key uncertainties in net impacts of disease burden in...

Key uncertainties in net impacts of Key uncertainties in net impacts of disease burden in Bangladesh due to disease burden in Bangladesh due to the substitution of groundwater for the substitution of groundwater for alternative drinking water suppliesalternative drinking water supplies

George Adamson* and David George Adamson* and David PolyaPolya**

SEAES, The University of Manchester, M13 9PL, UKSEAES, The University of Manchester, M13 9PL, UK

••[email protected]@yahoo.com; ; ••Corresponding author: Corresponding author: [email protected]@manchester.ac.uk

mailto:[email protected]

mailto:[email protected]

http://www.manchester.ac.uk/

AcknowledgementsAcknowledgementsFundingFunding–– European Commission / ESFEuropean Commission / ESF

MSc studentship to AdamsonMSc studentship to Adamson

–– EPSRCEPSRCResearch grant to Research grant to PolyaPolya, Lloyd, Vaughan & , Lloyd, Vaughan & WogeliusWogelius

DiscussionsDiscussions–– Raymond Raymond AgiusAgius (University of Manchester)(University of Manchester)–– DebapriyaDebapriya MondalMondal (University of Manchester)(University of Manchester)–– Sandy Sandy CairncrossCairncross (LSHTM) (LSHTM) –– MartieMartie van van TongerenTongeren (Edinburgh University)(Edinburgh University)–– DipankarDipankar Chakraborti (Chakraborti (JadavpurJadavpur University)University)

http://www.manchester.ac.uk/

Average Arsenic Average Arsenic Concentration in Concentration in

Bangladeshi Shallow Bangladeshi Shallow (< 150 m) Wells(< 150 m) Wells

Reproduced with acknowledgement from Peter Ravenscroft’s

“Arsenic Contamination and the Chemical Composition of Groundwater in Bangladesh”

based upon data collected during the “Groundwater Studies for Arsenic Contamination in

Bangladesh” project during 1998 and 1999. The project was funded by DFID on behalf of the

Bangladesh’s DPHE, and carried out by Mott MacDonald Ltd and the British Geological Survey.

Impact of Impact of tubewellstubewells: : trade off with diarrhoeal diseasetrade off with diarrhoeal diseaseWells constructed since 1970s to reduce Wells constructed since 1970s to reduce mortality due to diarrhoeal diseasesmortality due to diarrhoeal diseasesInfant mortality dropped:Infant mortality dropped:–– from 148 per 1000 live births in 1970from 148 per 1000 live births in 1970--19751975–– to to 59 per 1000 live births in 200059 per 1000 live births in 2000--20052005

Difficult to quantify impact of tubewells in Difficult to quantify impact of tubewells in reducing disease burden; howeverreducing disease burden; howeverCALCULATION OF IMPACTS ON DISEASE CALCULATION OF IMPACTS ON DISEASE BURDEN MUST TAKE INTO ACCOUNT BURDEN MUST TAKE INTO ACCOUNT –– DIARRHOEAL DISEASEDIARRHOEAL DISEASE–– ARSENICARSENIC--RELATED DISEASE BURDENRELATED DISEASE BURDEN

Quantifying Impact of Mitigation Quantifying Impact of Mitigation Solely Involving Transfer of Water Solely Involving Transfer of Water

Supplies from Supplies from TubewellsTubewells to to Alternative WatersAlternative Waters

Pathogenic microbes Pathogenic microbes related disease burdenrelated disease burden–– Negative impactNegative impact–– Independent of arsenic Independent of arsenic

concentration of well concentration of well waterswaters

ArsenicArsenic--related related disease burdendisease burden–– Positive impactPositive impact–– Strongly dependent Strongly dependent

upon arsenic upon arsenic concentration of well concentration of well waterswaters

Ideal ModelIdeal Model

Arsenic Concentration in Groundwater

Disease Burden Change in DALYs due to

improvements in sanitation and treatment of pathogenic microbes

(POSITIVE IMPACT)

Change in DALYs due to higher assumed value

for threshholdconcentration for As-

linked diseases

(POSITIVE IMPACT)

DALYs due to diseases related to pathogenic

microbes

(NEGATIVE IMPACT)

DALYs due to diseases related to arsenic

exposure from drinking water

(POSITIVE IMPACT)

