|Home | About | Journals | Submit | Contact Us | Français|
Previous studies have shown military physical therapists (PT) to have comparable clinical diagnostic accuracy (CDA) and interobserver agreement to orthopaedic surgeons (OS). However, no studies have examined hip pathology or used intraoperative findings as the reference standard for diagnosis.
To compare the CDA of physical examination findings among a PT, an OS, and two surgical orthopaedic residents (ORs) for hip labral tears.
Thirty-six patients (15 males, 21 females) aged 18–47 (mean + SD, 31.4 + 8.1 years) with 37 symptomatic hips were enrolled in a prospective study and underwent a standardized clinical examination followed by hip arthroscopy. A PT, an OS, and two ORs independently performed history and examinations with the emphasis of diagnosis on the results of six special tests.
Thirty-two of 37 individuals (86%) had labral tears to the hip at arthroscopy. Analysis of agreement between clinical diagnosis and intra-operative findings of a labral tear produced a CDA of 85.3% (29/34 correct) for the PT, 84.4% (27/32 correct) for the OS, and 80.0% (24/30 correct) for ORs. No significant difference in CDA occurred in comparing the PT, OS, and ORs.
Using arthroscopy as the reference standard, hip labral tears were clinically suspected with 80–85% accuracy. The clinical diagnostic accuracy of the PT, OS, and ORs was high with no significant difference between examiners. In this study, an experienced PT, an OS, and two ORs demonstrated similarly high diagnostic skills.
United States Army physical therapists (PTs) have been practicing in orthopaedic management roles since the Vietnam War1–3 and their primary role is to provide evaluation and treatment to alleviate or prevent physical impairments stemming from injury, pre-existing problems, or disease.2,4,5 In their role as physician extenders, Army PTs can also gain privileges to evaluate patients without physician referral; order radiographs, bone scans, magnetic resonance imaging (MRI), and computed tomography scans; order certain lab tests; refer patients to medical specialty clinics; perform electromyographic and nerve conduction studies; restrict service members to their living quarters for up to 72 hours; restrict work and training for up to 30 days; and prescribe certain medications.2,6–8 It is well documented that Army PTs have performed successfully as physician extenders in the evaluation and treatment of patients with neuromusculoskeletal dysfunction.1,2,4,5,9–12 Further, evidence points towards minimal risk for negligent care when patients are evaluated and managed by PTs, through direct access or by referral.13
Diagnostic accuracy is fundamental to direct access providers. Previous studies in the military health care system have shown that PTs have comparable clinical diagnostic accuracy and interobserver agreement to orthopaedic surgeons (OSs) comparing MRI findings for multiple conditions or radiographs following patients with acute ankle sprains.11,12 However, no research currently exists which prospectively compares the accuracy of diagnosis between PTs and OSs from a battery of clinical examination tests and compares it to the gold standard for orthopaedic diagnosis-intraoperative findings.
One condition for which limited evidence exists in the accuracy of the clinical examination is hip labral pathology.
The acetabular labrum functions to both enhance joint stability and decrease contact stresses between the acetabular and femoral cartilage.14,15 A patient with a labrum tear can be symptomatic and may require open or arthroscopic debridement, or possibly repair.14,16–21
Table 1 describes the clinical signs and symptoms of acetabular labral tears. An examiner would not rely on the finding of just one clinical item in isolation, but should use information gathered from the history, site of pain, mechanical symptoms, and physical examination to determine a diagnosis. Mechanisms of injury and risk factors noted in the literature include hip hyperabduction, twisting, falling, motor vehicle accidents, sports (especially those that require hip external rotation or hyperextension such as soccer, karate or ballet), a direct blow, or hip dislocation.22 The anterior inguinal area is the most common site of pain in patients with labral pathology; this sign is highly sensitive and pain is typically rated as moderate to severe.22 For mechanical symptoms, some patients with labral pathology report clicking, catching, or locking of the hip with motion, though the significance of these signs is questionable.22 Currently, research has not demonstrated sufficient specificity of individual or clusters of clinical tests to confidently rule in a diagnosis of hip labral lesion; but high sensitivity of many tests allows a negative finding to increase confidence that a hip labral lesion is absent.22 The decision, therefore, to perform hip arthroscopy on a patient suspected of having labral pathology is typically based on a number of factors to include patient history, conservative treatment results, clinical examination, magnetic resonance arthrography (MRA), and response to intraarticular injection of anesthetic. While the value of MRA and intraarticular injections in diagnosis has been shown, the accuracy of clinical examination tests in detecting labral tears is less well defined.15,16,18,19,21–26
Physical therapists in orthopaedic and sports medicine practices manage patients with suspected hip labral tears and are trained to perform physical examinations using clinical tests shared by orthopaedics and physical therapy practice. It is important for a PT to determine how his or her clinical diagnostic accuracy (CDA) compares with an OS and ORs working in the same facility in order to diagnose, treat, or refer patients most appropriately during the conservative treatment phase. Therefore, the purpose of this study was to prospectively assess the CDA of physical examination findings for hip labral pathology among a PT, an OS, and ORs using arthroscopy as the definitive diagnosis. The hypothesis to be tested was that all providers would have similar CDAs.
