Dear Editor

The article by Harrison and colleagues’[1] reports on a relationship between personal and static microenvironment air sampling for carbon monoxide and nitrogen dioxide and for PM10 which include the addition "of a personal cloud increment." Static sampling is also commonly referred to as area or stationary sampling.[2,3] These relationships are important because static sampling is more easily achieved than personal measurements and is generally less costly. To achieve a relationship for personal and static sampling they must be collected from the same pollutant population.[4-7] Thus, in establishing a microenvironment or personal cloud increment, there must be a relationship within the sampling location for the pollutant.

Previous occupational studies have noted no relationship[2,4,8-11] and a relationship[12,13] between personal and static sample measurements. As mentioned by Harrison et al., personal samples are generally higher in concentration than static samples because of people being closer to the source and spending more time within the source location, or in the emission pathway.[4,14] When static samplers are placed at the source location or emission pathway they are similar to the values reported for personal samples,[2,3] and in some incidents may exhibit a higher concentration.[4,13,15]

The relationship reported by Harrison et al., for CO and NO₂ is likely a result of these pollutants being a gas, their ability to diffuse, low reactivity, and similarly in concentration between indoor and outdoor environments. A personal cloud factor must be incorporated into the PM₁₀ measurement because of greater variability of concentration from location to location.[16] A microenvironment represents a similar location and the personal cloud is a correction factor extrapolating for the static exposure to personal measurements. It must be noted that this adds a degree of uncertainty in extrapolating exposure from one sampling method to the other. Even though static samples may be reported as similar, they will ultimately exhibit a lower concentration than personal measurements.

Harrison et al, provided summation of their data in the form of arithmetic mean (AM) and standard deviation. When data from Tables 2 and 3 were evaluated for form of distribution, using the Shapiro-Wilk test,[17] most exhibited a non-normal distribution (Table). However, due to the small number of samples in Harrison’s data the actual form of distribution cannot be determined. It is suggest[2,18] that the logarithmic form best represents airborne pollutants, including Harrison’s data. When providing pollutant data, it has been suggested to include summary statistics that representative it’s form of distribution.[2] Data should be shown as AM, standard deviation, range, geometric mean, and geometric standard deviation (GSD).[2,12] It has been suggested[19,20] that health effects from exposure are more closely related to AM values, especially for those that are chronic in nature, making AM an important summary value to report. Reporting all summary statistics will allow future investigators to select summary data most relevant to their purpose.

Tables: Form of distribution for data reported in Harrison et al., Tables 2 and 3

Table 2

Table 3

Since many environmental pollutants are distributed throughout a location, like homes, modeling will prove useful in establishing a relationship between personal and static samples. However, this relationship may not only depend on sampling locations and emission pathways, but the actual pollutant as well.[6]

Variability among samples must also be considered when predicting exposure levels. Most sample populations exhibit a GSD (day-to-day variability) of 2.0 to 3.0.[2] The probability of samples with this variability being “related” is about 28% to 17%.[21] The GSD for the data reported by Harrison et al, ranged from 1.4 to 2.6. Thus, sample variability raises issues with the predictability of accuracy in exposure estimation.[21] This variability may also skew modelling as well resulting in fallacious interpretations; although as mentioned in Harrison et al, when the population sample becomes larger or uses pooled data these influences may become diminished.

Historically, most inferred that there is no relationship between personal and static exposures,[2-4,6,9-11] while studies such as that provided by Harrison et al, question this concept. Establishment of a relationship between these two sampling methods will allow incorporation of additional data into occupational, environmental and epidemiological studies,[16] although caution must be applied in interpreting any relationship based on previous findings.[2,4] Thus, care must be exercised when evaluating studies that solely use static sampling as the method of estimating personal exposure.[7]

References

(1) Harrison RM, Thornton CA, Lawrence RG, Mark D, Kinneisley RP, Ayres JG. Personal exposure monitoring of particulate matter, nitrogen dioxide, and carbon monoxide, including susceptible groups. Occp Environ Med 2002; 59:671-9.

