Case study 1 : Silica in the construction industry

Hello colleagues.

Our very first case study for the expostats discussion forum. This idea came while discussing a case with an IH colleague who then kindly accepted the situation to be used to help users better understand some issues related to the use of Expostats.

Here is the link to the case study : link

In this case study, we discuss some issues related to what data to input together or not into expostats, how to interpret non detects when the LOQ is close to the OEL, and how to interpret the riskband or BDA charts.

Feel free to discuss !

1 Like

I love this. It’s very relevant to discuss these kinds case studies as I have see similar situations. I have a few strong opinions on this, mostly in agreement with the commentary, some not. Hopefully good fuel for further discussion!

General / Non-ExpoStats Comments

  1. There needs to be a clear problem statement / question / hypothesis before conducting an exposure assessment. What is the purpose of the monitoring? That will dictate how the monitoring should be undertaken and inform whether the results are relevant to the question being asked.
  • Are we trying to estimate a full-shift exposure profile, task-specific exposure, evaluation of control effectiveness
  • What was the rational for static monitoring and their locations? How far where they? Are they upwind?
  1. Exposure monitoring data without observations and contextual information is of very limited value. This is particularly true for small sample sizes such as these case study. Some example observations that seem relevant in the interpretation for me include:
  • for how long was the activity undertaken during the sample period?
  • What controls were in place (what is the ‘dust extractor’?)?
  • How well is the control zone established, are other workers likely to intrude on that zone?
  • Was this inside, outside?
  • Is this set up representative of future scenarios?
  • What are the specific locations for the monitoring?
  • Is it expected that cristobalite and tridymite would be present in the concrete?
    So many more questions that I believe that should be considered when analysing the numbers.
  1. Area verses Personal - OELs are indeed set for personal samples. So comparing these results to OELs for the purpose of legal compliance verification is inappropriate. But again, its not clear what the intention actually is

  2. Censored results above AL or OEL - Agreed with case comments. No useful conclusion can be reached through direct comparison, unless you assume the worst
 but even then not sure of the value of that.

General results comments

  1. Q+C doesn’t seem to make sense. Surely sample one should be <45 and sample three should be <16. Theoretically, the cristobalite could be at the limit of reporting. Therefore all of the Q+C are censored.

  2. There may not be a “total crystalline form” OEL, but I believe its common practise to combine exposure level percentages where there are several additive contaminants. (ref: ACGIH TLV guide etc).

  3. Total respirable dust may be a usual proxy for assessment


  1. What does the distribution represent by these combined results? I would argue nothing! All areas are misrepresented (either over or under). And I would assume that no person would spend time in each of the sample locations in proportion to the number of samples taken in that location.

I think the ExpoStats outputs for these combine samples are meaningless and should not be considered at all.

  1. “This statement (BDA) is wrong” - agreed! Actual exposure distributions verses samples distributions are often seen as exactly the same thing.

My Interpretation of Results

  1. The near-field result is greater than the OEL. Also, I assume the operators breathing zone would be closer to the source than the static sample. Given this, the grinding is poorly controlled. Detailed analysis and answering the complex questions is not needed here. First fix the problem.

  2. No need to further monitor control zone, or far zone at this stage. Controlling the source will further reduce these concentration levels.

Hope I was able to convey my thoughts clearly enough. Would love to hear other peoples thoughts. I think cases studies like this is a brilliant idea. Was so excited by it, I had trouble sleeping thinking about it last night haha!



1 Like

Excellent comments John, no disagreement with the points you raised ; these make the case study even more interesting. As you noted, we lack some important ancillary information, but we had to make do. Thanks for taking the time !

1 Like