Reply to the Letter by Dr Peter Griffin and Prof. Andrew Curran, “Response to Article by Prof. Hans Kromhout, Hygiene Without Numbers”

<span class="paragraphSection"><a href="#CIT0005" class="reflinks">Griffin and Curran (2016)</a> claim that I misunderstood the HSE’s message from the presentation entitled ‘Hygiene without numbers’ at a meeting in Manchester in October, 2015. They claim that HSE provides task-specific information to support small- and medium-sized enterprises (SMEs) in achieving control, based upon workplace studies including substantial measurement activities. They substantiate this evidence-based approach by pointing to a program launched by the HSE over a decade ago in 2004 (ending in 2007) among spray painters using isocyanate-based paints in the motor vehicle repair industry (<a href="#CIT0008" class="reflinks">Piney et al. 2015</a>). The evidence that this approach worked to reduce exposures came from an internal evaluation of the program showing that ‘18 per cent of body shops in the UK attended the SHADs (Safety and Health Awareness Days) and 90 per cent of the delegates expressed an “intention to act” to improve exposure control measures. A local assessment showed that at least 50 per cent of body shops improved exposure control measures.’ (<a href="#CIT0008" class="reflinks">Piney et al. 2015</a>). The evidence of the approach’s success was based amongst others on biological monitoring data showing that post-SHAD exposures were lower. Details of this biomonitoring study can be found in the <a href="#CIT0008" class="reflinks">Piney et al. (2015)</a> paper, but was previously published in the Annals of Occupational Hygiene by <a href="#CIT0006" class="reflinks">Jones et al. (2013)</a>. A close look at the design of the biomonitoring study shows that the researchers made a classic mistake in a longitudinal study and are actually describing a phenomenon called ‘regression to the mean’ that was first described by Sir Francis <a href="#CIT0004" class="reflinks">Galton (1886)</a> in a paper called ‘Regression to mediocrity in hereditary stature’. By only offering retests to workers whose initial biomonitoring results were positive, the repeat samples would by definition be lower than the mean of the initial positive results. With 83% of all initial samples under the limit of quantification, the average of the follow-up measurements of workers with positive samples had to be under the limit of quantification as well. This finding was therefore likely not the result of the HSE intervention (SHADs for small and medium-sized body shops where exposure to isocyanates occurred) but rather a case of classical regression to the mean.</span>

from #AlexandrosSfakianakis via Alexandros G.Sfakianakis on Inoreader http://ift.tt/2qPj3Th
via IFTTT