Physical therapists' use of interventions with high evidence of effectiveness in the management of a hypothetical typical patient with acute low back pain.

Background and Purpose. Evidence-based practice aims to improve patient care and service delivery, particularly in the management of individuals with low back pain (LBP), the largest client group seen by outpatient physical therapists. The purpose of this study was to determine the prevalence of use of interventions with evidence of effectiveness in the management of acute nonspecific LBP by physical therapists. Subjects. A multicenter cross-sectional study was conducted on 100 physical therapists working with patients with LBP. Methods. Using a telephone-administered interview, therapists described their current and desired treatment practices for a typical case of LBP. Each intervention reported was coded according to its evidence of effectiveness (strong, moderate, limited, or none). Information on clinician, workplace, and client characteristics also was obtained. Results. The prevalence of use of interventions with strong or moderate evidence of effectiveness was 68%. However, 90% to 96% of therapists also used interventions for which research evidence was limited or absent. Users of interventions with high evidence of effectiveness, as compared with nonusers, had graduated more recently and had taken a higher number of postgraduate clinical courses. Discussion and Conclusion. Although most therapists use interventions with high evidence of effectiveness, much of their patient time is spent on interventions that are not well reported in the literature. The results indicate the need for improvement in the quality of clinical research as well as its dissemination and implementation in a way that is appealing to therapists, such as through practice-related courses. [Mikhail C, Korner-Bitensky N, Rossignol M, Dumas JP. Physical therapists' use of interventions with high evidence of effectiveness in the management of a hypothetical typical patient with acute low back pain. Phys Ther. 2005;85:1151-1167.]

Key Words: Back pain, Evidence-based practice, Health care surveys, Practice guidelines.

In recent years, much of the focus in health care has shifted toward evidence-based practice, which is defined as "the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients, integrating individual clinical expertise with the best available external clinical evidence from systematic research." (1) A challenge for health care professionals is to offer the highest quality care in the most cost-effective way based on current research. This has become especially important in the care of people with low back pain (LBP), where chronicity and cost are a great burden on the health care system. Indeed, it is estimated that 79.2% of the general population will experience some form of back pain throughout their lifetime. (2) In adults aged less than 45 years, LBP is the most prevalent cause of disability. (3) Back pain is also a social and economic problem for stakeholders, accounting for $25 billion, in the United States, in overall annual medical costs and is the second leading cause of workdays lost after the common cold. (4) Many people with LBP seek treatment from a physical therapist. Indeed, LBP is the most common condition (25% of caseload) managed by physical therapists in outpatient settings in the United States. (5)

Four English-language studies have described physical therapists' current management of LBP. (6-9) Battie et al (6) conducted a survey on physical therapists' management of a hypothetical patient with acute LBP (no sciatica), as described in a vignette. In a more recent study by Foster et al, (7) physical therapists were asked to identify which methods of treatment were available to them and to rank how often they used specific treatment methods. They also were asked to identify interventions that they would prefer to use under more favorable circumstances. A Canadian study by Li and Bombardier (8) surveyed physical therapists regarding assessments and interventions they reportedly use for acute and sub-acute lumbar impairments. Their questionnaire was based on 3 hypothetical case scenarios depicting 3 "typical" patients with LBP (acute, subacute, and acute with sciatica). From a list of assessment and treatment approaches, the therapists identified those that they would use for each of the 3 cases. Although open-ended questions were used to determine the therapists' use of other assessment or treatment approaches not included in the list, the results were dichotomized into whether therapists used a certain intervention or not. Information on frequency of use of each intervention was not ascertained. Gracey et al (9) conducted a study investigating clinical practice for LBP among 157 physical therapists and 1,062 patients in Northern Ireland. During the first treatment session, the therapists recorded their patients' physical findings and sociodemographic data. During subsequent visits, interventions and modalities that the therapists used were recorded, along with the time spent on each in 5-minute units.

These 4 studies (6-9) used different methods to ascertain interventions for LBP. The results found by Foster et al (7) were based on neither a hypothetical patient nor a real patient and, therefore, do not allow exploration of variations in treatment for a given case scenario. The studies by Battie et al (6) and Li and Bombardier (8) controlled for this by providing therapists with a typical case of LBP, so that therapists would be questioned regarding an identical client. Finally, Gracey et al, (9) by direct observation, were able to eliminate the potential bias associated with self-reporting by clinicians but introduced variability based on patient differences and preferences. Indeed, the use of case studies has been supported by a number of authors who have demonstrated that this is a valid form of treatment ascertainment. (10,11) In addition, much of medical training currently is designed using problem-based or case-based learning. The use of a vignette permits evaluation of variations in practice patterns, while keeping the patient profile constant. This is especially important in the investigation of individuals with nonspecific LBP, because this condition commonly results in a very heterogeneous group of clients having a typical clinical presentation.

These studies (6-9) also suggest that interventions used by physical therapists have changed in the last decade, evolving from stretching and aerobic exercises to the McKenzie approach, manual therapy, and electrotherapy. Although the reasons for this shift in approach remain unclear, possible explanatory factors include the development of new techniques, updates in research studies, and the emergence of evidence-based practice principles in academia. The literature on LBP is extensive, including evidence of interventions that are and are not effective, but it is unknown to what extent effective interventions are being used by clinicians. Using Green's model as a framework, Tamblyn and Battista (12) have classified factors affecting clinical practices into predisposing, enabling, and reinforcing factors. Predisposing factors include such things as the practitioner's knowledge, skills, and attitudes as well as sociodemographic characteristics, including age, sex, and training. Enabling and reinforcing factors include elements such as the characteristics of the practice setting and the patient population. However, there is a lack of understanding of the factors that may contribute to the uptake of evidence-based practice. Therefore, the objectives of this study were: (1) to determine the prevalence of use of interventions with evidence of effectiveness in the management of acute nonspecific LBP by physical therapists, (2) to identify the clinician, environment, and client factors associated with being a user of interventions with high evidence of effectiveness, and (3) to describe clinicians' desired use of interventions in the management of acute nonspecific LBP and the perceived barriers to use. Acute nonspecific low back pain is defined as pain of 6 weeks' duration or less between the gluteal fold and the uppermost lumbar vertebra without a specified cause, such as nerve root compression, trauma, infection, or tumor. (5,13,14) This definition excludes postsurgical LBP.

Methods

Research Design

A cross-sectional survey was undertaken to investigate practice behaviors of physical therapists who work with patients with LBP. Participants were 100 physical therapists working in the province of Quebec, Canada. Data were collected via a structured telephone interview based on a 6-part questionnaire. Therapists were asked questions regarding their current treatment practices in the management of a "typical" individual with acute LBP, as depicted in a clinical vignette (Appendix). *

Study Population

Physical therapists were recruited through the 2003 publicly available listings provided on the Web site of the provincial licensing body, the Ordre Professionel de la Physiotherapie du Quebec (OPPQ). As of June 2003, all physical therapists are required to register as regular members in the OPPQ in order to practice in the province of Quebec. The sampling frame for this study consisted of physical therapists working in public or private practice in orthopedics or rheumatology with adult clients in Quebec. Clinicians were first stratified according to private practice (which includes home care or private clinic) or public practice (which includes general hospital, rehabilitation center, or centre local de service communautaire [CLSC, translated as "local center for community service"]). Potential respondents were selected randomly in proportion to the number of therapists working within each of the 16 administrative regions in Quebec. The inclusion criteria were physical therapists in the private or public setting currently working with an adult clientele ([greater than or equal to] 18 years of age) with LBP on an outpatient basis; having worked in the same setting for a minimum of 3 months; and having a caseload that consists of, on average, at least 1 patient with LBP per week. Sample size estimates were based on the conservative estimate that 20% of therapists would use interventions with high evidence of effectiveness. Using a 2-sided confidence interval of 95% and a desired precision of 8%, about 100 physical therapists would be required to allow stable estimates of prevalence of use.

Questionnaire Development and Content

The interview questionnaire was designed specifically for the purposes of this study using rigorous questionnaire design methodology. Specifically, following content creation, the questionnaire was pilot tested by having several people from different health-related backgrounds (physical therapists, epidemiologists, laypeople) read the questionnaire and suggest modifications to ensure clarity and completeness of the text. The questionnaire was translated into French and then back-translated by a bilingual physical therapist, again using rigorous methodology for both the forward and backward translations. The questionnaire components included:

1. Clinician characteristics--for example, clinical experience, educational background, full-time/part-time status.

2. Clinical vignette of a typical patient, Mrs C, with acute nonspecific LBP.

3. Actual practices--open- and closed-ended questions regarding interventions that therapists would typically use with the client described in the vignette.

4. Desired practices--open-ended questions regarding interventions that therapists desire to use more frequently as well as those that they do not use but would like to use in the ideal world.

5. Checklist of interventions with known evidence of effectiveness--list of 13 interventions for which therapists rated their frequency of use on a 5-point scale ranging from "never" to "always" in relation to a client similar to the one in the vignette.

6. Environment and client characteristics--for example, public/private practice, teaching/nonteaching hospital, typical age of clientele.

