household water disinfection in hurricane-affected...

American Journal of Public Health | Supplement 1, 2007, Vol 97, No. S1S130 | Research and Practice | Peer Reviewed | Ram et al.

RESEARCH AND PRACTICE

Objectives. Thousands of Louisiana residents were asked to boil water becauseof widespread disruptions in electricity and natural gas services after HurricaneRita. We sought to assess awareness of boil water orders and familiarity withhousehold water disinfection techniques other than boiling.

Methods. We conducted a cross-sectional survey in randomly selected mobilehome communities in Louisiana.

Results. We interviewed 196 respondents from 8 communities, which had boilwater orders instituted. Of 97 who were home while communities were still underorders to boil water, 30 (31%) were aware of the orders and, of those, 24 (80%)said the orders were active while they were living at home; of the 24, 10 (42%)reported boiling water. Overall, 163 (83%) respondents were aware of a methodof water disinfection at the household level: boiling (78%), chlorination (27%),and filtration (25%); 87% had a container of chlorine bleach at home.

Conclusions. Few hurricane-affected respondents were aware of boil water or-ders and of alternate water disinfection techniques. Most had access to chlorineand could have practiced household chlorination if disruption in natural gas andelectricity made boiling impossible. (Am J Public Health. 2007;97:S130–S135.doi:10.2105/AJPH.2006.094441)

Household Water Disinfection in Hurricane-Affected Communities of Louisiana: Implications for Disaster Preparedness for the General Public| Pavani K. Ram, MD, Elizabeth Blanton, MPH, Debra Klinghoffer, DrPH, Mary Platek, RD, Janet Piper, BS, Susanne Straif-Bourgeois, PhD, MPH,

Matthew R. Bonner, PhD, MPH, and Eric D. Mintz, MD, MPH

In 2005, a record-breaking 14 hurricanes struckthe Atlantic basin. Hurricane Katrina madelandfall along the US Gulf Coast on August29, 2005, and was responsible for more than1600 deaths and an estimated US$100 billionin property damage. Hurricane Rita struck theTexas and Louisiana coastlines on September24, 2005, also causing extensive propertydamage, albeit with minimal loss of life. In theaftermath of the hurricanes, thousands of resi-dents were left without basic municipal ser-vices, such as electricity and natural gas.1 Di-rect physical damage and prolonged poweroutages greatly affected municipal treatmentand distribution systems for drinking water, aswell as wastewater treatment and collection.

In areas affected by Hurricane Rita, watersystems that sustained a loss of electric poweror loss of pressure in the distribution systemplaced customers on orders to boil water. Re-gional health authorities provided water systemoperators with sample text for use in their com-munications with consumers regarding boilwater orders; operators were not mandated touse the text provided them (Figure 1). In addi-tion, the Louisiana Department of Health andHospitals updated an emergency news Website with boil water order information every24 hours.2 Local media, such as radio and tele-vision, also carried boil water order details.

During the early posthurricane periods,power outages made communicating boilwater orders by electronic media very difficult(Figure 2). Moreover, even if residents wereaware of the boil water orders, lack of electric-ity and natural gas service may have pre-vented them from complying. We sought toexamine the extent to which hurricane-affected residents in Louisiana were aware ofboil water orders imposed on their communi-ties, their compliance with the boil water or-ders, and knowledge of household water disin-fection techniques other than boiling.

METHODS

On the day after Hurricane Rita, approxi-mately 80 community water systems in Cal-casieu and Cameron Parishes in southwest-ern Louisiana placed customers on orders toboil water. We conducted a cross-sectionalsurvey among residents of southwesternLouisiana mobile home communities duringthe week of November 7, 2005, approxi-mately 6 weeks after the hurricane. We enu-merated mobile home communities in Cal-casieu and Cameron Parishes that wereinhabitable after Hurricane Rita. Of 26 sur-viving communities, 4 had fewer than 20mobile homes each and were excluded. Arandom-number generator was used to select8 mobile home communities at random fromthe remaining 22 for inclusion in the survey.Trained interviewers attempted to interview1 adult from each inhabited mobile homewithin the randomly selected mobile homecommunities. Respondents were eligible forparticipation in the survey if the following

criteria were met: she or he was aged 18years or older at the time of interview, theinterview location was the primary residenceof the respondent at the time of interview,and the respondent had been a resident ofCalcasieu or Cameron Parishes before Hurri-cane Rita made landfall on September 24,2005. If no adult was home at the time ofthe interviewer’s visit, 1 additional attemptwas made to interview an eligible respondent.

