Quantitative Exposure Assessment (Book Review)
TL;DR: An enlightening book on the philosophy and mathematics of exposure assessments. Recommended, but a bit of a grind due to mathematical notation and equations.
Statistics is the bane of many occupational hygienists. So it is unfortunate that foundational knowledge of statistics is essential to meaningful quantitative exposure assessments. There are thousands of textbooks on general statistics, and a handful of occupational hygiene resources with a chapter or two on the topic. There are very few books that discuss focus on statistics exclusively in the context of exposure assessments.*
This by itself makes Rappaport & Kupper’s Quantitative Exposure Assessment worthy of recommendation. I found myself solidifying ideas I was shaky on, and completely reevaluating other ideas (such as the logic for using the mean verses percentiles for decision making). The one hesitation I have in recommending this to you is that the authors, who are titans in this area, don’t realise how uncomfortable some of us practising hygienists are with mathematical notation. The maths isn’t too difficult (upper high school level), but it still took me a few re-reads of many sections to understand what was being said.
It's a book written originally in 2005, and acts a summation of Rappaport’s, Kupper’s and others’ research since 1980’s - 150-200 papers in total are referenced. But don’t let the slight age fool you. I think it is still completely relevant to hygiene today. The only change may be a greater focus on bayesian statistics (this book is solely uses frequentist statistics).
The book has four distinct parts:
History & principals
Models of analysis
Models of decision making
Connecting to disease risk
History and Principals
Although the start of the book has no statistics at all, I feel that these chapters are worth the price of the whole book alone. The first two chapters discuss the history of exposure monitoring and the development of (U.S.) exposure standards. It’s not just an interesting history lesson on the last 100 years; the authors also discuss their concern on the change in priorities of occupational hygienists.
One such idea is despite going numbers of hygienists, and technological improvements in monitoring and sample analysis, the depth of quantitative exposure assessments appears to be dropping! While previously assessments were scientific investigations into health risk, by focusing on exposure standards the role of hygienists is now more of managing corporate liability. A confronting thought. But one I have appreciated reflecting on.
Chapter 3 discusses sample collection strategy including (not so) similar exposure groups and random verses worse-case sampling approaches.
Models of Analysis
This is where it really kicks off. This series of chapters starts by describing the relationship between normal and log-normal distributions and their descriptive statistics.
From this models describing exposures are slowly built up. Starting from focusing on the group as a whole, then accounting for worker variances, and then exposure determinants generally. This is where is helps to be a little familiar with notation. For example, while it is explained in the book, I found myself reading out the notation long hand. That is, “mu(Yki)” is read as “the arithmetic mean of the logged transformed results for a given person in a given group”. Quite a mouth full.
The context is occupational hygiene but the stats are nothing special. What is described are things like linear regression and analysis of variance (ANOVA) etc. Things you would find in any introductory stats course, just explained in context.
Models of Decision Making
By now perhaps you know whether this is a book for you, so I will keep the detail light from here. This section discusses the pros, cons, limitations and underlying purposes of using the arithmetic mean and percentiles in making decisions on the acceptability of exposure profiles.
The authors also present an approach to deciding on what kind of controls - that is, controls that affect everyone or specific individuals - based on the variability of exposures.
Connecting to disease risk
Finally the authors discuss going beyond exposure quantification. They show how understanding the pharmacokinetics, such as through biomarkers, can improve the estimates to worker health.
It’s not a long book. It’s not a revolutionary book (anymore). But I think it has helped me to understand a lot of the fundamental ideas we often take for granted when using calculators like IHstats and ExpoStats.
Check it out if you are interested. If you do, please let me know your thoughts!
Link to purchase the book - but you may find it elsewhere
https://www.amazon.com/Quantitative-Exposure-Assessment-Stephen-Rappaport/dp/B00587HR2Y
*If you know any good ones, let me know!