ent 379 (2007) 167–175www.elsevier.com/locate/scitotenv
Science of the Total Environm
Screening of arsenic in tubewell water with field test kits:Evaluation of the method from public health perspective
Md. Jakariya a,d,f,⁎, Marie Vahter b, Mahfuzar Rahman c, M. Abdul Wahed c,Samar Kumar Hore c, Prosun Bhattacharya d, Gunnar Jacks d, Lars Åke Persson c,e
a Research and Evaluation Division, Bangladesh Rural Advancement Committee (BRAC), 75 Mohakhali, Dhaka 1205, Bangladeshb Institute of Environmental Medicine, Karolinska Institutet, SE-171 77 STOCKHOLM, Sweden
c ICDDR, B: Centre for Health and Population Research, Dhaka, Bangladeshd KTH-International Groundwater Arsenic Research Group, Department of Land and Water Resources Engineering,
Royal Institute of Technology (KTH), SE-100 44 STOCKHOLM, Swedene International Maternal and Child Health (IMCH), Uppsala University, SE-750 07 UPPSALA, Sweden
f NGO Forum for Drinking Water Supply and Sanitation, 4/6 Block E, Lalmatia, Dhaka-1207, Bangladesh
Received 25 May 2006; received in revised form 3 October 2006; accepted 23 November 2006Available online 29 January 2007
Keywords: Tubewell; Bangladesh; Drinkingwater; Field test kit; Analysis; Assessment; Specificity; Sensitivity; Predictive values; Exposure; Public health
Arsenic (As) contamination in tubewell (TW) water,which serves as the primary source of drinking water inBangladesh has now been recognized as a serious publichealth problem (Khan et al., 1997; Ahmad et al., 1998;
168 Md. Jakariya et al. / Science of the Total Environment 379 (2007) 167–175
NRC, 2001; Alam et al., 2002; Caldwell et al., 2003a,b;Yu and Ahsan, 2004; IARC, 2004; Rahman et al., 2006).In order to eliminate diarrheal diseases, caused bydrinking microbial contaminated surface water, 6–11 million TWs were installed across Bangladeshduring the past three decades. Most of the hand-pumpedTWs are installed at depths varying between 20–30 m(UNICEF, 2003). Among these TWs, approximately30–40% have been estimated to contain As at levelsexceeding the current Bangladesh drinking waterstandard (BDWS) value of 50 μg/L (BAMWSP,2006). As a consequence, out of a total population of144 million in the country, about 35 million are believedto be exposed to As concentration in drinking waterabove 50 μg/L, and about 57 million to concentrationgreater than 10 μg/L (Milton and Rahman, 1999;Bhattacharya et al., 2002a,b; Ahmed et al., 2004;Hoque et al., 2004; Dhaka Community Hospital Trust,2005; BAMWSP, 2006; Jakariya et al., in press).
Arsenic contamination in groundwater of Bangladeshis geogenic in nature, originating from sedimentaryaquifers of Ganges–Brahmaputra–Meghna (GBM) delta(Bhattacharya et al., 1997, 2002a,b; Nickson et al., 2000;Smedley and Kinniburgh, 2002). The concentration ofAs in TW water varies widely, both on a local and aregional scale (BGS/DPHE, 2001; van Geen et al., 2003;Ahmed et al., 2004; Bhattacharya et al., 2006; Rahmanet al., 2006). Thus, As contamination cannot be ac-curately predicted by sporadic testing of water from asmall number of TWs. Screening of water of the all TWsin the country is, therefore, needed to identify the extentand distribution as well as monitoring the seasonal andtemporal variations of As in the groundwater ofBangladesh (BGS/DPHE, 2001; Cheng et al., 2004).
