Appendix 1. Technical appendix and data inputs Appendix 1. Technical appendix and data inputs About Public Health England Public Health England exists to protect and improve the nation’s health and wellbeing, and reduce health inequalities. We do this through world-leading science, knowledge and intelligence, advocacy, partnerships and the delivery of specialist public health services. We are an executive agency of the Department of Health and Social Care, and a distinct delivery organisation with operational autonomy. We provide government, local government, the NHS, Parliament, industry and the public with evidence-based professional, scientific and delivery expertise and support. Public Health England Wellington House 133-155 Waterloo Road London SE1 8UG Tel: 020 7654 8000 www.gov.uk/phe Twitter: @PHE_uk Facebook: www.facebook.com/PublicHealthEngland Prepared by: UK Health Forum © Crown copyright 2018 You may re-use this information (excluding logos) free of charge in any format or medium, under the terms of the Open Government Licence v3.0. To view this licence, visit OGL. Where we have identified any third party copyright information you will need to obtain permission from the copyright holders concerned. Published May 2018 PHE publications PHE supports the UN gateway number: 2017858 Sustainable Development Goals 2 Appendix 1. Technical appendix and data inputs Contents About Public Health England 2 Data inputs 4 UKHF microsimulation methodology 29 UKHF tool methodology 45 References 49 3 Appendix 1. Technical appendix and data inputs Data inputs 1 Population data Demographic data was collected for England and Wales combined (no separate data for England was available), and for each UK local authority separately. Information was collected on the age and sex distribution of the population, the distribution of births by mother’s age, the total fertility rate and the distribution of deaths by age and sex. The data were processed as text files, in a format suitable for inclusion in the microsimulation programme. The data sources were as follows Table 1. Population data sources by geography Demography Geography Source Total England ONS. Population Estimates for UK, England and Wales, population and Wales Scotland and Northern Ireland: mid-2015. ONS; 2016.(1) by age and Local ONS. Population Estimates for UK, England and Wales, sex authority Scotland and Northern Ireland: mid-2015. ONS; 2016.(1) Births by England ONS. Birth Summary Tables – England and Wales. mothers age and Wales 2015. ONS; 2016 (2) Local ONS. Births by mothers' usual area of residence in the authority UK. ONS; 2016.(3) Total fertility England ONS. Birth Summary Tables – England and Wales. rate and Wales 2015. ONS; 2016 (2) Local ONS. Births by mothers' usual area of residence in the authority UK. ONS; 2016.(3) Deaths by England ONS. Deaths registed in Engalnd and Wales 2015. ONS; age and sex and Wales 2016 (4) Local ONS. Mortality statistics - underlying cause, sex and authority age. ONS; 2016.(5) 2 Disease data A number of air pollution-related diseases were modelled (see Table 2). The list of diseases modelled for each pollutant was determined after a review of the literature available on the dose-response relationship between exposure to air pollutants and risk of incidence of disease (Table 2). Decisions were made by the advisory committee to include or exclude diseases based on existence of a dose-response relationship and availablility of epidemiological data. For instance, while an association exists between lung function in children and both NO2 and PM2.5, this condition is not a fixed disease 4 Appendix 1. Technical appendix and data inputs with identifiable prevalence, incidence and mortality data, so this condition was excluded from the disease list. Pre-term birth was also identified, but low-birth weight was considered a better proxy of the process of intrauterine growth reswtriction, and excluding pre-term births may avoid double counting. Table 2. Characteristics of diseases modelled for each