Also included was a 17-item Practice Style Questionnaire (15) that classifies clinicians into 1 of 4 categories according to how they respond to new information: seekers, receptives, traditionalists, or pragmatists. ([dagger]) The interview questionnaire was first pretested on a convenience sample of 6 physical therapists to resolve format, duplication, and clarity issues.

Development of the Clinical Vignette

Previous work on LBP has made use of case studies describing a client with LBP. (6,8) Therefore, to elicit information on therapists' typical practice patterns, a clinical vignette (Mrs C) was developed by a focus group that was instructed to create a typical patient profile of a patient with acute nonspecific LBP (Appendix). The focus group was run using focus group methodology (16) and consisted of 5 participants, including 4 expert physical therapists who were knowledgeable in the treatment of people with LBP (2 from the public sector and 2 from both the public and private sectors).

Tracing and Recruitment of Study Participants

To ensure that not only "easy-to-reach" clinicians were recruited for the study, tracing was done in a systematic manner using a structured tracing sheet. The first tracing effort was made at their last recorded place of work. A second search on the OPPQ Web site occasionally yielded another location of employment for the same therapist, and the clinician was then contacted there. As an alternative, local telephone directories and Internet 411 Web sites also were used. Eight tracing attempts were made before declaring the therapist as "untraced." These combined methods ensured high tracing rates for physical therapists.

Once an individual was contacted, the study was

described and eligibility as well as willingness to participate was established. The research coordinator scheduled a convenient date and time for a 20-minute telephone interview. The research coordinator then sent (by e-mail or fax) the necessary documents to the therapist, including: the vignette describing a typical patient with LBP, the list of factors affecting the therapist's choice of interventions, an explanatory letter ensuring confidentiality would be preserved, and a request not to divulge any of the contents of the documents and interview. To avoid potential contamination resulting from clinicians sharing information about the vignette and their questionnaire responses, participants were traced and recruited in such a way that those from the same administrative region were interviewed within a short time frame. The recruitment process continued until the required number of participants was accrued within each stratum, that is, from public or private practice.

Because the validity of data from this study depended on high response rates, a specific interview guideline was used to achieve maximal participation: the Total Design Method. (17-19) This guideline has been shown to achieve high recruitment rates, ranging from 87% to 95%, for various subjects, including patients, families, and clinicians. It details the optimal format for asking questions, the process for choosing the mode of administration, and the optimal timing of contact with the respondent. This method also discusses the rigorous training procedures to be used in terms of styles of interaction, introductory statements, and the development of a personalized contact. Two bilingual physical therapists were trained to conduct the study in a standardized way in the respondents' preferred spoken language (English or French).

Levels of Evidence of Effectiveness

A classification was required to categorize each intervention used in the physical therapy management of LBP according to its level of evidence of effectiveness. Two databases (CINAHL and MEDLINE) were searched from the years 1995 to 2003, and the following key words and combinations were used: "evidence-based practice," "clinical practices," "physiotherapists," "physiotherapy," and "back pain." The literature regarding interventions used in the management of acute nonspecific LBP was closely examined. Most clinical practice guidelines (CPGs) and systematic reviews were found to draw conclusions on nonspecific LBP, and, in 90% of patients with LBP, no specific medical diagnosis is made. (20) After an extensive review of more than 15 guidelines, (21,22) we deemed that the Dutch Physiotherapy Guidelines (13) and the Cochrane Back Reviews (23-34) would be used to classify interventions according to evidence. These summaries were chosen for 3 reasons. First, they are the most recent and up-to-date, having been published in 2003. The second reason is that the Dutch Physiotherapy Guidelines are the only ones that are solely directed at physical therapists. Finally, it would not have been feasible to summarize the recommendations of all 12 guidelines because each guideline uses a different set of ratings of evidence of effectiveness and these ratings are not equivalent and cannot be converted. Both the Dutch Physiotherapy Guidelines and the Cochrane Back Reviews use the same classification scheme:

1. Strong evidence: these interventions are deemed to be either effective or ineffective, with strong support in the literature as determined by consistent findings or results in several high-quality randomized controlled trials (RCTs) or in at least 1 meta-analysis.

2. Moderate evidence: these interventions are deemed to be either effective or ineffective, with moderate support in the literature as determined by consistent findings or results in 1 high-quality RCT and 1 or several low-quality RCTs.

3. Limited/contradictory evidence: these are interventions with weak or conflicting support in the literature as determined by 1 RCT (high or low quality) or inconsistent findings among several RCTs.

4. No known evidence: these interventions have not been sufficiently studied in the literature in terms of effectiveness, and no RCTs have been done in this area.

Data Analysis

Therapists indicated the interventions that they would use regularly or occasionally (at initial visit, as needed, and at discharge) with the client described in the vignette. Prevalence of use of each intervention (irrespective of whether it was used regularly or occasionally) was calculated. Each reported intervention then was coded according to its level of evidence of effectiveness. Clinicians were classified as "users" or "nonusers" of interventions with high levels of effectiveness, with users being those who identified use of at least 1 intervention with a strong or moderate rating of effectiveness for the client (Mrs C) who was depicted in the vignette.

Univariate statistics were used to analyze differences between users and nonusers on the potential explanatory variables related to the clinician, environment, and client. For categorical or ordinal variables, chi-square tests and Fisher exact test were used. (35) In the case of ordinal data, the Cochran-Armitage trend (36) was used. For continuous variables, simple t tests were used to compare means for the users and nonusers. Continuous data also were transformed into categorical data with arbitrary cutoff points based on the frequency distributions of the data. For example, the continuous variable "number clients with LBP seen per day" was classified as 0 to 2, 3 to 5, and >5. Based on these procedures, potentially important variables were identified using the following criterion: a significant association was considered to exist at P<.05. A Bonferroni correction was used to account for multiple comparisons, and the level of significance was set at P = .01.

Additional analyses were performed using forward conditional logistic regression to investigate the contribution of explanatory variables that were bivariately associated with being a user or nonuser of interventions with strong or moderate ratings of effectiveness. A dichotomous outcome variable was created to indicate whether the therapist was a user or nonuser. The level of significance for entering an explanatory variable was set at P = .05.

Results

Comparison of Participants and Nonparticipants

A total of 176 physical therapists were randomly selected within the strata formed by public and private practices and contacted to participate in this survey; of these, 2 therapists (1.1%) were untraced and 116 therapists (65.9%) were eligible.

Of the eligible physical therapists, 86.2% (n = 100) agreed to participate. Sixteen physical therapists (13.8%) refused to participate. A comparison of the nonrespondents with the respondents indicated that there were no differences between the 2 groups with respect to sex, administrative region, urban or rural status, practice setting, clinical experience, and workload status. Two people refused to provide baseline characteristics.

Table 1 describes the characteristics of the entire sample of therapists. They were mostly female (67%), working full-time (71%), 61% had greater than 10 years of clinical experience, and 57% had greater than 10 years of experience with clients with LBP. Tables 2 and 3 describe the characteristics of the physical therapists' workplace and their typical clientele. On average, there were 2.63[+ or -]1.55 information resources available in the respondents' workplace, with the most typical being a computer with Internet access (Tab. 2). Therapists reported seeing an average of 3.64[+ or -]2.66 clients with LBP on a typical day, and four fifths of the therapists indicated an average treatment time of 30 to 60 minutes (Tab. 3).

Prevalence of Use of Interventions and Their Timing of Use

When asked to describe, using the open-ended format, the interventions they would typically use for the client depicted in the vignette, the therapists indicated a total of 43 different interventions. Table 4 shows the prevalence of each intervention, according to its level of evidence of effectiveness. Vertebral mobilizations, ice, and ultrasound were the most popular interventions reported, regardless of whether they were used regularly or occasionally at any point during the course of treatment. Of the 100 physical therapists, 3% reported using at least 1 intervention with strong evidence of effectiveness, that is, recommendations to stay active or to refrain from bed rest. About two thirds of the therapists (68%) reported using at least 1 of the 2 interventions with moderate evidence of effectiveness, and 93% of the therapists reported using at least 1 of the 10 interventions with moderate evidence of ineffectiveness. For 56 of these 93 therapists, exercise therapy was reported without mentioning any other interventions with moderate evidence of ineffectiveness. Use of interventions with no known evidence of effectiveness was reported by 96% of the therapists.

The top 4 interventions that physical therapists reported they would use on a regular basis with Mrs C were exercise therapy (88.0% of therapists), spinal manipulative therapy (50%), ultrasound (42%), and electrotherapy (37%). At initial visits, the top 3 interventions were ice (21%), education (19%), and posture correction (17%). On an as-needed basis, the top 3 interventions were ice (25%), electrotherapy (21%), and heat (21%). At or around discharge, the top 3 interventions were exercise therapy (24%), home exercise program (15%), and functional or work conditioning (10%).

When therapists were asked to indicate the typical treatment frequency for Mrs C, 58% stated they would treat her 3 times per week, 35% said once or twice per week, and 7% said 5 times per week. The typical length of treatment lasted 2 to 5 weeks, with 63.0% reporting this category. The average treatment duration was 50.21 [+ or -] 18.61 minutes per session. When asked to report on the duration of each intervention within a typical session, interventions with strong evidence of effectiveness took an average of 5.67 [+ or -] 1.15 minutes, although only 3 therapists used them. Interventions with moderate evidence of effectiveness lasted an average of 10.68 [+ or -] 6.23 minutes. On average, interventions with limited or contradictory evidence of effectiveness (excluding back school), moderate evidence of ineffectiveness, and no known evidence lasted greater than 22.16 [+ or -] 13.21 minutes.

