View source for Changing ambient UVR and future skin cancer in London, Rome and Helsinki: non-melanoma skin cancer (BCC and SCC)

{{assessment}}
:''This case study was carried out as part (work package 3.7) of the EU-funded INTARESE project. The contents were obtained from [http://www.integrated-assessment.eu/content/changing_ambient_uvr_and_future_skin_cancer_london_rome_and_helsinki_non_melanoma_skin_cance IEHIAS toolbox].

It is widely accepted that ambient UVR is carcinogenic (Lucas et al 2006). This assessment will focus on the impact on non-melanoma skin cancer: basal cell carcinoma (BCC) and squamous cell carcinoma (SCC).

The international mechanism for protecting the ozone layer is the “Montreal Protocol on Substances That Deplete the Ozone Layer” that came into force in 1989 and its subsequent amendments (UNEP 2006). The production of the most harmful ozone depleting substances (ODS) are now phased out worldwide. The atmospheric concentrations of these gases is anticipated to decline over the next decades (IPCC 2005). As a result stratospheric ozone is expected to increase and, subsequently, ambient UVR is anticipated to decrease. Hence, a decrease in non-melanoma skin cancer is expected, but other factors may play an important role as well such as the aging of the population.



==Scope==

'''Description'''

We explore the effects of changing exposure to ambient ultraviolet radiation (UVR) on non-melanoma skin cancer (BCC and SCC), accounting for the recovery of the ozone layer due to decreasing emission of ODS and future emissions of CH4 and N20 (IPCC 2005). We also account for future demographic change.

[[File:Summary_report_fig_1_0.jpg|thumb|400px|Causal diagram of the case study.]]


'''Scenario(s) and Type of Assessment''' 

Type of assessment: prognostic.

UVR scenarios: IPCC (2000) SRES A2 and B1: future population & emissions/concentrations of ODS, CH4 and N2O.

SRES population scenarios were downscaled from the regional to the city-level.


{|{{prettytable}}
!IPCC Senario
!SRES-A2
!SRES-B1
|-----
|     ODS Emission
|     Decreasing trend 
Same B1
|     Decreasing trend 
same as A2 
|-----
|     CH<sub>4</sub> Emission
|     Higher Increase 
|     Lower Increase 
|-----
|     N<sub>2</sub>O Emission 
|     Higher Increase 
|     Lower Increase 
|-----
|     Population
|     Higher Growth
|     Lower growth
|-----
|     Age Structure
|     Slower Aging
|     Higher Aging 
|}


'''Geographical and temporal scope''' 

*Study area(s): City of Helsinki, Greater London, City of Rome
*Populations: 5-year age groups up to 85+; males/females
*Timeframe: 2001, 2030, 2050


'''Environmental and health factors''' 

*Source: emissions of ODS (decreasing), N2O (increasing) and CH4 (increasing) ,affecting ozone fields
*Environmental hazard: ambient UVR (erythemal dose).
*Other risk factor: age
*Health outcomes: BCC and SCC incidence/mortality (rates)


'''Stakeholders:''' 


{|{{prettytable}}
! '''Stakeholder'''
! '''Interest'''
! '''Role'''
|-----
| Health authorities, policy/decision makers and health promotion agencies
| Healthy population
| Issue recommendations on solar exposure
|-----
| Health care providers
| Disease prevention, treatment
| Treatment and advice particularly to groups at risk
|-----
| Societies (cancer, osteoporosis, etc)
| Support member interests
| Issue public information
|-----
| Advocacy groups
| Protecting vulnerable groups
| Reduce health impacts related to UVR.
|-----
| Researchers
| Research
| Further knowledge of UVR-related health effects

|}

==Results==


'''Main findings''' 


'''UVR exposure''' UVR exposures are higher (and ozone recovery slower) in SRES-B1 compared to SRES-A2. The larger increase in ozone column in SRES-A2 can probably be explained by the higher abundance of CH4in this scenario.


'''Health impact model 1''' only accounts for future change in population size and structure, without accounting for future changes in incidence and mortality rates. The growing and ageing population results in a future increase in skin cancer incidence/mortality in both scenarios. As population growth is highest in SRES-A2, the total number of cases and deaths is also higher in this scenario compared to SRES-B1. For Helsinki and London, skin cancer incidence and mortality rates in the total population, however, are higher in the SRES-B1 scenario compared to the results for SRES-A2; this can be explained by the faster ageing of the population in SRES-B1.


