristics for Patients Referred by
Primary Care Physicians (PCPs) for a One-Time Physical Therapist
Consultation With Patients Listed in the National Database of Patients
Referred by PCPs for Receiving Physical Therapy Intervention

                                Consultation     Treatment
                                (n = 352)        (n = 17,201)

Characteristic                  No.     %        No.        %

Sex
   Male                         187     53        9,357     54
   Female                       165     47        7,844     46

Age (y) (a,b)
   0-24                          94     27        2,425     14
   25-44                        143     41        6,743     39
   45-64                         87     25        5,246     31
   65+                           24      7        2,769     16

Health insurance (public        263     75       12,643     74
   health insurance)

Education (b,c)
   Low                          150     43       10,427     61
   Middle                       154     44        5,428     32
   High                          42     12        1,347      8

Employed (yes) (b)              172     49        5,785     34

Marital status (single) (b)      33      9        2,563     15

(a) Data missing for 4 patients referred for consultation.

(b) Chi-square test, P<.01.

(c) Data missing for 6 patients referred for consultation.

Table 5.
Comparison of Health Status-Related Characteristics for Patients
Referred by Primary Care Physicians (PCPs) for a One-Time Physical
Therapist Consultation With Patients Listed in the National Database
of Patients Referred by PCPs for Receiving Physical Therapy
Intervention

                                          Consultation   Treatment
                                          (n = 352)      (n = 17,201)

Characteristic                            No.        %   No.        %

Duration of complaints (wk) (a,b)
   <l                                      15        4   2,597     15
   1-12                                   137       39   9,512     58
   13-52                                   99       29   1,496     18
   >52                                     97       28   1,565      9

Prior physical therapy, same               56       16   6,261     36
   complaint (yes)

Prior medical diagnostic/therapeutic
   actions (b)

   Radiography                            103       29   5,659     33
   Laboratory                              22        6     189      1
   Surgery                                 12        3   1,772     10
   Medication/injection                    58       16   7,172     42
   Cast/tape                               12        3     722      4

Type of disorder in terms of ICPC (b,c)

   Knee (L15)                              61       17     533      3
   Back (L02)                              58       16   1,049      6
   Lower extremity (L14)                   30        9     722      4
   Shoulder (L08)                          27        8   1,376      8
   Neck (L01)                              22        6   1,686     10
   Upper extremity (L09)                   16        5     756      4
   Thigh (L13)                             14        4     206      1
   Ankle (L16)                             11        3     189      1
   Other complaints (L29)                  11        3     189      1
   Elbow (L10)                              9        3      68      1

Total                                     259       74   6,774     39

(a) Data missing for 6 patients referred for consultation.

(b) Chi-square test, P<.01

(c) ICTP = International Classification of Primary Care (top 10
referrals for consultation). (29)

Table 6.
Management Decisions and Referral Patterns of Primary Care Physicians
(PCPs) Before and After a One-Time Physical Therapist Consultation (a)

                                 Before Consultation

                                                Referral to
                                 Continue Own   Medical
                                 Management     Specialist

After consultation

Continue own management          10              11
Referral to medical specialist    9              21
Referral to physical therapist   23              67
Other                             2

Total before consultation        42 (12.7%)     101 (30.4%)

                                   Before Consultation

                                   Referral to
                                   Physical
                                   Therapist     Other

After consultation

Continue own management             16           2           39 (11.7%)
Referral to medical specialist      24           1           55 (16.6%)
Referral to physical therapist     137           5          232 (69.9%)
Other                                4                        6 (1.8%)

Total before consultation          181 (54.5%)   8 (2.4%)   332 (100%)

(a) No distinction is made in the table for medical specialist referral
for treatment or consultation after a one-time physical therapist
consultation.

References

(1) Hendriks HJM. Physiotherapist's Consultation in General Practice [thesis]. Maastricht, the Netherlands: Maastricht University, 2000.

(2) Geijer RMM, Burgers JS, van der Laan JR, eds. NHG-Standaarden Voor de Huisarts, Deel 1 [National Practice Standards of the Dutch College of General Practitioners (NHG), Part 1]. Maarssen, the Netherlands: Elsevier/Bunge; 1999.

(3) van de Lisdonk EH, van den Bosch WJHM, Huijgen FJA, et al. Ziekten in de Huisartsenpraktijk (3e Druk) [Epidemiological Data on Diseases in Primary Care Physicians' Practice]. Maarsen, the Netherlands: Elsevier/ Bunge; 1999.

