Funding sources This research was supported by the Swiss National Accident Insurance Organization (SUVA) and the French Agency for Environmental and Occupational Health Safety (Afsset).
Estimating the contribution of occupational solar ultraviolet exposure to skin cancer†
Article first published online: 13 JAN 2014
© 2013 British Association of Dermatologists
British Journal of Dermatology
Volume 170, Issue 1, pages 157–164, January 2014
How to Cite
Milon, A., Bulliard, J.-L., Vuilleumier, L., Danuser, B. and Vernez, D. (2014), Estimating the contribution of occupational solar ultraviolet exposure to skin cancer. British Journal of Dermatology, 170: 157–164. doi: 10.1111/bjd.12604
Conflicts of interest None declared.
Plain language summary available online
- Issue published online: 13 JAN 2014
- Article first published online: 13 JAN 2014
- Accepted manuscript online: 26 AUG 2013 03:10AM EST
- Manuscript Accepted: 21 AUG 2013
- Swiss National Accident Insurance Organization
- French Agency for Environmental and Occupational Health Safety
Exposure to solar ultraviolet (UV) radiation is the main causative factor for skin cancer. Outdoor workers are at particular risk because they spend long working hours outside, may have little shade available and are bound to take their lunch at their workplace. Despite epidemiological evidence of a doubling in risk of squamous cell carcinoma (SCC) in outdoor workers, the recognition of skin cancer as an occupational disease remains scarce.
To assess occupational solar UV doses and their contribution to skin cancer risk.
A numerical model (SimUVEx) was used to assess occupational and lunch break UV exposure, and to characterize exposure patterns and anatomical distribution. Risk of SCC was estimated from an existing epidemiological model.
Horizontal body locations received 2·0–2·5 times more UV than vertical locations. The dose associated with having lunch outdoors every day was similar to that from doing outdoor work 1 day per week, but only half that of a seasonal worker. Outdoor work is associated with an increased risk of SCC and also with frequent acute episodes.
Occupational solar exposure contributes greatly to overall lifetime UV dose, resulting in an excess risk of SCC. The magnitude of the estimated excess in risk supports the recognition of SCC as an occupational disease.
Exposure to solar ultraviolet (UV) radiation is the main causative factor for skin cancer. Different patterns of exposure are associated with different types of skin cancer. While cutaneous malignant melanoma (MM) and basal cell carcinoma (BCC) are predominantly related to intermittent and acute UV exposure, and often occur on occasionally exposed anatomical sites, squamous cell carcinoma (SCC) results mainly from chronic exposure, such as the exposure of outdoor workers, and prevails on unprotected body sites.[2, 3]
In Europe, between 29% and 51% of subjects exposed to hazardous substances at the workplace are exposed to UV radiation. This makes solar radiation the most frequent occupational carcinogenic agent in many countries, particularly in regions where agriculture and fishing industries are important, and it affects, according to the CAREX (CARcinogen EXposure) database, about 9 million European workers. Risk of solar UV overexposure during work is increased by the fact that activities usually take place regardless of the ambient radiation, and repeated tasks performed in the same posture favour chronic overexposure of specific anatomical sites (e.g. neck overexposure for a vineyard worker bent over the vine several hours per day). It is estimated that outdoor workers receive annually about 10% (7–18%) of the available ambient UV radiation, which represents two to nine times the UV dose that indoor workers get. Depending on the job, the location and the season, much higher fractions of ambient UV radiation received have been documented in various occupational populations.[8-10] In Denmark, exposure of gardeners exceeded 10 standard erythemal doses (SED) for 160 days per year, and in Australia, exposure of the hands of mail delivery personnel exceeded 15 SED for 6 months. These doses are well above the 0·3 SED recommended by the International Commission on Non-Ionizing Radiation Protection.