(2) Lange JH. A statistical evaluation of asbestos air concentrations. Indoor-Built Environ 1999; 8:293-303.

(3) Corn M. Assessment and control of environmental exposure. J Allergy Clin Immunol 1983; 72:231-241.

(4) Lange JH, Kuhn BD, Thomulka KW, Sites SLM. A study of matched area and personal airborne asbestos samples: evaluation for relationship and distribution. Indoor and Built Environ 2000; 9:192-200.

(5) Esmen NA, Hall TA. Theoretical investigation of the interrelationship between stationary and personal sampling in exposure estimation. Appl Occup Environ Hyg 2000; 15:114-119

(6) Liu LJS, Koutrakis P, Suh HH, Mulik JD, Burton RM. Use of personal measurements for ozone exposure assessment: a pilot-study. Environ Health Perspectives 1993; 101:318-324.

(7) Edwards RD, Jurvelin J, Koistinen K, Saarela K, Jantunen M. VOC source identification from personal and residential indoor, outdoor and workplace microenvironmental samples in EXPOLIS-Helsinki, Findland. Atmospheric Environ 2001; 35:4829-4841.

(8) Lange JH, Thomulka KW. Air sampling during asbestos abatement of floor tile and mastic. Bull Environ Cont Tox 2000; 64:497-501

(9) Linch AL, Weist EG, Carter MD Evaluation of tetraethyl lead exposure by personal monitoring surveys. Am Ind Hyg Assoc J 1970; 31:170- 179.

(10) Stevens DC. The particle size and mean concentration of radioactive aerosols measured by personal and static air samples. Ann Occup Hyg 1969; 12:33-40.

(11) Linch AL, Pfaff HV. Carbon monoxide: evaluation of exposure potential by personal monitor surveys. Am Ind Hyg Assoc J 1971; 32:745- 752.

(12) Breslin AJ, Ong L, Glauberman H, George AC, LeClare P. The accuracy of dust exposure estimates obtained from conventional air sampling. Am J Ind Hyg Assoc J 1967; 28:56-61.

(13) Lange JH, Lange PR, Reinhard TK, Thomulka KW. A study of personal and area airborne asbestos concentrations during asbestos abatement: a statistical evaluation of fibre concentration data. Ann Occup Hyg 1996; 40:449-466

(14) Leung P-L, Harrison RM. Evaluation of personal exposure to monoaromatic hydrocarbons. Occup Environ Med 1998; 55:249-257.

(15) Lange JH, Thomulka KW. Airborne exposure concentration during asbestos abatement of ceiling and wall plaster. Bull Environ Cont Tox 2002; 69:712-718.

(16) Cherrie JW. How important is personal exposure assessment in the epidemiology of air pollutants? Occup Environ Med 2002; 59:653-654.

(17) Shapiro SS, Wilk MB. An analysis of variance test for normality. Biometrika 1965;52:591-611.

(18) Esmen NA, Hammad Y. Log-normality of environmental sampling data. J Environ Sci Hlth 1977; A12:29-41.

(19) Seixas NS, Robins TG, Moulton LH. Use of geometric and arithmetic mean exposures in occupational epidemiology. Am J Ind Med 1998; 14:465-477.

(20) Armstrong BG. Confidence intervals for arithmetic means of lognormality distribution exposures. Am Ind Hyg Assoc J 1992; 53:481-485.

(21) Leidel NA, Busch KA, Lynch JR Occupational exposure sampling strategy manual. DEHW (NIOSH) Publication Number 77-173, National Technical Information Service Number PB-274-792. Cincinnati, Ohio: National Institute for Occupational Safety and Health, 1977.

Dear Editor

We were interested to read the recent rapid response to our short report by Leman and colleagues,[1] but were surprised that they seem to have taken issue with the findings, particularly as the purpose of audit is to evaluate practice and, with the application of the iterative audit cycle, improve the quality, effectiveness, and efficiency of care provided.