Water testing of all TWs of the country is a gigantictask that involves technical, institutional, and socialchallenges. Given the scale of the problem, screeningand regular monitoring of all the existing TW water isalmost impossible using the laboratory method by AASthat is generally taken as the “gold standard”. This is dueto the limited availability of facilities, transportation ofacidified water samples to laboratories and poor econom-ical situation of the country (Rahman et al., 2006).Importantly, a scientific inter-laboratory comparisonstudy conducted by the WHO in Bangladesh, revealedthat the test results of less than one-third of the parti-cipating laboratories were within 20% of the expectedvalues (Aggarwal et al., 2001). Therefore, simple, low-cost methods for As determination, such as the field testkits have proved to be most suitable for performing theTW screening in shortest possible time. Several commer-cial field test kits are available for determination of As in
TW water (Rahman et al., 2002; Khandaker, 2004;Cherukurii and Anjaneyulu, 2005; Deshpande and Pande,2005; van Geen et al., 2005; Steinmaus et al., 2006). Fieldkits provide semi-quantitative results and the reliability ofseveral field kits are questioned because of poor accuracy(Rahman et al., 2002). While on the other hand, van Geenet al. (2005) suggests that field kits could be usedsuccessfully to detect As in TW water. Thus, there is aneed for further evaluation of the screening results by thefield kit, prior to its recommendation for wide scale use inBangladesh and elsewhere in the world.
An overall risk assessment including component ofmitigation for As contamination should be based onaccurate determination of As levels in TW water usingeconomically viable methods for As screening. Thispaper presents a comparative study of field test kit andlaboratory method for As determination in the TWwaterof Matlab Upazila in southeastern Bangladesh, with anaim to identify the accuracy and technical effectivenessof field test kit results for screening and regularmonitoring of TW water. The paper also identifies thespecific As concentration categories especially atconcentration levels close to the limits of the WHOand BWDS drinking water standards where emphasisneed to be given in order to get more accurate field kitresults, which would greatly contribute to minimise Asexposure as well as better groundwater management.
2. Materials and methods
2.1. Study area
Matlab Upazila is situated about 50 km southeast ofDhaka. The geological setting of Matlab is characterizedby a thick succession of alluvial sediments of Holoceneage, deposited by the Meghna River and its tributaries(Jakariya et al., 2005; von Brömssen et al., 2007). Since1966, ICDDR,B has been running a health anddemographic surveillance system in 142 villages ofthe Matlab Upazila, encompassing a population of220,000 on 18,386 ha of land. The operational area ofICDDR,B at Matlab is situated both inside and outsideof a flood control embankment (Fig. 1).
2.2. Screening of TW water and sampling forlaboratory analyses
All the TWs in the study area, distributed over theseven blocks of the study area, were tested with Merckfield test kit and analysed in the laboratory (Fig. 1). TheTWs in the study area were assigned a unique numberand the geographic coordinates of these TWs were
Fig. 1. Location map of the study area showing the prevalence of tubewells (TWs) with As concentrations≥50 μg/L (red circles) and those b50 μg/L(green circles). Inset is the map of Bangladesh showing the location of Matlab Upazila.(For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)
169Md. Jakariya et al. / Science of the Total Environment 379 (2007) 167–175
determined with hand-held Global Positioning System(GPS) receivers. A four-member team performed Asdetermination in TW water using the Merck (Merck UroLab, Germany) field test kit.
Prior to the field screening, a week-long training wasgiven to the field testers that included instructions aboutthe procedure for testing TWwater, as well as guidelinesfor the dissemination of As related messages during
screening (Chowdhury and Jakariya, 1999). Each TWwas purged 20/30 strokes before testing with the Merckkit and collection of water samples in two 20-mlpolyethylene vials (Merck Eurolab) marked with thetube well ID and pre-treated with acid to preventprecipitation of iron and co-precipitation of As andtransported to the Matlab hospital laboratory. Thesamples were kept at −20 °C until analysis in the
170 Md. Jakariya et al. / Science of the Total Environment 379 (2007) 167–175
laboratory at the ICDDR,B, Dhaka. A detailed descrip-tion of the water sample collection and Merck kitanalytical procedures have been described elsewhere(Jakariya et al., 2003). Field testers moved village-wiseto test TW water in different parts of the study area.