Despite compelling epidemiological evidence and the fact that many outdoor occupational activities lead to exposure to UV radiation above the recommended exposure limits, sun-induced skin cancer is not officially listed as an occupational disease, either by the International Labour Office or by national authorities. This situation leads to a large underestimation of the burden of SCC as an occupational disease. For instance, in 2008, skin cancers represented 1% of all occupationally recognized cancers in Europe. In Switzerland, where no fewer than 4500 SCCs are diagnosed yearly, only three cases of SCC were recognized as occupational disease between 2000 and 2007.
A recent meta-analysis estimated a nearly twofold relative risk (RR) of SCC (RR 1·8, 95% confidence interval 1·4–2·2) for outdoor workers compared with indoor workers after adjusting for other SCC risk factors. In most countries this doubling in risk meets the threshold for the recognition of an occupational disease, even though occupational disease recognition laws differ across countries. Considering a hypothetical increase of 2·5% in SCC risk for every percentage increase in UV dose received, doubling the risk corresponds to a 40% (100/2·5) increase in total UV dose received. Thus, an occupational exposure leading to a 40% increase in UV dose could arguably result in a twofold risk of SCC for an outdoor worker compared with an indoor worker.
The contribution of occupational exposure to the total cumulative dose is central for the recognition of SCC as an occupational disease. This is more complex for MM and BCC, for which the occupational risk is often confounded by concurrent recreational exposure to UV.
Exposure to UV radiation at the workplace is a health concern where a better understanding of the determinants is needed to improve means of UV protection. Exposure data for different outdoor occupational groups are needed in order to perform risk analysis. As suggested by Kütting and Drexler, accurate models for determining occupational exposure may also help workers to receive compensation by the statutory accident insurance.
This study aims to assess the contribution of occupational UV exposure, in terms of annual (chronic) anatomical dose and acute events, to the risk of SCC for various occupational scenarios. A validated simulation tool was used to obtain daily and yearly UV doses.
Materials and methods
Exposure model and data source
The SimUVEx model has been used to assess solar UV. SimUVEx uses three-dimensional computer graphics techniques to compute anatomical UV doses on the basis of postural information and ambient UV (erythemally weighted) measurements. The model takes into account body surface inclination and orientation to the sun, as well as the shading from other body parts. SimUVEx results can be expressed either numerically, as anatomical doses (J m−2) over 45 body sites, or graphically, using a coloured virtual manikin (false-colour scale). The principles and a validation of the SimUVEx model in field conditions have been detailed elsewhere.
Five input parameters are required by the model: direct irradiance (W m−2), diffuse irradiance (W m−2), ground reflected irradiance (W m−2) and sun position [defined by its azimuth p(t) and zenith d(t) angles]. Ambient UV irradiance data were obtained from the MeteoSwiss Payerne station (46·8°N, 6·9°E, altitude 491 m), which is part of the Baseline Surface Radiation Network of the World Meteorological Organization, World Climate Research Program. Ambient UV measurements are performed every minute at this facility using broadband UV radiometers (Biometer 501A, Solar Light Company Inc., Glenside, PA, U.S.A.) with filters mimicking the erythemal response (erythemally weighted irradiance). These broadband radiometers undergo strict quality assurance procedures, including regular calibrations traceable to the European Ultraviolet Calibration Center, and the overall uncertainty of the measurement is estimated at 10%.
Ground irradiance data collected for the entire year 2012 were used in this study (5 25 600 measurements). Due to maintenance of the measuring device, 0·9%, 1·4% and < 0·1% missing/aberrant values were recalculated for direct, diffuse and reflected measurements, respectively, using ground global irradiance. Data were treated and analysed using Stata/IC 12.0 (StataCorp LP, College Station, TX, U.S.A.).
A manikin was placed standing up, arms down and rotating on itself at 24° min−1 (complete revolution time of 15 min). This posture predominates among construction workers, for example, where activities typically require frequent movements without a specific orientation. The computed daily doses assumed unprotected skin, so that the results indicate upper dose estimates.
