Abstract
Background:
A strong understanding of diagnostic imaging has been advocated for physical therapists. There have been recent changes in physical therapy curricula and increased opportunities to utilize imaging during clinical practice
Purpose:
The aim of this study was to explore the ability of practicing clinicians to accurately identify selected musculoskeletal conditions on plain‐film radiograph (X‐ray), magnetic resonance imaging (MRI), and computed tomography scan (CT scan). Further, to determine whether improvements in identification of pathology occur when the clinical scenario is added to the imaging and whether there are related training/exposure factors.
Methods:
A cross‐ sectional electronic survey was sent out to physical therapists in the state of Ohio. Participants were asked to identify conditions (cervical fracture, anterior cruciate ligament tear, and avascular necrosis of the femoral head) first given diagnostic images only, and then given the images and a clinical scenario.
Results:
Eight hundred sixty‐six surveys of the 7537 sent out were eligible for analysis. With clinical scenarios, 61.3% of respondents were correct with the ACL injury identified on MRI, 36.4% for identification of the cervical spine fracture on CT and 25.6% for identification of avascular necrosis on plain film. The accuracy significantly improved (p<0.01) with the addition of the clinical information for all three of the diagnoses. The most remarkable improvement was seen with the AVN diagnosis on plain film radiograph (365.5% improvement), followed by the ACL injury on MRI (27.2% improvement) and cervical fracture diagnosis on CT scan (17.8% improvement). Finally, formal and informal training, board certification through the APTA and to a lesser extent, degree level, all improved diagnostic accuracy.
Conclusions:
A clinical scenario paired with images notably improved identification of pathology. Physical therapists were better at identifying the ACL pathology that was presented on MRI. This is a common diagnosis to physical therapists and was paired with a relatively common imaging modality. This study suggests that physical therapists can improve accuracy with identifying pathologies on diagnostic images through a physical therapy curriculum or post‐graduation through certifications and continuing education.
Level of Evidence:
Level 4
Keywords: certifications, degree, diagnostic imaging, education, survey
Introduction
The American Physical Therapy Association's (APTA) Vision 2020 statement was created as a foundation for physical therapists' autonomous practice.1 Professional progression toward autonomous practice requires knowledge advancement and competencies in areas such as diagnostic imaging, pharmacology, and surgical applicability; since patients commonly consult with their physical therapist on care‐related matters beyond the therapist's traditional scope of practice. Advancements in these areas, specifically imaging, have occurred through formal curricular training, through extra‐curricular and extra‐clinical training, and direct exposure/responsibilities in clinical practice.
As early as 1999, Threlkeld et al2 underscored the importance of diagnostic imaging in Doctorate of Physical Therapy (DPT) curriculums in order to aid in the patient evaluation, diagnosis, and management.2 As of 2007, most DPT curricula had adopted formal imaging training within required coursework.3 Training may be embedded in practice‐based courses or may consist of a dedicated class that is focused on identification of structures and pathologies on multiple imaging modalities to aid in clinical decision making. The importance of training related to diagnostic imaging has been quantified via survey, in which 97% of the academic and clinical physical therapist respondents felt that diagnostic imaging should be included in entry‐level doctorate of physical therapy curricula.3
Practicing physical therapists also feel that extra‐curricular and extra‐clinical knowledge and understanding of imaging could potentially improve patient care skills.4 Despite this, there is an understanding that most physical therapists rely solely on an external interpretation of an image (e.g., a radiologist's interpretation), rather than viewing the image to gain further understanding of their patient's presentation.5 While the authors agree that the responsibility of interpreting the image belongs to the radiologist, viewing the image and being able to recognize relevant pathology can aid in management of patients' conditions. Deyle6 emphasized that physical therapists (regardless of imaging privileges) need to be aware of imaging indications, risks and benefits and should consider requesting an imaging study with appropriate rationale. Imaging knowledge could improve a clinician's ability to refer for imaging when results are poor, when imaging may influence management decisions and when a positive finding for imaging may prompt the need for proper surgical intervention.6
Direct exposure and responsibilities have also elevated imaging knowledge among physical therapists. Many physical therapists are employed in direct access environments and have capabilities to order diagnostic images.7‐11 Military‐based physical therapists serve as physician extenders, practice direct access, and have the ability to order diagnostic images and certain medications as well as restrict work and activity.8 Civilian physical therapists in the United States as well as other countries are also practicing in this manner. 9,11‐15 Evidence exists that when physical therapists refer patients for imaging they aid in preventing overutilization of imaging techniques, as physical therapists tend to be appropriately conservative. 8,12 In addition to ordering images in appropriate circumstances,10,11,16 physical therapists have better clinical diagnostic accuracy than all providers except orthopaedic surgeons (with whom they were shown not to be different). 16 Additionally, when patients are managed by physical therapists working as physician extenders versus management by orthopedic surgeons, there is no clinically or statistically difference in outcomes, except an improvement in patient satisfaction for those being managed by physical therapists.12
McKinnis5 emphasizes benefits that exist when physical therapists view the images themselves: 1) viewing provides a more comprehensive evaluation and 2) it provides a unique perspective from a profession (physical therapy) that often assesses information on an image differently than a referring physician or radiologist. Past studies have shown that when clinical information is added to the assessment of an image for medical physicians, diagnoses are more accurate.17 To the authors' knowledge, there are no studies that have evaluated whether identification of pathology on imaging improves with the addition of clinical information.
