Blunders in vestibular and audiological testing

Timothy C. Hain, MD Page last modified: July 6, 2011

There are a very large number of possibilties for error in vestibular and audiological testing. The purpose of this document is to briefly discuss a menagerie of errors that we have encountered in both our own and other laboratories over the years.

Before starting -- let me first say that everybody makes errors (we are only human, etc), and that a small number of errors in a laboratory is to be expected.

• A good laboratory will have a process in place to catch errors, will make reasonable efforts to prevent and correct errors, and admit when there is an error. When there is an error, a good laboratory will repeat the test at no charge, and do it right the 2nd time.
• A "bad" laboratory will deny the possibility that an error could ever occur, have no monitoring method (i.e. use an "open loop" method where the same person who does the test also monitors for error), and refuse to document errors (perhaps in fear of losing payment or credibility). This type of behavior is close to fraud.
• As is the case in bookkeeping, no laboratory should have a single person doing everything -- in other words, there should always be two eyes to check for errors.

1. ENG (electronystagmography) errors

There are so many ENG errors that we have a separate (large) section on them.

Failure to do ice water on "dead" ears -- especially with air technique, and oculomotor (eye tracking) testing on blind eyes (really !) are the two most common errors.

2. Rotatory Chair errors

• Wrong value for distance to eye calibration array. When there is a mismatch, there can be a peculiar pattern where mid-frequency VOR gain is too high.
• Light leak -- this causes too high gain at all frequencies, including low frequency. Phase is 0 at all frequencies.
• Slip ring problem -- causes noise.
• Uncalibrated chair velocity, or uncalibrated OKN velocity. This can be spotted with the VVOR test as gain should be 1.0 at low frequencies.
• Wrongly calibrated eyes, causing incorrect OKN (too fast)

3. VEMP testing errors

• Uncalibrated or too soft sound generator. Too small responses (or no responses). The insert earphones often wear out because VEMP's drive them at high volumes. Generally, VEMP earphones are not calibrated either. This is an obvious problem. We think that equipment needs to be improved.
• Wrong ear - -operator testing R ear, and recording it as L ear. This can be hard to detect as the equipment does not have a method of detecting the side. Here, we think a web-cam that takes a snapshot of the electrode hookup would be a good addition to the software that does VEMP testing.
• Poor electrode placement. Too small responses.
• Improper head positioning. Again, a web-cam would help.
• Improper technique allowing patient to relax neck before test is finished. Causes too small responses. Hard to detect as depends on operator admitting that allowed subject to "cheat".
• Peak picking - -operator chooses responses that fit their bias and ignores others. Causes responses to appear in "dead" ears. Here, we think software should be modified to automatically print out how many runs were obtained, vs how many were displayed.
• Impossible latency responses.

4. Posturography errors

• Poor instruction
• Too close holding of patient so no sway (too good responses).
• Too quick stopping of the machine so that no sway is recorded (just falls)

5. Hearing Testing errors

• Poor instruction
• Unreliable
• Inconsistency
• Lack of masking or improper masking
• Lack of comprehension of operator of masking dilemma (makes it appear patient has conductive hearing loss when has no hearing at all).
• Missing conductive hyperacusis in SCD.

6. ECochG

• Electrode not on ear drum (no response or noisy response)
• Excessive noise (nothing useful) --images/ecog noisy.jpg
• High frequency hearing loss makes recording impossible (no resposne)
• Peak picking -- operator chooses responses to fit their bias concerning patient's disorder (usually responses where no response exists)

7. ABR -- see this example.

• Excessive noise -- usually due to poor overly high impedance
• Peak picking -operator chooses responses to fit their bias concerning patient's disorder (usually peaks that have little relationship to the trace)
• Overly aggressive conclusions -- ABR's cannot "rule out' acoustic neuromas