Two types of doses were computed for each day of the year 2012, leading to 732 simulation runs: (i) workday doses, defined as 08·00–12·00 h and 13·00–17·00 h exposure, and (ii) lunch breaks doses, defined as exposure between 12·00 h and 13·00 h.
The dataset generated during simulations was used to establish cumulative yearly doses corresponding to exposure situations for different body locations. For convenience, semiquantitative exposure categories were used (high, moderate and low) to describe occupational and lunch exposure.
As shown in Table 1, the exposure bands were defined in terms of number of days per week or month spent outdoors. These exposure bands cover a variety of occupational activities, from individuals working mostly indoors (e.g. office workers taking their lunch outside) to full-time outdoor workers (e.g. construction work, farming). A 1 day per week exposure has been considered for intermediate situations (moderate occupational exposure). A typical example of moderate exposure is a ‘white-collar’ worker in the construction sector supervising outdoor work on a regular basis. A seasonal scenario has also been considered, and is defined as the same amount of time spent outdoors as a moderate exposure scenario (416 h per year for occupational activity and 52 h per year for lunch), but concentrated during the summer holiday period (from 6 July to 26 August).
|Activity||Exposure period||Exposure intensity||Exposure frequency||Hours per year|
|Occupational||08·00–12·00 h + 13·00–17·00 h||High||5 days per week||2088|
|Moderate||1 day per week||416|
|Low||1 day per month||96|
|Lunch||12·00–13·00 h||High||5 days per working week||261|
|Moderate||1 day per working week||52|
|Low||1 day per working month||12|
|Occupational||08·00–12·00 h + 13·00–17·00 h||Seasonal||6 July 2012 to 26 August 2012||416|
|Lunch||12·00–13·00 h||High||5 days per working week||52|
|Moderate||1 day per working week||8|
|Low||1 day per working month||2|
Distributions of daily doses and cumulative yearly doses were computed for various anatomical parts. The cumulative yearly doses were expressed either in SED or exposure ratios (ERs) corresponding to the percentage of ambient radiation received. The latter indicator is of interest as it is independent of the geographical location where the measurement is performed. The frequency of acute events, expressed as the number of days per year for which a dose threshold was exceeded, was also computed.
Estimating squamous cell carcinoma risk
The model, first proposed by Schothorst et al. and Slaper et al., was used to estimate the influence of occupational outdoor activity and lunch break on the risk of SCC. In this model, the risk of SCC is expressed as a function of age and cumulative UV dose:
where α and β are the respective age-dependent and biological amplification factor constants and UVtot is the cumulative UV dose received from occupational exposure, lunch exposure and recreational exposure (integrating residential exposure and exposure during recreational activities).
From this equation, we can express the RR of SCC for two people of the same age, exposed to two different cumulative UV doses (A and B, respectively) as:
The RR of SCC was estimated at age 60 years with an outdoor work history of 25 years (cumulative UV dose B), the reference being an indoor worker who took their lunch indoors (cumulative UV dose A). A value of 2·5 was considered for the biological amplification factor.
The cumulative UV dose at age T = 60 years is expressed as the sum of exposures during work and lunch during y years of occupational activity and recreational time from 0 to T:
UVocc and UVlunch were obtained with SimUVEx, while UVrecr was based on a behavioural model applied to time–microenvironment–activity–diary data collected in a large European study (Expolis).[26, 27] An annual dose of 297 SED for the face was used as a reference for recreational exposure. This value corresponds to the recreational dose for office workers in Milan, the surveyed Expolis location where annual ground irradiance was the closest to Payerne (6941 SED and 6296 SED, respectively). The sensitivity of the model to the recreational dose was assessed by varying the recreational dose (297 SED) by 10% either way (267·3–326·7 SED).
The exposure situations and categories of exposure frequencies considered are summarized in Table 1. As the annual recreational UV dose has been shown to be independent of age, the recreational dose was considered to be the same during the 60 years. The occupational UV dose was considered to be constant during the 25 years of work according to the chosen scenario.