Change in DALYs due to higher assumed link

between As concentration and As-related disease

(POSITIVE IMPACT)

Necessary to determine level of Necessary to determine level of uncertainty within parametersuncertainty within parameters

Unfortunately insufficient data available to Unfortunately insufficient data available to produce linear doseproduce linear dose--response relationships for response relationships for arsenic exposed populationarsenic exposed populationHowever, Lokuge et al (2004) have produced However, Lokuge et al (2004) have produced model which compares mitigation to 10 ppb As model which compares mitigation to 10 ppb As and to 50 ppb As and to 50 ppb As Important given current controversy over Important given current controversy over threshold values for arsenic in groundwater threshold values for arsenic in groundwater (Smith and Smith, 2004; Chakraborti, pers. (Smith and Smith, 2004; Chakraborti, pers. comm.)comm.)

Model of Model of LokugeLokuge et al (2004)et al (2004)Calculation of number of deaths and Disability Adjusted Calculation of number of deaths and Disability Adjusted Life Years (DALYs) in Bangladesh attributable to arsenic Life Years (DALYs) in Bangladesh attributable to arsenic contaminated drinking water through:contaminated drinking water through:–– Skin Cancer;Skin Cancer;–– Tracheal, Bronchial and Lung Cancer;Tracheal, Bronchial and Lung Cancer;–– Bladder Cancer;Bladder Cancer;–– Kidney Cancer;Kidney Cancer;–– Ischemic Heart Disease; andIschemic Heart Disease; and–– Diabetes MellitusDiabetes Mellitus

DALYs: objective measure of one lost healthy life year DALYs: objective measure of one lost healthy life year due to death or incapacity from Global Burden of due to death or incapacity from Global Burden of Disease study (Murray and Lopez, 1996)Disease study (Murray and Lopez, 1996)–– Displayed as annual figuresDisplayed as annual figures

Lokuge et al (2004) MethodLokuge et al (2004) MethodCalculation of populations exposed to arsenic at 0-10ppb, 10-50ppb, 50-100ppb, 100-300ppb, 300-600ppb and >600ppb

Multiplication of overall disease burden and deaths in Bangladesh in the covered sequelas with fraction of population within exposure categories = number of deaths and lost DALYs within each exposure category caused by diseases with a known link to arsenic

Calculation of Population Attributable Fractions due to arsenic contamination at various exposure categories (stratified for cancers, unstratified for diabetes mellitus and ischemic heart disease), using epidemiological data from Taiwanese populations

Multiplication of PAFs with diseases and DALYs in each exposure category to give number of diseases attributable to arsenic

Calculation of trade-off between DALY and deaths saved due to arsenic mitigation and 20.8% increase in diarrhoeal disease for populations without access to adequate sanitation

Lokuge et al (2004) contLokuge et al (2004) cont……Data on exposure levels taken from BGS study Data on exposure levels taken from BGS study ((SmedleySmedley and and KinniburghKinniburgh (2002))(2002))Standardised Mortality Ratios (SMRs) for disease Standardised Mortality Ratios (SMRs) for disease attributable to arsenic taken from studies on attributable to arsenic taken from studies on Taiwanese Populations (Smith et al (1992) and Tsai Taiwanese Populations (Smith et al (1992) and Tsai et al (1999))et al (1999))Average increase in diarrhoeal risk for populations Average increase in diarrhoeal risk for populations without access to adequate sanitation taken as without access to adequate sanitation taken as 20.8%, from 20.8%, from PrussPruss et al (2002)et al (2002)Existing disease burden in Bangladesh from Global Existing disease burden in Bangladesh from Global Burden of Disease 2002Burden of Disease 2002

Adamson and Adamson and PolyaPolya (2007)(2007)Attempt to determine level of uncertainty within Lokuge et al MoAttempt to determine level of uncertainty within Lokuge et al Modeldel–– Two papers published since 2004 by NavasTwo papers published since 2004 by Navas--Acien et al which Acien et al which

question the quality of the Tsai et al (1999)question the quality of the Tsai et al (1999)66 study, from which study, from which SMRs for Diabetes Mellitus and Ischemic Heart Disease were SMRs for Diabetes Mellitus and Ischemic Heart Disease were derivedderived