>From a list of 10 options provided to clinicians regarding what factors most influenced their choice of interventions (Appendix), the top 3 were known effectiveness for LBP (35% of therapists), familiar and learned during professional training (28% of therapists), and learned about it at a conference, seminar, meeting, or course (12% of therapists).

Therapists' Recommendations Regarding Bed Rest and Return to Work

Based on the prompted question "What recommendations would you make to this client in terms of bed rest?", 73% of the therapists recommended avoiding bed rest, 23% of the therapists recommended intermittent, short periods of bed rest during the day, and 4% of the therapists recommended bed rest in the first 24 to 48 hours if there was severe pain. Based on the prompted question "What recommendations would you make to this client in terms of return to work?", 22% of therapists used terms such as "return to work as soon as possible," 44% recommended a progressive return to work (in workload and duties), 13% recommended a return to work according to the patient's signs and symptoms (decrease in pain, improved function), and 21% recommended a progressive return to work according to the patient's signs and symptoms.

Diagnostic Classification of Mrs C

When clinicians were asked about the diagnostic classification of the client in the vignette, 2 therapists classified her as having specific LBP and 1 therapist indicated a radicular syndrome. Of the 97 therapists who correctly classified Mrs C as having nonspecific LBP, 72 therapists required clarification on the definitions of specific LBP, nonspecific LBP, and radicular syndrome.

Users of Interventions With High Evidence of Effectiveness

Users were defined as therapists who reported use of at least 1 intervention that was coded to have a strong or moderate rating of effectiveness for Mrs C. As shown in Table 5, 68 therapists were classified as users of interventions with high evidence of effectiveness and 32 therapists were classified as nonusers. Users, when compared with nonusers, were more likely to have practiced for less than 15 years and had graduated more recently, having an average year of graduation of 1990 [+ or -] 8.03 versus 1985 [+ or -] 11.47 (P=.0098). Users also were significantly more likely to have taken more postgraduate courses than nonusers. Users took an average of 4.26 [+ or -] 2.63 courses as compared with nonusers, who took an average of 2.88 [+ or -] 1.95 courses (P=.0091). Three other variables approached significance: years of experience with LBP, highest diploma received, and participation in LBP research. Specifically, although only 6 individuals indicated having graduate degrees, all 6 were users. Of the 6 variables related to the environment (administrative region, rural/urban setting, public/private setting, teaching/nonteaching hospital, training of students, and available information resources) and of the 5 client variables (number of clients with LBP per day, typical client age, typical treatment duration, treatment frequency, and length of treatment), none were significantly associated with being a user.

When this relationship was further explored using logistic regression (Tab. 6), only 1 variable and 1 interaction were significantly associated with being a user: the year of graduation (P=.006) and the interaction between the year of graduation and the number of postgraduate courses taken (P=.005). Specifically, more recent graduation and a higher number of courses taken increased the probability of being a user of evidence-based interventions. The number of courses taken was not, by itself, significantly associated with the outcome (P=.712). When the year of graduation and its interaction with continuing education were entered in the model, no other variable considered could improve the prediction of being a user.

Desired Practices

When clinicians were asked to indicate their desired use of interventions given an ideal world, 33 therapists responded that they would have liked to use additional interventions for a client like Mrs C. Of the 16 desired interventions, all were reported by under 10% of clinicians, exercise therapy (7%) and functional or work conditioning (7%), to name a few. The top 3 barriers to using these desired interventions were unavailability of equipment, unavailability of services, and requirement of special training. As for interventions therapists already used and would have liked to use more frequently, 28 therapists reported 1 intervention, and 9 therapists reported 2 interventions. The top 2 interventions therapists would have liked to use more frequently were exercise therapy and vertebral mobilizations. Of the 37 therapists who reported a desire to provide an intervention "more frequently," the top 2 barriers to use were time constraints and unavailability of equipment.

Frequency of Use of Interventions With Known Evidence of Effectiveness

Therapists rated how frequently they used 13 specified interventions for a client with a condition similar to that of Mrs C. The top 3 interventions rated to be used always or often were exercise therapy (99% of therapists), advice to stay active (94%), and spinal manipulative therapy (66%). The top 3 interventions that were indicated to be used rarely or never were acupuncture (94%), back school (84%), and advice to take bed rest (84%).

Discussion

Prevalence of Interventions With Evidence of Effectiveness The findings of this study suggest that there is a low prevalence (3%) of use of interventions with strong evidence of effectiveness but that about two thirds of therapists use interventions with moderate evidence of effectiveness. The interventions with clear evidence demonstrating effectiveness for LBP were those that require no technical skills, that is, advice to stay active and advice against bed rest. Almost all therapists indicated using interventions that are classified as having limited or contradictory evidence of effectiveness, no known evidence of effectiveness, and even moderate evidence of ineffectiveness.

Of the 43 interventions mentioned by the therapists, 12 were classified as having limited or contradictory evidence of effectiveness and 18 were classified as having no known evidence of effectiveness. These findings demonstrate the paucity of high-quality research on the effectiveness of interventions that are widely used by clinicians. Therapists cannot be expected to draw clinical conclusions on the effectiveness of many of the interventions they use based on the available research. Alternatively, these findings may indicate that physical therapists are using interventions that are not well studied because they are unaware of the literature, disagree with the literature, are content with their patient outcomes, and find that changing their practice is too difficult. (37)

Interestingly, of the 10 interventions with moderate evidence of ineffectiveness, exercise therapy encompassed 8 interventions and was the most commonly used. One plausible explanation is that exercise therapy is recommended, although inconsistently, by several CPGs for acute nonspecific LBP and is recommended by almost all CPGs for subacute LBP. (21) Because the evidence on specific exercises shows a great deal of variation, it would be difficult for therapists to make any clinical decisions based on the literature, and they would have to rely on clinical experience. Therapists who use back stabilization exercises might argue that the literature evaluating that intervention's effectiveness is promising but is still in its preliminary stages (as described later). Furthermore, although some studies have shown no difference between the McKenzie approach and other therapeutic approaches, this finding may be explained, in part, by controversies over the appropriate dosage. Indeed, studies included in the Cochrane Back Review by Van Tulder et al (32) have evaluated McKenzie exercises used 3 times per day, whereas the frequency recommended by McKenzie and May (38) is 10 repetitions every hour. Van Tulder et al's conclusion is that exercise is beneficial, but the evidence does not distinguish between different types or intensities of exercise. (32) Therefore, taking the uncertainties into account, if exercise therapy is classified as having unclear evidence of effectiveness, then the prevalence of use of interventions with moderate evidence of ineffectiveness would be lower in this sample, that is, 37% rather than 93%.

The only 2 interventions with moderate evidence of effectiveness were spinal manipulation and vertebral mobilization, which were used by 68% of the therapists. This classification was based on results in a systematic review done by the Cochrane Back Review Group. (34) However, recently, there has been a change in the level of effectiveness for spinal manipulation. In a recent meta-analysis through collaboration with the Cochrane Back Review Group, the authors concluded that there is no evidence that spinal manipulative therapy is superior to other standard interventions for patients with acute or chronic LBP. (39) Based on this latest work, the authors concluded that the intervention has no statistical or clinical benefit in comparison with general practitioner care, analgesics, exercises, or back school. Indeed, spinal manipulation was superior only to sham therapy or to modalities and interventions judged to be ineffective or even harmful, such as traction, corset, bed rest, home care, topical gel, diathermy, massage, and placebo. This systematic review incorporated the results of 53 articles (with a total of 5,486 patients), representing 39 studies, and included the more recent RCTs published after 1998, which had not been included in the review by Mohseni-Bandpei et al. (34) Assendelft et al (39) acknowledged that the main limitation of their systematic review was the uneven quantity and quality of the RCTs included in the meta-analysis. This factor might explain, in part, the difference between this work and 8 previous systematic reviews that showed favorable outcomes of spinal manipulative therapy for people with acute or chronic LBP. (40-47) Nonetheless, the discrepancy between previous Cochrane Review results and those of this newer review is striking, considering there was only a 5-year difference between publication dates.

Although no previous studies have Specifically examined LBP practices according to their level of evidence of effectiveness, the results of this project are similar to those of previous surveys with respect to the prevalence of use of exercise, spinal mobilizations, and electrotherapeutic modalities. (6-9) However, the type of exercise varied across these studies. In a study based on data from 1990, the most common type of exercise was stretching. (6) In more recent studies, McKenzie exercises and back stabilization exercises were found to be the most popular. (8,9) In our study, the most popular exercise was back stabilization exercise (60%). This shift in exercise prescription may be associated with the studies done on the use of stabilization exercises for LBP, which have been emphasized in undergraduate curricula only recently. These specific spinal exercises have been shown to be effective in reducing recurrence rates for patients with acute idiopathic LBP, as compared with a control group. (48,40) Nevertheless, these studies are in their preliminary phases, and no specific recommendations for CPGs have been made regarding these interventions.