For physical therapists, a strong understanding of all aspects of diagnostic imaging has previously been advocated.18 Since recent changes in physical therapy formal curricula and opportunities to utilize imaging during clinical practice have notably increased, the authors were interested in the ability of practicing clinicians to accurately identify selected musculoskeletal conditions on plain‐film radiograph (X‐ray), magnetic resonance imaging (MRI), and computed tomography scan (CT scan). Secondarily, the authors were interested in examining whether improvements in identification of pathology occur when the clinical scenario is added to the imaging and whether there are training/exposure‐related factors associated with the accuracy of the diagnosis. In particular, training/exposure‐related factors of: a) imaging training, b) different levels of educational training, and c) value of certification of specialization within the APTA were of interest. Findings may help assess appropriateness of physical therapists' capacity to view diagnostic images in clinical practice and may identify areas of training/exposure that are superior to others. To the authors' knowledge there are no studies that have evaluated the ability of licensed physical therapists to identify pathology on selected images. Nor have studies investigated the value of the clinical scenario or training/exposure factors and how they related to physical therapist's ability to identify pathology on images. Therefore, the purpose of this study was to explore the ability of practicing clinicians to accurately identify selected musculoskeletal conditions on plain‐film radiograph (X‐ray), magnetic resonance imaging (MRI), and computed tomography scan (CT scan) and to determine whether improvements in identification of pathology occur when the clinical scenario is added to the imaging, and whether there are related training/exposure factors.
Methods
Design
The study involved a cross sectional, electronic survey completed in November of 2013. In order to complete the survey, respondents were required to provide electronic informed consent. The study was approved by the Walsh University Human Subject Review Committee (13‐43).
Survey Development
An 18‐question survey was developed by two of the authors (AM and AH). The survey has three primary sections, with the first section asking the respondent to rate how confident he was in his ability to interpret diagnostic imaging results on either plain film radiograph, MRI or CT scan. The next section involved multiple images (plain film radiograph, an MRI, and a CT scan) that were representative of three actual diagnoses (cervical fracture, anterior cruciate ligament tear, and avascular necrosis of the femoral head). (The actual images can be found in the appendix.) First, the participant was given the image only and asked to type a diagnosis in the space provided. Permission for the images was received by Radiopaedia.org and MedPix (rad.usuhs.edu). Order of viewing was not randomized and included 1) anterior cruciate ligament injury then 2) cervical fracture and then 3) avascular necrosis of the femoral head. Each of the three clinical scenarios and survey questions are represented in Appendix I.
The second part to each question included the same images but also included a clinical scenario associated with those images. The clinical scenario provided the respondent with the mechanism of injury (tennis match, MVA, insidious onset). The clinical scenarios are included in Appendix 1. Once again the respondent was asked to type a diagnosis in the available space. The last section of the survey was designed to outline the exposure, training and work environment of the respondent.
Case Content Validity
The lead author wrote the cases and found published images to use. Information on mechanisms of injury, signs and symptoms, risk factors, and differential diagnosis were extracted from the literature and used to formulate the cases. The cases were then passed on to the other authors who are content experts in orthopedic physical therapy. These authors reviewed the cases to ensure appropriate content and clarity. One content expert has 24 years of clinical experience specializing in orthopedics and orthopedic research with over 170 peer reviewed publications in that area. The other content expert has been treating patients with orthopaedic dysfunction for 15 years, has achieved a manual therapy and orthopaedic clinical specialization and has been teaching orthopaedic content for 5 years.
Respondents
Convenience sampling was used to recruit respondents. Because of the availability of a database, which included all practicing physical therapists, Ohio‐based physical therapists were targeted. In order to be eligible for the study, the physical therapist had to have an active license to practice physical therapy as well as a valid email address on file with the Ohio Occupational Therapy, Physical Therapy, and Athletic Trainers Board. Respondents were recruited via email.
Procedure
The survey was distributed through SurveyMonkey, Inc. (US). The survey was open for 5½ weeks; four reminder emails with the survey link were sent during that period requesting participation. Survey responses were confidential and could not be traced back to the participant. To further reduce the risk of bias during the coding process each variable in question was coded in an isolated fashion.
Data Analysis
Descriptive statistics, diagnostic accuracy of imaging detection, with and without clinical scenarios, and associations related to exposure, training and confidence levels of the respondents were calculated using SPSS 22.0. Descriptive statistics were first derived using frequencies, to produce the population frequencies. Chi‐Square analyses were used to determine the difference in correctness in interpretation for each set of images with and without the clinical vignette. Accuracy of the participants' written responses was determined by the lead author. The radiologists' published interpretation of the images utilized in this study was used as the standard for comparison. Regression analyses, using binary logistic regressions were used to predict exposure, training and confidence levels of the respondents. In all three regression analyses, dependent variables included the accuracy of the plain‐film radiograph, the MRI, and the CT scan, all with the clinical scenarios included. In all three categories the reference category was an “inaccurate” diagnosis. For all regression analyses, beta values with standard deviations, odds ratios and 95% confidence intervals, and p values were reported. A p value of <0.05 was used for all analyses for determining statistically significant differences.
Results
A total of 7537 surveys were sent out to practicing Ohio physical therapists. One thousand one hundred forty (1140) surveys were returned for a response rate of 15.1%. One hundred and eleven (111) subjects opted out of the survey, and 47 emails were reported as “undeliverable”. Of the 1140 surveys completed, 274 did not complete the imaging questions. Subsequently, 866 responses were included in data analysis for a final response rate of 11.5%. There were no missing values in the final 866 respondents included in the data analyses.