Of studies which relevant to the Lokuge et al model, Tsai et al Of studies which relevant to the Lokuge et al model, Tsai et al (1999)(1999)66 rated most poorly for both diseases, e.g.:rated most poorly for both diseases, e.g.:–– No internal comparisons within study participants,No internal comparisons within study participants,–– No adjustment for smoking, hypertension, lipids, BMINo adjustment for smoking, hypertension, lipids, BMI–– Intensity of search not independent of exposure statusIntensity of search not independent of exposure status

Additionally:Additionally:–– Population structure updated to 2001 census of BangladeshPopulation structure updated to 2001 census of Bangladesh–– Disease burden for Bangladesh only recently published Disease burden for Bangladesh only recently published ––

previously estimated from SEARpreviously estimated from SEAR--D global subD global sub--region, which was region, which was recognised by Lokuge et al as being nonrecognised by Lokuge et al as being non--representative of rural representative of rural populationspopulations

Results using updated population Results using updated population and disease burden dataand disease burden data

Pop. IntervenedPop. Intervened Cancer thresholdCancer threshold DeathsDeaths DALYsDALYs

> 10> 10µµg lg l--11 nonenone 5,1415,141 --12,28112,281

> 10> 10µµg lg l--11 > 50> 50µµg lg l--11 4,1564,156 --23,14623,146

> 10> 10µµg lg l--11 > 100> 100µµg lg l--11 3,5033,503 --30,46130,461

> 50> 50µµg lg l--11 nonenone 5,3895,389 18,42118,421

> 50> 50µµg lg l--11 > 100> 100µµg lg l--11 4,7374,737 11,10511,105

Results using updated population Results using updated population and disease burden data contand disease burden data cont……

Arsenic Concentration in Groundwater

Disease Burden

DALYs due to diseases related to pathogenic

microbes

(NEGATIVE IMPACT)DALYs due to diseases

related to arsenic exposure from drinking

water

(POSITIVE IMPACT) NET BENEFITS FOR DEFINED MITIGATION AT As CONC. HIGHER

THAN THIS THRESHHOLD

10ppb

Comparison of mitigation to 50ppm and 10 Comparison of mitigation to 50ppm and 10 ppmppmModelModel Population supplied Population supplied

with interventionwith interventionAll those exposed to All those exposed to

arsenic levels > 10 arsenic levels > 10 µµg/Lg/L

All those exposed to All those exposed to arsenic levels > arsenic levels >

50 50 µµg/Lg/L

NameName DescriptionDescriptionThreshold for arsenicThreshold for arsenic--related lung, bladder related lung, bladder and kidney cancerand kidney cancer

No No thresthres--holdhold >50 >50 µµg/Lg/L >100 >100

µµg/Lg/LNo No thresthres--

holdhold>100 >100 µµg/Lg/L

DeathsDeaths 77777777 66086608 58265826 78837883 70657065

DALYs discounted at 3 %DALYs discounted at 3 % 2854728547 1607716077 78277827 5182351823 4357243572

DeathsDeaths 50645064 41564156 35683568 53895389 48014801

DALYs discounted at 3 %DALYs discounted at 3 % --1358013580 --2314623146 --2926429264 1842118421 1230212302

DeathsDeaths 1444114441 1353313533 1294512945 1340213402 1281312813


DeathsDeaths 53855385 44774477 38893889 57115711 51235123


DeathsDeaths 59085908 50005000 44124412 62336233 56455645

DALYs discounted at 3 %DALYs discounted at 3 % 23402340 --72267226 --1334413344 3434134341 2822328223

DeathsDeaths 1790217902 1699416994 1640516405 1822718227 1763917639


DeathsDeaths 2290722907 2199921999 2141121411 1692716927 1633916339


DeathsDeaths 50115011 41034103 35153515 53365336 47484748


N/IHDN/IHD33

This study, NULL MODEL modified byThis study, NULL MODEL modified bySMRs for ischemic heart disease SMRs for ischemic heart disease fromWufromWu

N/IHDN/IHD22

This study, NULL MODEL modified byThis study, NULL MODEL modified bySMRs for ischemic heart disease from TsengSMRs for ischemic heart disease from Tseng

N/IHDN/IHD11

This study, NULL MODEL modified byThis study, NULL MODEL modified bySMRs for ischemic heart disease from LinSMRs for ischemic heart disease from Lin