Comparison of Spontaneous Versus Prompted Responses

For particular interventions, there was a discrepancy between what therapists spontaneously reported in open-ended questions regarding Mrs C's case and how they responded to closed-ended questions on an intervention's use for a typical client such as Mrs C. For example, the prevalence for advice to stay active and advice against bed rest was found to be very low (3%). It may be argued that therapists did not mention giving "advice" to this client because they did not necessarily classify it as an intervention. However, clinicians were provided with the definition of "intervention," which included "any educational interventions." Indeed, when clinicians were specifically prompted, most recommended that bed rest should be avoided and that they would advise the patient to stay active "often" or "always." In reflection over the differences between spontaneous responses and prompted responses, we thought that therapists would have mentioned these interventions for Mrs C had they felt they were important enough. Almost half of therapists mentioned the terms "advice" or "education," but when asked to specify, they did not report "advice to stay active" or "advice against bed rest." It may be that the prompted responses were subject to social desirability bias (50); that is, respondents may have unknowingly had a tendency to report the use of interventions they thought were correct, such as interventions requiring education.

About two thirds Of the therapists gave sound advice on return to work, which was consistent with occupational CPGs. (22) Indeed, these guidelines emphasize the importance of early return to work, with temporary adaptations of work duties (if needed), even in the presence of LBP. The remaining one third of physical therapists recommended return to work most often according to the patient's pain, advice that is discouraged by these guidelines.

Timing of Use of Interventions and Their Duration

To our knowledge, this is the first study to have examined which interventions are used at different periods during the rehabilitation of a client with acute nonspecific LBP. In this study, we found that ice, education, and posture correction played the biggest roles at the initial visit. At or around discharge, we found that therapeutic approaches tended to have a more active component, emphasizing exercise therapy, home exercise programs, and functional or work conditioning. This active approach is in line with the recommendations in all of the CPGs, because they specifically consider exercise therapy as being useful 6 weeks after onset of an LBP episode. Of the 36 therapists who said they would monitor Mrs C for 6 to 10 weeks, 24 therapists reported the use of exercise therapy at discharge. However, all of these therapists also used exercise on a regular basis, that is, from the acute stage until discharge. This finding demonstrates that these therapists made no distinction on the use of exercise therapy for the acute or subacute stage, and we therefore considered that they could not be classified as users of interventions with high evidence of effectiveness.

Interestingly, the interventions that are well supported in the literature take very little time and are not always done (eg, advice to remain active). This finding suggests that time constraints are not the limiting factor for the use of effective interventions.

Factors Affecting Choice of Interventions

Surprisingly, the top factor affecting the choice of an intervention was that it had known effectiveness for LBP. However, it was clear that most interventions used by clinicians in this study had no known effectiveness. A plausible explanation for this inconsistency is that therapists interpret the term "effectiveness" as having clinically observable results based on their experience, rather than based on research results. Indeed, only 5 therapists mentioned literature searches and only 4 therapists indicated textbooks or journals as the primary factor affecting their choice of interventions. Because this sample of therapists was mainly French-speaking, it is possible that they might have been less likely to seek out the literature in English. However, these findings are consistent with a previous study that interviewed English-speaking therapists. Turner and Whitfield (51) found that original education, prior experience, and practice-related courses were the top 3 reasons for choosing specific interventions.

Diagnostic Classification of Mrs C's Low Back Pain

Although nearly all of the therapists correctly classified Mrs C as having nonspecific LBP, almost three quarters of them required clarification on the definitions of specific LBP, nonspecific LBP, and radicular syndrome. Because these classification terms are in all of the CPGs, including the Dutch Physiotherapy Guidelines, as well as the Cochrane Back Reviews, this finding may suggest that many therapists are not familiar with this literature. Indeed, only 14 therapists stated they were at least somewhat aware of the therapeutic recommendations made by the Cochrane Back Review Group, and only 7 therapists were aware of the Dutch Physiotherapy Guidelines.

Knowledge Translation

Clinicians are expected not only to seek out knowledge but also to critically appraise the evidence in order to judge its applicability within their own practice. Providing evidence-based practice for individuals with LBP is a major challenge to the physical therapist, who must combine up-to-date scientific knowledge with clinical experience and judgment. The current scientific knowledge on interventions for LBP is confusing and does not easily provide the answers that physical therapists need regarding clinical questions, such as which intervention to use with a specific client. Indeed, even if therapists strictly adhere to the CPGs, the recent example of spinal manipulative therapy shows that what appears to be sound evidence may, soon after, be disproved.

According to the Practice Style Questionnaire classification, half of the physical therapists were found to be pragmatists, that is, their practice is a function of time, patient flow, and patient satisfaction, whereas few therapists were found to be seekers of knowledge. (15) One conceivable explanation is the idea of poor knowledge translation, which is the bidirectional process of sharing knowledge between researchers and clinicians. There seems to be a gap between the current interventions that physical therapists use and those that are being studied by researchers. Perhaps the main issue is the lack of successful dissemination or implementation of the results of LBP research. In a workshop at the Fourth International Forum on Low Back Pain Research in Primary Care, it was decided that merely publishing research findings is ineffective for the uptake of research evidence by clinicians. (52) In this forum, 10 potentially effective ingredients to successful implementation were proposed; 2 are highlighted here. The first is the inclusion of a clear and strong evidence base in CPGs, which is not yet the case for physical therapy interventions. The second is communication with all relevant stakeholders (patients, professional organizations, and policy makers) in order to create guidelines that relate to current daily practice. It was suggested that all stakeholders should be "represented in the group developing the guidelines and should have the opportunity to comment on the guidelines before publication." (52)(pE125)

Users and Nonusers of Interventions With High Evidence of Effectiveness

Two thirds of the physical therapists were categorized as users of interventions with high evidence of effectiveness. This classification must be interpreted very carefully because nonusers may or may not have been aware of research evidence and, if they were aware, may have made a conscious clinical decision not to use certain interventions based on their experience. However, the results suggest that few therapists were aware of the recommendations in the Dutch Physiotherapy Guidelines and the Cochrane Back Reviews.

Recent graduation was significantly associated with being a user of interventions with high evidence of effectiveness. This finding may be explained by the recent integration of evidence-based practice principles into undergraduate physical therapy curricula. Interestingly, in the logistic model, the number of courses taken was not, by itself, significantly associated with being a user. Rather, it was the interaction of year of graduation with continuing education. This finding is consistent with the results of the study by Barnard and Wiles, (53) who found that younger physical therapists reported having the skills necessary to appraise research literature, as compared with more senior physical therapists. Both participation in LBP research and completion of a graduate degree were associated with being a user, but this relationship was not significant because too few therapists had those characteristics. Of interest, users had taken significantly more postgraduate courses than nonusers had taken. The specific courses that seemed to be related to being a user were osteopathy and manual therapy courses, both of which focus on hands-on techniques. This finding was not surprising because therapists were classified as users almost always because of their use of vertebral mobilization. Contrary to what was hypothesized, awareness of the recommendations made in the literature (by the Cochrane Back Review Group and the Dutch Physiotherapy Guidelines) was not associated with use of interventions with high evidence of effectiveness. Indeed, few physical therapists had knowledge of these guidelines.

Unexpectedly, having information resources available at work (for example, journal clubs) was not associated with being a user. This finding suggests that therapists are not accessing these resources, are not changing their practices in consequence to new knowledge, or do not have the skills to appraise the literature. It may be that availability of resources is not enough; therapists might require designated time during their workday if they are to be encouraged to seek best practice information.

Desired Practices and Perceived Barriers to Use

Overall, there was a low prevalence of desired use of interventions. It seems that most physical therapists were not dissatisfied with their practice. The most prevalent barriers were typically associated with the work environment and the physical therapist's training. For example, the reason given for not using spinal manipulation was that it requires special training.

Strengths and Limitations of This Study

Prior to recruitment, therapists were randomly selected based on their practice setting (public or private), as described in the OPPQ Web site. This strategy allowed stratification on this variable, which further ensured that therapists were equally represented in the public and private sectors. A response rate of 86.2% was achieved in this study. Comparisons of the nonparticipants with the participants did not suggest any differences in the baseline clinician characteristics.

It is important to note that all of the evidence behind physical therapy interventions is based on research of a stand-alone treatment and not as part of a comprehensive treatment program. Categorizing therapists as users or nonusers of interventions with high evidence of effectiveness was deemed not to be an ideal measure of evidence-based practice. Indeed, an evidence-based practitioner is one who integrates both research- and clinical-based evidence. Still, there were some interesting and highly plausible differences between users and nonusers, indicating that this classification may have had some validity.

One important variable that surprisingly was not associated with being a user of interventions with high evidence of effectiveness was physical therapists' practice style. A plausible explanation for this finding is that, for the purposes of this study, some alterations were made to the original questionnaire. When developed, this questionnaire was self-administered by the clinician, but for this study, it was interviewer-administered. The interviewer-administered format may have been more subject to social desirability bias, which may have artificially inflated the number of seekers, that is, those who are evidence-based practice oriented. In contrast to the results of the present study, Green et al (15) found that 2.5% of physicians were seekers, 57.0% were receptives, 12.6% were traditionalists, and 27.9% were pragmatists. Still, there were about double the number of pragmatists, suggesting that social desirability bias may not have been the only factor affecting scores on the Practice Style Questionnaire. The second major modification to the questionnaire was the language of administration, due to the high proportion of French-speaking physical therapists in the province of Quebec. Indeed, the French version of the tool has not been formally tested for validity and reliability.