Descriptive characteristics of the participants are shown in Table 1. Half the respondents are doctorally trained. The majority work in an outpatient setting and are employed full time. Twelve percent (12%) of the respondents earned an APTA certification or a different certification that was not sanctioned by the APTA. The most commonly acquired imaging training was received as a requirement for an entry level physical therapy degree.
Table 1.
Descriptive Statistics.
| Variable | Frequency (percentage) |
|---|---|
| Highest Degree Earned | DPT/PhD = 50.0% |
| MPT/BSPT= 50.0% | |
| Entry Level Physical Therapy Degree | DPT= 33.6% |
| MPT/BSPT= 66.4% | |
| Practice Setting | Outpatient (Hospital‐ Based, Private Outpatient, Health and Wellness Facility) = 53.1% |
| Inpatient (Acute care, Subacute, SNF/ECF/ICF) = 27.5% | |
| Other (Home Care, School System, Academic Institution, Research Center, Industry, All Other) =19.4% | |
| Specialization | OCS and SCS = 8.3% |
| Non OCS and SCS = 18.1% | |
| APTA Certifications = 12.0% | |
| Non APTA Certifications = 14.7% | |
| Employment Status | Full‐time = 73.2% |
| Part‐time = 17.3% | |
| PRN=5.1% | |
| Retired or Unemployed= 4.3% | |
| Imaging Training | Requirement for Entry Level Physical Therapy degree = 45.3% |
| Requirement for Post Professional degree = 11.1% =11.1% | |
| CEU Course = 6.1% | |
| Other = 23.1% |
Table 2 demonstrates the accuracy levels of the participants in identifying the correct pathology on MRI, plain films, and CT scan with and without the clinical scenario. The highest level of accuracy was noted with the ACL injury noted on MRI followed by the cervical spine fracture on CT and finally avascular necrosis on plain film. In all three cases, the accuracy significantly improved (p<0.01) with the addition of the clinical information. The most remarkable improvement was seen with the AVN diagnosis on plain film radiograph where accuracy improved by 365.5% when provided with the clinical scenario. Accuracy improved by 27.2% for the ACL injury on MRI and by 17.8% for the cervical fracture diagnosis on the CT scan.
Table 2.
Accuracy of Identification of pathology on Imaging and with Imaging plus Clinical Scenarios.
| Item | With Imaging Only | With Imaging and Clinical Scenario | P value |
|---|---|---|---|
| Anterior Cruciate | 417 (48.2%) = correct | 531 (61.3%) = correct | <0.01 |
| Ligament Rupture (MRI) | 449 (51.8%) = incorrect | 335 (38.7%) = incorrect | |
| Avascular Necrosis of the Hip (Radiograph) | 48 (5.5%) = correct | 222 (25.6%) = correct | <0.01 |
| 818 (94.5%) =incorrect | 644 (74.45.8%) = incorrect | ||
| Cervical Spine (CT) | 268 (30.9%) = correct | 315 (36.4%) = correct | <0.01 |
| 598 (69.1%) = incorrect | 551 (63.6%)= incorrect |
MRI= Magnetic resonance image; CT= Computerized tomography
Although participants demonstrated the greatest number of correct responses with the MRI (ACL) question, there was a wide array of incorrect answers. When presented with the image only, over 25% of participants responded with a diagnosis of a variety of different fractures. Other incorrect diagnoses included meniscus tears, posterior cruciate ligament injury, arthritis and degenerative joint disease.
When presented with the image only for the CT‐scan (cervical fracture) question participants responded the most, with either spondylosis or spondylolisthesis (20.7%), upper cervical fractures (17.1%), and degenerative disc disease (11.7%). When presented with the clinical scenario in addition to the image respondents continued to respond with spondylosis or spondylolisthesis and upper cervical fractures as the leading responses, but less respondents diagnosed the patient with degenerative disc disease. Forms of cervical nerve root impingement, specifically between C4‐C6, were also incorrect answers provided when with the image and clinical scenario were given.
The plain film radiograph (hip AVN) had similar incorrect responses for both the image only and image and clinical scenario questions. Over half of all incorrect answers were some form of arthritis. Other prevalent incorrect responses were degenerative joint disease and various fractures.
All forms of training improved diagnostic accuracy for diagnosis of the ACL, Hip AVN, and cervical fracture images and clinical scenarios. The most robust improvements depended on the imaging scenario as imaging training during continuing education courses (CEU) was the strongest predictor for accuracy for the MRI (ACL), other formal or informal imaging training was strongest for accuracy of the CT‐Scan (Cervical Fracture) and imaging training during entry level program was the strongest for the plain‐film radiograph (hip AVN) (Table 3).
Table 3.