N/DMN/DM33

This study, NULL MODEL modified byThis study, NULL MODEL modified bySMRs for diabetes mellitus from WangSMRs for diabetes mellitus from Wang

N/DMN/DM22

This study, NULL MODEL modified byThis study, NULL MODEL modified bySMRs for diabetes mellitus from RahmanSMRs for diabetes mellitus from Rahman

N/DMN/DM11

This study, NULL MODEL modified byThis study, NULL MODEL modified bySMRs for diabetes mellitus from Lai SMRs for diabetes mellitus from Lai

NULLNULL This study, NULL MODEL, parameter values as This study, NULL MODEL, parameter values as described in the textdescribed in the text

LQMLQMThis study, parameter values as used by Lokuge This study, parameter values as used by Lokuge

where possible, otherwise as described in the where possible, otherwise as described in the texttext

NavasNavas--Acien et al (2005 and 2006)Acien et al (2005 and 2006)

Relevant studies (i.e. lowest exposure Relevant studies (i.e. lowest exposure category 300ppb or lower) rated most category 300ppb or lower) rated most highly were Lai et al (1994) for Diabetes highly were Lai et al (1994) for Diabetes Mellitus and Tseng et al (2003) for Mellitus and Tseng et al (2003) for ischemic heart diseaseischemic heart disease

These studies stratified exposure into 3 These studies stratified exposure into 3 categories, whereas Tsai et al (1999) only categories, whereas Tsai et al (1999) only compared compared ‘‘exposedexposed’’ (>300ppb) with (>300ppb) with ‘‘unexposedunexposed’’..

Model using SMRs from alternative Model using SMRs from alternative studies for Diabetes Mellitusstudies for Diabetes Mellitus

0

200

400

600

800

1000

1200

1400

1600

Lai et al, 1994 Wang et al, 2003 Rahman et al,1999

Study from which SMRs were derived

perc

enta

ge d

iffer

ence

in o

vera

ll re

sult

from

nul

l mod

el

DeathsDALYs

Model using SMRs from alternative Model using SMRs from alternative studies for Ischemic Heart Diseasestudies for Ischemic Heart Disease

-200

0

200

400

600

800

1000

1200

1400

Tseng et al, 2003 Lin and Yang,1988

Wu and Tsai

Study from which SMRs were derived

perc

enta

ge d

iffer

ence

in o

vera

ll re

sult

from

nul

l mod

el

DeathsDALYs

ConclusionsConclusionsNavasNavas--Acien studies have identified Diabetes Mellitus Acien studies have identified Diabetes Mellitus and Ischemic Heart disease as key uncertainties within and Ischemic Heart disease as key uncertainties within the model, with very large errorsthe model, with very large errorsNecessary for linear dose response relationships Necessary for linear dose response relationships between arsenic exposure and arsenicbetween arsenic exposure and arsenic--related diseases related diseases to be established, as to be established, as

more exposure categories = greater risk exposedmore exposure categories = greater risk exposedControversy surrounding mitigation to 50 ppb Controversy surrounding mitigation to 50 ppb vsvs 10 ppb 10 ppb renders reduction in errors more crucialrenders reduction in errors more crucialCould be major consequences if models are used by Could be major consequences if models are used by policy makers without further workpolicy makers without further work–– N.B. degree of error only checked for two endN.B. degree of error only checked for two end--points, likely to be points, likely to be

much higher if cancers are also taken into account!much higher if cancers are also taken into account!

Final PointFinal PointUse of 20.8% in contentionUse of 20.8% in contentionUsed as a blanket value in Lokuge et al as Used as a blanket value in Lokuge et al as quantity of water seen as important as quality, quantity of water seen as important as quality, and household handling may be a major source and household handling may be a major source of contamination (of contamination (EsreyEsrey et al (1991), et al (1991), HoqueHoque et al et al (1989).(1989).Clearly this is not the case, as demonstrated by Clearly this is not the case, as demonstrated by seminars given today at RGS 2007seminars given today at RGS 2007Studies currently ongoing to reassess this claim Studies currently ongoing to reassess this claim (see (see ClasenClasen and and CairncrossCairncross (2004) Refining the (2004) Refining the dominant paradigm)dominant paradigm)

References References –– Page 1 of 2Page 1 of 2Adamson, G.C.D.; Adamson, G.C.D.; PolyaPolya, D.A. Critical pathway analysis to determine key uncertainties , D.A. Critical pathway analysis to determine key uncertainties in net impacts on in net impacts on

disease burden in Bangladesh of arsenic mitigation involving thedisease burden in Bangladesh of arsenic mitigation involving the substitution of surface water for substitution of surface water for groundwater drinking water supplies. groundwater drinking water supplies. Environmental Science and Health Part A,Environmental Science and Health Part A, 42, in press.42, in press.