The annual UV doses and ER for the face, back of the neck, shoulders, top of shoulders and forearm related to various occupational exposure scenarios are given in Table 2. These anatomical sites (Fig. 1) were of special interest because they are often left uncovered, have various orientations and have been associated with BCC and SCC (at least for the face and neck). Computed doses provided upper boundary limits, as unprotected skin was assumed for the whole year.
|Activity||Exposure intensity||Yearly dose, SED (ER, %)|
|Face||Forearm (right)||Shoulder (right)||Back of neck||Top of shoulders|
|Occupational||High||1604 (25·5)||1696 (26·9)||2219 (35·2)||2373 (37·7)||3438 (54·6)|
|Moderate||311·0 (4·9)||329·3 (5·2)||427·3 (6·8)||457·8 (7·3)||655·0 (10·4)|
|Low||76·2 (1·2)||80·6 (1·3)||103·8 (1·6)||111·1 (1·8)||157·3 (2·5)|
|Lunch||High||311 (4·9)||327·3 (5·2)||433·7 (6·9)||464·8 (7·4)||683·8 (10·9)|
|Moderate||62·1 (1·0)||65·3 (1·0)||85·5 (1·4)||91·5 (1·5)||132·8 (2·1)|
|Low||14·9 (0·2)||15·8 (0·3)||20·2 (0·3)||21·6 (0·3)||30·3 (0·5)|
|Occupational||Seasonal||578·1 (9·2)||612·2 (9·7)||824·0 (13·1)||880·3 (14·0)||1323 (21·0)|
|Lunch||High||99·8 (1·6)||105·4 (1·7)||145·5 (2·3)||156·6 (2·5)||240·6 (3·8)|
|Moderate||14·0 (0·2)||14·7 (0·2)||20·5 (0·3)||22·3 (0·4)||34·8 (0·6)|
|Low||2·0 (0·05)||2·0 (0·05)||2·8 (0·07)||3·2 (0·1)||4·8 (0·1)|
The facial exposure of full-time outdoor workers (lunch excluded) was 1604 SED, corresponding to a mean of 5·8 SED per working day. Higher doses were observed in anatomical sites more horizontally oriented, such as the shoulders (2219 SED) the neck (2373 SED) or the top of the shoulders (3438 SED). On average, the most horizontally oriented site considered (top of shoulders) received consistently 2·0–2·1 times more of the ambient radiation than the most vertically oriented site (face) during work-time periods. During the lunch period (12·00–13·00 h), when the sun is at its highest zenithal angle, this ratio only marginally increases, with the top of the shoulders receiving 2·0–2·2 times more ambient radiation than the face.
As expected, seasonal workers were more exposed than year-round workers due to the elevated ambient radiation in summer. Anatomical doses in seasonal workers were about twice the anatomical doses endured by year-round workers exposed for the same cumulative duration (corresponding to moderate occupational activity). The difference in exposure between horizontally and vertically oriented body locations increased with decreasing zenithal angle. The top of the shoulder received 2·3–2·5 times more ambient radiation than the face for seasonal activity.
Daily doses (08·00–12·00 h and 13·00–17·00 h) were used in this study as an indicator of acute exposure. Daily doses above the minimum erythemal dose for an average skin phototype were considered as overexposures. Typically, the minimum erythemal doses for skin types II and III, the most common phototypes in white populations, are 2·5 and 3·0 SED, respectively. The monthly distribution of daily doses for different body locations is presented in Fig. 2. Similarly to ambient irradiation, personal exposure varies substantially within a year, exhibiting a typical bell-shaped curve. The profile was pronounced for horizontally oriented anatomical sites: the top of the shoulders received daily doses up to 37 SED. A strong attenuation was observed for vertically oriented locations, such as the face, which exhibited a maximum of 15 SED.
