A Visual Field That Looks Worse but May Not Be Reliable
Mrs. Okafor, 61, is a glaucoma suspect being followed for a slightly enlarged optic cup in the right eye. Today's Humphrey 24-2 SITA Standard field on the right eye was obtained after a 45-minute wait in a crowded lobby; she has a cold and kept shifting in the chair. Results: fixation losses 2/13, false-positive errors 4%, false-negative errors 40%, foveal threshold 34 dB, mean deviation -6.8 dB, pattern standard deviation 3.9 dB, and glaucoma hemifield test "outside normal limits." Eight months ago her mean deviation was -1.2 dB with false-negative errors of 8%. She told you several times that she "wasn't sure when to click" and once asked whether the test had stopped. The physician is comparing the two printouts and asks you whether this represents progression.
1. Which reliability indices do you report, what does each mean, and which finding here most undermines the test?
Read the worked answer
Report all three indices with the raw counts, not just the percentages:
- Fixation losses 2/13 (~15%) are within the usual acceptable range (roughly under 20%).
- False-positive errors 4% are acceptable; high false positives suggest a trigger-happy patient and can artificially raise sensitivity.
- False-negative errors 40% are well above the commonly used 15-20% cutoff. A false negative means the patient failed to respond at a point where the machine had already confirmed they could see a brighter stimulus. High false negatives can mean genuine dense field loss, but they also occur with fatigue, inattention, misunderstanding, or poor button technique.
The most undermining finding is the 40% false-negative rate combined with a normal foveal threshold of 34 dB. If the mean deviation of -6.8 dB reflected true widespread loss, central sensitivity would usually be reduced as well. A normal foveal threshold with high false negatives and a large drop from the prior field points strongly toward an unreliable test. The glaucoma hemifield test is calculated from the data, so it inherits the unreliability; it cannot be read as proof of progression. The fixation losses and false positives are reassuring, which makes inattention rather than a machine or set-up problem most likely.
2. How would you set up and coach the repeat test so the comparison is valid?
Read the worked answer
Repeat on a different day when she is rested and her cold has settled if the clinical urgency allows, because a same-day repeat after a 45-minute wait may reproduce the fatigue. Match the test strategy (SITA Standard 24-2) and the same eye coverage so the printouts are comparable.
Set-up: quiet room, door closed, phone silenced, comfortable chair height, correct trial lens for near fixation in place, patch over the nontested eye, chin and forehead against the rests, and the fellow eye monitored through the telescope. Confirm she can see the central fixation target and understands that she should look straight ahead, blink normally, and press only when she sees a light, even a faint one. Tell her not to worry about whether the light is "real" or about being fast; guessing is acceptable only when she thinks she sees something. Allow a bathroom break before starting and a short rest between eyes.
Coaching during the test matters: remind her to keep looking at the center, not to follow the lights, and to blink. Do not coach about specific results or tell her she is "failing." Note the room conditions, her symptoms, and any breaks. If the repeat is still unreliable, ask the physician about a shorter strategy or a different test rather than simply repeating the same conditions.
3. What do you tell the physician, and what should not be concluded from this field?
Read the worked answer
Communicate the data and the circumstances plainly: this field shows a 40% false-negative rate, a normal foveal threshold, a mean deviation that dropped from -1.2 to -6.8 dB, and a patient who was symptomatic, fatigued, and unsure when to respond. State that the indices and the observation make the result unreliable and that a repeat is needed before any change in diagnosis or treatment. Offer the practical details the physician cannot see on the printout: the 45-minute wait, the cold, the chair shifting, and her comments.
Do not conclude progression, do not label her as having a new dense defect, and do not recommend escalating therapy on this single field. Also do not dismiss her as a difficult tester. The honest formulation is "unreliable field, cannot interpret change; repeat and correlate with optic nerve examination and imaging."
If a reliable repeat confirms a defect, the physician may consider a 10-2 field or comparison with optic nerve OCT, but that is their decision. Your documentation should include the repeat plan, the education given, and the fact that the prior field was reliable, so the next technician and physician understand the baseline.