ClasenClasen, T.F.; , T.F.; CairncrossCairncross, S. Household water management: refining the dominant paradigm., S. Household water management: refining the dominant paradigm. Tropical Tropical Medicine and International Health Medicine and International Health 2004, 2004, 99,187,187--191191

EsreyEsrey, S.A.; Potash, J.B.; Roberts, L.; , S.A.; Potash, J.B.; Roberts, L.; ShiffShiff, C. Effects of improved water supply and sanitation on , C. Effects of improved water supply and sanitation on ascariasisascariasis, diarrhoea, , diarrhoea, dracunculiasisdracunculiasis, hookworm infection, , hookworm infection, schistosomiasisschistosomiasis, and trachoma. Bulletin of , and trachoma. Bulletin of the World Health Organisation the World Health Organisation 1991, 1991, 6969, 609, 609--621621

HoqueHoque, B.A.; , B.A.; HuttlyHuttly S.R.; S.R.; AzizAziz, K.M.; , K.M.; PatwaryPatwary, M.Y; , M.Y; FeachemFeachem, R.G. , R.G. TubewellTubewell water consumption and its water consumption and its determinants in a rural area of Bangladesh. Journal of Tropical determinants in a rural area of Bangladesh. Journal of Tropical Medicine and Hygiene Medicine and Hygiene 19891989, , 9292, 197, 197--202202

Lai., M.Lai., M.--S.; S.; HsuehHsueh, Y., Y.--M.; Chen, C.M.; Chen, C.--J.; J.; ShyuShyu, M., M.--P.; Chen, S.P.; Chen, S.--Y.; Y.; KuoKuo, T., T.--L.; Wu, M.L.; Wu, M.--M.; Tai, T.M.; Tai, T.--Y. Ingested Y. Ingested Organic Arsenic and Prevalence of Diabetes Mellitus. American JoOrganic Arsenic and Prevalence of Diabetes Mellitus. American Journal of Epidemiology urnal of Epidemiology 1994,1994, 139,139,484484--492492

Lin, S.M.; Yang, M.H. Arsenic, selenium, and zinc in patients wiLin, S.M.; Yang, M.H. Arsenic, selenium, and zinc in patients with th blackfootblackfoot disease. Biological Trace disease. Biological Trace Elements Research Elements Research 19881988, 15, 213, 15, 213--221221

LokugeLokuge, K.M.; Smith, W.; Caldwell, B.; Dear, K.; Milton, A.H. The Effe, K.M.; Smith, W.; Caldwell, B.; Dear, K.; Milton, A.H. The Effect of Arsenic Mitigation Interventions ct of Arsenic Mitigation Interventions on Disease Burden in Bangladesh. Environmental Health Perspectivon Disease Burden in Bangladesh. Environmental Health Perspectives es 2004,2004, 112112,, 11721172--11771177

Murray, J.L.; Lopez, D. Murray, J.L.; Lopez, D. ((edseds)) The Global Burden of DiseaseThe Global Burden of Disease. 1996. 1996 Published by The Harvard School of Published by The Harvard School of Public Health on behalf of The World Health Organisation and ThePublic Health on behalf of The World Health Organisation and The World BankWorld Bank

PrPrüüssss, A.; Kay, D.; , A.; Kay, D.; FewtrellFewtrell, L.; Bartram, J. Estimating the Burden of Disease from Water, S, L.; Bartram, J. Estimating the Burden of Disease from Water, Sanitation, and anitation, and Hygiene at a Global Level. Environmental Health Perspectives Hygiene at a Global Level. Environmental Health Perspectives 2002,2002, 110110,, 537537--542542