The ratio in annual doses previously described between horizontally and vertically oriented anatomical sites appears to be time dependent. The shoulders were about 1·7 times more exposed than the face in the middle of winter (December/January) and about 2·3 times more exposed in early summer (June/July). Exposure variability appears also to be greater for horizontally oriented anatomical sites.
Overall, predicted daily doses for unprotected skin were found to be high for all anatomical sites and, except in winter, above the minimal erythemal doses for skin types II and III.
Episodes of acute exposure were examined more closely by computing the number of days per year for which a given dose could be exceeded (Fig. 3).
For both the face and the top of the shoulders, the daily dose exceeded the minimal erythemal doses for skin types II and III more than 200 days per year. Opportunities for overexposure above 10 SED occurred 80 days per year for the face and around 190 days per year for the top of the shoulders. No opportunity for overexposure above 20 SED was observed for the face. The top of the shoulders potentially received above 20 SED on about 90 days per year.
Relative risk of squamous cell carcinoma
The estimated RR of SCC for the face is presented in Table 3. The RR compares outdoor workers with various exposure patterns to an individual with the same recreational activities (thus the same recreational UV dose) but with no occupationally related exposure. The reference value (RR 1·0, bottom right of Table 3) corresponds to a nonexposed individual, such as an indoor worker, who does not take his or her lunch outside. The sensitivity analysis indicated a maximum change of 22% in RR for 10% variation in recreational dose around the reference value of 297 SED.
|Outdoor lunch||Outdoor occupational activity, RR of SCC (+10 to −10%)|
|High||26·1 (22·0–31·7)||4·8 (4·3–6·9)||5·3 (4·7–6·1)||3·0 (2·7–3·3)||2·5 (2·3–2·7)|
|Moderate||20·3 (17·3–24·5)||2·9 (2·6–3·8)||4·5 (4·1–5·1)||1·6 (1·5–1·6)||1·2 (1·2–1·3)|
|Low||19·4 (16·4–23·3)||2·6 (2·4–3·3)||4·4 (4·0–5·0)||1·4 (1·3–1·4)||1·1 (1·0–1·1)|
|None||19·0 (16·2–22·9)||2·5 (2·3–3·1)||4·4 (4·0–5·0)||1·3 (1·3–1·3)||1·0 (1·0–1·0)|
Because the exposure model assumes long working hours outdoors, no shade, no protective clothing and no tanning effect, RR estimates should be considered as upper values. Still, only a few scenarios lead to less than a twofold RR of SCC. In particular, having an activity with high or moderate occupational exposure or a summer outdoor occupation over 25 years confers a substantially increased risk of SCC, regardless of exposure during the lunch break. Apart from seasonal workers, frequent outdoor lunches (one to five times per week over a whole year) contribute significantly to the increase in SCC risk, which is consistent with prevention messages advocating avoidance of peak sun exposure.
When unprotected against solar radiation, outdoor workers could be exposed to high UV doses, up to 25·5–54·6% of the annual ambient irradiance depending on the anatomical site. In practice, this potential exposure is mitigated by occasional shade (e.g. surrounding trees or buildings, indoor working periods) and protection from clothing or absence from worksite (e.g. travelling, sick leave), factors that were not accounted for in our model. Comparing results across studies is difficult due to discrepancies in study design, geographical location, body sites and dose types. Annual doses previously reported for regular outdoor work are between 54 and 669 SED on the wrists of gardeners in Denmark, around 1097 SED on the sides of the hands of mountain guides in the Alps, and 1920 SED on the chest, 3880 SED on the hands and 2830 SED on the back of postmen in Australia (ambient = 15 100 SED). These observations suggest that a realistic scenario would be between the high and moderate outdoor exposure situations considered in Table 1.
