Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Clinical Comparison of the Proview Eye Pressure Monitor With the Goldmann Applanation Tonometer and the TonoPen—Reply

Clinical Comparison of the Proview Eye Pressure Monitor With the Goldmann Applanation Tonometer... In reply We thank Drs Leung and Lam for their interest in our article.1 The following are answers to their questions. We chose to measure intraocular pressure (IOP) in the fixed order (Goldmann, TonoPen, and Proview) because we believe that Goldmann and TonoPen tonometries by the physician have very minimal if any effect on subsequent IOP measurements due to massaging the globe. On the other hand, use of Proview by the patient may have a very significant effect due to the nature of the technique. For the second question, none of our study patients had previous incisional eye surgery. Third, we used the mean of 2 measurements from each method for every patient in the analysis. To use the median would require more measurements, longer testing time, and more manipulation of the eye, which may lead to inaccuracy of the data. We agree with the fourth point of Drs Leung and Lam that the correlation coefficient alone may not be an appropriate way to compare data sets obtained by 2 different measurements when the issue is agreement. Although we did present plots of IOP by one method vs another method in our article (Figures 5, 6, and 7), suggesting Pearson correlations, we also reported the intraclass correlations, which are typical measures of agreement for continuous variables. Possibly, these should have been emphasized more than the plots. Table 1 presents the intraclass correlations again for each pair of methods. These intraclass correlations suggest very good agreement between Goldmann and Tonopen. However, Proview did not agree with either Goldmann or Tonopen. Table 1. View LargeDownload Intraclass Correlations We are grateful to Drs Leung and Lam for pointing out the article by Bland and Altman,2 in which an alternative assessment of agreement is presented. We have explored this method, called the “difference against mean” plot. To produce it, the mean and the difference of the measurements made by the 2 methods are computed for each patient. Then the average mean difference and the SD of the differences are calculated. The differences are plotted against the means, with reference lines on the graph for the overall mean difference and the values that are ±2 SD of the mean. The authors state that if the average difference ±2 SD is not clinically important, the measurements could be used interchangeably. Analysis of our data using this method yielded compelling, and confirming, results. Table 2 shows the mean difference and SD for each comparison, along with the values that are ±2 SD of the mean. Table 2. View LargeDownload Mean Differences and SDs Between Methods The tables confirm our findings based on the intraclass correlations. The Goldmann may measure an IOP 4.2 mm Hg below or 9.8 mm Hg above Proview. Proview may measure 7.5 mm Hg below or 4.9 mm Hg above Tonopen. However, Goldmann may be 2.1 mm Hg below or 5.1 mm Hg above Tonopen. These results indicate that only the Goldmann and Tonopen may be used interchangeably. The fifth point made by Drs Leung and Lam is that IOPs in glaucoma patients are unlikely to be normally distributed and that nonparametric tests may be more appropriate. We did examine the distributions and found them on visual inspection not to be skewed and not to depart substantially from normal. However, as a precaution, we recomputed the intraclass correlations based on the ranks of the data, rather than the data values themselves. Results shown in Table 3 confirm our previous findings. Table 3. View LargeDownload Intraclass Correlations Based on Ranks Correspondence: Dr Herndon, Glaucoma Service, Duke University Eye Center, PO Box 3802, Erwin Road, Durham, NC 27710-3802 (hernd012@mc.duke.edu). References 1. Li JHerndon LWAsrani SGStinnett SAllingham RR Clinical comparison of the Proview eye pressure monitor with the Goldmann applanation tonometer and the Tonopen. Arch Ophthalmol 2004;1221117- 1121PubMedGoogle ScholarCrossref 2. Bland JMAltman DG Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1307- 310PubMedGoogle ScholarCrossref http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Archives of Ophthalmology American Medical Association

Clinical Comparison of the Proview Eye Pressure Monitor With the Goldmann Applanation Tonometer and the TonoPen—Reply

Download PDF
1 page

Clinical Comparison of the Proview Eye Pressure Monitor With the Goldmann Applanation Tonometer and the TonoPen—Reply

Abstract

In reply We thank Drs Leung and Lam for their interest in our article.1 The following are answers to their questions. We chose to measure intraocular pressure (IOP) in the fixed order (Goldmann, TonoPen, and Proview) because we believe that Goldmann and TonoPen tonometries by the physician have very minimal if any effect on subsequent IOP measurements due to massaging the globe. On the other hand, use of Proview by the patient may have a very significant effect due to the nature of the...
Loading next page...
 
