In Reply

    loading  Checking for direct PDF access through Ovid

Excerpt

Thank you for giving us the opportunity to respond to Mr Taylor’s letter and address his concerns. We are disappointed to hear that Reichert Technologies personnel who received our mail and phone inquiries did not forward him our questions regarding the TonoPen. Nonetheless, we are happy to learn that Mr Taylor would have been willing to “share certain information” with us. We hope it is not too late to accept his kind offer, and refer him to our contact details provided in the manuscript. We would further suggest that he share this information with all readers of this journal, as we are certain that it would be of interest to the reading public. Sadly, at present, the information that Reichert Technologies provides regarding TonoPen technology is very limited. We read the instruments’ instruction manuals as well as the information provided on the company’s website, and to the best of our knowledge, nowhere is it disclosed that “certain measurements (outliers, etc.) are disregarded as part of the Tono-Pen averaging process.” This statement by Mr Taylor is a revelation, and indeed somewhat worrying. Both researchers and clinicians are trained to consider and evaluate all raw data, and to use their training and judgment in deciding whether to reject some of it. We are concerned to learn that such a commonly used instrument “disregards” readings using a secret algorithm, in effect taking over the decision-making process of the user. We call on the company to allow this algorithm to be evaluated by outside experts.
Until the company is more forthcoming about the instruments’ algorithm, we are forced to rely on the information it makes available to the general public. According to the TonoPen XL instructions manual “each single valid IOP reading is digitally displayed on the LCD. When four valid readings are obtained, the mean IOP and the standard deviation for those readings are shown on the LCD.” This text is at odds with Mr Taylor’s statement that “one should not expect that the individual readings which appear on the screen … are all used in the final displayed IOP.” Our interpretation of the instructions manual text is that, contradictory to Mr Taylor’s letter, “each valid reading is displayed” and that after displaying four such readings, “the mean IOP and standard deviation for those readings are shown.” This brief description of the instrument’s mode of action formed the basis for our investigation.1
Mr Taylor is concerned that our results and conclusions also indirectly condemn the AVIA model (which we did not evaluate), even though it “uses significantly different methodology in its calculation of IOP values.” This revelation about the AVIA technology (which, to the best of our knowledge, is not made public on the Reichert Technologies website) begs an obvious question—if TonoPen’s reliability has been established “through dozens of publications over the past 30+ years,” why did the company feel compeled to use significantly different methodology in its new model?
We are well aware of the publications cited by Mr Taylor regarding TonoPen’s accuracy, but draw his attention to one of his own references. According to Frenkel et al,2 the TonoPen “lacks good correspondence [with the Goldmann applanation tonometer] in the 31 to 45 mm Hg interval,” which are the highest values evaluated by these researchers. Similarly, Iester et al3 found significant differences between TonoPen and Goldmann values for individuals with intraocular pressure (IOP)>20 mm Hg, and, when compared with manometry, TonoPen underestimated IOP in the higher pressures of rabbits, dogs, cats, and horses.
    loading  Loading Related Articles