Based on the field kit results, the wells with Asconcentrations ≥50 μg/L were painted red to indicatethat the water contain unsafe level of As, and those withlevels b50 μg/L were painted green. People wereencouraged to take water for drinking and cookingpurposes from the green marked TWs.
2.3. Validation of field kit results
Validation of the field testing results was done at twolevels: re-testing certain percentage of TW water byField Supervisors using the same field kit method toassess Field Testers performance and re-testing of all theTW water of the study area at laboratory for cross-checking field kit results.
2.3.1. Re-testing by field supervisorsAs a first control of the field test kit results, 400 TWs
(representing about 3% of the total screened TWs inMatlab Upazila) were selected randomly and re-testedby field supervisors for As levels without any prior biasof the former screening results using the same kit andprocedure. Later on, the results were also crosscheckedwith the laboratory results for further validation.
2.3.2. Laboratory analysesTotal As was measured by hydride generation atomic
absorption spectrophotometer (HG-AAS, ShimadzuModel AA-6800), equipped with an auto sampler(ASC-6800). The lower limit of detection was b1 μg/L.Each sample was assayed twice. Accuracy and precisionwere carefully tested using a certified reference material(NIST 1643d). To evaluate the results of As analyses atthe ICDDR,B laboratory, about 2% of the water sampleswere randomly selected and re-analyzed at the laboratoryof Karolinska Institute in Sweden for cross-checking ofICDDR,B results. The comparative result of the twolaboratories was found to be highly significant (pb0.01).The detailed laboratory analytical procedure and inter-laboratory comparison results were presented elsewhere(Wahed et al., 2006).
2.4. Data analysis and validation
The data were analyzed using SPSS Software. Inorder to compare and evaluate the field kit results withthe laboratory analysis, As test results were grouped into
categories of 0–9.9, 10–24, 25–49, 50–99, 100–499and ≥500 μg/L according to the colour scale range forthe Merck kit. Three indices such as sensitivity,specificity, and predictive values of positive resultshave been widely used for the validation of screeningtests (Grimes and Schulz, 2002; Persson and Wall,2003). Prevalence, sensitivity, specificity, and positivepredictive values (PPV) were calculated for bothBangladesh Drinking Water Standard (BDWS) andWHO drinking water guideline value in order tocomment on the effectiveness of the field test kit resultsand their implications on the management of risk for Asexposure and related public health issues.
2.5. Sensitivity and specificity and the positive predictivevalue (PPV)
2.5.1. Sensitivity and specificitySensitivity (also termed as the detection rate) is the
ability of the field test kit to find those wells withconcentrations of As at cut-off levels at BDWS andWHO guideline values. Sensitivity of the test kitscould be calculated using the formula (A /A+C)(Grimes and Schulz, 2002; Persson and Wall, 2003;Linn, 2004). In contrast, specificity denotes the abilityof a test to identify the TWs without the condition (i.e.exceeding the drinking water standards). The specific-ity is calculated by using the formula (D /B+D),Fig. 2).
2.5.2. Predictive values (PV)In order to interpret results obtained from the
screening test using the field test kit, it is important toknow the predictive values (PV), particularly thepositive predictive value (PPV) of the test results thathas significant implications from public health perspec-tive. The PPV is calculated using the relation (A /A+B)(Grimes and Schulz, 2002; Persson and Wall, 2003;Linn, 2004, Fig. 2, Table 2).
3. Results and discussions
3.1. Screening of TW water
The 12,532 TWwater samples were tested using bothMerck kit and AAS methods. A comparison of theoverall results of As determinations by the field test kitand the AAS in terms of distribution of TWs in thedifferent As categories, is presented in (Fig. 3). Nearly64% of the total water samples indicated As atconcentration exceeding the BDWS of 50 μg/L, and72% exceeded the WHO guideline value of 10 μg/L.
