Study design: Maximal and submaximal (feigned) cervical motions in healthy patients were compared.
Objective: To test the efficiency of the coefficient of variation in differentiating maximal (sincere) from submaximal (feigned) cervical motion in healthy patients.
Summary of background data: Although limitation of cervical motion is a recognized impairment, no well-founded approach for verifying the degree to which a patient is maximizing his or her performance is available currently.
Methods: A new ultrasound-based system for three-dimensional motion analysis of the head was used to test 25 healthy patients (22 women and 3 men) ages 26 to 48 years. During the first test, (Test 1), the patients were asked initially to move the head maximally at a self-determined velocity in all the primary directions: flexion, extension, right and left rotation, and right and left lateral flexion. They then were presented with a short vignette describing a fictitious accident and asked, using the same protocol, to perform the same types of cervical motions as if they had experienced an injury. No further instructions were provided. A retest (Test 2) in reverse order (feigned effort first) took place 1 to 16 weeks (mean, 3.3 weeks) later.
Results: The range of motion and average velocity were significantly smaller (P = 0.0001) in the feigned than in the maximal performance in all directions and on both tests. Feigned range of motion and velocity also were significantly reduced in Test 2 (P = 0.006), as compared with those in Test 1 (P = 0.0001). The range of motion coefficient of variation in the feigned performance (CVf) remained stable on the retest, but was significantly larger on both tests (P < 0.001) than the corresponding CV derived from maximal performance (CVm). Furthermore, a case-by-case analysis showed that whereas the American Medical Association guidelines identified only 16% of the feigned cases, the use of CVf and CVm resulted in a corresponding rate of 87%. Tolerance intervals at 95% and 99% indicated total separation between the distributions of CVf and CVm, respectively. The average velocity-based CVf and CVm were not of a comparable differentiating power.
Conclusions: This study indicates that in healthy patients, feigned performance may be differentiated from maximal (sincere) performance effectively and reproducibly using the coefficient of variation.