This survey used a hypothetical case scenario. Previous articles(10,11) have addressed concerns about the use of vignettes to elicit information about clinicians' practices in that they may not accurately reflect therapists' actual practices. Alternatively, chart audits have been used. Although chart audits may more accurately reflect actual practice, they would not be superior because any discrepancies that are reported in practice may be related to differences in the client rather than to actual discrepancies in therapist's choices. The use of a vignette allowed for the standardization of the patient into a typical acute presentation. It also allowed the identification of interventions used when certain cues were included in the vignette.

Conclusion

It is encouraging to find that two thirds of therapists use interventions with moderate evidence of effectiveness. Interestingly, therapists tended to use many interventions for which the evidence is limited, unclear, or absent, and there appeared to be a lack of familiarity with research evidence. Further research is warranted to study interventions having a high prevalence of use but unclear or absent evidence. It also appears that targeted knowledge translation and implementation studies are needed in the area of LBP because publishing the results of effectiveness studies is not sufficient to change physical therapist practice.

Appendix.

Low Back Pain Vignette

Mrs C is a 35-year-old nurse who is seen at your outpatient physical therapy clinic 5 days after injuring her lower back while at work. She recalls feeling a twinge of pain while trying to catch a client who had fainted. The pain increased during the day such that she went home without finishing her shift. During the next 2 days, she experienced severe stabbing pain across the low back, relieved only by lying down. She did not report any pain down her leg. That evening, she went to a walk-in clinic where she was given anti-inflammatories and muscle relaxants and was referred for physical therapy with a diagnosis of lumbar strain. There were no significant findings on x-ray films.

Since her injury, she has been off work and is covered by CSST (Quebec workman's compensation board). She is a parent of 2 children and enjoys gardening. She reports having difficulty doing housework, driving, and lifting her 3-year-old daughter. Although she usually attends a gym twice a week, she has been unable to do so since her injury. The pain increases with bending, sitting, and turning. Her pain decreases with walking and standing and when she lies down on her back with pillows under her knees. She reports having had a similar episode of low back pain 2 years ago, which resolved within a few days: the only treatment she used was Motrin. (a)

On evaluation, you observe that this patient has a decreased lumbar lordosis with no shift. You also note that her range of motion is decreased in lumbar flexion and extension. She has no neurological signs in that dermatomes, myotomes, and reflexes are all normal. Your assessments of the hip and sacroiliac joints, as well as the straight leg raise and the femoral nerve stretch, do not reproduce the patient's pain.

(a) McNeil Consumer & Specialty Pharmaceuticals, Div of McNeil-PPC Inc, Camp Hill Road, Fort Washington, PA 19034.

References

(1) Bury T, Mead J. Evidence-Based Healthcare: A Practical Guide for Therapists. Oxford, United Kingdom: Butterworth-Heinemann; 1998:11.

(2) Walker BF, Muller R, Grant WD. Low back pain in Australian adults: prevalence and associated disability. J Manipulative Physiol Ther. 2004; 27:238-244.

(3) Loney PL, Stratford PW. The prevalence of low back pain in adults: a methodological review of the literature. Phys Ther. 1999;79:384-396.

(4) Frymoyer JW. Predicting disability from low back pain. Clin Orthop. 1992;279:101-119.

(5) Philadelphia Panel Evidence-Based Clinical Practice Guidelines on Selected Rehabilitation Interventions for Low Back Pain. Phys Ther. 2001;81:1641-1674.

(6) Battie MC, Cherkin DC, Dunn R, et al. Managing low back pain: attitudes and treatment preferences of physical therapists. Phys Ther. 1994;74:219-226.

(7) Foster NE, Thompson KA, Baxter GD, Allen JM. Management of nonspecific low back pain by physiotherapists in Britain and Ireland: a descriptive questionnaire of current clinical practice. Spine. 1999;24: 1332-1342.

(8) Li LC, Bombardier C. Physical therapy management of low back pain: an exploratory survey of therapist approaches. Phys Ther. 2001; 81:1018-1028.

(9) GraceyJH, McDonough SM, Baxter GD. Physiotherapy management of low back pain: a survey of current practice in Northern Ireland. Spine. 2002;27:406-411.

(10) Jones TV, Gerrity MS, Earp J. Written case simulations: do they predict physicians' behavior? J Clin Epidemiol. 1990;43:805-815.

(11) Langley GR, Tritchler DL, Llewellyn-Thomas HA, Till JE. Use of written cases to study factors associated with regional variations in referral rate. J Clin Epidemiol. 1991;44:391-402.

(12) Tamblyn R, Battista R. Changing clinical practice: which interventions work? J Contin Educ Health Prof. 1993;13:273-288.

(13) Bekkering GE, Hendriks HJM, Koes BW, et al. Dutch physiotherapy guidelines for low back pain. Physiotherapy. 2003;89:82-96.

(14) Clinical Guidelines for the Management of Acute Low Back Pain. London, United Kingdom: Royal College of General Practitioners; 1999.

(15) Green LA, Gorenflo DW, Wyszewianski L. Validating an instrument for selecting interventions to change physician practice patterns: a Michigan consortium for family practice research study. J Fam Pract. 2002 ;51:938 -942.

(16) Morgan DL, Krueger RA. The Focus Group Kit. Vols 1-6. London, United Kingdom: Sage Publications; 1998.

(17) Dillman DA. Mail and Telephone Surveys: The Total Design Method. New York, NY: Wiley-Interscience; 1978.

(18) Gore-Felton C, Koopman C, Bridges E, et al. An example of maximizing survey return rates: methodological issues for health professionals. Eval Health Prof. 2002;25:152-168.

(19) Sutherland HJ, Beaton M, Mazer R, et al. A randomized trial of the total design method for the postal follow-up of women in a cancer prevention trial. Eur J Cancer Prey. 1996;5:165-168.

(20) Nachemson AL. Newest knowledge of low back pain: a critical look. Clin Orthop. 1992;279:8-20.

(21) Koes BW, Van Tulder MW, Ostelo RWJG, et al. Clinical guidelines for the management of low back pain in primary care: an international comparison. Spine. 2001;26:2504-2514.

(22) Staal JB, Hlobil H, Van Tulder MW, et al. Occupational health guidelines for the management of low back pain: an international comparison. Occup Environ Med. 2003;60:618-626.

(23) Furlan AD, Brosseau L, Imamura M, Irvin E. Massage for Low Back Pain (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(24) Hagen KB, Hilde G, Jamtvedt G, Winnem M. Bed Rest for Acute Low Back Pain and Sciatica (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(25) Hilde G, Hagen KB, Jamtvedt G, Winnem M. Advice to Stay Active as a Single Treatment for Low Back Pain and Sciatica (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(26) Milne S, Welch V, Brosseau L, et al. Transcutaneous Electrical Nerve Stimulation (TENS)for Chronic Low Back Pain (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(27) Ostelo RWJG, de Vet HCW, Waddell G, et al. Rehabilitation After Lumbar Disc Surgery (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(28) Schonstein E, Kenny DT, Keating K, Koes BW. Work Hardening, Work Conditioning and Functional Restoration for Workers With Back and Neck Pain (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(29) Van Tulder MW, Cherkin DC, Berman B, et al. Acupuneturefor Low Back Pain (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(30) Van Tulder MW, Esmail R, Bombardier C, Koes BW. Back Schools for Non-specific Low Back Pain (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(31) Van Tulder MW, Jellema P, Van Poppel MNM, et al. Lumbar Supports for Prevention and Treatment of Low Back Pain (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(32) Van Tulder MW, Malmivaara A, Esmail R, Koes BW. Exercise Therapy for Low Back Pain (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(33) Van Tulder MW, Ostelo RWJG, Vlaeyen JWS, et al. Behavioural Treatment for Chronic Low Back Pain (Cochrane Review) [Update software]. Oxford, United Kingdom: The Cochrane Library; 2003:2.

(34) Mohseni-Bandpei MA, Stephenson R, Richardson B. Spinal manipulation in the treatment of low back pain: a review of the literature with particular emphasis on randomized controlled clinical trials. Physical Therapy Reviews. 1998;3:185-194.

(35) Armitage P, Berry G. Statistical Methods in Medical Research. Oxford, United Kingdom: Blackwell Scientific Publications; 1994:1234.

(36) Agresti A. Categorical Data Analysis. New York, NY: John Wiley & Sons Inc; 1990:100-102.

(37) Pinnington MA. Why are we finding it so hard to change our approach to low back pain? Physiotherapy. 2001;87:58-59.

(38) McKenzie R, May S. The Lumbar Spine: Mechanical Diagnosis and Therapy. Waikanae, New Zealand: Spinal Publications New Zealand Ltd; 2003.