Fig. 2. Generic approach for the validation of the screening results using Merck field test kit and the As concentrations measured in the laboratory byAAS at different arsenic cut-off levels: a) BDWS (50 μg/L), and b) WHO drinking water guideline value (10 μg/L). The values in parenthesesrepresent the total number of cases in the respective cut-off levels of field test kit results as well as their percentages.
171Md. Jakariya et al. / Science of the Total Environment 379 (2007) 167–175
Among the total number of TWs, a distinct bimodaldistribution could be observed for the water Asconcentrations (Fig. 3). Block-wise distribution of thestatus of As in TW water is presented in Fig. 4.
3.2. Evaluating the field screening procedure
In order to check accuracy of the field screeningresults a total of about 400 TWs (representing about 3%of the total tested TWs in Matlab) tested by the Merckfield test kit, were randomly selected and re-testedimmediately following the same testing procedure.
These results were later compared with the laboratoryresults for cross validation. The results of the fieldtesters agreed with those of the Field Supervisors forabout 95% of the total screened TWs. The falsenegative results, i.e. the cases where As contaminatedTWs (≥50 μg/L) were painted green by the fieldtesters, represented 8.9% of the total TWs tested to beb50 μg/L. Similarly, the false positive results, the caseswhere a TW being marked red turned out to haveb50 μg/L in the Field Supervisors' testing, were about3.1% of the total high As TW water samples screenedby the field testers.
Table 1Comparison of Merck field kit results with laboratory (AAS) results atthe different As concentration categories
⁎ The values in parentheses represent the percentage of TWs inrespective categories.
Fig. 3. Comparison of the results of the determination of Asconcentrations in TWs of Matlab Upazila with Merck test kit andAAS methods (n=12,532).
172 Md. Jakariya et al. / Science of the Total Environment 379 (2007) 167–175
3.3. Comparison of analytical results using field test kitand AAS methods
The comparison of the TW screening results by theMerck field test kit in relation to the different categoriesof As concentrations obtained by AAS results and thetwo water As guideline values (i.e. 10 and 50 μg/L) arepresented in Table 1.
In general, overall agreement between the Merck andthe laboratory results was found to be statisticallysignificant for Bangladesh drinking water standard(k=0.92, pb0.01) and WHO guideline value (k=0.91,pb0.01). However, it is observed that the agreementwas below 90% for the categories of 25–49 μg/L (87%)and 50–99 μg/L (70%) when compared with theBWDS, and in the categories of 10–24 μg/L (only47%) and 25–49 μg/L (86%) for comparison with theWHO guideline value. On the other hand, agreementbetween Merck kit data with the respective laboratoryresults was found to be comparatively poor (k=0.69,
Fig. 4. Status of As distribution in TWs at different blocks of MatlabUpazila at cut-off level of BDWS (50 μg/L) as revealed by AASmeasurements.
pb0.01) for all As concentration categories except for0–9.9 μg/L (91%), as in this category field kits usuallydo not develop any color on the test strip.
3.4. Effectiveness of TW screening and validation
3.4.1. Comparison with Bangladesh Drinking WaterStandard level (BDWS, 50 μg/L)
The overall correct identification of the Merck kit atthis standard level was more than 98% of the total testedTWs in the study area, except for the two categories of25–49 μg/L (87%) and 50–99 μg/L (70%) (Fig. 2). Thepercentage of false negative detections (classified asb50 μg/L although AAS showed ≥50 μg/L) was only2.3% of the total number of analyzed water samples. Itshould be mentioned that people in this category wouldstill continue drinking water from contaminated sourceswithout even knowing the fact. More than 71% of thetotal false negative cases were identified in the categoryof 25–49 μg/L.
Table 2Assessment of the effectiveness of field test kit results (n=12,532) andtheir validation in Matlab Upazila
Indices forvalidation of thefield kit results
Basis forcalculation inTable 1
BWDS cut-off level(50 μg/L)
WHO limitcut-off level(10 μg/L)
Prevalence (%) – 64 74False positive (%) B / total
positive results4.4 3.6
False negative (%) C / totalnegative results
2.3 5.1
Sensitivity A / (A+C) 0.99 0.98Specificity D / (B+D) 0.92 0.91PPV A / (A+B) 0.96 0.96NPV D / (C+D) 0.98 0.95
Data based on Fig. 2.