Imaging Training associated with Accuracy of identification of pathology on imaging plus Clinical Scenario Assessment
| Variable | Odds Ratio (95% CI) | p‐ value |
|---|---|---|
| Correctness of ACL Diagnosis | ||
| Imaging training during entry level program | 3.26 (2.30, 4.62) | <0.01 |
| Imaging training post‐graduate degree | 2.72 (1.63, 4.52) | <0.01 |
| Imaging training during CEU courses | 3.72 (1.82, 7.60) | <0.01 |
| Other formal or informal imaging training | 2.99 (2.00, 4.48) | <0.01 |
| Correctness of AVN Diagnosis | ||
| Imaging training during entry level program | 3.46 (2.22, 5.39) | <0.01 |
| Imaging training post‐graduate degree | 2.73 (1.56, 4.79) | <0.01 |
| Imaging training during CEU courses | 3.22 (1.71, 6.05) | <0.01 |
| Other formal or informal imaging training | 4.24 (2.60, 6.92) | <0.01 |
| Correctness of Cervical Fracture Diagnosis | ||
| Imaging training during entry level program | 5.28 (3.46, 8.05) | <0.01 |
| Imaging training post‐graduate degree | 5.27 (3.07, 9.04) | <0.01 |
| Imaging training during CEU courses | 4.82 (2.50, 9.29) | <0.01 |
| Other formal or informal imaging training | 3.76 (2.35, 6.01) | <0.01 |
Highest degree earned and primary degree received associated with accuracy are seen in Table 4. Possessing a DPT as a primary degree versus an undergraduate degree was a strong predictor for accuracy for the MRI (ACL), but was not significant for the plain‐film radiograph (hip AVN) or CT‐ Scan (cervical fracture). Having a primary masters degrees versus an undergraduate degree did not improve accuracy. Again, possession of a DPT as the highest degree earned was a predictor for accuracy for the MRI (ACL), but not for the other forms of imaging.
Table 4.
Primary Degree and Highest Degree obtained associated with Accuracy of identification of pathology on Imaging plus Clinical Scenario assessment
| Variable | Odds Ratio (95% CI) | p value |
|---|---|---|
| Correctness of ACL Diagnosis | ||
| Primary Physical Therapy Degree DPT* | 0.51 (0.35, 0.75) | <0.01 |
| Primary Physical Therapy Degree Masters Degree* | 0.78 (0.53, 1.16) | 0.23 |
| Highest Degree Earned | 1.66 (1.20, 2.30) | <0.01 |
| Correctness of AVN Diagnosis | ||
| Primary Physical Therapy Degree DPT* | 0.95 (0.65, 1.4) | 0.80 |
| Primary Physical Therapy Degree Masters Degree* | 0.89 (0.60, 1.3) | 0.57 |
| Highest Degree Earned | 1.12 (0.81, 1.56) | 0.50 |
| Correctness of Cervical Fracture Diagnosis | ||
| Primary Physical Therapy Degree DPT* | 0.88 (0.62, 1.25) | 0.47 |
| Primary Physical Therapy Degree Masters Degree * | 0.91 (0.62, 1.33) | 0.62 |
| Highest Degree Earned | 1.11 (0.82, 1.51) | 0.49 |
BS degree was the reference variable
Table 5 provides information on the utility of APTA and non‐APTA sanctioned certifications. Orthopedic (OCS) and sports (SCS) board certifications within the APTA, along with “other” APTA certifications (including OCS and SCS) was associated with increased accuracy for the MRI (ACL) scenario, CT‐scan (cervical fracture) and the plain‐film radiograph scenario (hip AVN) where non APTA certifications was associated with increased accuracy for only the MRI (ACL) scenario. The highest odds ratios involve those with OCS and SCS certifications.
Table 5.
Certifications or Specializations associated with Accuracy of identification of pathology on Imaging plus Clinical Scenarios (Univariate Analyses)
| Variable | Odds Ratio (95% CI) | p value |
|---|---|---|
| Correctness of ACL Diagnosis | ||
| OCS and SCS* | 4.88(2.39, 9.94) | <0.01 |
| Non OCS and SCS† | 1.39 (0.96, 2.00) | 0.08 |
| APTA Certifications‡ | 2.44 (1.50, 3.97) | <0.01 |
| Non APTA Certifications§ | 1.57 (1.05, 2.37) | 0.03 |
| Correctness of AVN Diagnosis | ||
| OCS and SCS* | 1.87 (1.13, 3.11) | 0.02 |
| Non OCS and SCS† | 1.38 (0.94, 2.02) | 0.10 |
| APTA Certifications‡ | 1.88 (1.22, 2.91) | <0.01 |
| Non APTA Certifications§ | 1.23 (0.80, 1.88) | 0.35 |
| Correctness of Cervical Fracture Diagnosis | ||
| OCS and SCS* | 2.36 (1.45, 3.84) | <0.01 |
| Non OCS and SCS† | 1.22 (0.85, 1.17) | 0.28 |
| APTA Certifications‡ | 1.74 (1.15, 2.62) | <0.01 |
| Non APTA Certifications§ | 1.30 (0.89, 1.91) | 0.18 |
OCS and SCS = Orthopedic Certified Specialist and Sports Certified Specialist
Non OCS and SCS = Other American Board of Physical Therapy Specialties (Cardiac, Electrotherapy, Geriatric, Pediatric, Neurologic, Women's Health), Fellow of the American Academy of Orthopedic Manual Physical Therapy, Mechanical Diagnosis and Therapy, Certification in Orthopedic Therapy, Constraint Induced Movement Therapy, Manual Therapy Certification,
APTA Certifications = Other American Board of Physical Therapy Specialties (Cardiac, Electrotherapy, Geriatric, Orthopedic, Pediatric, Neurologic, Sports, Women's Health)
NON APTA Certifications = Fellow of the American Academy of Orthopedic Manual Physical Therapy, Mechanical Diagnosis and Therapy, Certification in Orthopedic Therapy, Constraint Induced Movement Therapy, Manual Therapy Certification,
Table 6 provides information on practice setting. Working in an outpatient setting significantly increased accuracy only for the MRI (ACL scenario). Treating a population that is at least 75% orthopaedic was significant for increased accuracy for the MRI (ACL) and CT‐scan (cervical fracture) scenarios.
Table 6.
