Journal of Bone & Joint Surgery - American Volume:
Letters to the editor
Corresponding authors: Yuan-Ya Liao, Department of Surgery, Chung Shan Medical University Hospital, No. 110, Sec. 1, Jianguo N. Road, Taichung 40201, Taiwan. Yu-Min Lin, Department of Orthopaedics, Taichung Veterans General Hospital, No. 160, Sec. 3, Chung-Kang Road, Taichung 40705, Taiwan, e-mail: email@example.com
To The Editor:
We read with great interest the well-written article entitled “Perioperative Testing for Joint Infection in Patients Undergoing Revision Total Hip Arthroplasty” (2008;90:1869-75), by Schinsky et al. We were confused while reading the results of Tables V and VI. According to the laws of probability and set theory, the union is greater than or equal to the intersection. We noticed that the authors reported that the numbers of patients with both an elevated erythrocyte sedimentation rate and an elevated C-reactive protein level (n = 79) were greater than the numbers of patients with either an elevated erythrocyte sedimentation rate or an elevated C-reactive protein level (n = 60). In addition, the authors did not provide the sensitivities and specificities of synovial fluid white blood-cell counts of >3000 and >9000 white blood cells/mL, which makes utilization, recalculation, and confirmation of their diagnostic test results difficult. Furthermore, in Tables V and VI, we observe strange results contrary to the general principles while utilizing multiple tests. Given two parallel tests, A and B, assuming the sensitivity and specificity of one test are independent of the results of the other test, there are two methods in which the tests can be combined in parallel1. The first is the OR rule, in which the diagnosis is positive if either A or B is positive. Both A and B must be negative for the diagnosis to be negative. The second is the AND rule, in which the diagnosis is positive if both A and B are positive. Either A or B can be negative for the diagnosis to be negative. With the OR rule, the sensitivity of the combined result is higher than either test individually, but the specificity is lower than either test individually. With the AND rule, the specificity is higher than either test individually, but the sensitivity is lower than either test individually. However, the authors reported otherwise in Tables V and VI. Between Tables IV and VI, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy are different for all patients (n = 201) concerning a white blood-cell differential of >80% polymorphonuclear cells. We hope that the authors can clarify and correct the statistical errors.
Disclosure: The authors did not receive any outside funding or grants in support of their research for or preparation of this work. Neither they nor a member of their immediate families received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity.
1. Zhou XH Obuchowski NA McClish DK. Statistical methods in diagnostic medicine. New York: Wiley-Interscience; 2002. p 15–56.