Dear Editor,

It is disturbing that the Interphone study group first publishes several papers purportedly finding negative results but only now publishes a validation study showing that the methods used to measure exposure are so deeply flawed that it was unlikely the previously published studies would detect an increase in risk of brain tumour in mobile phone users(1).

In the validation study of 672 volunteers recall of phone use over a 3-8 months period about six months later, was compared with objective data from phone bills and use of special soft-ware modified phones. The study found large random errors in recall which can be expected to have a “large impact” and bias risk estimates toward a null effect. There was also a wide variation between countries which further emphasises the randomness of recall (and the difficulty in conducting a meta-analysis of results from participating countries). Moreover the validation study design minimises the errors of recall for several reasons. It largely used "colleagues and acquaintances of the investigators". These volunteers would have been well informed of the intent of the study and the importance of accurately recalling mobile phone usage, and are therefore different from the naïve subject in the actual studies whose recall would be even more random. Also the validation study only sought recall of mobile phone use for 3-8 months about six months later, whereas in the actual studies recall for up to 10 years was sought which is likely to lead to even more randomness in recall. The validation study did not address the problem of impaired recall in the cases with brain tumours who are likely to have their memory affected from their tumour and subsequent surgery, radiotherapy, chemotherapy and psychological distress, let alone the accuracy of surrogates (spouses) used in some studies.

The validation study reveals at a minimum the flawed methodology being used in the Interphone studies which would lead to Type2 (false negative) errors thus lessening the ability to except with confidence the published negative papers regarding the health effects of mobile phones(2). The late publication of a key methods paper after publication of many results papers raises concerns about the processes within the Interphone study group.

Yours

Dr Bruce Hocking.

References

1. Vrijheid M, Cardis E, Armstrong BK, Auvinen A, Berg G, Blaasaas KG, Brown J, Carroll M, Chetrit A, Christensen HC, Deltour I, Feychting M, Giles GG, Hepworth SJ, Hours M, Iavarone I, Johansen C, Klaeboe L, Kurttio P, Lagorio S, Lonn S, McKinney PA, Montestrucq L, Parslow RC, Richardson L, Sadetzki S, Salminen T, Schuz J, Tynes T, Woodward A; Interphone Study Group. Validation of short term recall of mobile phone use for the Interphone study. Occup Environ Med. 2006 Apr;63(4):237-43.

2. Hocking B. Mobile phone use and risk of acoustic neuroma. Br J Cancer. 2006 May 8;94(9):1350; author reply 1352-3.

Dear Editor,

In their publication, the authors postulate that effects upon health and performance cannot be ruled out despite a low exposure to high-frequency electro-magnetic fields, effects far below the WHO threshold values. Unfortunately, this paper has substantial methodological problems.

1. There may well have been a clustering in the choice of test- subjects’ addresses in relation to the location of the mobile phone base stations. For example, in the sense that other environmental influences may have affected everyone in the same cluster. In this case, the independence of the people within the cluster may no longer be given, which in turn may lead to an overestimation of the effective sample size, and therefore to progressive (results are too often significant) decisions. In order to control this effect, the Intra-Class-Correlation should have been calculated to give an impression of the effective sample size. In the light of the few, tightly significant results, such effects may have played a decisive role.

2. It still remains unclear as to whether there are a priori differences as to the degree of “concerns about base station“ in relation to the expositional groups. This was unfortunately not presented in table 1. What was the percentage of test-persons who were concerned, and to what extend? How is this distributed throughout the groups? In addition, the degree of concern was measured on a ranking-scale, despite the fact that the co-variance analysis demands an interval scaled co-variant.

3. The Zerssen Scale was evaluated dichotomously, whereby even very slight complaints (“1“) were already registered as complaints. This then raises the question as to whether the named effects in the Zerssen scale would still be evident if one had pooled the groups 0 and 1 (no, and/or light complaints), and the complaints-group 2 and 3. Who doesn’t sometimes suffer of headaches, dizziness, and appetite disorders, etc? It would be interesting to see how the complaints - before a dichotomization – were distributed amongst the exposition-groups.

4. In Table 2, 13(!) covariance analyses were carried out, from which the p-value of one main effect reaches tendency level (p<0.1). After the implementation of a Bonferroni-Holm-Correction, which strictly seen should also include the p-values inside the covariance analyses, there are no effects to be found whatsoever. In this case, it is then not permitted to omit the non- significant co-variables, even in the co-variance analysis. The model for the covariance analysis must always be set, a priori, before testing; otherwise a Type I-error-inflation will ensue. The attained significance would only then be credible if confirmed on the basis of a further sample.

5. If one would adjust the p-values in table 4 according to the Bonferroni- or Bonferroni-Holm-convention, then none of the symptoms would be significant anymore. The same would apply to the effects shown in table 5.