While we accept that the paper has several limitations, and that Dr Leman and his emergency department (ED) colleagues make some useful points, we still feel there are important lessons to be learnt for both occupational health (OH) and the emergency departments, and would like to respond in detail to the comments made by the ED physicians.

We agree that the title could have been different and, while we accept that there may be factors which may influence out-of-hours performance, we disagree with their assertion that there is a significant difference between the groups of staff seen in the ED and the OH department; there is no information to suggest that the two groups of staff (that is those seen in-hours and those seen out-of-hours) are so different that the results were significantly biased.

We agree that the major limitation in this study is that non-recording of data was used as a surrogate for omission in procedures and this was naturally discussed in the paper. However, we would like to point out that adequacy of records is an important part of clinical practice, and many audits and research are based on this. Furthermore, incompleteness of proforma or records cannot be excused both from a clinical governance or from a medicolegal perspective, and the fact that ED proforma are not being completed in itself warrants some further evaluation.

In addition, it seems that Dr Leman and colleagues feel that our conclusions are based purely on non-recording of data. This is not the case; our conclusions were also based on incorrect or inappropriately recorded data. Furthermore, data such as whether appropriate recipient and source blood testing had been arranged, or whether hepatitis B post-exposure prophylaxis or HIV post-exposure prophylaxis had been prescribed were corroborated by checking objective information such as blood test results, and batch numbers or completed drugs charts documenting administration of drugs or vaccines.

There also seems to be an intimation by the ED physicians that the paper was biased because of non-blinding and because the authors used their own criteria for assessing adequacy of risk assessments and management. The issue of non-blinding was discussed in our paper, and we do not accept that non-blinding would have seriously affected the results. The criteria used for assessing adequacy of assessment and management were based on both trust policy and national standards – not made up on the whims of the authors.

As stated in our paper, there are shortcomings in the management of occupational body fluid exposures both in and out-of-hours. The deficiencies in the OH Department are being addressed and extra staff training has been instituted. We are confident that our management of body fluid exposures has improved as a result of this.

We disagree that it is ‘difficult to draw many conclusions about the HIV PEP data.’ The data extracted from the notes is shown in table 2, and the validity of our concerns is selfevident in this table, and shows that even with incomplete risk assessments HIV PEP should not have been prescribed in a number of ED cases. ,

We were aware that some of the individuals presenting to the ED with body fluid exposures were not staff, and as this was clearly recorded in the ED records we were able to specifically exclude non-employees from the data analysis. Therefore, there was no patient misclassification, and the poor attendance rate at the OH department was an accurate, and non-biased finding.

We agree that ED staff are experienced in clinical risk assessment; however, in our paper we were referring to risk assessment in the context of occupational health. While this was not stated explicitly in our paper, it is something that would have been clear to the readers of Occupational and Environmental Medicine - who will appreciate that clinical risk assessment for managing disease in ‘patients’ is very different to risk assessment in healthy staff in the context of disease prevention.

Our paper mentions several solutions to the shortcomings in present procedures in our trust – one being the provision of out-of-hours OH telephone advice for management of body fluid exposures, another being the development of management algorithms. We provided examples of two similar trusts where this system seems to work efficiently. Our suggestion that using management algorithms may help is based on the fact that they seem to have been useful in many aspects of medicine. Of course, an intervention study would be the best way of ascertaining this, and would be an ideal way of completing the audit cycle.

The final point in the response from Dr Leman and colleagues is astonishing: the authors seem to think that audit should be subject to the same consent and ethics committee review processes as research. The General Medical Council guidelines on confidentiality and research state quite clearly that guidance relating to research does not apply ‘to clinical audit which involves no experimental study...’ Moreover, in the case of audit the guidance states that disclosure of anonymised data does not require patient consent.

