Research Review By Dr. Dana Lawrence©

Audio:

Download MP3

Date Posted:

February 2011

Study Title:

Bias in the physical examination of patients with lumbar radiculopathy

Authors:

Suri P, Hunter DJ, Katz JN & Rainville J

Author's Affiliations:

Department of Physical Medicine and Rehabilitation, Harvard Medical School, Boston MA

Publication Information:

BMC Musculoskeletal Disorders 2010;11:275 DOI:10:1186/1471-2474-11-275.

Background Information:

In clinical practice, patients often obtain diagnostic imaging prior to their initial consult if they have consulted their family physician. In many cases, manual therapists may review these imaging results prior to, or during the patient history. This may not represent a best practice strategy however, based on the growing body of literature that calls into question the clinical correlation of many common imaging findings with patient symptomatology. Many findings, particularly on lumbar MRI, are incidental and should not necessarily guide treatment (not to mention history and examination) approaches. To date, no studies have examined systematic bias in the musculoskeletal physical examination. The objective of this study was to assess the effects of bias due to prior knowledge of lumbar spine magnetic resonance imaging findings (MRI) on perceived diagnostic accuracy of the physical examination for lumbar radiculopathy.

Pertinent Results:

  • Average age, leg pain, back pain, and comorbidity were comparable between the independent and non-independent groups.
  • The main conclusion by the authors was that the perceived sensitivity of pinprick sensory testing was higher when physicians had prior knowledge of the MRI results (20% v. 36%, p=0.005). Sensitivities and specificities for exam components otherwise showed no statistically significant differences.
  • At baseline, the groups were essentially equivalent in all important factors; however, there were fewer females and a shorter duration of symptoms in the independent group.
  • There were no differences between groups in exam components of provocative testing, motor testing, and reflex testing.

Clinical Application & Conclusions:

Diagnostic testing is critically important in health care and for the chiropractic profession. Physicians wish to minimize bias and variation in the tests they use, because these affect the internal and external validity of the results. MRI is commonly used in spinal care and clinical decision making. Often, patients come with MRI already available from an earlier medical examination and these are reviewed prior to planning treatment, or even prior to speaking with or examining the patient. This may influence the decision making process, though given how often incidental findings may occur on such imaging this may be potentially problematic. This study looked at whether having such knowledge impacts or affects diagnostic accuracy.

The study did not present evidence of such bias save for when there was abnormal results on sensory testing. As the authors note, the physical examination is the most frequently used diagnostic testing strategy and has the benefit of being low cost and low risk. Bias can affect how we interpret test results, and this study is one of the first to look at this question for low back diagnosis.

Clinical practitioners rarely consider how they may be influenced by previous diagnosis or by information the patient brings from another healthcare practitioner. The risk here is that a diagnostic impression may be skewed by that information, without the physician being aware of that happening. This paper demonstrates that, in the case of patients with radiating leg pain, having information from an earlier MRI seems not to have that impact save for the specific case of pinprick sensory testing. In that sense, it is a positive finding, indicating that the heuristics physicians use may not be significantly impacted by earlier diagnostic information.

Study Methods:

This study was designed as a cross-sectional comparison of performance characteristics of the physical examination when examiners were blinded to MRI results (the “independent group”) with performance in the situation where the physical examination was not blinded to the MRI results (the “non-independent group”). The reference standard was the final diagnostic impression of nerve root impingement by the examining physician. Patients were recruited from a hospital outpatient clinic, were at least 18 years of age, and were seeking care for lower extremity radiating pain of 12 weeks duration or less. 154 subjects were enrolled, and all had lumbar disc herniation confirmed by lumbar MRI. Sensitivity and specificity were calculated in both groups for the 4 components of the radiculopathy examination, including provocative testing, motor strength testing, pinprick sensory testing and deep tendon reflex testing.

Study Strengths / Weaknesses:

The authors note several limitations to this work. First, their use of a composite reference standard of final clinician diagnosis may be seen as less than ideal. However, they argue that it does reflect the process of diagnosis as used by physicians in actual clinical practice.

Second, they note the possibility of incorporation bias, where the result of the index test is used to establish the final diagnosis, as well as test-retest bias (where there may be lack of blinding in the reference standard). However, these biases should normally affect both groups equally.

Third, aspects of the design may have oversimplified situations which are clinically much more complex, and this might overestimate accuracy or affect variability.