To sum up: A scientifically founded evaluation would have had to have seen corrections according to Bonferroni-Holm (or some other). Had this happened, there would be no significant result, and thereby no provable effects. Certain information was not included in the tables, such as the nature of the dichotomization and choice of sample. This creates the impression that hypothesis finding and hypothesis testing were not done quite independently of each other, or at least, that there is little in this study to revoke this impression. However, the results beg a prompt replication with all details, hypotheses, and corresponding evaluation criteria fixed a priorily.

Dear Editor,

It has been clear for years, based on much published research, that symptoms in office workers are associated with a number of environmental factors in office buildings and also, independently, with psychosocial stressors at work. So we were surprised to see a recent article by Marmot et al. (1) report that, in offices in the Whitehall II Study, “raised symptom levels appear to be largely due to a working environment characterized by poor psychosocial conditions”(p. 288). The article concluded that the physical environment in the offices had a small and unimportant influence on these symptoms. The analyses, however, had substantial limitations that were not mentioned. Furthermore, the conclusions were inconsistent with much of the current scientific literature, but the discussion cited only other studies that agreed with the findings and none of the substantial literature that disagreed. We expand on these points below.

1) Key environmental measurements and interpretations used by Marmot et al. (1) in the 1991-1993 data collection are no longer considered relevant by most indoor environmental scientists. Single metrics of total volatile organic compounds that lump all compounds together have long been considered inappropriate for predicting human response, because irritancy and odor vary among specific volatile organic compounds by orders of magnitude (2). Metrics based on counts of culturable airborne fungi and bacteria do not detect most indoor microbial matter and “provide little information about the microbial status of an indoor environment” (3)(p. 58). Also, many of the thresholds for acceptability used by Marmot et al. are not considered relevant for studying building-related symptoms: e.g., dry bulb temperature between 19-24°C; carbon dioxide (CO2) ≤500 parts per million (ppm); or any particular number for total airborne fungi or bacteria or volatile organic compounds. In addition, the lumping together of extreme high and low levels for many of the parameters (e.g., combining very hot and very cold temperatures in one category and comparing to a broad middle range of temperatures) is inappropriate – high and low temperature (or high and low humidity) may have quite different, even opposite, effects. Researchers using more current or precise metrics have reported consistent associations between building- related symptoms in office workers and both lower ventilation rates (6, 7) and temperatures (8, 9). There is also a substantial literature showing that visible dampness and mold, but not traditional airborne mold counts, are consistently associated with asthma exacerbation and respiratory symptoms in building occupants (3, 4).

2) The authors do not report the association of passive tobacco smoke exposure at work with symptoms, although it is included in their models as a confounding factor (1) and was strongly correlated with increased symptoms (p = 0.004) in prior analyses of their data (5). The current article does not consider passive tobacco smoke exposure to be an indoor environmental risk factor, although it reports risk estimates for other indoor air pollutants (1).

3) The paper cites prior studies that agreed with its findings, but is inexplicably silent about the many prior studies that have disagreed (1). Ventilation rate (6, 7) and temperature (8, 9), both objectively assessed indoor environmental risk factors, have been significantly and independently associated with symptoms in multiple prior office studies. A 1999 review of studies reported between 1986 and 1999 found that (a) in 16 studies using measured ventilation rates, 20 of 27 comparisons of different ventilation rates found increased symptoms associated with lower measured office ventilation rates, and (b) 9 of 18 studies using CO2 measurements as simpler but less accurate indicators of ventilation rate also found such associations (10). Nine articles on associations between temperature and symptoms in offices, published between 1989 and 2004, show overall that symptom prevalence increased systematically as temperatures increased between about 21 and 24.5 ºC, almost entirely within the “acceptable” reference level used in Marmot et all (1, 11).

We suggest further analyses of the Whitehall II data by Marmot et al. to refine their findings and help resolve discordant results: a) the use of environmental metrics based on current knowledge that symptoms in office workers generally increase as temperatures increase above 21 or 22 ºC, as indoor CO2 increases above about 600 to 800 ppm, with presence of passive tobacco smoke, and in buildings with air-conditioning or humidification; b) statistical adjustment for season of study; c) the inclusion of the additional psychosocial variables available in Whitehall II (5); and d) separate analyses for outcomes of biologically related symptom subgroups – rather than continuing to use the very nonspecific “sick building syndrome” metric (for instance, lumps rash/itch together with cold/flu) which may be sensitive to stress-related over-reporting but insensitive to specific biologic effects.

In conclusion, substantial evidence suggests that psychosocial and physical factors in indoor environments, as well as biological and chemical factors, all influence the symptoms experienced by office workers, through multiple mechanisms that we still do not understand. Dismissing the importance of any of these indoor environmental risk factors is not useful or, based on all that we know, justified. Ultimately, researchers in many disciplines will be needed to help us understand causation and prevention of this problem.

Mark J. Mendell and William J. Fisk, Indoor Environment Department, Lawrence Berkeley National Laboratory

Competing interests: none.

References

1. Marmot AF, Eley J, Stafford M, Stansfeld SA, Warwick E, Marmot MG. Building health: an epidemiological study of "sick building syndrome" in the Whitehall II study. Occup Environ Med 2006;63:283-289.