(4) Rijken PM, van Heugten CM, Dekker J. Brancherapport Paramedische Zorg [Branch Report of Dutch Allied Health Care]. Utrecht, the Netherlands: Netherlands Institute for Health Care Research (Nivel); 1997.

(5) Centraal Bureau Voor de Statistiek (CBS). Statistisch Jaarhoek [Statistics in Health Care]. Voorburg, the Netherlands: CBS; 1997.

(6) Uunk W, Dekker J, Groenewegen P. Verwijzingen van Huisartsen Naar Fysiotherapeuten, Morbiditeitsspecifieke Verwijs-Percentages: Deelproject Huisarts en Fysiotherapeutische Zorg, Basisgegevens uit de Nationale Studie van Ziekten en Verrichtingen in de Huisartsenpraktijk [Primary Care Physicians' Referrals to Physical Therapists]. Utrecht, the Netherlands: Netherlands Institute for Health Care Research (Nivel); 1991.

(7) Hulme JB, Wackernagel B, Lewis JW. Communication between physicians and physical therapists. Phys Ther. 1988;68:26-31.

(8) Anderson J, Campbell SK, Gardner HG. Correlates of physician utilization of physical therapy, Int J Technol Assess Health Care. 1992;8:10-19.

(9) Ritchey FJ, Pinkston D, Goldbaum JE, Heerten ME. Perceptual correlates of physician referral to physical therapists: implications for role expansion. Soc Sci Med. 1989;28:69-80.

(10) Higgs J, Titchen A. Research and knowledge. Physiotherapy. 1998;84:72-80.

(11) Beckerman H, Bouter LM, van der Heijden GJMG, et al. Efficacy of physiotherapy for musculoskeletal disorders: what can we learn from research? Br J Gen Pract. 1993;43:73-77.

(12) Newman DJ. Physiotherapy for best effect. Physiotherapy. 1997;83:5-11.

(13) Koes BW. Now is the time for evidence based physiotherapy [editorial]. Physiotherapy Research International. 1997;2:iv-v.

(14) Hendriks HJM, Brandsma JW, Wagner C, et al. Experiences with physiotherapist consultation: results of a feasibility study. Physiotherapy Theory & Practice. 1996;12:211-220.

(15) Uunk WJG, Groenewegen PP, Dekker J. Verwijzingen van huisartsen naar fysiotherapeuten: een verklaring en analyse van verschillen tussen huisartsen [Primary care physicians' referrals to physical therapists: explaining variation of primary care physicians in number and type of referrals]. Mens Maatschappij. 1992;67:389-411.

(16) Yoshida KK, Allison KR. Sociodemographic and health-related characteristics of Ontarians making contact with physical therapists: analysis of the 1990 Ontario health survey. Physiotherapy Canada. 1997;49:214-221.

(17) Wellman B. Lay referral networks: using conventional medicine and alternative therapies for low back pain. Research in Social Health Care. 1995;12:213-238.

(18) Carey TS, Garrett J, Jackman A, et al. The outcomes and costs of care for acute care low back pain among patients seen by primary care practitioners, chiropractors and orthopedic surgeons. N Engl J Med. 1995;333:913-917.

(19) Carey TS, Evans AT, Adler NM, et al. Acute severe low back pain: a population-based study of prevalence and care-seeking. Spine. 1996;21:339-344.

(20) Mieleuz TJ, Carey TS, Dyrek DA, et al. Physical therapy utilization by patients with acute low back pain. Phys Ther. 1997;77:1040-1051.

(21) Moncur C. Perceptions of physical therapy competencies in rheumatology: physical therapists versus rheumatologists. Phys Ther. 1987;67:331-339.

(22) Guide to Physical Therapist Practice. 2nd ed. Alexandria, Va: American Physical Therapy Association; 2001.

(23) Hendriks HJM, de Bie RA, van der Zee J. Consultation between general practitioners and medical consultants in the Dutch health care system: what can we learn from the literature? In: Hendriks HJM. Physiotherapist's Consultation in General Practice [thesis]. Maastricht, the Netherlands: Maastricht University; 2000:57-74.

(24) Wilkin D, Smith A. Explaining variation in general practitioners referrals to hospital. Fam Pratt. 1987;4:160-169.