Practice Setting Associated with Accuracy of identification of pathology on Imaging plus Clinical Scenarios (Univariate Analyses).
| Variable | Odds Ratio (95% CI) | p value |
|---|---|---|
| Correctness of ACL Diagnosis | ||
| Outpatient Setting versus other | 1.57 (1.13, 2.17) | <0.01 |
| Population is 75% orthopedic or higher | 1.86 (1.31, 2.62) | <0.01 |
| Correctness of AVN Diagnosis | ||
| Outpatient Setting versus other | 0.98 (0.71, 1.36) | 0.91 |
| Population is 75% orthopedic or higher | 1.08 (0.77, 1.52) | 0.63 |
| Correctness of Cervical Fracture Diagnosis | ||
| Outpatient Setting versus other | 1.16 (0.86, 1.57) | 0.33 |
| Population is 75% orthopedic or higher | 1.37 (1.01, 1.87) | 0.04 |
Discussion
The purpose of the current study was three fold: First, to determine the overall ability of licensed physical therapists to identify pathology using three different imaging modalities with conditions of the hip, cervical spine and knee. Second, to investigate whether providing a concurrent clinical scenario improved the accuracy of the identification. Third, to evaluate factors associated with accuracy when both the imaging and clinical scenario were provided. In particular, the associative value of a) imaging training, b) different levels of educational training, and c) value of certification of specialization within the APTA were analyzed. The current findings suggest that although diagnosis using imaging notably improves with a clinical scenario, it appears that the diagnosis may be dependent on both the clinical condition and the imaging modality. Further, the current findings suggest that formal and informal training, board certification through the APTA and to a lesser extent, degree level, can also influence diagnostic accuracy. There are likely several reasons for these findings.
Correct diagnoses (identification of pathology) on each of the three imaging scenarios with clinical information ranged from a low of 25.6% (hip AVN, plain film radiograph) to a high of 61.3% (ACL tear, MRI). Whether these values are similar to other health providers with image viewing privileges is unknown. The lower accuracy scores for the hip and cervical conditions may be reflective of the challenging natures of both conditions. It is well known that a diagnosis of AVN is an exceptionally difficult diagnosis to make in the early stages of the condition,19and that staged MRI's may be most useful during the diagnosis20 followed by CT Scan.21 In a radiology study avascular necrosis was misdiagnosed 42.27% of the time by radiologists on plain film radiograph; the author found that relevant clinical history helped to reduce the rate of misdiagnosis, which supports the current findings.22 The imaging modality used in this study was plain film radiograph which exhibits low to moderate sensitivity values in detection of hip AVN.23
Cervical spine fractures are often overlooked on plain film radiographs by radiologists and with clinical examination by primary care physicians.24 Absence of initial neurological defects25 or less common mechanism of injury (low velocity versus high velocity MVA) further complicates cervical spine fracture detection.26Plain film sensitivity of cervical spine fracture is 52% compared with 98% for CT scan.27 Although CT scan is an excellent study for detection of cervical spine fracture,28 the fracture was shown to be difficult for physical therapists to identify in this study. This may due to the challenging nature of reading CT scans, a decreased familiarity with CT scans or the fact that only three images were provided in this study.
Prevalence may also have played a role in picking the diagnosis from the image. For example, a lateral view of an MRI of the knee may have prompted the selection of an ACL injury since this is the most commonly performed imaging modality and position used for ACL tear detection.29 Recently physical therapists have had a number of opportunities to view MRIs in clinical practice thus increasing exposure to this diagnostic modality.8,16,30‐32
Imaging‐only accuracy ranged from 5.5% correct responses (using a radiograph and the diagnosis of hip AVN) to 48.2% correct responses (diagnosis of ACL tear using an MRI). In all three scenarios, accuracy improved with the addition of a clinical scenario. Changes in identification ranged from a low of 17.8% improvement for the cervical spine fracture on a CT scan to a high of 365.5% improvement for identifying hip AVN with a radiograph. This finding is consistent with the medical literature on providers reading and interpreting radiographs and CT scans. Loy and Irwig17 reported that clinical information improved imaging reading accuracy, in conditions such as digitized chest, wrist, and general bone assessment. Their review17 found no instances in which clinical information provided deleterious results.
The authors were able to identify a number of training/exposure opportunities that were associated with improved accuracy in imaging detection when the clinical scenario was provided including: training through an entry‐level program, training through a post‐graduate degree, training during a CEU course, and through other informal imaging training. No known studies that have assessed the accuracy of physical therapists for use of imaging. Higher educational training (having a doctorate of physical therapy rather than a master's or bachelor's degree in physical therapy) only seemed to influence the accuracy of the ACL diagnosis. To the authors' knowledge, studies have not investigated the impact the DPT has on clinical decision making compared to other entry‐level degrees.
In addition, board certification from the APTA was associated with increased accuracy. This reveals that orthopaedic and sports clinical certification increases accuracy with diagnosing based on imaging and clinical scenarios compared to all other certifications both APTA and non‐APTA. Board certification is a formal recognition for physical therapists with advanced clinical knowledge, experience, and skills in a special area of practice. To be eligible for board certification, clinicians must submit evidence of 2000 hours of direct patient care and/or exhibit evidence of residency or fellowship training.33 These factors may also expose clinicians to additional images in practice. Childs' et al34 also found that when evaluating the results of a standardized exam that had been previously been validated in the literature,35 physical therapists who were board certified in orthopedic or sports physical therapy achieved significantly higher scores and passing rates than their non‐ board certified colleagues. Jette36 discovered that physical therapists with an OCS were almost twice as likely to make a correct determination on management for patients with critical medical and musculoskeletal conditions. This is consistent with our findings that those holding the specific APTA certifications OCS and SCS performed with increased accuracy. This demonstrates that physical therapists holding APTA certifications were more accurate with diagnosing than those who held any other specialty certifications. Having a certification other than an APTA certification increased accuracy for only the MRI diagnosis. This finding supports that physical therapists have recently had more exposure to MRI's,8,16,30‐32 which corresponds to the increased accuracy with the diagnosis of the MRI case.