For moderate and high occupational exposure scenarios, the elevated UV doses emphasize the need for targeted prevention among outdoor workers. Moreover, lunchtime exposure was estimated separately from work-time exposure. Whether lunchtime exposure should be considered as work related or not depends on the context. An office worker may choose to have their lunch break on an open terrace, while a building worker on a remote construction site may have no other choice but to take their lunch break onsite, where shading is not necessarily available. The former exposure is not intrinsically occupationally related, whereas the latter exposure is work related and should be added to the work-time exposure when estimating a risk. Our estimates suggest that moderate and high lunch exposure intensities (at least once a week), which amounted to 1·0–2·1% and 4·9–10·9% of annual ambient irradiance, respectively, contribute significantly to the annual dose. The UV yearly dose of workers taking their lunch outdoors on every workday was comparable with the exposure dose of someone working outdoors 1 day per week. These results support the importance of seeking shade during ‘peak’ periods (typically 11·00–15·00 h), as advocated in public prevention messages. However, such a recommendation seems hardly practicable at some work sites.
With regards to anatomical variability, annual exposure of horizontally oriented body parts was found to be 2·0–2·5 times higher than that of vertically oriented body parts. This anatomical variability concurs with previous studies. According to Wright et al., vertical surfaces received 13–76% of the exposure of the vertex of the head, with an average of 38%, while horizontal surfaces received 47–82% with an average of 60%. Moreover, a ratio of 2 in average values received for vertical (38%) and horizontal (75%) surfaces was also found by Diffey. The fact that diffuse radiation is the major contributor to the annual exposure doses probably explains the limited anatomical variability found in this study. However, it should be pointed out that anatomical variability is often greater with shorter exposure (e.g. daily exposure) or short periods of exposure during the day (e.g. lunch period). Sun position and thus posture and orientation to the sun are indeed key factors in the anatomical distribution of exposure. On the one hand, vertically oriented body parts receive a larger amount of the direct normal irradiance when the sun is low (high zenithal angle), such as early or late during the day or early or late in the year; on the other hand, direct sunlight goes through a larger amount of the ozone layer when the sun is low, corresponding to higher UV absorption. The two effects may compensate for each other. For horizontal surfaces, the zenithal angle effect is reversed, but the atmospheric path effect of the radiation is still the same and the two effects add up.
Most mathematical models assessing cumulative doses use ambient exposure data (e.g. UV index) and ambient exposure fraction. This approach should be used carefully, particularly when estimating anatomical exposure. SimUVEx does take into account shaded body areas and can be used for different postures.
Considering acute events, most of the year (> 200 days) the daily dose endured by unprotected skin potentially exceeds the minimal erythemal doses for skin types II and III. Opportunities for severe sunburn are frequent for outdoor workers. In practice, sunburn occurs noticeably less frequently because the simplifying hypotheses of the model (long working time outdoors, without shading or protective clothing, no tanning attenuation effect) leads to upper exposure estimates. However, the high frequency of potential sunburns suggests that high chronic doses are also associated with a high number of opportunities for acute, intermittent overexposure. While sunburn (acute exposure) is a risk marker for MM and possibly BCC, chronic exposure has been associated with SCC and, to a lesser extent, BCC. Our results underline the somewhat arbitrary distinction made between chronic and intermittent exposure, a finding that questions the potential contribution of occupational exposure to MM.
The model used for estimating the risk of SCC, although widely applied and supported by epidemiological evidence, implicitly assumes constant annual exposure. Use of a more elaborate model that accounts for varied patterns of annual occupational exposure over time (such as that proposed by Slaper et al. for melanoma) reduced overall the magnitude of the RR of SCC, but did not significantly affect the gradients observed in RR with the frequency of outdoor lunch and outdoor occupational activity (data not shown). These observations further indicate that although our results represent in absolute terms an upper boundary of UV exposure doses, they seem to reflect fairly well the relative variations in risk according to occupational UV doses.