In addition, the assertion that this work was performed without the consent of the ED is similarly surprising: one of the signatories gave their permission and support for this analysis to be done. Furthermore, at least three of the signatories saw an early draft of this paper which was sent to one of the ED physicians for comment. The less than constructive criticism from three of the signatories actually made us question whether this paper was worth submitting for publication. Accordingly, we asked three senior consultants (in other relevant specialties) within the same trust to provide an objective overview of the paper. All three (one a professional epidemiologist) felt that, despite the acknowledged limitations, our findings were worth reporting. The peer review process instituted by OEM confirmed this, including as it did a helpful critique from the US where they are in no doubt of the importance of this issue. As a consequence of our original experience with the ED, we did not feel that asking for their consent for publication would serve any useful purpose.

The response by Leman and colleagues shows quite clearly that they have not appreciated the main message of our paper - namely, that present procedures do not serve the best interests of staff, and there are implications for both the OH department and ED. Moreover, the suggestion that an occupational health department could run as a telephone advisory service indicates their lack of understanding of occupational health practice. We agree that audit should be collaborative in order to develop and support best practice.

Dr Dipti Patel (Occupational Physician)
Dr Ira Madan (Occupational Physician)
Dr David Snashall (Occupational Physician)

Reference

(1) Leman P, O'Connor N, Terris J, Goudie A and Lacy C. Occupational Health Department role in bodily fluid exposures [electronic response to Patel D et al., Out of hours management of occupational exposures to blood and body fluids in healthcare staff] occenvmed.com 2002http://oem.bmjjournals.com/cgi/eletters/59/6/415#39

Dear Editor

In a recent interesting study published in your journal, Hoogendoor et al.[1] determined that high physical work load and job dissatisfaction increase the risk of sickness absence due to low back pain. I would like to focus on the job satisfaction variable.

It is to be noted that the above study was performed in a prospective fashion with employed workers who had no recent history of low back pain injury. As such, I would like to familiarize the readership with a series of studies performed with chronic low back pain (CLBP) patients treated in a pain facility. The results of the studies described below resonate with Hoogendoor's results[1] and point to the importance of perceived job stress and job dissatisfaction and their importance to job function.

In a series of four papers, Fishbain and colleagues have attempted to determine if pre-injury job satisfaction impacts on "intent" to return to work to the pre-injury job after pain facility treatment. In the first report, Fishbain et al[2] demonstrated that chronic pain patients not intending to return to work after pain facility treatment were more likely to complain of job dissatisfaction. In the second report from this group, Rosomoff et al.[3] demonstrated that an association between non-intent to return to work after pain facility treatment and pre-injury job dissatisfaction was similarly found across Workers' Compensation and non- Workers' Compensation chronic pain patients. In the third report, Fishbain et al.[4] looked at actual return to work after pain facility treatment in relation to these variables. They found that actual return to work was predicted at one month "by intent", perceived job stress and job like (job dissatisfaction plus other variables). At 36 months, return to work was predicted by "intent" and by perceived job stress plus other variables. In the final study, Fishbain et al.[5] attempted to predict "intent" to return to work after pain facility treatment in relation to actual return to work. "Intent" was predicted by perceived pre-injury job stress plus other variables. In addition, those chronic pain patients who intended to return and did not, were predicted by whether there was a job to go back to. Also chronic pain patients not intending to go back to work to the pre-injury job initially, but doing so later, were predicted by having a job to go back.

Overall, this series of studies points to a strong relationship between pre-injury work variables such as job dissatisfaction and "intent" to return to that job after treatment. In addition, these studies indirectly support the findings of Hoogendoor et al.[1] It seems that in trying to understand the low back pain injury and recovery process, it is important to take into account work related perceptions such as those of perceived job dissatisfaction and job stress.

References

(1) Hoogendoorn WE, Bongers PM, de Vet HCW, Ariens GAM, van Mechelen, Bouter LM. High physical work load and low job satisfaction increases the risk of sickness absence due to low back pain: results of a prospective cohort study. Occup Environ Med 2002;59:323-328.