Fig. 5. Plot showing the relationship of Positive Predictive Values(PPV) with prevalence of TWs with As concentration ≥50 μg/L.
173Md. Jakariya et al. / Science of the Total Environment 379 (2007) 167–175
False positive determinations (classified as≥50 μg/Lalthough AAS analyses revealed As concentrationb50 μg/L), on the other hand, represented about 4.4%of the screened TWs by the field test kits. About 82% ofthose TWs identified by Merck kit fell above the WHOdrinking water guideline value of 10 μg/L. Similarly,more than 77% of the total false positive TWs fell in thecategory of 50–99 μg/L of As. In the case of BDWS cut-off level, the false positive and false negative TWs weredistributed throughout the study area and thus indicatedmore likely field kit methodological error than humanerror. At this cut-off level, the correct identification of Ascontaminated TWs is better than As-safe TWs with ahigh sensitivity and specificity of 0.99 and 0.92,respectively (Table 2).
3.4.1.1. Comparison with WHO guideline value fordrinking water (10 μg/L). Merck kit did not provideacceptable results in the lower As concentration ranges(i.e. 10–25 and 25–50 μg/L) if the WHO drinking waterguideline value of. 10 μg/L is considered. The correctdetection was found to be lowest for the 10–25 μg/Lcategory (47%). The false negative and false positivecases at this cut-off level were 5.1% and 3.6%respectively, of the total tested TW water, in spite ofhigh sensitivity of 0.98 and comparably low specificityof 0.91 (Table 2). A similar trend was also observed forall the geographical blocks of the study area, except forone, where the percentage of correct screening of TWwater was comparatively poor. The lowest prevalence ofAs (54%) in TW water in this block (see Wahed et al.(2006) for details) might be one of the main reasons forthis poor performance. The distribution of false positiveand false negative TWs identified at this cut-off levelwas also spread over the entire study area.
3.4.2. Predictive values (PV) and prevalenceThe prevalence of tubewell As contamination in a
certain location affects the screening test performance.When the number of true positives is low (i.e. in a regionwith an overall low prevalence of As contaminatedTWs), the PPV will be lower, indicating a lowerreliability of each positive result and vice versa. Thistrend indicates that the PPV increases with theprevalence of As contamination in TW water (Fig. 5).
4. Discussion and conclusion
The present study reveals that the Merck field test kitproduced false negative results for water samples at50 μg/L concentration levels in 2.3% of the total 12,532TW water samples tested in the study area, which isundesirable from a public health point of view. Thestudy indicates that the TWs with As concentrationbetween (10–24) for the 10 μg/L and (50–99) for the50 μg/L cut-off levels, respectively, are mostly mis-classified that need to be re-analyzed by Field Super-visors and verified by laboratory analyses for crossvalidation. The percentage of false negative tested TWsat the WHO guideline value of 10 μg/L level was almostdouble (i.e. 5.1%) than that of 50 μg/L, indicating poorperformance of the kit at lower As concentration ranges.On the other hand, the false positive cases identified byMerck kit were only 4.4% and 3.6% of the total testedTW water samples for the 50 μg/L and 10 μg/L of Ascut-off levels, respectively. It is to be mentioned herethat As concentrations in about 70% of such falsepositive TWs at the BDWS level were found to behigher than that of the WHO cut-off value. Suchidentification of the Merck kit is beneficial for its users,considering that long-term exposure to As contaminatedwater even at 10 μg/L that increases the risk of varioushealth hazards (Smith et al., 2000; IARC, 2004; Kapajet al., 2006).