Thirty-six consecutive military health care beneficiaries presenting to the orthopaedic sports medicine clinic at a tertiary military medical center with hip pain were recruited by ORs. All subjects provided informed consent to their participation and the rights of the subjects were protected as governed by the Clinical Investigation and Human Use Committees of the Department of Clinical Investigation at Walter Reed Army Medical Center. Subjects included active-duty military members or Department of Defense beneficiaries who were between 18–47 years of age and who were seeking treatment for hip pain refractory to conservative treatment. Subjects who were pregnant or with previous hip surgery were excluded. Subjects with a primary diagnosis of hip osteoarthritis, congenital hip pathology (i.e. dysplasia), avascular necrosis, or femoral neck stress fracture were also excluded.
Before initiation of the study, a PT with 19 years of experience, two ORs (one with 4 years, one with 5 years surgery experience) and an OS (with 7 years experience as a fellowship-trained sports surgeon) who performed all hip arthroscopies participated in a 30 minute practice session to standardize the following clinical examination techniques: Thomas hip flexion-to-extension maneuver (aka McCarthy Sign), internal rotation load/grind, Fitzgerald Test, eccentric hip flexion, resisted straight leg raise (SLR), and resisted SLR in external rotation.
Patients were examined independently by the PT, one of two ORs, and the OS in varied order based on provider availability. Physical examinations were performed first and the results recorded prior to gathering clinical histories and radiographic findings. Each examiner was blinded to the results of the other providers. For the purposes of this study, the test was considered positive if the patient had one or more of these symptoms during the test: click, clunk, or pain in the groin region which reproduced their chief complaint. The final diagnosis was not algorithmically derived, instead the diagnosis was driven by clinical reasoning based on meaningful interpretation of all the factors (pain, location, mechanical symptoms) integrated across all six tests. A description of physical examination tests follows.
In supine, the subject fully flexed both hips (Figure 1), then the examiner slowly and passively extended the subject's lower extremities with hips going into external rotation (ER) (Figure 2A). This test was repeated, but with the subject's hip going into internal rotation (IR)(Figure 2B). Sensitivity and specificity of this test has yet to be published.22,27
In supine, the examiner flexed the subject's hip passively to approximately 100 degrees and then rotated the subject's hip from IR to ER while pushing along the long axis of the femur through the knee to cause “grind” (axial compression of the femoral head in the acetabulum through knee) (Figure 3). This movement mimics, and is very similar to, the flexion-internal rotation-axial compression test, which has a reported specificity of 0.43 and a sensitivity of 0.75.28
To test the anterior labrum, the examiner started with the subject's hip in maximum flexion, ER, and full abduction (Figure 4A); then extended the subject's hip while placing it into full IR, and adduction (Figure 4B). To test the subject's posterior labrum, the examiner started with the subject's hip in maximum flexion, IR, and adduction (Figure 5A); then extended the subject's hip while placing it into full ER and abduction (Figure 5B). Sensitivity is reported to be 1.00.18 For inter-rater reliability of the flexion-internal rotation-adduction-impingement test, which is described like the Fitzgerald test for anterior labral tears, Kappa was 0.58 with a 95% confidence interval of (0.29–0.87).29
While supine, the subject lifted the lower extremity into full hip flexion with knee extended, then slowly lowered the leg to the table, reporting any clicks, clunks, or pain. This test was used to identify possible iliopsoas tendon snapping.