(25) Pool JB, Hingstman L. Cijfers uit de Registratie van Beroepen in de Eerstelijns-gezondheidszorg: Statistische Gegevens per 1 January 1994 Over. Huisartsen en Verloskundigen [Statistical Data on Primary Cure Physicians and Midwives]. Utrecht, the Netherlands: Netherlands Institute for Primary Health Care (Nivel); 1994.

(26) Pool JB, Hingstman L. Cijfers uit de Registratie van Beroepen in de Eerstelijns-gezondheidszorg 1994: Statistische Gegevens per 1 January 1994 Over in de Eerste Lijn Werkzame Fysiotherapeuten [Statistical Data on Primary Health Care Professionals]. Utrecht, the Netherlands: Netherlands Institute for Primary Health Care (Nivel); 1995.

(27) The Dutch College of General Practitioners (NHG). National Practice Guideline: the referral letter to medical specialists. Huisarts Wet. 1988;32:102-105.

(28) Richtlijnen Voor de Fysiotherapeutische Verslaglegging [Guidelines for Documentation]. Amersfoort, the Netherlands: Royal Dutch Physical Therapy Association (KNGF); 1993; updated 2003.

(29) Lamberts H, Wood M. International Classification of Primary Care. Oxford, United Kingdom: Oxford University Press; 1987.

(30) International Classification of Impairments, Activities and Participation: A Manual of Dimensions of Disablement and Functioning (ICIDH-2). Geneva, Switzerland: World Health Organization; 1997.

(31) Heerkens YF, Brandsma JW, Lakerveld-Heyl K, van Ravensberg CD. Impairments and disabilities, the differences: proposal for Adjustment of the International Classification of Impairments, Disabilities, and Handicaps. Phys Ther. 1994;74:430-442.

(32) Hendriks HJM, Brandsma JW, Heerkens YF, et al. Intraobserver and interobserver reliability of assessments of impairments and disabilities. Phys Ther. 1997;77:1097-1106.

(33) van der Valk R, Dekker J, Boschman M. Basisgegevens Extramurale Fysiotherapie 1989-1992: Gegevens uit het Project "Beleidsgericht Evaluatie-en Effectonderzoek Fysiotherapie" [Epidemiological Data on Extramural Physical Therapy Care]. Utrecht, the Netherlands: Netherlands Institute for Primary Health Care (Nivel); 1995.

(34) Dekker J, van Baar ME van, Curfs EC, Kerssens JJ. Diagnosis and treatment in physical therapy: an investigation of their relationship. Phys Ther. 1993;73:568-580.

(35) Rasbach J, Woodhouse G. MLA Command Reference. London, United Kingdom: Institute of Education, University of London; 1995.

(36) Bryk AS, Raudenbusch SW. Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, Calif: Sage Publications; 1992.

(37) Goldstein H. Multilevel Statistical Models. 2nd ed. New York, NY: Halsted Press; 1995.

(38) Kerssens JJ, Sluijs EM, Verhaak PFM, et al. Back care instructions in physical therapy: a trend analysis of individualized back care programs. Phys Ther. 1999;79:286-295.

(39) Prescot-Clarke P, Brooks T, Machray C. Focus on Health Care: Surveying the Public in Four Health Care District. London, United Kingdom: Royal Institute of Public Administration and Social and Community Planning Research; 1988.

(40) Pouwer F, van der Ploeg HM, Bramsen I. Dwalingen in de methodologie: bias door vragenlijsten [Straying through methodology, II: questionnaire bias]. Ned Tijdschr Geneeskd. 1998;142:1556-1558.

(41) Sudman S, Bradburn NM, Schwarz N. Thinking About Answers: The Application of Cognitive Processes to Survey Methodology. San Francisco, Calif: Jossey-Bass; 1996.

(42) Stokx LJ, Glocrich ABM, Kersten TJJMT. Kostenbesparing Door Kwaliteitsbevordering [Cost Efficiency by Improvement of Quality of Care]. Utrecht, the Netherlands: Netherlands Institute for Primary Health Care (Nivel); 1992.

(43) Stokx LJ, de Bakker DH, Delnoy DMJ, et al. Verwijscijfers Belicht [Investigation on Primary Care Physicians' Referral Figures]). Utrecht, the Netherlands: Netherlands Institute for Primacy Health Care (Nivel); 1992.

(44) Hendriks HJM, Oostendorp RAB, Bernards ATM, et al. The diagnostic process and indication for physiotherapy: a prerequisite for treatment and outcome evaluation. Physical Therapy Reviews. 2000;5:29-47.