The study significantly enhances the arguments forthe use of field test kits as the preferable tool for massscale screening of TW water for As in Bangladesh aswell as in other geographical regions of the world withsimilar problems, as compared to the earlier studies (viz.Rahman et al., 2002; van Geen et al., 2005). The mainadvantage of the field kit is that it allows screening ofTW water in relatively short period of time. Besides, theanalysis of TW water with the field kit in the courtyardof the TW owner in the presence of enthusiastic crowd,also helps to raise people's awareness about the healthhazards of As.
Another major outcome of this paper is that it hasidentified specifically the As concentration categories
174 Md. Jakariya et al. / Science of the Total Environment 379 (2007) 167–175
where special measure needs to be taken to obtain moreaccurate filed kit results. Positive Predictive Value(PPV) of the tubewell test results showed highlydependent on the prevalence of As concentration inTW water, the risk of false detection by the Merck kit iscomparatively higher where the prevalence of As is lowand vise versa.
However, necessary precautionary measures aretaken for the specific As concentration categoriescritical from the public health perspective, where theperformance of the field test kit was comparatively poor.Considering the limitations of the field test kits, furtherimprovement of the existing Merck kit method, wouldbe necessary for better performance in order to mitigatethe risks of As exposure from the drinking TW water.
In order to reduce false identification during TWsscreening by field test kit, several issues need to beconsidered: i) commitment and knowledge of the testersabout the As problem; ii) training of field testers fordeveloping skills for proper colour perception; iii)development of digital colour comparator at low Asconcentration levels to minimize the human errors; iv)checking the quality of testing reagents since a field testkit once opened needs to be used within 15–20 days ashumidity (above 90%) and temperature (above 25 °C)make the chemicals inactive; and iv) options forverification to assess validity of the field kit results. Inaddition, different field test kits available commercially,also need to be compared for accuracy, reliability, costand their “user friendliness”.
Acknowledgements
The authors are grateful to the Swedish InternationalDevelopment Agency (Sida), Dhaka, Bangladesh fortheir financial support through the AsMat project(Donor reference U11 BB/1.5.5–3, 1998–05440 andU11 BB/1.5.5–3/A). The authors thank Ms. DulalyChowdhury for the analyses of As at the laboratories ofICDDR,B laboratory and all the field workers whotested As in tubewell water in the field. We deeplyappreciate the comments of Deoraj Caussy, JochenBundschuh and Colin Neal for their suggestions on theearlier drafts of this manuscript.
References
Aggarwal PK, Dargie M, Groening M, Kulkarni KM, Gibson JJ. Aninter-laboratory comparison of arsenic analysis in Bangladesh.International Atomic Energy Agency, Vienna, Austria; 2001. Draftreport. URL: http://www-naweb.iaea.org/napc/ih/BGD-Inter-comp-IAEA.pdf (Accessed on May 29, 2006).
Ahmad SA, FaruqueeMH, SayedMH, KhanMH, Jalil MA, Ahmed R,et al. Chronic arsenicosis: management by vitamin A, E, Cregimen. J Prev Soc Med 1998;17:19–26.
Ahmed KM, Bhattacharya P, Hasan MA, Akhter SH, Alam SMM,Bhuyian MAH, et al. Arsenic contamination in groundwater ofalluvial aquifers in Bangladesh: an overview. Appl Geochem2004;19(2):181–200.
AlamMGM, Allinson G, Stagnitti F, Tanaka A,WestbrookeM. Arseniccontamination in Bangladesh groundwater: a major environmentaland social disaster. Int J Environ Health Res 2002;12(3):235–53.
Bangladesh Arsenic Mitigation Water Supply Project (BAMWSP). 2006.URL: http://www.bamwsp.org (Accessed on 15 February, 2006).
BGS/DPHE (British Geological Survey/Department of Public HealthEngineering). Arsenic contamination of groundwater in Bangla-desh. Technical Report WC/00/19 2001. Keyworth, UK: BritishGeological Survey, Natural Environment Research Council,Department for International Development, Government of thePeople's Republic of Bangladesh.
Bhattacharya P, Chatterjee D, Jacks G. Occurrence of arseniccontaminated groundwater in alluvial aquifers from Delta Plains,Eastern India: options for safe drinking water supply. Int J WaterRes Manag 1997;13:79–92.