While supine, the subject actively raised the lower extremity to 30 degrees of hip flexion with the knee fully extended. The subject held the lower extremity while the examiner applied resistance to the ankle. The resisted SLR is thought to load the joint antero-superiorly and to cause anterior groin pain if an intra-articular lesion is present.30 Sensitivity and specificity of this test has yet to be published.22,27
The test was repeated, but subject's hip was in ER (Figure 6). This test is thought to “wind up” the iliopsoas and place more tension at the labrum. Sensitivity and specificity of this test has yet to be published.22,27
Descriptive statistics, frequency tables, and Cochran's Q test31–32 (repeated measures test for dichotomous data with three or more independent variables- examiners in this case) were used to determine whether a PT could demonstrate a comparable degree of CDA with an OS and one of two ORs when conducting hip examination tests targeting labral pathology (alpha level set at p<.05). Clinical diagnostic accuracy has been used in a previous study11 and is a ratio represented by the number of correct diagnoses as the numerator and the total number of diagnoses made as the denominator (number correct diagnoses/total number diagnoses).33 The number of false negatives (condition in which the examiner diagnosed a subject without a tear but a tear existed) and false positives (condition in which the examiner diagnosed a subject with a tear but no tear existed) were also determined for each examiner. All statistics were performed with SPSS 9.0 for Windows (SPSS Inc., Chicago, IL).
Over an 18 month period, 36 patients (15 male and 21 females) aged 18–47 (mean 31.4 + SD 8.1 years) with 37 symptomatic hips were enrolled in this study. All 36 enrolled subjects completed the study. The PT, OS and ORs independently examined the patients, documented results of their clinical tests, and then made a diagnosis based on their findings. Table 2 shows the frequency of positive clinical findings among the examiners. Only one test (Fitzgerald test going into external rotation and abduction) showed a large variability between providers.
In this cohort, 34 patients had complete examination and diagnosis data by the PT, 32 patients by the OS, and 30 patients by the ORs. The ORs examined approximately 15 patients each. All 36 patients underwent hip arthroscopy, one bilaterally. Results from the hip arthroscopy provided the definitive diagnosis. Thirty-two of 37 subjects (86%) had acetabular labral tears at the hip confirmed at the hip during hip arthroscopy. Some subjects had multiple tears at the hip and the locations noted were: 11 tears located anterior, 21 anterior-superior, six superior, and one tear posterior-superior.
Analysis of agreement between clinical diagnosis and intra-operative findings of a labral tear produced a CDA of 85.3% (29/34) for the PT (five false positives), 84.4% (27/32) for the OS (five false positives), and 80.0% (24/30) for ORs (four false positives, two false negatives). No significant difference in CDA existed among all three examiners (Q= 2.00, p = .999). A remarkable observation was that the PT and OS each had five false positives which were on the same five patients.
The findings from this study support the hypothesis regarding CDA by demonstrating that the PT, OS, and ORs practicing at a tertiary military medical center during the period of this study demonstrated a high degree of CDA on hip labral pathology diagnosis, confirmed with arthroscopy, for patients referred with hip pain. Though responses of individual tests during the clinical exam showed some variability between providers (i.e. Fitzgerald test), the overall interpretation of all clinical exam tests combined yielded very similar diagnoses. As a result, no significant difference occured in diagnostic accuracy between the PT, OS and ORs.
The results of this study are comparable to previously published literature. Other studies which have addressed PT clinical accuracy have found similar interobserver or CDA agreements.11,12 Moore et al11 conducted a retrospective review of 560 patients with musculoskeletal injuries referred for magnetic resonance imaging and compared CDA of PTs to OSs and non-orthopaedic providers. The authors reported no significant difference in CDA between PTs (74.5%; 108/145) and OSs (80.8%; 139/172) across a variety of orthopaedic conditions, though hip labral tears were not assessed. Physical therapists and OSs both showed significantly higher CDAs than nonorthopaedic providers (35.4%; 86/243). Therefore, the current study reinforces the diagnostic accuracy of PTs. Further, the present study design improves on the validity of these findings by overcoming two limitations Moore et al11 describe: 1) the need for a prospective analysis of CDA, and 2) a reference standard based on surgical confirmation of the diagnosis.