pollutant Duration Terminal Age Pollutant category NO2 PM2.5 Respiratory outcomes Asthma (children) Chronic Yes Child X X Asthma (adults) Chronic Yes Adult X COPD Chronic Yes Adult X Cardiovascular outcomes CHD Chronic Yes Adult X Stroke Chronic Yes Adult X Diabetes Chronic No Adult X X Cancer and other outcomes Dementia Chronic Yes Adult X Low birth weight Acute No Adult X X Lung cancer Chronic Yes Adult X X All diseases except for low birth weight were lifelong, chronic diseases, so once acquired, were prevalent for the duration of an individual’s life (see section Module 2: Microsimulation model on modelling birth weight for further details). Individuals could develop more than one diseases, but these were considered independent of one another. All diseases apart from diabetes and low birth weight were terminal. Epidemiological data on each disease’s incidence, prevalence, mortality and survival and dose-response was collected (see Table 3). When a parameter, eg Survival was not available from the literature or national statistics, this was computed – see Module 2: Microsimulation model section Approximating missing disease statistics for methods. 2.1 Summary of data sources Table 3. Summary of disease data sources Diseases Incidence Prevalence Mortality Survival Relative Risk Asthma BLF BLF Asthma ONS, Deaths Computed NO : Khreis et al. 2016 (7) 2 Asthma Statistics (6) Registrations from In children =<6 years : OR Statistics Summary prevalence 1.08 (1.04; 1.12) per (6) Statistics, and mortality England and 4µg/m3Converted to Wales, 2015 (4) OR 1.212 (1.103; 1.328) per 10µg/m3REDUCED by 60%  1.08 (1.01; 1.12) per 10µg/m3 5 Appendix 1. Technical appendix and data inputs In children >6 years: OR 1.03 (1.00; 1.06) per 4µg/m3Converted to OR 1.08 (1.00; 1.16) per 10µg/m3 REDUCED by 60%  1.03 (1.00; 1.06) per 10µg/m3 Jaquemin et al. 2015 (8) In adults: OR 1.10 (0.99;1.21) per 10µg/m3 REDUCED by 60%  1.04 (0.996; 1.08) per 10µg/m3 PM : Khreis et al. 2016 (7) 2.5 In children >6 years: OR 1.04 (1.02; 1.07) per 1µg/m3 Converted OR 1.48 (1.22 ; 1.97) per 10µg/m3 COPD Computed PHE ONS, Deaths Computed PM : COMEAP 2016 (10) 2.5 from modelled Registrations from COMEAP recommend using PM 10 prevalence estimates, Summary prevalence estimate based on Cai et al. 2014 and 2008 (9) Statistics, and mortality estimate for chronic phlegm in never mortality. England and smokers in sensitivity analyses: Wales, 2015 (4) OR 1.32 (1.02; 1.71) per 10µg/m3 of PM  scale to PM using the 10 2.5 conversion factor of PM -> PM : 2.5 10 0.7 (or PM -> PM :1.42) recently 10 2.5 used in the air quality index, COMEAP: Converted to 1.49 (1.03; 2.14) per 10µg/m3 of PM 2.5 CHD Smolina et BHF, ONS, Deaths Computed PM : Cesaroni et al. 2014 (13) 2.5 al 2012. Cardiovasc Registrations from Estimate used in CAPTOR tool from Corrected ular Summary prevalence subgroup analysis of participants data on Disease Statistics, and mortality with additional information on CVD incidence Statistics England and risk factors: and 2014 (12) Wales, 2015 (4) HR 1.19 (1.01; 1.42) per 5µg/m3 mortality in Converted to 1.41 (1.00 - 2.01) per 2013 (11) 10µg/m3 Diabetes Personal National Non-terminal Non-terminal NO Eze et al. 2015 (15) 2: communica Diabetes RR 1.12 (1.05; 1.19) per tion with dr Audit 2015- 10µg/m3REDUCED by 60%  Craig Curry 2016(14) 1.05 (1.02; 1.07) per 10µg/m3 from Cardiff PM : Eze et al. 2015 (15) 2.5 University RR 1.10 (1.02; 1.18) per 10µg/m3 Stroke BHF, BHF, ONS, Deaths Computed PM : Scheers et al. 2015 (17) 2.5 stroke Cardiovasc Registrations from HR 1.064 (1.021; 1.109) per 5µg/m3 6 Appendix 1. Technical appendix and data inputs statistics ular Summary prevalence Converted to 1.13 (1.04; 1.23) per 2009 (16) Disease Statistics, and mortality 10µg/m3 Statistics England and 2014 (12) Wales, 2015 (4) Dementia Computed Dementia ONS, Deaths Computed NO : Oudin et al. 2016(19) 2 from UK 2014 Registrations from HR 1.08 (1.00; 1.16) per 10µg/m3 prevalence (18) Summary prevalence NOx. Scaling factor: NOx → NO2: and Statistics, and