Bhattacharya P, Frisbie SH, Smith E, Naidu R, Jacks G, Sarkar B.Arsenic in the environment: a global perspective. In: Sarkar B,editor. Handbook of Heavy Metals in the Environment. New York:Marcell Dekker; 2002a. p. 145–215.
Bhattacharya P, Jacks G, Ahmed KM, Routh J, Khan AA. Arsenic ingroundwater of the Bengal delta plain aquifers in Bangladesh. BullEnviron Contam Toxicol 2002b;69:538–45.
Bhattacharya P, Ahmed KM, Broms S, Fogelström J, Jacks G, SracekO, et al. Mobility of arsenic in groundwater in a part ofBrahmanbaria district, NE Bangladesh. In: Naidu R, Smith E,Owens G, Bhattacharya P, Nadebaum P, editors. Managing Arsenicin the Environment: From Soil to Human Health. Melbourne:CSIRO Publishing; 2006. p. 95–115.
Caldwell BK, Caldwell JC, Mitra SN, Smith WT. Tubewells andarsenic in Bangladesh: challenging a public health success story.Int J Popul Geogr 2003a;9:23–38.
Caldwell BK, Caldwell JC, Mitra SN, Smith WT. Searching for anoptimum solution to the Bangladesh arsenic crisis. Soc Sci Med2003b;56:2089–96.
Cheng Z, van Geen A, Jing C, Meng X, Seddique AA, Ahmed KM.Performance of a household-level arsenic removal system during4-month deployments in Bangladesh. Environ Sci Technol2004;38:3442–8.
Cherukurii J, Anjaneyulu Y. Design and development of low-cost,simple, rapid and safe, modified field kits for the visual detectionand determination of arsenic in drinking water samples. Int JEnviron Res Publ Health 2005(2):322–7.
Chowdhury M, Jakaria M. Testing of water for arsenic in Bangladesh.Science 1999;284:1622.
Deshpande LS, Pande SP. Development of arsenic testing field kit— atool for rapid on-site screening of arsenic contaminated watersources. Environ Monit Assess 2005;101(1–3):93–101.
Dhaka Community Hospital Trust. Arsenic Problem in Bangladesh;2005.URL: http://www.dchtrust.org/arsenic_problem.htm (Accessedon 7 May, 2005).
Grimes DA, Schulz KF. Uses and abuses of screening tests. Lancet2002;359(9309):881–4.
Hoque BA, Hoque MM, Ahmed T, Islam S, Azad AK, Ali N, et al.Demand-based water options for arsenic mitigation: an experiencefrom rural Bangladesh. Public Health 2004;118(1):70–7.
175Md. Jakariya et al. / Science of the Total Environment 379 (2007) 167–175
IARC. Some Drinking Water Disinfectants and Contaminants,including Arsenic. Monograph, vol. 84. Lyon, France: IARC;2004. 512 pp.
Jakariya M, Chowdhury AMR, Hossain Z, Rahman M, Sarker Q,Khan RI, et al. Sustainable community-based safe water options tomitigate the Bangladesh arsenic catastrophe— an experience fromtwo Upzilas. Curr Sci 2003;85(2):141–6.
Jakariya M, Rahman M, Chowdhury AMR, Rahman M, Yunus M,Bhiuya MA, et al. Sustainable safe water options in Bangladesh:experiences from the Arsenic Project at Matlab (AsMat). In:Bundschuh J, Bhattacharya P, Chandrasekharam D, editors.Natural Arsenic in Groundwater: Occurence, Remediation andManagement. London: Balkema; 2005. p. 319–30.
Jakariya M, von Brömssen M, Jacks G, Chowdhury AMR, AhmedKM, Bhattacharya P. Searching for sustainable arsenic mitigationstrategy in Bangladesh: experience from two upazilas. Int J EnvironPollut, in press.
