We read with great interest the article by Chemla et al (October 2004)1on predicting mean pulmonary artery pressure (mPAP) from the systolic pulmonary artery pressure (SPAP) and have a few comments. We believe that correlation coefficients are not the best way to determine whether two measurement methods are clinically interchangeable. While the correlation coefficient for the measured SPAP and the calculated mPAP using the equation proposed by Chemla et al,1 and the data from all 31 patients is 1.0, these two hemodynamic values are clearly not equivalent. The analysis proposed by Bland and Altman2 would be more appropriate to make the determination of the accuracy of the proposed equation. Applying Bland-Altman analysis to the data, we find that the bias (ie, the mean difference between the calculated mPAP and the measured mPAP) is − 0.63 mm Hg. However, the 95% limits of agreement (ie, the spread of the differences) are − 6.18 and 4.92 mm Hg. When using the standard formula to calculate the mPAP (ie, 1/3 × pulse pressure + diastolic pressure), the bias is − 1.53 mm Hg with 95% limits of agreement of − 4.34 and 1.3 mm Hg. Therefore, the standard formula is easier and arguably more accurate. The advantage of the formula of Chemla et al,1 is that it only requires a measurement of the SPAP to calculate the mPAP, which can be estimated by Doppler echocardiography. However, given the significant inaccuracy of the Doppler echocardiography-derived estimates of SPAP,3 we caution against the use of this technique to diagnose pulmonary hypertension.