2. Molhave L, Clausen G, Berglund B, et al. Total Volatile Organic Compounds (TVOC) in Indoor Air Quality Investigations. Indoor Air 1997;7:225-240.

3. Institute of Medicine Committee on Damp Indoor Spaces and Health. Damp Indoor Spaces and Health. Washington, D.C.: National Academies Press, 2004.

4. Park JH, Schleiff PL, Attfield MD, Cox-Ganser JM, Kreiss K. Building- related respiratory symptoms can be predicted with semi-quantitative indices of exposure to dampness and mold. Indoor Air 2004;14:425-33.

5. Marmot AF, Eley J, Nguyen M, Warwick E, Marmot MG. Building health in Whitehall: An epidemiological study of causes of SBS in 6,831 civil servants. In: Woods JE, Grimsrud DT, Boschi N, eds. Proceedings of Healthy Buildings/ IAQ '97. Washington, DC, 1997:483-488.

6. Stenberg B, Eriksson N, Hoog J, Sundell J, Wall S. The Sick Building Syndrome (SBS) in office workers. A case-referent study of personal, psychosocial and building-related risk indicators. Int J Epidemiol 1994;23:1190-7.

7. Jaakkola JJ, Miettinen P. Ventilation rate in office buildings and sick building syndrome. Occup Environ Med 1995;52:709-14.

8. Mendell MJ, Fisk WJ, Petersen MR, et al. Indoor particles and symptoms among office workers: results from a double-blind cross-over study. Epidemiology 2002;13:296-304.

9. Jaakkola JJ, Heinonen OP. Sick building syndrome, sensation of dryness and thermal comfort in relation to room temperature in an office building: Need for individual control of temperature. Environment International 1989;15:163-168.

10. Seppanen O, Fisk WJ, Mendell MJ. Association of ventilation rates and CO2 concentrations with health and other responses in commercial and institutional buildings. Indoor Air 1999;9:226-252.

11. Seppanen O, Fisk WJ. Some quantitative relations between indoor environmental quality and work performance or health. ASHRAE Research Journal 2006;(in press).

Dear Editor:

We read with interest the article by Alamgir et al (1) regarding the use of hospital discharge records in occupational health in the April issue of Occupational and Environmental Medicine. The paper adds to the evidence that these records represent an alternative and independent source of information for serious work-related injuries. In their introduction, the authors also make the point that, to their knowledge, studies in Canada using this data source have not validated its use against other available indicators.

We would like to bring to the attention of readers a previous Canadian study conducted by us (2) in which we attempted to confirm the diagnoses in the hospital records. In an initial study (3), we examined the association of pneumoconiosis and cor pulmonale in Ontario using hospital discharge data available from the Hospital Medical Records Institute (HMRI, now the Canadian Institute for Health Information) and the Ontario Ministry of Health (3). In the follow-up report (2), we attempted to confirm the validity of the coding of the diagnoses of a subset of the hospital discharges from the original study, and to identify work exposure (occupation and industry) information available in hospital records. We wrote to hospitals providing abstraction forms to be completed for 521 subjects hospitalized for pneumoconiosis, cor pulmonale or both requesting information regarding diagnoses, occupation and industry data. We received completed abstraction forms from 151 (76%) of the hospitals contacted, representing 720 (76%) of the 944 discharges and 421 (81%) of the 521 patients. The HMRI diagnoses were confirmed for 97% of those with silicosis and for 96% of those with asbestosis, and there was very good agreement between the two sources for the presence or absence of these conditions (kappa of 0.84 and 0.86, respectively). We also examined occupation and industry information available in the charts for which silicosis was a diagnosis. Of the 242 charts with silicosis, 122 had occupation/ industry information that could be classified regarding exposure to a specific dust type, and of these 89 (73%) were classified as consistent with silica exposure.

Furthermore, as a check against a second available indicator, of 34 individuals in this data set known from the Ontario Ministry of Labour’s Chest Clinic x-ray surveillance program of miners to have silicosis, 33 (97%) were diagnosed by the hospitals as having pneumoconiosis, and all but two were silicosis. These findings indicated that at least for pneumocioses (conditions that are pathognomonic of occupation), we could confirm the diagnoses and the hospital records frequently contained information about the responsible exposures.

Gary M Liss, MD, MS, FRCPC, Robert A. Kusiak, MSc, Ontario Ministry of Labour, Toronto, ON, Canada

References

1. Alamgir H, Koehorn M, Ostry A, Tompa E, Demers P. An evaluation of hospital discharge records as a tool for serious work related injury surveillance. Occup Environ Med 2006; 63:290-296.

2. Liss GM, Kusiak RA, Gailitis MM. Hospital records: an underutilized source of information regarding occupational diseases and exposures. Am J Ind Med 1997; 31:100-106.

3. Kusiak RA, Liss GM, Gailitis MM. Cor pulmonale and pneumoconiotic lung disease: An investigation using hospital discharge data. Am J Ind Med 1993; 24:161-173.