'''Health impact model 2''' also accounts for the future recovery of the ozone layer. The modelling shows a decrease in age-specific skin cancer incidence rates due to the recovery of the ozone layer. As expected, skin cancer incidence/mortality (rates) are lower in model 2 compared to model 1. The SRES-B1 scenario (i.e. the alternative with the lower recovery of the ozone layer) has somewhat higher UVR levels and, consequently, higher age-specific incidence rates than SRES-A2.

Similar conclusions for model 1 and model 2 can be drawn for the DALY calculations.


The '''sensitivity analyses''' shows that the choices regarding the PAF (BCC high estimate is 0.9; low estimate is 0.5/SCC high estimate is 0.7; low estimate is 0.5 ) had a large influence on the health outcomes as expected. Hence for policy purposes, it might be recommended to present all results using both the high and low PAF estimates. Applying age weights and discounting simultaneously to the DALY calculations decreased the number of DALY’s with more circa 50% for BCC and circa 56% for SCC.  The choice regarding the imr showed a very large effect on the outcomes for SCC (almost 240% increase in DALYs compared to standard setting), but not so much for BCC (circa 12%). The sensitivity analyses also shows that the additive effect of an 1°C increase in summer temperature (Van der Leun et al 2008) in 2030 (compared to baseline) could counterbalance (results for BCC) or even outweigh (results for SCC) the effect of the decreasing UVR levels on age-specific incidence and mortality rates.


==Assessment Methods==

'''Exposure assessment''' 


Modelled current and future exposure:

Based on modelled future percental change in exposure (SRES-A2 and SRES-B1 storylines and associated emission modelling). UVR exposures were calculated by applying ozone fields (estimated from SRES data on atmospheric concentrations of ODS, N2O and CH4) from the Oslo Chemical Transport Model (Rummukainen et al., 1999) complemented by atmospheric fields from the European Centre for Midrange Weather Forecasting (ECMWF). These model outcomes have been made available to INTARESE by Bjorg Rognerud, University of Oslo. These ozone fields were translated to ambient UVR assuming otherwise current atmospheric conditions.

All simulations are done using the accurate libRadtran software (www.libradtran.org).


'''Health effect assessment''' 

Exposure-response function (ERF):

Changes in ambient UVR (annual erythemal dose) are combined with exposure-response functions (ERF) based on Lucas et al (2006). Lucas et al (2006) used ambient erythemally weighted UVR as the “exposure” to develop exposure-disease relationships for SCC and BCC in the following steps: 1) spreadsheets were developed (Microsoft Excel) to record data on incidence rates by sex and age group (recording all data by geographical position of the study region and year of publication for studies from 1979 to 2003; and 2) using these data, population-level exposure-response curves (annual ambient erythemal UVR vs. incidence rate) were constructed for each WHO age group, for lightly pigmented populations. SCC/BCC DRFs are available for males and females and for the following age groups: 0-4, 5-14, 15-29, 30-44, 45-59, 60-69, 70-79, 80+.

We received these unpublished DRFs (from R. Lucas), but cannot distribute them freely within INTARESE. These DRFs are not linear.



Health effect modelling:

*In the baseline model baseline incidence rates and baseline population estimates are applied.
*In model 1, baseline incidence rates and future population estimates are applied.
*In model 2, future incidence rates are calculated by applying the BAF and the future percental change in modelled ambient UVR to the baseline incidence rates. Results are combined with future population estimates.

Mortality (rates) (baseline/future) are calculated using the age-specific incidence:mortality ratios (imr), based on the incidence and morality rates in the EURA region as presented by Lucas et al (2006).


Attributable incidence and mortality are calculated using the upper estimate of population attributable fractions (PAF for BCC=0.9; PAF for SCC=0.7) as estimated by the WHO (Lucas et al. 2006).


UVR: Model Runs

{|{{prettytable}}
! Health Model
! Demographic Change
! Decreasing Ambient UVR, due to Future Stratospheric Ozone Recovery
|-----
| Baseline
| -
| -<sup>*</sup>
|-----
| Model<sup>*</sup> <sup>*</sup>
| X
| -<sup>*</sup>
|-----
| Model<sup>*</sup> <sup>*</sup>
| X
| X
|}


<sup>*</sup> using baseline incidence rates per age group and gender.

<sup>*</sup> <sup>*</sup> Selected scenarios: SRES-B1 2030, SRES-B1 2050, SRES-A2 2030, SRES-A2 2050.


