Reply to Fenna Visser et al.

Abstract

We thank Dr. Visser and colleagues for their interest in our study [1]. Dr. Visser and colleagues correctly observe that the percentage reported as sensitivity is not sensitivity. After statistical analysis was performed, positive predictive value and negative predictive value (not prevalence) figures were placed in the article instead of sensitivity and specificity. This was an oversight while the values were being transferred from the calculation window to the text and unfortunately, while proofreading the text, the disparity was not noticed. We apologise for this error. We revise sensitivity of clinical examination to 28.3% and specificity of clinical examination to 94.5%. The following sentence appears in our discussion: “As for the sensitivity of clinical examination in detecting DDH, it should be kept in mind that many cases of DDH defined by ultrasonography can go undetected by clinical examination.” As understood from this sentence, it is obvious that the calculated sensitivity of physical examination for detection of developmental dysplasia of the hip (DDH) is low. But the same mistake disseminated and the negative predictive value was placed in parenthesis instead of sensitivity. Sensitivity measures the proportion of actual patients who are correctly determined by the test. Tests which have high sensitivity recognise most of the patients as patients. Specificity measures the proportion of healthy people that are correctly determined by the test. Tests which have high specificity recognise most of the healthy people as healthy. A highly sensitive test is ideal for a screening test while a highly specific test is best in a confirmatory role. Any test which has low sensitivity is not useful as a screening test regardless of its specificity because it misdiagnoses most of the patients as healthy [2]. Ultrasonography was found to detect a high number of false positive hips, which means it has a low positive predictive value. Although there are a limited quantity and quality of papers on this topic, let us assume that ultrasonography (USG) has a low positive predictive value as reported previously. Roovers et al. detected that USG screening has 88.5% sensitivity and a 61.5% positive predictive value [3]. That means only 61.5% of hips detected by USG as pathological were actually DDH; others were normal. If we regulate our data by the positive predictive value reported previously, 126 of 208 hips detected as DDH were actually DDH. Let us accept that the other 82 hips were detected as DDH by mistake; thus, if we suggest an optimistic assumption that all of the 82 hips would be determined to be normal by physical examination, the data table would look like this: We previously calculated the sensitivity of clinical examination as 28.3%. Sensitivity of the clinical examination calculated by this table is 46.8%. The last calculated sensitivity rate is not acceptable for a screening test either. Clinical examination which has 46.8% sensitivity recognises 53.2% of the patients as healthy. Clinical examination is an essential part of DDH diagnosis. We suggest that although the clinical examination is not reliable to detect DDH, it has a good value in DDH diagnosis as a confirmatory test due to its high specificity (94.5%).