The goal of a risk based exposure monitoring program is to obtain an estimate of a worker's actual exposure to a health hazard

One of our goals at GCG is to share technical aspects of occupational hygiene in a way that any health and safety professional or business/organisation can understand. This post on the Upper Confidence Limit (UCL) aims to share what this statistical measure is, explaining why and how we use it in exposure monitoring programs.

Assessing a worker’s exposure to a risk is most commonly achieved by sampling a representative portion of a group of workers, called a similar exposure group (SEG). All results for the SEG are then compared to the relevant occupational exposure limits (OELs).

This is a recognised approach internationally for chronic health hazards and is much more practical and cost efficient than assessing all workers each day.

Let’s review an example.

Should 5 workers perform the same job, they could be classified as the same SEG. Say for example that the group of workers in a SEG have over 1200 shifts worked per year, it is not feasible to monitor each working shift to find the true average exposure to workers. Therefore, a more realistic approach is taken that may require as few as 6-10 samples per year for that SEG.

This approach can mean that less than 1% of possible exposures are measured within a normal working year. As such, there is a strong reliance on a conservative measure to estimate the average exposure to workers.

Grouping all personal monitoring results for a SEG together and assessing the potential exposure to those workers will require statistical analysis to estimate not only the average, but also an indication of the spread (distribution) of the results.

In occupational hygiene, this spread is commonly measured using the geometric standard deviation (GSD).

Given a realistic approach is taken and sampling only occurs during a small portion of the shifts worked each year, a large GSD will heavily influence an estimation of the potential exposure to workers.

This generally means that the more variability in the sample results, the less certain we are of the actual average. Therefore, upper estimates of the average exposure are amplified.

When we compare a SEGs exposure against an occupational exposure limit, the statistical interpretation is rarely identical to what the actual exposure would be if we measured each and every shift.

One of the tradeoffs in taking a pragmatic approach to estimating exposures for a SEG, is that a more conservative approach is required due to the uncertainty associated with such a small number of samples.

The 95% Upper Confidence Limit (95%UCL), also commonly called UCL, is a frequently used statistical measure to compare grouped results against the occupational exposure limit. In short, it’s a tool that is used in occupational hygiene to help improve the data and support decision making. 

To put this all into perspective, because you’re generally sampling a small portion of shifts worked in a year, using only an average can be highly inaccurate. Instead, using a measure that considers the confidence we have in that data is a more conservative approach in light of the fact that we’re considering the health of our workers. The UCL is used internationally for this purpose and tells us with 95% confidence what the true average exposure to workers will be less than.

Let’s take a quick break – The definition of the UCL along with other terms spoken about in this article have been summarised in table 1.

Table 1 – Definitions of terms used in this article

Term Definition
SEG Similar Exposure Group (SEG). A grouping of workers who perform similar tasks and who are likely to have a somewhat comparable exposure to occupational health hazards.
OEL Occupational Exposure Limit (OEL) / Workplace Exposure Standard (WES). Airborne concentrations of a particular chemical or substance in the workers’ breathing zone that should not cause adverse health effects or cause undue discomfort to nearly all workers.
MVUE / AM Minimum variance unbiased estimator (MVUE) of the Arithmetic Mean (AM). The preferred estimate of the mean of a lognormal distribution.
GSD Measure of the spread of a dataset.
UCL This represents the value, below which we are 95% confident, lies the true value of the SEGS mean exposure.

In practice, it is likely the worst case mean and can be directly compared to the OEL to determine the potential risk to worker health.

For example, if the UCL is less than the occupational exposure limit, one can be 95% confident that the actual average worker exposure of the population is less than the occupational exposure limit. Being highly influenced by the spread of data on which it is calculated, the UCL can be misleading and overstate the risk when the:

  1. dataset has a small sample size, and/or;
  2. variation in individual sample results is high.

It is important to note that the use of any statistical measure is limited by the assumptions and parameters guiding its use.

There are many other factors that can influence exposure monitoring results. As such, statistical analysis should always be interrogated and interpreted by a competent person.

The information contained within this article is intended to be a basic guide to the UCL for professionals not in the occupational hygiene field. If you would benefit from additional information, please feel free to contact your local GCG office to speak with a Senior Hygiene Consultant.

Ready to chat more about UCL?

Get in touch. We’re here to help.

Australia wide: 1300 424 474 

dust monitoring GCG

Every day more people choose us, contact us today

Stay up to date with the latest in WHS and occupational hygiene news.