Glaucoma: Diagnosis and Therapy

Re: Coefficient of variation

Janos Nemeth M.D.

On Mon Dec 7, Hiroshi Ishikawa, M.D. wrote
>Thank you for the very interesting poster, 'Accuracy of Orbital Color Doppler Flow Velocity Measurements in Glaucoma Patients'.
>To my knowledge, reproducibility can be said good only when the coefficient of variation (CV) is less than 10%. Over 10%, reproducibility is not poor, but at the same time it is not good. I would say reproducibility is poor if CV is over 15%.
>I saw many parameters showing CV more than 10% in your results, especially in minor vessels, and you stated that reproducibility is relatively good. If you have any reference supporting your statement, please let me know.
>Did you detach and attach the probe while obtaining repeated images?

Thank you very much for your comment and question.
Concerning the coefficient of variation (COV), I agree with you. We found good reproducibility (COV below 10%) only in some arterial flow velocity parameters like peak systolic velocity, acceleration time and resistivity index and we suggested to use these parameters for diagnostic and follow-up purposes. Other parameters are less reliable, especially in ciliary vessels.
In case of central retinal vein, the COV values were around 10 % which is not bad. Different is the situation in the case of the ophthalmic vein where we expected a far lower reproducibility but we got COV values around 15%, at the limit of accaptability. It was a surprise as our earlier experiences suggested very high variability so we are not measuring routinely in this vessel.
Concerning your question: Yes, after obtaining an image we always detached the probe during the calculation and documentation of the results. So we attached again the probe for the next image about 1-2 minutes after the detachment.
Janos Nemeth M.D.

[ Previous ] [ Next ] [ Index ]           Sat Dec 12