References References –– Page 2 of 2Page 2 of 2

Rahman, M.; Rahman, M.; TondelTondel, M.; Ahmad, S.A.; , M.; Ahmad, S.A.; ChowdhuryChowdhury, I.A.; , I.A.; FaruqueeFaruquee, M.H.; , M.H.; AxelsonAxelson, O. Hypertension and , O. Hypertension and Arsenic Exposure in Bangladesh. Hypertension Arsenic Exposure in Bangladesh. Hypertension 19991999, 33, 74, 33, 74--7878

SmedleySmedley P.L.; P.L.; KinniburghKinniburgh, D.G. A review of the source, behaviour and distribution of ars, D.G. A review of the source, behaviour and distribution of arsenic in enic in nautralnautralwaters. Applied Geochemistry, waters. Applied Geochemistry, 20022002, , 1717, 517, 517--568568

..Smith, A.H.; Smith, A.H.; HopenhaynHopenhayn--Rich, C.; Bates, M.N.; Rich, C.; Bates, M.N.; GoedenGoeden, H.M.; Hertz, H.M.; Hertz--PicciottoPicciotto, I.; Duggan, H.M.; Wood, , I.; Duggan, H.M.; Wood,

R.; R.; KosnettKosnett, M.J.; Smith, M.T. Cancer Risks from Arsenic in Drinking Water., M.J.; Smith, M.T. Cancer Risks from Arsenic in Drinking Water. Environmental Health Environmental Health Perspectives Perspectives 1992,1992, 9797, , 259259--267267

Smith, A.H.; Smith, M.M.H. Arsenic drinking water regulations inSmith, A.H.; Smith, M.M.H. Arsenic drinking water regulations in developing countries with extensive developing countries with extensive exposure. Toxicology exposure. Toxicology 2004,2004, 198198,, 3939--4444

Tseng, C.Tseng, C.--H.; H.; ChongChong, C., C.--K.; Tseng, C.K.; Tseng, C.--P.; P.; HsuehHsueh, Y., Y.--M.; M.; ChiouChiou, H., H.--Y.; Tseng, C.Y.; Tseng, C.--C.; Chen, C.C.; Chen, C.--J. LongJ. Long--term arsenic exposure and ischemic heart disease in term arsenic exposure and ischemic heart disease in arseniasisarseniasis--hyperendemichyperendemic villages in Taiwan. villages in Taiwan. Toxicology LettersToxicology Letters 2003,2003, 137137,, 1515--2121

Tsai, S.Tsai, S.--M.; Wang, T.M.; Wang, T.--N.; N.; KoKo, Y., Y.--C. Mortality for Certain Diseases in Areas with High Levels of AC. Mortality for Certain Diseases in Areas with High Levels of Arsenic in rsenic in Drinking Water. Archives of Environmental Health Drinking Water. Archives of Environmental Health 1999,1999, 5454,, 186186--193193

Wang, S.Wang, S.--L.; L.; ChiouChiou, J., J.--M.; Chen, C.M.; Chen, C.--J.; Tseng, C.J.; Tseng, C.--H.; Chou, W.H.; Chou, W.--L.; Wang, C.L.; Wang, C.--C.; Wu, T.C.; Wu, T.--N.; Chang, L.N.; Chang, L.--W. W. Environmental Health Perspectives Environmental Health Perspectives 20032003, 111, 155, 111, 155--159159

WHO WHO The World Health Report, 2002 The World Health Report, 2002 –– Reducing Risks, Promoting Healthy LifeReducing Risks, Promoting Healthy Life. . 20022002 Available from Available from http://www.who.int/whr/2002/en/index.html [online]http://www.who.int/whr/2002/en/index.html [online] Accessed 21/7/06Accessed 21/7/06

Wu, M.Wu, M.--M.; M.; KuoKuo, T., T.--L.; Hwang, Y.L.; Hwang, Y.--H.; Chen, C.H.; Chen, C.--J. DoseJ. Dose--Response Relation between Arsenic Response Relation between Arsenic Concentration in Well Water and Mortality from Cancers and VascuConcentration in Well Water and Mortality from Cancers and Vascular Disease. American Journal of lar Disease. American Journal of Epidemiology Epidemiology 19891989, 130, 1123, 130, 1123--11321132

key uncertainties in net impacts of disease burden in...

Documents