The results are in agreement with previous literature on the diagnostic accuracy of Army PTs and their orthopaedic colleagues. The high agreement in exam findings and diagnosis translates into better consistency in management of orthopaedic-related conditions and ultimately benefits the patient and the medical system. The strength of the accuracy may be explained by several factors, some intrinsic to the Army physical therapy education, training, and credentialing model, and others related to the nature of the Army medical system.
The Army has always emphasized training physical therapists in strong orthopaedic evaluation skills to serve as physician extenders. Almost all Army PTs attend a postgraduate training course in which specialized skills including orthopaedic examination, advanced diagnostic imaging, and pharmacological management are emphasized. Additionally, as part of the process for PTs to become credentialed as neuromusculoskeletal evaluators, PTs will typically shadow physician colleagues in orthopaedics for up to a week at a time, and shadow radiologists and primary care physicians, as well.
The Army medical system further strengthens the evaluation skill sets of PTs by the close relationship in most facilities where PTs and OSs routinely see patients together in combined clinics to manage nonsurgical or perioperative conditions. Combined training is critical and greatly emphasized in a deployed theater, where the PT's role in managing the large volume of nonsurgical orthopaedic conditions frees the OS to concentrate on individuals with complicated trauma and surgical cases.
The subjects enrolled in this study were all selected from a tertiary-level orthopaedic sports medicine clinic. Since this clinic is completely referral based, all patients presenting for evaluation would have been previously evaluated by a PT or physician. Further, the OS has developed a specialization in hip arthroscopy for the treatment of patients with labral pathology, which is known to the referring providers. As a result, this sample is biased towards hip labral pathology, as other etiologies for hip pain, which may indeed be more common, were typically excluded prior to final referral to the clinic. Therefore, it is not known how many people might have similar hip symptoms or complaints who were never referred to the clinic. A limitation known as “spectrum bias” could have occurred which could have improved the overall diagnostic accuracy by eliminating patients with conditions in which the physical examination tests assessed in this study are less discriminate. Results of this study may be different if the PT, OS, and ORs had evaluated the subjects in a general practice setting prior to any other evaluations and interventions.
Additionally, the use of arthroscopy as the gold standard reference in this study significantly improves validity, but at the consequence of furthering spectrum bias. Spectrum bias can cause an overestimation when diagnostic accuracy is studied in samples in which the vast majority of subjects have the disease in question. These studies tend to overstate the accuracy when applied to the general population. Regardless, this population bias effect would be expected to equally impact each of the examiners. Thus, spectrum bias may have artificially elevated the CDA; however, it should not have impacted the finding of equivalent CDA for PT, OS and ORs conducting physical exams of the subject's hip to detect labral tears.
Lastly, the use of a single PT, OS, and two ORs limits the generalizability of the findings. This limitation could have been overcome by having more than one of each type of provider perform examinations. While the external validity of our results may have improved, such a study was not practical in the present clinic setting.
Army PTs frequently perform initial evaluations for a myriad of orthopaedic and sports injuries while serving in a physician extender role. Recent studies have shown the effectiveness of using Army PTs as primary neuromusculoskeletal screeners during peace and war, including during deployments to Operations Desert Shield and Storm, Bosnia, and Operation Iraqi Freedom.9,10 This study provides further evidence that military PTs demonstrate competency in making sound, independent clinical judgments regarding the evaluation and management of patients with hip labral pathology.
These findings warrant further studies to evaluate CDA between PTs and other health care providers in a variety of settings for the patients with the most common musculoskeletal conditions across the full spectrum of a disease or injury process. In addition, prospective studies involving PTs with varying levels of clinical experience, board certification, and fellowship training will provide important data to further conclusions regarding the abilities of PTs to manage patients in a direct access environment.
Using arthroscopy as the reference standard, hip labral tears in the subjects were clinically suspected with 80–85% accuracy among the examining clinicians. Clinical diagnostic accuracy of an experienced physical therapist, orthopaedic surgeon, and orthopaedic residents on patients with hip labral pathology was excellent with no significant difference among examiners. This study further strengthens the evidence that the use of Army PTs in the role of managing, evaluating, and treating patients with neuromusculoskeletal dysfunction is a successful model.
The views expressed in this manuscript are those of the authors and do not reflect the official policy of the Department of Army, Department of Defense, or U.S. Government. All authors are employees of the United States government. This work was prepared as part of their official duties and as such, there is no copyright to be transferred.