'''Impact Assessment''' 

Disability Adjusted Life Years (DALY’s)

Our DALY worksheets/spreadsheet builds on the following templates/studies:

*Disease models derived from the 2006 WHO study on UVR attributable disease burden by Lucas:[[http://www.who.int/uv/health/solaruvradfull_180706.pdf]]

*WHO standard life tables:e (a) (life expectancy at age a), West level 26: [[http://www.who.int/healthinfo/bodreferencestandardlifetable.xls]]

*The WHO Global Burden of Disease Study  (BoD) DALY calculations template worksheet (excel):[[http://www.who.int/healthinfo/bodreferencedalycalculationtemplate.xls]]


'''Uncertainties''' 


Main assumptions

*Recovery of the ozone layer is influenced by two factors: 1) future changes in the emissions of ozone depleting substances (ODS) and 2) future atmospheric levels of relevant GHGs (CH4, N2O).
*The erythemal dose is a proper proxy exposure indicator for skin cancer effects.
*Everybody will experience the same decreasing UVR exposure.
*No changes in exposure behaviour ormedical treatment.
*If data is only available for larger age groups, an even distribution over the corresponding 5-year age groups is assumed.
*Several assumptions were made in the downscaling of the SRES-population scenarios.
*ERF and PAF for lightly pigmented populations apply to the whole population.
*We did not account for lag-time between exposure and health effect.
*Several assumptions were made in estimating baseline incidence rates.
*Incidence: mortality ratios based on Lucas et al.(Lucas, McMichael et al. 2006)are assumed to be representative for  our cases as well; no future change in these incidence:mortality ratios.
*Standard DALY calculation are based on the WHO standard life table, uniform age weighing, no discounting and no adjustments for co-morbidity. Selected disease models apply to the case-study population.


'''UVR scenarios''' In order to deal with uncertainties in future emissions and demographic trends, we use two different SRES scenarios from the IPCC: SRES-A2 and SRES-B1.


'''UVR sensitivity analyses''' Standard model setting versus alternative model setting for ambient UVR, cloud cover, PAF, incidence: mortality ratios, additive effect of increasing summer temperature, and DALY calculations (age weights, discounting; adjustment for co-morbidity). The sensitivity analysis is applied to model 2 for Helsinki, SRES B1-2030.

==Appraisal==

'''Implications''' 

*Decreasing UVR levels result in lower age-specific incidence and mortality rates (model 1 versus model 2).
*For the population as a whole, this positive effect of decreasing ambient UVR is offset by increasing population size and further ageing of the population. Hence, policies to prevent or treat non-melanoma skin cancer will still be relevant in the future, despite the recovery of the ozone layer.
*Based on the sensitivity analyses for Helsinki 2030 B1 model 2, it can be concluded that the positive effect of the recovery of the ozone layer could be counterbalanced or even outweighed by the possible additive effect of 1°C increase in summer temperature between 2000 and 2030,which is not very unlikely given the IPCC climate change scenarios (the national projections for Finland by Jylhä et al. (2004) indicated a temperature rise of 1-3˚C in 2010-2039 compared to 1960-1990).


'''Lessons learned''' 

*The inclusive full-chain model stimulated discussion about the risk assessment context. The exclusive framework stimulated discussion about excluded system factors.
*You often have to work with what is available.  This facilitated discussion on choices/assumptions made, improving the (transparency of) the assessments. The sensitivity analyses explored the impact of varying relevant input assumptions, in order to determine how ‘sensitive’ a model is to changes in the value of the parameters of the model. Further discussion with stakeholder might add to this as well, but will be very time-consuming.
*One of the biggest challenges was the downscaling of the SRES-population scenarios  from the OECD region to the city level (downscaling).
*There is no one-size fits all integrated assessment approach. General principles, methods and frameworks have to be made meaningful with the specific context of the case study. This requires: creativity, transparency, open attitude, and learning by doing.
*Our inventory of important literature, supporting materials and other useful (data) sources might be useful in future assessments.





==See Also==

Case study summary reports

#[[Changing ambient UVR and future skin cancer in London, Rome and Helsinki: melanoma skin cancer (CMM)]]
#[[Changing ambient UVR and future skin cancer in London, Rome and Helsinki: non-melanoma skin cancer BCC and SCC]]


Worked examples - complete assessments

#[[Changing ambient UVR and future melanoma and non-melanoma skin cancer in London, Rome and Helsinki]]



Worked examples of assessment techniques

#[[Scoping the assessment: ultraviolet radiation (UVR) and future skin cancer]]
#[[Creating scenarios: an example for UVR and skin cancer]]
#[[Estimating health effects: UVR and skin cancer]]
#[[Downscaling IPCC SRES-population scenarios from OECD- to the city-level: London, Rome, Helsinki]]

==References==

<references/>

==Related files==

{{mfiles}}

{{eracedu}}