(45) Hendriks HJM, Bekkering GE, van Ettekoven H, et al. Development and implementation of national practice guidelines: a prospect for continuous quality improvement in physiotherapy--an introduction to the method of guideline development (part I). Physiotherapy. 2000;86:535-547.

(46) May B, Dennis JK. Expert decision making in physical therapy: a survey of practitioners. Phys They: 1991;71:190-202.

(47) Sahrmann SA. Diagnosis by the physical therapist: a prerequisite for treatment: a special communication. Phys Ther. 1988;68:1703-1706.

(48) Verbrugge LM, Jette AM. The disablement process. Soc Sci Med. 1994;38:1-14.

(49) Jette AM. Physical disablement concepts for physical therapy research and practice. Phys Ther. 1994;74:380-386.

(50) Jette DU, Manago D, Medved E, et al. The disablement process in patients with pulmonary disease. Phys Ther. 1997;77:385-394.

(51) Cott CA, Finch E, Gasner D, et al. The movement continuum theory of physical therapy. Physiotherapy Canada. 1995;47:87-95.

(52) Johnston M. Models of disability. Physiotherapy Theory & Practice. 1996;12:131-141.

(53) Sahrmann SA. The Twenty-Ninth Mary McMillan Lecture: Moving precisely? Or taking the path of least resistance? Phys Ther. 1998;78:1208-1218.

(54) Schultz IZ, Crook JM, Berkowitz J, et al. Biopsychosocial multivariate predictive model of occupational low back disability. Spine. 2000;27:2720-2725.

(55) Fritz JM, George SZ. Identifying psychosocial variables in patients with work-related low back pain: the importance of fear-avoidance beliefs. Phys Ther. 2002;82:973-983.

EJM Hendriks, PT, PhD, is a physical therapist in private practice, the Klepperheide, Druten, the Netherlands. He is also Senior Researcher and Manager, Program for Development and Implementation of National Clinical Practice Guidelines for the Allied Health Professions, Department of Research and Development, Dutch National Institute of Allied Health Professions (NPi), and Senior Researcher (Health Scientist and Clinical Epidemiologist), Department of Epidemiology and Centre of Evidence-Based Physiotherapy, Maastricht University, Maastricht, the Netherlands. Address all correspondence to Dr Hendriks at Department of Research and Development, Dutch National Institute of Allied Health Professions (NPi), PO Box 1161, 3800 BD Amersfoort, the Netherlands ([e-mail]).

JJ Kerssens, PhD, is Psychologist and Senior Researcher, Netherlands Institute of Health Services Research (Nivel), Utrecht, the Netherlands.

J Dekker, PhD, is Clinical Psychologist and Professor in Allied Health Care, Department of Rehabilitation Medicine, Free University (VU) Medical Centre, Amsterdam, the Netherlands.

RM Nelson, PT, PhD, FAPTA, is Physical Therapist, Professor of Physical Therapy, and Chair, Department of Physical Therapy, Lebanon Valley College, Annville, Pa.

RAB Oostendorp, PT, MT, PhD, is Physical Therapist, Manual Therapist, and Professor in Allied Health Care Research, Nijmegen University Medical Centre, Centre of Quality of Care Research, Nijmegen, the Netherlands. He is also Director, Dutch National Institute of Allied Health Professions (NPi).

J van der Zee, PhD, is Medical Sociologist and Professor in Primary Health Care Research, Department of Medical Sociology, Maastricht University, and Director, Netherlands Institute of Health Services Research (Nivel). Dr Hendriks, Dr Dekker, and Dr van der Zee provided concept/idea/ research design.

Dr Hendriks, Dr Kerssens, and Dr Dekker provided writing and data collection and analysis. Dr Hendriks and Dr Dekker provided project management and subjects. Dr Hendriks, Dr Dekker, Dr Oostendorp, and Dr van der Zee provided fund procurement. Dr Dekker, Dr Oostendorp, and Dr van der Zee provided facilities/equipment and institutional liaisons. Dr Nelson, Dr Oostendorp, and Dr van der Zee provided consultation (including review of manuscript before submission). The authors thank Jenny de Fouw, MSc, and Marielle Jans, PhD, for their valuable comments on the text of this article and the Dutch Ministry of Health, Welfare and Sports, who funded the study.

The National Medical Ethics Committee and Review Board approved the study.