Practice setting and patient population did not provide as robust of findings as seen above. The frequency of anterior cruciate ligament tears being seen in outpatient orthopedic setting could have contributed to the accuracy of respondents working in that practice setting. Although all three diagnoses were orthopaedic in nature, the anterior cruciate ligament tear is the most commonly seen.
Limitations
There are limitations to this study. Although the researchers looked at primary physical therapy degree and highest degree earned years of experience or age of clinicians was not examined. The survey was only sent out to Ohio Physical Therapists for convenience and cost reasons; therefore, it is possible that the results of this survey are not representative of physical therapists across the country and world. The clinical vignettes were written by one of the authors. While the information included in the scenarios was taken from the literature, experts outside of the authorship team did not validate these scenarios. Only one imaging modality was chosen for each clinical scenario. Therefore, a conclusion cannot be drawn regarding the appropriateness of one imaging modality over another for each of the clinical scenarios. Finally, there is a test order effect. The order of the cases was not randomized for each participant. On a whole, the accuracy rates declined for each subsequent case. It cannot be determined if this was due to a test order effect or the clinical scenario and images.
Conclusion
Physical therapists ability to identify pathology on images greatly improves with a clinical scenario, although it is apparent that the clinical condition and imaging modality affects the accuracy of that diagnosis. In addition formal and informal training, board certification through the APTA and to a lesser extent, degree level, can also influence diagnostic accuracy, therefore demonstrating that education through a physical therapy curricula or later in forms of certifications and continuing education can prepare physical therapists for the increased utilization of imaging in the clinic for improved patient care.
Acknowledgements
We would to thank the following for the contribution of the diagnostic images: Rob Liottta, Steven Goldstein, Hani Alsalam, and Radiopaedia
Appendix 1
Survey
*1. I am confident in my ability to interpret diagnostic imaging results on a(n)
| Strongly Agree | Agree | Disagree | Strongly Disagree | |
|---|---|---|---|---|
| Plain Film Radiograph | ||||
| MRI | ||||
| CT Scan |
*2. Please type a diagnosis based on the following images found for case 1.
Figure 1:
Figure 2:
Figure 3:
*3. Given the same images and the following clinical information, please type in the correct diagnosis for this patient. The patient is a 16 year old who plays high school tennis. During a match one week ago, she approached the net to block a ball and fell to the ground with immediate left knee pain. The knee immediately swelled and she was unable to bear weight. The athletic trainer did an initial examination and recommended she go to the emergency room. The emergency room physician had an xray taken which came back negative. He recommended rest, elevation and ice and to follow up with an orthopedic doctor or physical therapist that week. The patient was able to walk and felt “almost perfect” two days later. She was ready to get back to tennis; however, her parents made her see a physical therapist. The physical therapist had concerns. Although the patient stated she felt fine, she also admitted to buckling episodes a few times a day. The therapist realized the patient was extremely (maybe too) motivated to return to tennis as conferences were approaching. Due to her concerns, the physical therapist discussed further imaging with the patient's physician and an MRI was ordered.
*4. Please type a diagnosis based on the following images found for case 2.
Figure 4:
Figure 5:
Figure 6:
*5. Given the same images and the following clinical information, please type in the correct diagnosis for this patient.The patient is a 52 year old male with a history of a MVA six months ago. He went to the ER immediately following the accident and was sent home. Six months later he is experiencing pain bilaterally through his neck and shoulders which he describes as a 6/10 on the visual analog scale. He states his pain is daily and feels like a constant throb. He has come to physical therapy for evaluation and treatment of what he describes as whiplash. The physical therapist does not note any myelopathic signs, no decreases in strength, but there are decreased reflexes in his upper extremities and decreased cervical range of motion. The PT is hesitant to continue with this patient based on the history of the MVA and constant pain since. He recommends an appointment with his PCP to rule out any contraindications to physical therapy. The PCP ordered an X ray which showed abnormal alignment so a CT was ordered for further investigation.
*6. Please type a diagnosis based on the following images found for case 3.
Figure 7:
Figure 8:
*7. Given the same images and the following clinical information, please type in the correct diagnosis for this patient.The patient is a 50 year old male who has sickle cell anemia. He has a two yearold granddaughter who he watches for a few hours a day, three days a week. Three months ago he came to your clinic because he started to have pain in his right hip with weight bearing activities, and this was making it difficult to watch his granddaughter. Now, three months later his pain is continuing to increase with activities, but he also notes pain with lying in bed. He use to describe his pain as dull, achy pain and now he says it feels like its shooting down his thigh and into is groin. He has become stiff and is demonstrating decreased ROM in his hip. He is expressing frustration since he is “getting worse” and can no longer watch his granddaughter without being in more pain. The PT recommends he follow up with his PCP since he does not appear to be getting any relief from physical therapy. His PCP sent him for a plain film radiograph.