(39) Assendelft JJ, Morton SC, Yn EI, et al. Spinal manipulative therapy for low back pain: a meta-analysis of effectiveness relative to other therapies. Ann Intern Med. 2003;138:871-881.

(40) Van Tulder MW, Koes BW, Bouter LM. Conservative treatment of acute and chronic non-specific low back pain: a systematic review of randomized controlled trials of the most common interventions. Spine. 1997;22:2128-2156.

(41) Shekelle PG, Adams AH, Chassin MR, et al. Spinal manipulation for low back pain. Ann Intern Med. 1992;117;590-598.

(42) Koes BW, Assendelft wJ, Van der Hejden GJ, Bouter LM. Spinal manipulation for low back pain: an updated systematic review of randomized clinical trials. Spine. 1996;21:2860-2871.

(43) Ottenbacher K, Di Fabio RP. Efficacy of spinal manipulation/ mobilization therapy: a meta-analysis. Spine. 1985;10:833-837.

(44) Di Fabio RP. Efficacy of manual therapy. Phys Ther. 1992;72: 853-864.

(45) Anderson R, Meeker WC, Wirick BE, et al. A meta-analysis of clinical trials of spinal manipulation. J Manipulative Physiol Ther. 1992;15: 181-194.

(46) Bronfort G. Spinal manipulation: current state of research and its indications. Neurol Clin. 1999;17:91-111.

(47) Van der Weide WE, Verbeek JH, Van Tulder MW. Vocational outcome of intervention for low back pain. Scand J Work Environ Health. 1997;23:165-178.

(48) HidesJA, Richardson CA, Jull GA. Multifidus inuscle rehabilitation decreases recurrence of symptoms following first-episode low back pain. In: Proceedings of the National Congress of the Australian Physiotherapy Association, Brisbane, Queensland, Australia, 1996. Melbourne, Victoria, Australia: Australian Physiotherapy Association; 1996.

(49) Hides JA, Jull GA, Richardson CA. Long-term effects of specific stabilizing exercises for first-episode low back pain. Spine. 2001;26: E243-E248.

(50) Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide to Their Development and Use. Oxford, United Kingdom: Oxford University Press; 1995:72-75.

(51) Turner P, Whitfield TWA. Physiotherapists' use of evidence based practice: a cross-national study. Physiother Res Int. 1997;2:17-29.

(52) Van Tulder MW, Croft PR, Van Splunteren P, et al. Disseminating and implementing the results of back pain research in primary care. Spine. 2002;27:E121-E127.

(53) Barnard S, Wiles R. Evidence-based physiotherapy: physiotherapists' attitudes and experiences in the Wessex area. Physiotherapy. 2001;87: 115-124.

C Mikhail, MSc, is Physical Therapist, Faculty of Medicine, School of Physical and Occupational Therapy, McGill University, 3630 Promenade Sir William Osler, Montreal, Quebec, Canada H3G 1Y5 ([e-mail]). Address all correspondence to Ms Mikhail.

N Korner-Bitensky, PhD, is Associate Professor, Faculty of Medicine, School of Physical and Occupational Therapy, McGill University.

M Rossignol, MD, MSc, FRCP, is Consulting Physician, Department of Occupational and Environmental Health, Direction de la Sante Publique, Montreal, Quebec, Canada.

JP Dumas, MSc, is Physical Therapist and Faculty Lecturer, Faculty of Medicine, School of Physical and Occupational Therapy, McGill University.

Ms Mikhail designed the study, conducted the focus group, collected and entered the data, analyzed the results, and prepared the publication. Dr Korner-Bitensky assisted with the focus group, provided methodological advice, and assisted with the analyses and preparation of the publication. Dr Rossignol consulted repeatedly on the design and coding of the low back pain interventions, assisted in the interpretation of the results, and reviewed the manuscript. Mr Dumas provided valuable clinical input that was used in designing the protocol, interpreting the results, and reviewing the manuscript. Special appreciation is extended to those who helped carry out this study: Pierre Allard, Natali Mahdavian, and Amelia Gaglietta for participating in the focus group; Eve Saucier for help with interviews; and Mary Mikhail for help with data entry. This research could not have been possible without the 100 physical therapists who graciously gave of their time to complete the interviews. Aliki Thomas provided valuable information and resources on knowledge translation.

Ethics approval for this study was provided by the Institutional Review Board, Faculty of Medicine, McGill University.

The first author was supported by a Master's Research bursary provided by the provincial licensing body of physical therapists, the Ordre Professionel de la Physiotherapie du Quebec (OPPQ).

This article was received January 14, 2005, and was accepted April 4, 2005.

* The interview questionnaire (English version), including the clinical vignette, is included as Appendix 1 in the article on the Physical Therapy Web site (http://www.ptjournal.org).

([dagger]) The 17-item Practice Style Questionnaire is included as Appendix 2 in the article on the Physical Therapy Web site (http://www.ptjournal.org).

Table 1.
Description of Clinician Characteristics According to Practice Setting

                                 Public                Private
                                (n = 41)              (n = 47)

Clinician Characteristics        n      %             n      %

Sex
  Female                         30   73.2            28   59.6
  Male                           11   26.8            19   40.4

Workload status
  Full-time                      26   63.4            34   72.3
  Part-time                      15   36.6            13   27.7

Clinical experience
  < 1 y                           1    2.4             0
  1-5 y                           5   12.2            10   21.3
  6-10 y                          7   17.1            10   21.3
  11-15 y                        10   24.4            13   27.7
  >15 y                          18   43.9            14   29.8

LBP (a) experience
  < 1 y                           2    4.9             1    2.1
  1-5y                            6   14.6            10   21.3
  6-10 y                          8   19.5            10   21.3
  11-15 y                         9   22.0            12   25.5
  >15 y                          16   39.0            14   29.8
Highest diploma
  Physical therapy diploma        1    2.4             2    4.3
  BScPT (b)                      38   92.7            42   89.4
  Graduate degree                 2    4.9             3    6.4

Year of graduation
  Before 1975                     6   14.6             2    4.3
  1975-1984                      12   29.3             6   12.8
  1985-1994                      13   31.7            28   59.6
  After 1994                     10   24.4            11   23.4

Average ([+ or -] SD) no.       3.34 [+ or -] 2.50   4.30 [+ or -] 2.53
of postgraduate courses taken

Teaching experience
  Academic teaching (c)           5   12.2             5   10.6
  None                           36   87.8            42   89.4

Participation in LBP research
  Primary investigator            1    2.4             0
  Other roles (d)                 1    2.4             9   19.1
  None                           39   95.1            38   80.9

Awareness of recommendations
made by Cochrane Back Review
Group
  Yes                             4    9.8             1    2.1
  Somewhat                        5   12.2             3    6.4
  No                             32   78.1            43   91.5

Awareness of recommendations
made by Dutch Physiotherapy
Guidelines
  Yes                             0                    0
  Somewhat                        3    7.3             3    6.4
  No                             38   92.7            44   93.7

Practice style (e)
  Seeker                          5   12.2             8   17.0
  Receptive                       8   12.2            11    8.5
  Traditionalist                  6   14.6             6   12.8
  Pragmatist                     22   53.7            22   46.8

                                  Both                  Total
                                (n = 12)              (N = 100)

Clinician Characteristics        n      %             n      %

Sex
  Female                          9   75.0            67   67.0
  Male                            3   25.0            33   33.0

Workload status
  Full-time                      11   91.7            71   71.0
  Part-time                       1    8.3            29   29.0

Clinical experience
  < 1 y                           0                    1    1.0
  1-5 y                           5   41.7            20   20.0
  6-10 y                          1    8.3            18   18.0
  11-15 y                         1    8.3            24   24.0
  >15 y                           5   41.7            37   37.0

LBP (a) experience
  < 1 y                           0                    3    3.0
  1-5y                            5   41.7            21   21.0
  6-10 y                          1    8.3            19   19.0
  11-15 y                         2   16.7            23   23.0
  >15 y                           4   33.3            34   34.0
Highest diploma
  Physical therapy diploma        0                    3    3.0
  BScPT (b)                      11   91.7            91   91.0
  Graduate degree                 1    8.3             6    6.0

Year of graduation
  Before 1975                     0                    8    8.0
  1975-1984                       3   25.0            21   21.0
  1985-1994                       3   25.0            44   44.0
  After 1994                      6   50.0            27   27.0

Average ([+ or -] SD) no.       3.58 [+ or -] 2.31   3.82 [+ or -] 2.51
of postgraduate courses taken

Teaching experience
  Academic teaching (c)           3   25.0            13   13.0
  None                            9   75.0            87   87.0

Participation in LBP research
  Primary investigator            1    8.3             2    2.0
  Other roles (d)                 0                   10   10.0
  None                           11   91.7            88   88.0

Awareness of recommendations
made by Cochrane Back Review
Group
  Yes                             1    8.3             6    6.0
  Somewhat                        0                    8    8.0
  No                             11   91.7            86   86.0

Awareness of recommendations
made by Dutch Physiotherapy
Guidelines
  Yes                             1    8.3             1    1.0
  Somewhat                        0                    6    6.0
  No                             11   91.7            93   93.0

Practice style (e)
  Seeker                          2   16.7            15   15.0
  Receptive                       2    8.3            21   21.0
  Traditionalist                  2   16.7            14   14.0
  Pragmatist                      6   50.0            50   50.0

(a) LBP = low back pain.