mortality 0.44 which was developed by mortality England and Anderson et al. based on the ratio Wales, 2015 (4) that fell midway between the average or roadside vs urban background monitoring sites in London for 2001 (see Online Supp 2) (20) Converted from NOx to NO2:HR 1.03 (1.00; 1.07) REDUCED by 60%  1.01 (1.01; 1.03) per 10µg/m3 of NO2 Low ONS Birth Considered Non-terminal Non-terminal NO : Pedersen et al. 2013 (22) 2 birth Characteris equivalent OR 1.09 (1.00; 1.19) per weight tics, 2015 to incidence 10µg/m3REDUCED by (21) 60%  1.04 (1.00; 1.07) per 10µg/m3 PM : Pedersen et al. 2013 (22) 2.5 OR 1.18 (1.06; 1.33) per 5µg/m3Converted OR 1.39 (1.12; 1.77) per 10µg/m3: Lung CRUK, Not required CRUK, 2012-14 1, 5 year: NO : Hamra et al. 2015 (26) 2 cancer 2012-14 in model as (23) ONS, 2010-14 RR 1.04 (1.01; 1.08) per (23) model uses (24); 10 year: 10µg/m3REDUCED by 60%  incidence ONS, 2008-12 1.02 (1.00; 1.03) per 10µg/m3 (25) PM : Hamra et al. 2014 (27) 2.5 RR 1.09 (1.04; 1.14) per 10µg/m3 All NO2 relative risks reduced by 60% following COMEAP recommendations(28) 7 Appendix 1. Technical appendix and data inputs 2.2 Incidence, Prevalence, Mortality data by disease Asthma Table 4. Asthma epidemiological data (per 100,000 population) Incidence Prevalence Mortality BLF Asthma Statistics(6) BLF Asthma Statistics(6) ONS 2015(4) Data sourced from the THIN Data sourced from the THIN ICD 10: J45-J46 Database, ICD codes unclear Database, ICD codes unclear Age group Both genders Age group Both genders Age group Male Female 0-5 929.0 0-5 3114.0 <1 0.0 0.0 1-4 0.1 0.0 6-10 561.0 6-10 10079.0 5-14 0.2 0.2 11-15 356.0 11-15 15899.0 16-20 170.0 16-20 20180.0 15-24 0.2 0.2 21-30 150.0 21-30 17306.0 25-34 0.2 0.2 31-40 180.0 31-40 13286.0 35-44 0.4 0.3 41-50 201.0 41-50 11751.0 45-54 0.5 1.0 51-60 204.0 51-60 10903.0 55-64 1.1 1.4 61-70 231.0 61-70 10848.0 65-74 1.5 2.9 71-80 194.0 71-80 11526.0 75-84 5.2 12.3 81+ 111.0 81+ 10135.0 85+ 32.3 56.0 8 Appendix 1. Technical appendix and data inputs Chronic obstructive pulmonary disease (COPD) COPD incidence was estimated from prevalence and mortality data, – see Module 2: Microsimulation model section Approximating missing disease statistics for methods. Table 5. COPD epidemiological data (per 100,000 population) Incidence Prevalence Mortality UKHF Derived estimates PHE Modelled estimates (2008) (9) ONS 2015(4) Computed from Prevalence and COPD based on FEV1 ICD 10: J40-J44 Mortality measurements in the Health Survey for England 2001 using British Thoracic Society criteria Age group Male Female Age group Male Female Age group Male Female 0-4 1.2 0.0 0-4 4.2 5.9 <1 0.0 0.0 1-4 0.0 0.0 5-9 2.6 3.4 5-9 10.4 6.0 5-9 0.0 0.0 10-14 2.1 1.6 10-14 23.2 23.1 15-19 0.0 0.0 15-19 33.7 31.3 15-24 0.0 0.0 20-24 1.8 1.3 20-24 27.6 12.9 25-29 3.4 3.4 25-29 36.8 19.6 25-34 0.1 0.1 30-34 7.9 9.8 30-34 54.0 36.5 35-39 25.0 28.2 35-39 93.3 85.5 35-44 1.0 0.5 40-44 52.0 60.4 40-44 218.0 226.3 45-49 106.0 121.2 45-49 477.1 527.0 45-54 6.4 4.4 50-54 205.2 174.5 50-54 1003.6 1128.5 55-59 298.0 235.4 55-59 2015.4 1988.1 55-64 32.0 28.2 60-64 491.9 326.9 60-64 3466.8 3136.4 65-69 437.5 216.8 65-69 5817.7 4709.4 65-74 126.6 101.4 70-74 334.8 228.0 70-74 7860.1 5737.9 75-79 263.4 0.0 75-79 9392.4 6807.6 75-84 367.7 283.2 80-84 0.0 0.0 80-84 10579.6 6783.3 85+ 0.0 0.0 85+ 9857.29 5300.6 85+ 939.45 558.4 9 Appendix 1. Technical appendix and data inputs Coronary heart disease (CHD) Table 6. CHD epidemiological data (per 100,000 population) Incidence Prevalence Mortality Smolina et al. 2012(11) BHF CVD Stats 2014(12) ONS 2015(4) ICD 10: I21-I22 ICD 10: I21 ICD 10: I21-I22 Age group Male Female Age group Male Female Age group Male Female 0-44 60.0 30.0 <1 0.0 0.3 1-4 0.0 0.0 5-9 0.0 0.0 15-24 0.1 0.0 25-34 0.8 0.2 30-54 88.1 21.2 35-44 4.9 1.4 45-54 1070.0 430.0 45-54 21.2 5.2 55-64 317.0 90.3 55-64 4510.0 1240.0 55-64 52.2 14.2 65-74 533.0 237.0 65-74 8660.0 2960.0 65-74 109.4 44.6 75-84 1017.0 597.0 75+ 14780.0 6960.0 75-84 281.0 146.0 85+ 1987.0 1395.0 85+ 692.0 454.7 10