J Khandalavala1, T A Van Geem. 1. Department of Obstetrics and Gynecology, Creighton University School of Medicine, Omaha, Nebraska 68131, USA.
Abstract
OBJECTIVE: To determine the accuracy, in a clinical setting, of two commercial pH papers compared to a hand-held digital pH meter. STUDY DESIGN: Vaginal specimens from 30 women, ages 17-40, both asymptomatic and symptomatic, pregnant and nonpregnant, were evaluated for vaginal pH using pHydrion paper, ColorpHast paper and a reference hand-held, battery-operated pH meter. Pearson product-moment correlation analysis and concordance correlation analysis were performed comparing each of the pH papers to the pH meter. RESULTS: Pearson product correlation coefficients suggested a strong correlation between the pH papers compared to the commercial pH meter; however, concordance correlation coefficients were fair (< 97%). If a pH cutoff of 4.5 had been used as one of the diagnostic tools for the evaluation of bacterial vaginosis (i.e., pH > 4.5), the ColorpHast paper would have resulted in a theoretical false negative rate of 21%, and the pHydrion paper would have resulted in a false negative rate of 24%. There were no false positives. Using a single pH readout of 5.0 could have resulted in a correct value, ranging from 3.85-6.15 pH units with pHydrion paper and a range of 4.32-5.68 using the ColorpHast pH paper. The accuracy of ColorpHast paper was better than that of pHydrion paper. A 1 SD range for the mean pH difference for pHydrion paper was 1.054-0.854 and for ColorpHast was 0.619-0.501. The correlation coefficient for the pHydrion paper was .87, and the correlation coefficient for ColorpHast paper was .88. CONCLUSION: Two pH papers had questionable accuracy in a clinical setting as compared to the hand-held, battery-powered pH meter. There was a theoretical 24% false negative rate if a pH cutoff of 4.5 had been used for pHydrion paper and a 21% false negative rate for ColorpHast paper. Although correlation coefficients were 88%, concordance correlations were inadequate for both papers.
OBJECTIVE: To determine the accuracy, in a clinical setting, of two commercial pH papers compared to a hand-held digital pH meter. STUDY DESIGN: Vaginal specimens from 30 women, ages 17-40, both asymptomatic and symptomatic, pregnant and nonpregnant, were evaluated for vaginal pH using pHydrion paper, ColorpHast paper and a reference hand-held, battery-operated pH meter. Pearson product-moment correlation analysis and concordance correlation analysis were performed comparing each of the pH papers to the pH meter. RESULTS: Pearson product correlation coefficients suggested a strong correlation between the pH papers compared to the commercial pH meter; however, concordance correlation coefficients were fair (< 97%). If a pH cutoff of 4.5 had been used as one of the diagnostic tools for the evaluation of bacterial vaginosis (i.e., pH > 4.5), the ColorpHast paper would have resulted in a theoretical false negative rate of 21%, and the pHydrion paper would have resulted in a false negative rate of 24%. There were no false positives. Using a single pH readout of 5.0 could have resulted in a correct value, ranging from 3.85-6.15 pH units with pHydrion paper and a range of 4.32-5.68 using the ColorpHast pH paper. The accuracy of ColorpHast paper was better than that of pHydrion paper. A 1 SD range for the mean pH difference for pHydrion paper was 1.054-0.854 and for ColorpHast was 0.619-0.501. The correlation coefficient for the pHydrion paper was .87, and the correlation coefficient for ColorpHast paper was .88. CONCLUSION: Two pH papers had questionable accuracy in a clinical setting as compared to the hand-held, battery-powered pH meter. There was a theoretical 24% false negative rate if a pH cutoff of 4.5 had been used for pHydrion paper and a 21% false negative rate for ColorpHast paper. Although correlation coefficients were 88%, concordance correlations were inadequate for both papers.