(2) Fishbain DA, Rosomoff HL, Cutler R et al. Do chronic pain patients' perceptions about their preinjury jobs determine their intent to return to the same type of job post-pain facility treatment? Clin J Pain 1995;11:267 -278.

(3) Rosomoff HL, Fishbain DA, Cutler R et al. Do chronic pain patients' perceptions about their pre-injury jobs differ as a function of worker compensation and non-worker compensation status? Clin J Pain 1997;12:297- 306.

(4) Fishbain DA, Cutler R, Rosomoff HL, Khalil T, Steele-Rosomoff R. Impact of chronic pain patients' job perception variables on actual return to work. Clin J Pain 1997;13(3):197-205.

(5) Fishbain D, Cutler B, Rosomoff HL, et al. The prediction of chronic pain patient "intent", and "discrepancy with non-intent" for return to work post pain facility treatment. Clin J Pain 1999;15:141-150.

Dear Editor

I write in response to the article by Hans Kromhout[1] which sets out the case for exposure monitoring and proposes robust strategies for collecting data. He acknowledges that exposure monitoring may be expensive, but justifies it on the grounds that it is needed to ensure worker protection and data can be used for multiple purposes (hazard evaluation, control and epidemiology). All this ignores the variety of competences and numbers of firms who use chemicals in the workplace.

We agree that good quality exposure data is extremely valuable for assessing the effectiveness of control measures, studies on health effects related to use of specific substances and for long term epidemiological studies. Now that workers do not normally remain in one job all their working life and may be exposed to many chemicals in different industries, the lack of well validated exposure measurements is a concern. It will limit our ability in the future to carry out meaningful epidemiological studies.

In the UK we estimate that over 1.3 million firms are using chemicals. It is not realistic to suggest that all these firms should be carrying out the type of sampling regimes the article suggests. The costs would be astronomical and there is no capacity to collect, analyse and interpret all the samples that would be generated. Recognising this and that small firms needed help to apply the risk assessment requirements of the Control of Substances Hazardous to Health (COSHH) Regulations led HSE, in collaboration with industry and trade unions, to the develop COSHH Essentials.

COSHH Essentials is not intended to replace the collection of well validated exposure data, where that is justified, rather it is intended to help firms, particularly small and medium sized firms, properly control the chemicals they are using. Inevitably, a generic system like COSHH Essentials which groups chemicals has to err on the side of caution, but the controls recommended by COSHH Essentials were peer reviewed by an expert group established by the British Occupational Hygiene Society and have the support of the industry and trade unions. COSHH Essentials has been used now for over 3 years by many firms. We have not had complaints that the controls are over precautionary. Thus, we reject the implication in the article that using COSHH Essentials leads to "ill advised control measures will arguably be even more costly in the long run, a classic case of being "penny wise but pound foolish".

The article misrepresents the purpose of the expert system, Estimation and Assessment of Substances Exposure (EASE). This was developed to help meet the requirement under the Dangerous Substances Directive for a risk assessment on new substances. Since workplace exposure data cannot be collected on new substances prior to release to the market place, EASE was developed to provide an exposure estimate for use in risk assessment. It is entirely appropriate that this should be precautionary. It is not a weakness as the article implies. EASE is not intended as a tool to help employers control exposures in the workplace.

The aim of chemical control is the protection of employees' health. This is best achieved with a range of tools. EASE has a valuable contribution to make before substances are released into the market place, COSHH Essentials is proving to be a valuable and welcome tool for many small and medium sized firms, helping them to establish suitable controls. The recently launched electronic version will be of even greater help to many small firms. In other circumstances structured data collection is needed. These tools all have a valuable role to play. They should be viewed as complementary, not as alternatives in the way article suggests.

Reference

(1) Kromhout H. Design of measurement strategies for workplace exposuresOccup Environ Med 2002;59: 349-354.