/lp/american-medical-association/clinical-comparison-of-the-proview-eye-pressure-monitor-with-the-RZZZDOhOgP

References (2)

Publisher
American Medical Association
Copyright
Copyright © 2005 American Medical Association. All Rights Reserved.
ISSN
0003-9950
eISSN
1538-3687
DOI
10.1001/archopht.123.4.579
Publisher site
See Article on Publisher Site

Abstract

In reply We thank Drs Leung and Lam for their interest in our article.1 The following are answers to their questions. We chose to measure intraocular pressure (IOP) in the fixed order (Goldmann, TonoPen, and Proview) because we believe that Goldmann and TonoPen tonometries by the physician have very minimal if any effect on subsequent IOP measurements due to massaging the globe. On the other hand, use of Proview by the patient may have a very significant effect due to the nature of the technique. For the second question, none of our study patients had previous incisional eye surgery. Third, we used the mean of 2 measurements from each method for every patient in the analysis. To use the median would require more measurements, longer testing time, and more manipulation of the eye, which may lead to inaccuracy of the data. We agree with the fourth point of Drs Leung and Lam that the correlation coefficient alone may not be an appropriate way to compare data sets obtained by 2 different measurements when the issue is agreement. Although we did present plots of IOP by one method vs another method in our article (Figures 5, 6, and 7), suggesting Pearson correlations, we also reported the intraclass correlations, which are typical measures of agreement for continuous variables. Possibly, these should have been emphasized more than the plots. Table 1 presents the intraclass correlations again for each pair of methods. These intraclass correlations suggest very good agreement between Goldmann and Tonopen. However, Proview did not agree with either Goldmann or Tonopen. Table 1. View LargeDownload Intraclass Correlations We are grateful to Drs Leung and Lam for pointing out the article by Bland and Altman,2 in which an alternative assessment of agreement is presented. We have explored this method, called the “difference against mean” plot. To produce it, the mean and the difference of the measurements made by the 2 methods are computed for each patient. Then the average mean difference and the SD of the differences are calculated. The differences are plotted against the means, with reference lines on the graph for the overall mean difference and the values that are ±2 SD of the mean. The authors state that if the average difference ±2 SD is not clinically important, the measurements could be used interchangeably. Analysis of our data using this method yielded compelling, and confirming, results. Table 2 shows the mean difference and SD for each comparison, along with the values that are ±2 SD of the mean. Table 2. View LargeDownload Mean Differences and SDs Between Methods The tables confirm our findings based on the intraclass correlations. The Goldmann may measure an IOP 4.2 mm Hg below or 9.8 mm Hg above Proview. Proview may measure 7.5 mm Hg below or 4.9 mm Hg above Tonopen. However, Goldmann may be 2.1 mm Hg below or 5.1 mm Hg above Tonopen. These results indicate that only the Goldmann and Tonopen may be used interchangeably. The fifth point made by Drs Leung and Lam is that IOPs in glaucoma patients are unlikely to be normally distributed and that nonparametric tests may be more appropriate. We did examine the distributions and found them on visual inspection not to be skewed and not to depart substantially from normal. However, as a precaution, we recomputed the intraclass correlations based on the ranks of the data, rather than the data values themselves. Results shown in Table 3 confirm our previous findings. Table 3. View LargeDownload Intraclass Correlations Based on Ranks Correspondence: Dr Herndon, Glaucoma Service, Duke University Eye Center, PO Box 3802, Erwin Road, Durham, NC 27710-3802 (hernd012@mc.duke.edu). References 1. Li JHerndon LWAsrani SGStinnett SAllingham RR Clinical comparison of the Proview eye pressure monitor with the Goldmann applanation tonometer and the Tonopen. Arch Ophthalmol 2004;1221117- 1121PubMedGoogle ScholarCrossref 2. Bland JMAltman DG Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1307- 310PubMedGoogle ScholarCrossref

Journal

Archives of Ophthalmology – American Medical Association

Published: Apr 1, 2005

Keywords: eye,tonopen tonometer

There are no references for this article.