Trial Information

Reliability of Hertel Exophthalmometer Measurements

Aim 1. In order to determine intrauser reliability of the Hertel exophthalmometer, patients
will have 2 independent exams using the Hertel exophthalmometer on the same day performed by
each of the same researcher (either DAC or HLP) in a blinded fashion. For each patient, we
will independently measure the base value and the values for the right and left eye
proptosis, respectively, using the same Hertel exophthalmometer. A randomization process
will be employed to determine which researcher (either DAC or HLP) will take the first set
of measurements. The distance of the observer from the patient will be standardized for all
measurements. Patient research data will be de-identified and recorded for each patient and
subsequently analyzed for interobserver reliability.

Aim 2. In order to determine interuser reliability of the Hertel exophthalmometer, patients
will have a second independent exam on the same day by the second observer as determined for
Aim 1 above. For each patient, we will independently measure the base value and the values
for the right and left eye proptosis, respectively, using the same Hertel exophthalmometer.
The distance of the observer from the patient will be standardized for all measurements.

Patient research data will be de-identified and recorded for each patient and subsequently
analyzed for interobserver reliability.

Aim 3. Variability in Hertel measurements may be related in large part to differences in the
base measurement. The base measurement is determined by the bony orbital structures and
should not be significantly affected over time by the soft tissue remodeling commonly seen
in TED and most other orbital disease processes. We will have the second observer re-measure
the same patient using the same Hertel exophthalmometer by beginning with the same base
measurement determined by the first observer. Again, all measurement distances will be
standardized. No other information will be provided to the second observer other than the
beginning base measurement. Patient research data will again be de-identified prior to
recording the information and subsequently analyzed to determine if beginning with the same
base improves interobserver reliability.