In most countries, doubling in risk meets the threshold for the recognition of an occupational disease. Indeed, all occupational scenarios, including seasonal work, led to an increased risk of SCC higher than twofold and higher than 40% of the occupational contribution to the annual dose, except when working outdoors only 1 day per month with a lunch break taken outdoors no more than 1 day per working week.
Again, a moderate exposure scenario is suggested to be the most common among outdoor workers, when compared with the SCC estimates obtained by Vishvakarman and Wong.
These results suggest that occupational exposure is a major contributor to SCC in outdoor workers and emphasize that (i) the recognition of SCC as an occupational disease for this population should be facilitated and systematized, and (ii) specific prevention messages towards chronic sun exposure at work are needed.
As expected, seasonal workers are particularly at risk, as ambient exposure peaks during the summer period. The behaviour of seasonal workers towards prevention strategies might be different from regular outdoor workers and could result in a potential increased risk of skin cancer. Prevention messages for outdoor workers should recognize these two types of working populations and include information about the consequences of acute and intermittent exposure during work periods.
This study shows the necessity of more accurate recreational exposure data, as well as an accurate estimation of the duration of occupational exposure. It also presents the use of an original tool to produce data that can help physicians in the recognition of skin cancers as an occupational disease. The approach used in this study could serve as a first estimate to evaluate the occupational contribution to the lifetime UV dose in the case of a claim for recognition of a skin cancer, replacing complicated tabulations. By taking into account exposure patterns during childhood, and recreational activities (in terms of corresponding UV doses), the relative part of the occupational cumulative dose can be estimated. A matrix is a convenient way to combine the different situations of exposure. It could support occupational physicians in estimating the contribution of work-related exposure and support decision making in occupational skin cancer recognition. It could also help to identify high-risk occupational activities, in order to implement prevention strategies to reduce SCC incidence in the population of occupational outdoor workers.
In conclusion, occupational activity is a major contributor to the overall UV doses in outdoor workers. The work-related doses calculated in this study clearly indicate an elevated risk of SCC for outdoor workers in high and moderate occupational exposure scenarios compared with indoor workers in high and moderate exposure scenarios. Additionally, the anatomical distribution of the dose indicates a nonuniform exposure of the body even when considering a simple posture. These results question the relevance of using the UV index as a prevention tool. Regular sun exposure also appears to favour increased episodes of acute exposure, raising concerns about the potential contribution of occupational UV exposure to other forms of skin cancers.
- 6Eurogip. Work-related cancers: what recognition in Europe? Available at: http://www.eurogip.fr/images/publications/EUROGIP_RecoCancerspro_49E.pdf (last accessed 15 October 2013).
- 12International Commission on Non-ionizing Radiation Protection (ICNIRP). Guidelines on limits of exposure to ultraviolet radiation of wavelengths between 180 nm and 400 nm (incoherent optical radiation). Health Phys 2004; 87:171–86.
- 13International Labour Office. List of occupational diseases (revised 2010). Identification and Recognition of Occupational Diseases: Criteria for Incorporating Diseases in the ILO List of Occupational Diseases, vol. 74. Geneva: ILO, 2010. Available at: . Available at: http://www.ilo.org/safework/info/publications/WCMS_150323/lang--en/index.htm (last accessed 29 October 2013).
- 14[Epidemiology of epithelial skin cancers]. Rev Med Suisse 2009; 5:882, 884–8 (in French)., , .
- 18Description of the use of a risk estimation model to assess the increased risk of non-melanoma skin cancer among outdoor workers in Central Queensland, Australia. Photodermatol Photoimmunol Photomed 2003; 19:81–8., .
- 21A reference action spectrum for ultra-violet induced erythema in human skin. In: Human Exposure to Ultraviolet Radiation: Risks and Regulations (International Congress) (Passchier WF, Bosnjakovich BFM, eds), Amsterdam: Elsevier Science Ltd, 1987, 83–7., .
- 31Anatomical distribution of ultraviolet solar radiation. South Afr J Sci 2004; 100:498–500., , .