Kapaj S, Peterson H, Liber K, Bhattacharya P. Human health effectsfrom chronic arsenic poisoning — a review. J Environ Sci HealthPart A 2006;41:2399–428. ___________________________doi:10.1080/10934520600873571.
Khan AW, Ahmad SK, SayedMH, Hadi SA, KhanMH, Jalil MA, et al.Arsenic contamination in ground water and its effect on humanhealth with particular reference to Bangladesh. J Prev Soc Med1997;16(1):65–73.
Khandaker NR. Limited accuracy of arsenic field test kit. Environ SciTechnol 2004;38(24):479A.
Linn S. A new conceptual approach to teaching the interpretation ofclinical tests. J Stat Educ 2004;12(3):1–9 URL: www.amstat.org/publications/jse/v12n3/linn.html (Accessed on 29 May, 2006).
Milton A, Rahman M. Environment pollution and skin involvementpattern of chronic arsenicosis in Bangladesh. J Occup Health1999;41:207–20.
Nickson RT, McArthur JM, Ravenscroft P, Burgess WG, Ahmed KM.Mechanism of arsenic release to groundwater, Bangladesh andWest Bengal. Appl Geochem 2000;15:403–13.
NRC. National Research Council: Arsenic in Drinking Water. 2001update. Washington: National Academy Press; 2001. p. 1–225.
Persson LA, Wall S. Epidemiology for public health. Department ofPublic Health and Clinical Medicine, Umea University; 2003. Printan Media, pp 272.
Rahman MM, Mukherjee D, Sengupta MK, Chowdhury UK, LodhDC, Roy S, et al. Effectiveness and reliability of arsenic fieldtesting kits: are the million dollar screening projects effective ornot? Environ Sci Technol 2002;36:5385–94.
Rahman M, Vahter M, Wahed MA, Sohel N, Yunus M, Streatfield PK,et al. Prevalence of arsenic exposure and skin lesions. A populationbased survey in Matlab, Bangladesh. J Epidemiol CommunityHealth 2006;60:242–8.
Smedley PL, Kinniburgh DG. A review of the source, behavior anddistribution of arsenic in natural waters. Appl Geochem2002;17:517–68.
Smith AH, Lingas EO, Rahman M. Contamination of drinking waterby arsenic in Bangladesh: a public health emergency. Bull WorldHealth Organ 2000;78:1093–103.
Steinmaus CM, George CM, Kalman DA, Smith AH. Evaluation oftwo new arsenic field test kits capable of detecting arsenic waterconcentrations close to 10 μg/L. Environ Sci Technol 2006;40(10):3362–6.
UNICEF. Arsenic mitigation in Bangladesh: media brief. Bangladesh:UNICEF; 2003.
van Geen A, Zheng Y, Versteeg R, StuteM, HornemanA, Dhar R, et al.Spatial variability of arsenic in 6000 tube wells in a 25 km2 area ofBangladesh. Water Res 2003;39:1140.
van Geen A, Cheng Z, Seddique AA, HoqueMA, Gelman A, GrazianoJH, et al. Reliability of a commercial kit to test groundwater forarsenic in Bangladesh. Environ Sci Technol 2005;39(1):299–303.
vonBrömssenM, JakariyaM, Bhattacharya P, AhmedKM,HasanMA,Sracek O, Jonsson L, Lundell L, Jacks G. Targeting low-arsenicaquifers in Matlab Upazila, Southeastern Bangladesh. Sci TotEnviron 2007;379:121–32. ____________________________doi:10.1016/j.scitotenv.2006.06.028.
Wahed MA, Chowdhury D, Nermell B, Khan SI, Ilias M, Rahman M,et al. A modified routine analysis of arsenic content in drinkingwater in Bangladesh by Hydride Generation-Atomic AbsorptionSpectrophotometry (HG-AAS). J Health Popul Nutr 2006;24(1):36–41.
Yu C, Ahsan H. Cancer burden from arsenic in drinking water inBangladesh. Am J Public Health 2004;94:741–4.
