What can be done to minimize or eliminate artifacts and uncertainties from my noise dosimetry results, and to determine whether or not these were caused intentionally?
Responding is Rob Brauch, business unit manager, Casella CEL Inc., Buffalo, NY.
The quality of dosimetry results often are taken for granted because modern dosimeters are reliable and easy to use, and need little more than a quick pre-calibration before sampling to obtain accurate data. However, this doesn’t always ensure the data collected is suitable for important decisions such as who is added to a hearing conservation program. If you see questionable results – like dose levels 2-3 times higher than expected – it’s appropriate to dig into the “why,” “how” and “what” before accepting that result as valid.
It’s important to apply common sense and critical thinking to confirm your noise exposure samples will withstand scrutiny – like being challenged by an expert witness in a hearing or court case. Poor-quality noise data is rarely the result of a fault in the instrument – it’s more often due to inadvertent or intentional human behavior – either of which forces a resample, an outcome that is never desirable.
Many measurement errors are “artifacts” (things observed in scientific investigations that are not normally present) resulting from deliberate actions by workers wearing the device or those in a position to create conditions not normally part of the worker’s daily routine. Common examples are whistling, shouting, singing, clapping, banging tools or running machinery beyond normal operating speeds – or it could simply be a machine malfunctioning. You may have artifacts from someone wearing the dosimeter on break, sitting in their car blasting music. Other documented cases of unrepresentative samples include workers refusing to wear a dosimeter and placing it in a quiet or noisy area. While reasons to influence or even sabotage the sampling process vary greatly, it’s up to the safety professional to recognize it and identify when it occurred. Unfortunately, that’s difficult to do for a number of reasons.
How can you identify artifacts that invalidate tests? Until recently, the common process was looking at the minute-by-minute download report to identify times when unexpectedly high (or low) levels occurred, and asking the worker “exactly what were you doing at 10:17?” But chances are they won’t remember or, worse, if their co-workers caused the artifact they may give misleading or false responses. Relying on the data alone may seem daunting, as sometimes the evidence is based on high decibel levels that only occurred for a moment or two, so having prior experience in noise data analysis helps in spotting these anomalies.
Newer-generation digital dosimeters add a number of capabilities increasing the safety professional’s ability to identify artifacts and their causes. Some dosimeters are able to capture digital recordings (.wav files) of actual sounds produced during loud events. These are replayed after download to clearly identify if the noise source was part of the expected normal working environment. Not all workplaces allow sound recording, for policy reasons or privacy issues, so features like audio recording must be turned off.
Modern dosimeters also offer “Octave Band Analysis,” which detects if a “tonal” sound (such as a whistle) was present and shows unusual patterns of frequencies unlike the machinery or tools being used. While this technique relies on the trained eye of more advanced practitioners, it’s another capability that helps ensure better-quality outcomes in your noise sampling program. It also gives precise information about what type of hearing protection devices to offer.
Motion sensing is another recent advancement for identifying unusual conditions, alerting the practitioner if the dosimeter was taken off and left on a toolbox or bench for part of the work shift.
Whether using older, basic or advanced new noise dosimeters, it’s important to always challenge the results before accepting them as valid, realizing that the most common cause of invalid samples isn’t the instruments – it’s the influence of behaviors and environments in which they were deployed.
Editor's note: This article represents the independent views of the author and should not be construed as a National Safety Council endorsement.