*8. What is the highest earned degree you hold in any area of study? (Select one)
Baccalaureate degree
Master's degree
PhD (or equivalent, e.g. EdD or ScD)
DPT
tDPT
PhD (or equivalent) and DPT
PhD (or equivalent) and tDPT
Other (please specify)
*9. What was your first (entrylevel) physical therapy degree, prior to taking the licensure exam?
Baccalaureate degree
Post baccalaureate certificate
Masters degree
DPT
Other (please specify)
*10. Have you received education in diagnostic imaging? Please check all that apply.
Requirement for an entry level PT degree
Requirement for a post professional degree
Taken as a continuing education course
Other (please specify)————
*11. Have you earned any of the following certifications? Please check all that apply.
CCS
ECS
GCS
OCS
PCS
NCS
SCS
Other (please specify)————
WCS
FAAOMPT
MDT
COMT
CIMT
MTC
*12. Using a total of 35 or more hours per week (at your primary position) as the definition of ‘fulltime’, which one of the following describes your current employment status?
Fulltime
Parttime
PRN
Retired
Unemployed/not seeking work
Unemployed/seeking fulltime employment
Unemployed/seeking parttime employment
Other (please specify)————
*13. Which of the following best describes the type of facility or institution in which you currently do all or most of your work (your primary position)?
Acute care hospital
Subacute rehab hospital (inpatient)
Health system or hospitalbased outpatient facility or clinic
Private outpatient office or group practice
SNF/ECF/ICF
Patient's home/home care n
School system (preschool/primary/secondary)
Academic Institution (postsecondary)
Health and Wellness Facility
Research Center
Industry
Other (please specify)————
*14. What percentage of your patients have an orthopaedic condition?
025%
2650%
5175%
76100%
Not applicable
*15. Which type of courses do you teach? Please check all that apply.
Orthopaedic
Diagnostic Imaging
Not applicable
Other (please specify)————
*16. How often do you view patient images?
0% of patients
110% of patients
1125% of patients
2650% of patients
5175% of patients
7699% of patients
100% of patients
Not applicable
*17. How often do you view
| Very Frequently | Frequently | Infrequently | Very Infrequently | Not Applicable | |
|---|---|---|---|---|---|
| Plain Film Radiographs | |||||
| Plain Film Radiograph Reports | |||||
| MRI | |||||
| MRI Reports | |||||
| CT Scan Images | |||||
| CT Scan Reports |
*18. I always use imaging information when available.
Strongly agree
Agree
Disagree
Strongly disagree
Not applicable
References
- 1.American Physical Therapy Association. Vision 2020. Available at: http://www.apta.org/Vision2020/ Accessed December 17, 2013
- 2.Threlkeld A Jensen G The Clinical Doctorate: A Framework for Analysis in Physical Therapist Education. Phys Ther [serial online]. June 1999;79(6):567 Available from: Academic Search Complete, Ipswich, MA. Accessed March 16, 2014.3. [PubMed] [Google Scholar]
- 3.Brudvig TJ Colbeck CL The doctor of physical therapy: clinical and academic physical therapists' perception of appropriate curricular changes. l Phys Ther Educ. 2007;21(1):3‐13 [Google Scholar]
- 4.Little T Lazaro R Physiotherapists' perceptions and use of medical imaging information in practice. Physiother res int. 2006;11(1):14‐23 [DOI] [PubMed] [Google Scholar]
- 5.McKinnis LN Fundamentals of Musculoskeletal Imaging. 3rd ed. Philadelphia, PA: F.A. Davis Company; 2010 [Google Scholar]
- 6.Deyle GD Musculoskeletal imaging in physical therapy practice. J Orthop Sports Phys Ther. 2005; 35(11): 708‐720 [DOI] [PubMed] [Google Scholar]
- 7.Threlkeld AJ Paschal KA Entry‐level physical therapist education in the United States of America. Phyl Ther Rev. 2007;12:156‐162 [Google Scholar]
- 8.Greathouse DG Schreck RC Benson CJ The United States Army physical therapy experience: evaluation and treatment of patients with neuromusculoskeletal disorders. J Orthop Sports Phys Ther 1994;19:261–6 [DOI] [PubMed] [Google Scholar]
- 9.Boyles RE Gorman I Pinto D, et al. Physical therapist practice and the role of diagnostic imaging. J Orthop Sports Phys Ther. 2011; 41(11):829‐837. 10.2519/jospt.2011.3556. [DOI] [PubMed] [Google Scholar]
- 10.Garber MB Diagnostic imaging and differential diagnosis in 2 case reports. J Orthop Sports Phys Ther. 2005 Nov;35(11):745‐54 [DOI] [PubMed] [Google Scholar]
- 11.Boissonnault WG Badke MB Powers JM Pursuit and implementation of hospital‐based outpatient direct access to physical therapy services: an administrative case report. Physl Ther. 2010;90(1):100‐109 [DOI] [PubMed] [Google Scholar]
- 12.Daker‐White G Carr AJ Harvey I, et al. A randomized controlled trial. Shifting boundaries of doctors and physiotherapists in orthopaedic outpatient departments. J Epidemiol Community Health. 1999 Oct;53(10):643‐50 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.McMeeken J Physiotherapy education in Australia. Phys Ther Rev. 2007;12:83‐91 [Google Scholar]