(b) BScPT = Bachelor of Science in Physiotherapy degree.

(c) Academic teaching included teaching the following
university-level courses: e1ectrotherapy, functional anatomy,
biomechanics, orthopedics, low back pain lecture, massage, taping,
manual therapy, and rheumatology.

(d) Other roles included chart abstraction, survey administration,
evaluation of patient outcomes, and filling out survey questionnaires.

(e) As cited by Green et al. (15)

Table 2.
Description of Environmental Characteristics According to Practice
Setting

                                  Public               Private
                                 (n = 41)              (n = 47)

Environmental Characteristics    n     %               n     %

Administrative region (a)
  <4 PTs (b) recruited            5   12.2             5   10.6
  4-6 PTs recruited              13   31.7            16   34.0
  >6 PTs recruited               23   56.1            26   55.3

Rurality
  Nonrural                       37   90.2            43   91.5
  Rural                           4    9.8             4    8.5

Practice setting (as per recruitment)
  General hospital               35   85.4             1    2.1
  Rehabilitation center           5   12.2             0
  CLSC (c)                        1    2.4             0
  Home care                       0                    2    4.3
  Private practice                0                   44   93.6

Type of hospital (d)
  Teaching hospital              22   53.7             0
  Nonteaching hospital           19   46.3             0

Training of PT students
  Provided                       34   82.9            20   42.6
  Not provided                    7   17.1            27   57.4

Available information resources
  Internet access                34   82.9            35   74.5
  University databases           22   53.7            13   27.7
  Journal clubs                  18   43.9            10   21.3
  On-site library                34   82.9             0
  Journal subscription           34   82.9            19   40.4

Average ([+ or -] SD) no. of    3.46 [+ or -] 1.32   1.66 [+ or -] 1.26
information resources
available

                                  Both                 Total
                                  (n = 12            (N = 100)

Environmental Characteristics     n     %             n     %

Administrative region (a)
  <4 PTs (b) recruited            3    25.0           13   13.0
  4-6 PTs recruited               1     8.3           30   30.0
  >6 PTs recruited                8    66.7           57   57.0

Rurality
  Nonrural                       10    83.3           90   90.0
  Rural                           2    16.7           10   10.0

Practice setting (as per recruitment)
  General hospital                7    58.3           43   43.0
  Rehabilitation center           1     8.3            6    6.0
  CLSC (c)                        0                    1    1.0
  Home care                       1     8.3            3    3.0
  Private practice                3    25.0           47   47.0

Type of hospital (d)
  Teaching hospital               8    66.7           30   56.6
  Nonteaching hospital            4    33.3           23   43.4

Training of PT students
  Provided                       11    91.7           65   65.0
  Not provided                    1     8.3           35   35.0

Available information resources
  Internet access                12   100             81   81.0
  University databases            5    41.7           40   40.0
  Journal clubs                   6    50.0           34   34.0
  On-site library                 9    75.0           43   43.0
  Journal subscription           11    91.7           64   64.0

Average ([+ or -] SD) no. of    3.58 [+ or -] 1.00   2.63 [+ or -] l.55
information resources
available

(a) Denotes the administrative regions from which fewer than 4
therapists were recruited, 4-6 therapists were recruited, or greater
than 6 therapists were recruited.

(b) PT = physical therapist.

(c) CLSC = centre local de service communautaire (translated as "local
center for community service").

(d) This only applies to therapists who work in the hospital setting.
The denominator includes the 53 therapists who work in either the
public sector only or both the private and public sectors.

Table 3.
Description of Client Characteristics According to Practice Setting

                                 Public               Private
                                (n = 41)             (n = 47)
Client Characteristics           n      %             n      %

Average ([+ or -] SD) no. of    2.07 [+ or -] 1.54   4.85 [+ or -] 2.61
clients with LBP (a) per day

Typical client age
  25-40 y                         7   17.1            21   44.7
  41-55 y                        21   51.2            25   53.2
  56-70 y                        12   29.3             1    2.1
  >70 y                           1    2.4             0

Typical treatment duration
  <30 min                         0                    1    2.1
  30-60 min                      31   75.6            38   80.9
  61-90 min                       9   22.0             8   17.0
  >90 min                         1    2.4             0

Typical treatment frequency     2.39 [+ or -] 0.63   2.40 [+ or -] 0.95
([bar.X]-SD) (b)

Typical treatment length
  <2 wk                           0                    0
  2-5 wk                         14   34.1            24   51.1
  6-10 wk                        23   56.1            21   44.7
  11-15 wk                        4    9.8             2    4.3
  >15 wk                          0                    0

                                 Both                 Total
                                (n = 12)             (N = 100)
Client Characteristics           n      %             n      %

Average ([+ or -] SD) no. of    4.25 [+ or -] 3.25   3.64 [+ or -] 2.66
clients with LBP (a) per day

Typical client age
  25-40 y                         7   58.3            35   35.0
  41-55 y                         4   33.3            50   50.0
  56-70 y                         1    8.3            14   14.0
  >70 y                           0                    1    1.0

Typical treatment duration
  <30 min                         1    8.3             2    2.0
  30-60 min                      10   83.3            79   79.0
  61-90 min                       1    8.3            18   18.0
  >90 min                         0                    1    1.0

Typical treatment frequency     2.33 [+ or -] 0.65   2.39 [+ or -] 0.79
([bar.X]-SD) (b)

Typical treatment length
  <2 wk                           0                    0    0
  2-5 wk                          5   41.7            43   43.0
  6-10 wk                         7   58.3            51   51.0
  11-15 wk                        0                    6    6.0
  >15 wk                          0                    0    0

(a) LBP = 1ow back pain.

(b) Denotes IIIc number of sessions per week of physical therapy
treatments.

Table 4.

Prevalence, Timing of Use, and Average Duration of Interventions
According to Level of Evidence of Effectiveness (a)

                                      Prevalence
Interventions Reported According      (95% CI)
to Their Level of Evidence            (N=100)

Strong evidence of effectiveness       3.0% (0.68.5)
  Advice to stay active                       3
  Advice against bed rest                     1
Moderate evidence of effectiveness    68.0% (57.9-77.0)
 Spinal manipulative therapy (c)
 Spinal manipulation                          3
  Vertebral mobilization                     68
Limited/contradictory evidence        90.0% (82.4-95.1)
of effectiveness
 Acupuncture                                  1
 Back school                                  2
 Electrotherapy
  Biofeedback                                 1
  Diadynamic current                          6
  Diathermy                                   1
  Interferential current                     56
  Microcurrent                                2
 Laser                                        5
 Lumbar supports                              0
 Massage
  Soft tissue massage                        28
  Myofascial release                          7
 TENS                                        20
 Ultrasound                                  61
No known evidence                     96.0% (90.1-98.9)
  Education (d)                              42
  Functional/work conditioning (e)           9/3
  Heat                                       45
  Home exercise program (f)                  22
  Ice                                        66
  Mulligan techniques (g)                     1
  Osteopathic techniques                      9
  OT referral                                 1
  Pool therapy (f)                            1
  Postural correction                        38
  Proprioceptive taping                       1
  Reassurance (h)
  RPG                                         5
  Sahrmann techniques (f)                     3
  Static positioning                          6
  Stretch and spray                           1

Moderate evidence of ineffectiveness  93.0% (86.1-97.1)
 Exercise therapy
  Aerobic conditioning                        8
  Back stabilization exercises (i)           60
  Gym program (i)                             5
  Lumbar ROM/stretches                       53
  Leg ROM/stretches                          13
  Lumbar strength exercises                  26
  Leg strength exercises                     11
  McKenzie approach                          36
Traction
  Manual                                     34
  Mechanical                                  3
Strong evidence of ineffectiveness
  Advice to take bed rest                     0

                                      Timing of Use (b) (N=100)

Interventions Reported According
to Their Level of Evidence            Reg   Initial   PRN   D/C

Strong evidence of effectiveness
  Advice to stay active                1       2
  Advice against bed rest                      1
Moderate evidence of effectiveness
 Spinal manipulative therapy (c)      50              19
 Spinal manipulation
  Vertebral mobilization
Limited/contradictory evidence
of effectiveness
 Acupuncture                           1
 Back school                                                 2
 Electrotherapy                       37       4      21     1
  Biofeedback
  Diadynamic current
  Diathermy
  Interferential current
  Microcurrent
 Laser                                 2               3
 Lumbar supports
 Massage                              22       2       9     1
  Soft tissue massage
  Myofascial release
 TENS                                  8       1      11
 Ultrasound                           42       6      15
No known evidence
  Education (d)                       19      19       6     3
  Functional/work conditioning (e)     1       1       2    10
  Heat                                23       3      21
  Home exercise program (f)            2       1       4    15
  Ice                                 28      21      25
  Mulligan techniques (g)              1
  Osteopathic techniques               9               1
  OT referral                                                1
  Pool therapy (f)                                     1
  Postural correction                 19      17       2     5
  Proprioceptive taping                                1
  Reassurance (h)                      1       1       1
  RPG                                  3               2
  Sahrmann techniques (f)              3
  Static positioning                   5       1
  Stretch and spray                                    1

