The Change Probability Map; Change In Mean Deviation - Zeiss HUMPHREY 720i User Manual

7-16

The change probability map

Change in mean
deviation
REV. B 02/03 PN 51680-1
STATPAC ANALYSIS & PRINTING RESULTS
for the two baseline tests. Just above them is printed a message indicating whether the results
of the glaucoma hemifield test (GHT) were within normal limits, outside normal limits, or
borderline. The mean deviation from normal for this test is printed between the graytone and
the total deviation plot.
For each of up to fourteen (14) follow-up tests there are two more test result analyses: the
change from baseline plot (third column), and its associated probability map (fourth col-
umn). The change from baseline plot in the third column subtracts the follow-up test result
from the baseline and indicates changes at each tested point in dB notation. If, for example, a
point is indicated with -6, this means that the tested point was 6 dB lower than the baseline for
the same point. A zero (0) means no change from baseline. All results use age-corrected
values over the follow-up period.
The change probability map in the fourth column gives the statistical significance of the
decibel changes shown in the change from baseline plot. It compares the changes
between the baseline and follow-up fields to the inter-test variability typical of stable glaucoma
patients and then shows a plot of point locations which have changed significantly. A solid
triangle identifies a degree of deterioration found less than 5% of the time at that location in
medically stable glaucoma patients, that is, deterioration significant at the 5% level. An open
triangle identifies improvement significant at the 5% level. Points not changing by a significant
amount are indicated by a single, solid dot.
An X signifies that the program was unable to determine whether the encountered change was
significant or not. This occurs mainly with deepening field defects which were already quite
deep at baseline. The finite amount of empirical data available to us and the practical limitation
of the maximum attainable brightness of the instrument made it difficult to obtain exact
significance limits for deterioration in points which are already highly depressed.
The very same reasons, finite amount of data and limits to the maximum brightness, also make
it difficult to determine change with certainty in fields where the mean deviation from normal
exceeds -15. In addition, variability is extremely large in highly disturbed fields (MD < -15).
Since variability increases with increasing MD in this range, STATPAC analysis can only verify
stability, not deterioration or improvement. STATPAC printouts carry the message AVERAGE
MEAN DEVIATION OF ALL TESTS TOO LOW when the average MD of baseline and follow-up
tests is lower than -15.
STATPAC also evaluates the significance of change in mean deviation over time. The objective
is to highlight those clinical cases where the MD changes by more than is typically observed
in stable glaucoma patients. The amount of change in decibels is printed under the message
MD Change. If the MD change is significant at the 10%, 5%, or 2.5% level, that p value is
printed along with a solid triangle to indicate degradation or an open triangle to indicate
improvement. If the amount of change is not judged to be significant, the words "Not Signifi-
cant" follow the decibel value.
HUMPHREY FIELD ANALYZER II -
i
series