Journal Information
Vol. 13. Issue 4.
Pages 276 (October - December 2020)
Share
Share
Download PDF
More article options
Vol. 13. Issue 4.
Pages 276 (October - December 2020)
Letter to the Editor
DOI: 10.1016/j.optom.2020.03.007
Open Access
Agreement analysis
Análisis de concordancia
Visits
...
Christian Faua,1,
Corresponding author
chfauf@gmail.com
cfau@fundacion2020.org

Corresponding author at: Fundación Oftalmológica 2020, Avd. Presidente Riesgo 5157, Dep. 212, Las Condes, Santiago 7560854, Metropolitana, Chile.
, Solange Nabzoa, Veronica Nasabunb
a Fundación Oftalmológica 2020, Iberoamerican Cochrane Network, Santiago, Chile
b Andrés Bello University, School of Nursing, Santiago, Chile
Related content
J Optom. 2020;13:27710.1016/j.optom.2020.05.002
Hassan Hashemi, Samira Heydarian, Abbas Ali Yekta, Mohamadreza Aghamirsalim, Mahin Ahmadi-Pishkuhi, Mehrnaz Valadkhan, Hadi Ostadimoghaddam, Ahmad Ahmadzadeh Amiri, Mehdi Khabazkhoob
Article information
Full Text
Bibliography
Download PDF
Statistics
Full Text

In this magazine in October, this article was published: “Agreement between Pentacam and handheld Auto-Refractor/Keratometer for keratometry measurement”,1 this study was conducted in 2015 in Iran, in which 579 right eyes were analyzed. The objective of this study according to the authors was: “we evaluated the agreement and interchangeability of handheld NIDEK ARK-30 and Pentacam systems”, after critically analyzing the article, we decided to present some reflections regarding the statistical errors that prevent establishing agreement Between the methods.

In the article it was calculated: Pearson's correlation coefficient, Student's t-test for paired data and Bland–Altman graphical analysis with 95% agreement limits, only the latter is useful for a concordance analysis between measuring instruments.

The authors realize that despite having high Pearson's correlation coefficients (r>0.9), these do not reflect the agreement between the measurement methods, I quote: “high correlation coefficient does not itself indicate a high agreement between the results of two devices,” but they failed to realize that the statistical analysis was wrong.

In this case, Pearson's correlation coefficient is not an adequate method to assess the degree of agreement between two variables, you can get an r=1, that is a perfect correlation, and despite this the measurement will be biased proportionally, with no agreement between the measurements, therefore, it does not provide information on the agreement between the two measurement methods, but only measures the linear association between two variables. Nor is the Student's t-test for paired data an adequate technique in this type of analysis, this test performs only a comparison of means, without comparing their distribution, which is what matters.

For a concordance analysis, an analysis of the variance must be used, so an ANOVA analysis is used, then based on this ANOVA analysis an intraclass correlation coefficient (ICC) is calculated, this is a parametric test, in the which coefficient estimates the average of the correlations between all possible pairs of observations available, like Pearson's r, this ICC ranges from 0 to 1, this coefficient is used to establish the agreement. The ICC can be assessed as follows: >0.90 very good; 0.90–0.71 good; 0.70–0.51 moderate; 0.50–0.31 slight; <0.30 poor or absent.

Therefore, in this study the opportunity to correctly establish the agreement between both methods was lost, due to a poor choice of statistical analysis methods.

Reference
[1]
H. Hashemi, S. Heydarian, A. Ali Yekta, et al.
Agreement between pentacam and handheld auto-refractor/keratometer for keratometry measurement.
J Optom, 12 (2019), pp. 232-239

http://www.miopia.cl.

Copyright © 2020. Spanish General Council of Optometry
Journal of Optometry

Subscribe to our newsletter

Article options
Tools

Are you a health professional able to prescribe or dispense drugs?

Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.