Moderate evidence of ineffectiveness
 Exercise therapy                     88              18    24
  Aerobic conditioning
  Back stabilization exercises (i)
  Gym program (i)
  Lumbar ROM/stretches
  Leg ROM/stretches
  Lumbar strength exercises
  Leg strength exercises
  McKenzie approach
Traction                              19              16
  Manual
  Mechanical
Strong evidence of ineffectiveness
  Advice to take bed rest

                                      Total
                                      Average
Interventions Reported According      Duration
to Their Level of Evidence            (min)([+ or -SD])

Strong evidence of effectiveness
  Advice to stay active               5.0 [+ or -] 0
  Advice against bed rest             2.0
Moderate evidence of effectiveness
 Spinal manipulative therapy (c)      10.6 [+ or -] 6.2
 Spinal manipulation
  Vertebral mobilization
Limited/contradictory evidence
of effectiveness
 Acupuncture                          13.0
 Back school                          330.0 [+ or -] 42.4
 Electrotherapy                       16.1 [+ or -] 8.1
  Biofeedback
  Diadynamic current
  Diathermy
  Interferential current
  Microcurrent
 Laser                                4.4 [+ or -] 1.5
 Lumbar supports
 Massage                              8.2 [+ or -] 5.5
  Soft tissue massage
  Myofascial release
 TENS                                 20.6 [+ or -] 3
 Ultrasound                           5.3 [+ or -] 1.5
No known evidence
  Education (d)                       10.1 [+ or -] 7.3
  Functional/work conditioning (e)    17.4 [+ or -] 9.0
  Heat                                17.5 [+ or -] 3.5
  Home exercise program (f)           11.5 [+ or -] 7.8
  Ice                                 15.2 [+ or -] 8.4
  Mulligan techniques (g)             20.00
  Osteopathic techniques              23.4 [+ or -] 16.9
  OT referral                         1.0
  Pool therapy (f)                    30.0
  Postural correction                 11.0 [+ or -] 7.6
  Proprioceptive taping               1.0
  Reassurance (h)                     4.3 [+ or -] 1.1
  RPG                                 21.0 [+ or -] 12.5
  Sahrmann techniques (f)             9.0 [+ or -] 5.0
  Static positioning                  8.3 [+ or -] 6.0
  Stretch and spray                   2.0

Moderate evidence of ineffectiveness
 Exercise therapy                     24.9 [+ or -] 11.7
  Aerobic conditioning
  Back stabilization exercises (i)
  Gym program (i)
  Lumbar ROM/stretches
  Leg ROM/stretches
  Lumbar strength exercises
  Leg strength exercises
  McKenzie approach
Traction                              9.0 [+ or -] 4.7
  Manual
  Mechanical
Strong evidence of ineffectiveness
  Advice to take bed rest

(a) Cl=confidence interval, TENS= transcutaneous electrical nerve
stimulation, OT=occupational therapy, RPG=reeducation postural'
globale (translated as "global postural re-education"), ROM=range of
motion.

(b) Reg=on a regular basis, initial-first few treatments, PRN=on an
as-needed basis, D/C=at discharge.

(c) Spinal manipulation involves a high-velocity, small-amplitude
thrust of the vertebral segments, whereas mobilization involves a
low-velocity, large-amplitude movement of the vertebral segments.

(d) When therapist  mentioned "education," they were asked to specify.
Their responses included the following: advice on lifting techniques,
explanation anatomical of cause for the low back
pain, postural
recommendations, biomechanical awareness during activities, sleep
positions, ergonomics at work, use of ice or heat at home, advice
on leisure activities, and prevention of recurrence by avoiding
activities that increase symptoms. When mentioned, activities to be
avoided included flexed positions, prolonged sitting, heavy lifting.
The "education" category did not include advice to stay active because
this was classified as a separate intervention with strong evidence of
effectiveness.

(e) Functional conditioning consists of in-clinic interventions, and
work conditioning is (lone at patient's workplace.

(f) These are interventions that can be considered as exercise.
However, because they were not formally included in the Cochrane Rack
Review, they as interventions with no known evidence.

(g) These techniques were classified as can he considered as manual
therapy. However, because they were not formally included in the
Cochrane Back Review, they interventions with no known evidence.

(h) Reassurance involves telling tire patient that nonspecific low back
pain is not harmful and that prognosis is goof.

(i) Exercises such as training of the transverus abdominus and
multifidus muscle with the use of a Swiss ball, a pressure biofeedback
unit, or no equipment were all classified under "stabilization"
exercises.

(j) Includes aerobic and resistance training.

Table 5. Clinician Variables According to Clinicians' Classification as
User-Nonuser of Interventions With High Evidence of Effectivess

                                                          % of Users
Clinician Variables                            n = 100      (n = 68)

Sex
  Female                                             67          64.2
  Male                                               33          75.8
Workload status
  Full-time                                          70          70.0
  Part-time                                          30          63.3
Clinical experience (b)
  0-5 y                                              21          71.4
  6-10 y                                             18          88.9
  11-15 y                                            24          75.0
  >15 y                                              37          51.4
Low back pain experience (b)
  0-5 y                                              24          66.7
  6-10 y                                             19          84.2
  11-15 y                                            23          78.3
  >15 y                                              34          52.9
Highest diploma received
  Basic PT training                                  94          66.0
  Graduate degree                                     6         100.0
Year of graduation (e)
  Before 1975                                         8          25.0
  1975-1984                                          21          57.1
  1985-1994                                          44          77.3
  After 1994                                         27          74.1
No. of courses taken (e)
  0-5                                                78          61.5
  [greater than or equal to]6                        22          90.9
Teaching experience
  Academic teaching                                  15          80.0
  No academic teaching                               85          65.9
Participation in low back pain research
  Yes                                                12          91.7
  No                                                 88          64.8
Awareness of recommendations made by
    Cochrane Back Review Group
  Yes/somewhat                                       14          64.3
  No                                                 86          68.6
Awareness of recommendations made by
    Dutch Physiotherapy Guidelines
  Yes/somewhat                                        7          71.4
  No                                                 93          67.7
Practice style (g)
  Seeker                                             15          73.3
  Receptive                                          21          76.2
  Traditionalist                                     14          64.3
  Pragmatist                                         50          64.0

                                          % of Nonusers
Clinician Variables                           (n = 32)    [p.sup.a]

Sex
  Female                                           35.8        .2432
  Male                                             24.2
Workload status
  Full-time                                        30.0        .5125
  Part-time                                        36.7
Clinical experience (b)
  0-5 y                                            28.6    .0296 (c)
  6-10 y                                           11.1    .0176 (d)
  11-15 y                                          25.0
  >15 y                                            48.6
Low back pain experience (b)
  0-5 y                                            33.3    .0729 (c)
  6-10 y                                           15.8    .0839 (d)
  11-15 y                                          21.7
  >15 y                                            47.1
Highest diploma received
  Basic PT training                                34.0        0.918
  Graduate degree                                   0.0
Year of graduation (e)
  Before 1975                                      75.0    .0200 (f)
  1975-1984                                        42.9    .0046 (d)
  1985-1994                                        22.7
  After 1994                                       25.9
No. of courses taken (e)
  0-5                                              38.5    .0091 (c)
  [greater than or equal to]6                       9.1
Teaching experience
  Academic teaching                                20.0        .2213
  No academic teaching                             34.1
Participation in low back pain research
  Yes                                               8.3        .0536
  No                                               35.2
Awareness of recommendations made by
    Cochrane Back Review Group
  Yes/somewhat                                     35.7        .7405
  No                                               31.4
Awareness of recommendations made by
    Dutch Physiotherapy Guidelines
  Yes/somewhat                                     28.6        .6024
  No                                               32.3
Practice style (g)
  Seeker                                           26.7        .7576
  Receptive                                        23.8
  Traditionalist                                   35.7
  Pragmatist                                       36.0

(a) Unless otherwise indicated, P values were the results of chi-square
testing or the Fisher exact test when chi-square values were not
considered valid. Italicized P values indicate a significant
association.

(b) Because there were very few therapists in the less than a year
category, less than a year and 1 to 5 years of experience were collapsed
under 0 to 5 years.

(c) Based on chi-square testing.

(d) Based on the Cochran-Arrnitage trend test.

(e) These are colnlnllons Variables that were transtormed into
categorical variables based on fretluency distributions of the
data for the purpose of analyzing cutoffs.

(f) Based on the Fisher exact test.

(g) As cited by Green et al. (15)

Table 6.
Regression Coefficients and Their Level of Significance for the
Variables Included in the Model (a)

               B        SE    Wald Test   df P (a)   Exp(B)

Step 2        0.069   0.025       7.478       1006    1.072
Step 1        0.000   0.000       7.717       1005    1.000
Constant   -138.392  50.550       7.495       1006    0.000

                     95% CI   for Exp(B)

           Lower     Upper

Step 2     1.020     1.127
Step 1     1.000     1.000
Constant

(a) Variables entered on step 1: number of postgraduate courses x
year of graduation. Variable entered on step 2: year of graduation.
CI=confidence interval.