- 14.Skinner MA Physiotherapy education in New Zealand. Phys Ther Rev. 2007;12:122‐128 [Google Scholar]
- 15.Murphy BP Greathouse D Matsui I Primary care physical therapy practice models. J Orthop Sports PhysTher. 2005; 35(11):699‐707 [DOI] [PubMed] [Google Scholar]
- 16.Moore JH Goss DL Baxter RE, et al. Clinical diagnostic accuracy and magnetic resonance imaging of patients referred by physical therapists orthopaedic surgeons, and nonorthopaedic providers. J Orthop Sports Phys Ther. 2005; 35(2):67‐71 [DOI] [PubMed] [Google Scholar]
- 17.Loy C, Irwig L Accuracy of Diagnostic Tests Read With and Without Clinical Information: A Systematic Review. JAMA [serial online]. October 6, 2004;292(13):1602‐1609 Available from: Academic Search Complete, Ipswich, MA. Accessed March 16, 2014. [DOI] [PubMed] [Google Scholar]
- 18.Donato EB DuVall RE Godges JJ, et al. Practice Analysis: Defining the clinical practice of primary contact physical therapy. J Orthop Sports Phys Ther. 2004; 34(6): 284‐304 [DOI] [PubMed] [Google Scholar]
- 19.Lucie RS Fuller S Burdick DC, et al. Early prediction of avascular necrosis of the femoral head following femoral neck fractures. Clin Orthop Relat Res. 1981 Nov. ‐Dec;(161):207‐14 [PubMed] [Google Scholar]
- 20.Zibis AH Karantanas AH, Roidis NT, et al. , The role of MR imaging in staging femoral head osteonecrosis. Eur J Radiol. 2007 Jul;63(1):3‐9 Epub 2007 Jun 6. [DOI] [PubMed] [Google Scholar]
- 21.Barille MF Wu JS McMahon CJ Femoral head avascular necrosis: a frequently missed incidental finding on multifactor CT. Clin Radio.l 2014;69(3):280‐5. 10.1016/j.crad.2013.10.012. Epub 2013 Dec 2. [DOI] [PubMed] [Google Scholar]
- 22.Wu XY Analysis of the causes of X‐ray misdiagnosis of avascular femur head necrosis. Zhongguo Gu Shang. 2013 Feb;26(2):162‐4 [PubMed] [Google Scholar]
- 23.Sugano N Kubo T Takaoka K Diagnostic criteria for non‐traumatic osteonecrosis of the femoral head. A multicenter study. JBJS. 1999;81(4):590‐5 [DOI] [PubMed] [Google Scholar]
- 24.Takami M Nohda K Sakanaka J, et al. Usefulness of full spine computed tomography in cases of high‐energy trauma: a prospective study. Eur J Ortop Surg Traumatol. 2013. 10.1007/s00590‐013‐1268‐0. Epub 2013 Jul 6. [DOI] [PubMed] [Google Scholar]
- 25.Rogers WA Fractures and dislocations of the cervical spine: an end result study. J Bone Joint Surgery 1957; 39‐A:341‐76 [PubMed] [Google Scholar]
- 26.Blackmore CC Emerson SS Mann FA, et al. Cervical spine imaging in patients with trauma: determination of fracture risk to optimize use. Radiology 1999;211(3);759‐65 [DOI] [PubMed] [Google Scholar]
- 27.Zabel DD Tinkoff G Wittenborn W, et al. Adequacy and efficacy of lateral cervical spine radiography in alert high‐risk blunt trauma patient. J Trauma 1997;43:952–6 [DOI] [PubMed] [Google Scholar]
- 28.Utz M Khan S O'Connor D, et al. MDCT and MRI evaluation of cervical spine trauma. Insights Imaging. 2014 Feb;5(1):67‐75. 10.1007/s13244‐013‐0304‐2. Epub 2013 Dec 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Vahey TN Meyer SF Shelbourne KD et al. MR imaging of anterior cruciate ligament injuries. Magn Reson Imaging Clin N Am. 1994;2(3):365‐80 [PubMed] [Google Scholar]
- 30.Benson CJ Schreck RC Underwood FB, et al. The role of Army physical therapists as nonphysician health care providers who prescribe certain medications: observations and experiences. Phys Ther 1995;75:380–6 [PubMed] [DOI] [PubMed] [Google Scholar]
- 31.James JJ Abshier JD The primary evaluation of musculoskeletal disorders by the physical therapist. Mil Med 1981;146:496–9 [PubMed] [PubMed] [Google Scholar]
- 32.James JJ Stuart RB Expanded role for the physical therapist. Screening musculoskeletal disorders. Phys Ther 1975;55:121–31 [PubMed] [DOI] [PubMed] [Google Scholar]
- 33.Learman K Showalter C Cook C Does the use of a prescriptive clinical prediction rule increase the likelihood of applying inappropriate treatments? A survey using clinical vignettes. Man Ther. 2012 Dec;17(6):538‐43. 10.1016/j.math.2012.05.011. Epub 2012 Jun 21. [DOI] [PubMed] [Google Scholar]
- 34.Childs JD Whitman JM Sizer PS, et al. A description of physical therapists' knowledge in managing musculoskeletal conditions. BMC Musculoskeletal Disorders. 2005; 6(32). 10.1186/1471‐2474‐6‐32.. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Freedman KB Bernstein J The adequacy of medical school education in musculoskeletal medicine. J Bone Joint Surg. 1998; 80(10).1421‐7 [DOI] [PubMed] [Google Scholar]
- 36.Jette D Ardleigh K Chandler K, et al. Decision‐Making Ability of Physical Therapists: Physical Therapy Intervention or Medical Referral. Phys Ther [serial online]. December 2006;86(12):1619‐1629 Available from: Academic Search Complete, Ipswich, MA. Accessed March 16, 2014. [DOI] [PubMed] [Google Scholar]