The questionnaire addressed demograph-ics, awareness of and compliance with boilwater orders, knowledge of household waterdisinfection techniques, knowledge regardingwaterborne and hand hygiene-related disease,and the use of waterless hand sanitizer forhand hygiene. No personal identifier informa-tion was captured on the survey instrument.Interviewers obtained verbal consent fromeach study participant and left an informationsheet regarding the study with participants.All of the questions were read aloud by theinterviewer, who recorded responses on thequestionnaire.

Supplement 1, 2007, Vol 97, No. S1 | American Journal of Public Health Ram et al. | Peer Reviewed | Research and Practice | S131


FIGURE 1—Boil advisory text recommended to water systems after Hurricane Rita,Louisiana, 2005.

RESULTS

We interviewed 196 respondents from 8mobile home communities (designated com-munities A—H) randomly selected for inclu-sion in the survey. The median age of re-spondents was 46 years (range: 19–90years); 64% were women, and 75% hadcompleted high school. Electric stoves wereused by 79%, and gas stoves were used by21%. At least 1 television was present in98% of homes, and 46% of respondents re-ported having access to the Internet at homebefore Hurricane Rita. A question aboutwhich diseases could be caused by drinkingcontaminated water yielded the following

responses: diarrhea (33%), parasites (7%),and skin infections (2%).

Nearly all of the respondents (97%) re-ported evacuating because of Hurricane Rita,with a median duration of evacuation of 15days (range: 2–60). In preparation for thehurricane, 119 respondents (61%) reportedpurchasing items such as bottled water (92 of 119; 77%) and canned foods (73 of119; 61%). There was no difference in self-reported use of waterless hand sanitizer be-fore (63%) and after (61%) the hurricane. Before Hurricane Rita, the community watersupply was the primary source of drinkingwater for 57% of respondents, and bottledwater was the primary source for 43%. After

the hurricane, the community water supplywas the primary drinking water source for31%, and bottled water for 69%.

Boil water orders were put into place in all8 of the mobile home communities on Sep-tember 24, 2005, the day that HurricaneRita made landfall; the duration of the boilwater orders varied from 14 to 43 days(Table 1).2 Overall, 39% of all respondentswere aware that their community had beenunder orders to boil water after HurricaneRita. The 77 respondents who were awareof the boil water orders reported obtainingthat information through a variety of media:leaflets (32%), word of mouth (23%), televi-sion (21%), radio (15%), newspaper (5%),and the Internet (1%). Many residents ofcommunity H, where 77% of respondentswere aware of the boil water orders, reportedthat the community manager walked fromhome to home to drop off leaflets and per-sonally explained the need for boiling water.Information provided on leaflets, as reportedby community members, varied from com-munity to community.

There was no association between gender,age group, years of education, or access to theInternet at home and the awareness of boilwater orders. Of the 196 respondents, all ofwhom were living in mobile home communi-ties that had boil water orders in place forsome time after Hurricane Rita, 77 (39%)were aware of the orders. Among the 77, 59(77%) said they were home when the boilwater orders were in place. Of those 59, 27(46%) reported boiling water. Twenty-three(39%) reported being able to boil water butdid not, and 9 (15%) reported being unable toboil water. Awareness of boil water orders var-ied significantly by mobile home community,ranging from 9% in community A to 77% incommunity H (P = .0005; Table 1). The rateof boiling water was highest in communitiesG and H at 100% and 63%, respectively; thedifference in the rate of boiling water betweenthe communities was not statistically signifi-cant. However, the number of respondents ineach community was small, ranging from just2 to 21. Bottled water was the primary watersource after the hurricane for 59% of respon-dents who reported boiling water and 81% ofrespondents who reported not boiling water(P =.07). Thus, a total of 25 respondents, or



FIGURE 2—Communication of availability of water and sewer services by signboardadjacent to hurricane-affected community, Louisiana, 2005.

TABLE 1—Awareness of Boil Water Orders by Mobile Home Community: Louisiana, 2005

No. of Duration of Boil Median Duration of Aware of Boil Boiled Water,Community Respondents Water Orders (d)a Evacuation (d) Orders, no. (%) no. (%)b

A 34 43 16 3 (9) 0 (0)

B 27 19 18 6 (22) 1 (33)

C 8 28 26 3 (38) 1 (50)

D 8 24 15 3 (38) 1 (50)

E 74 14 14 30 (40) 6 (29)

F 12 28 16 6 (50) 1 (20)

G 11 14 14 8 (73) 7 (100)

H 22 14 15 17 (77) 10 (63)

TOTAL 196 15 78 (40) 27 (47)

aAll boil water orders became effective on September 24, 2005.bThis proportion was calculated by dividing the number of respondents who reported boiling water by the number ofrespondents who were aware of boil water orders and who believed that the boil water orders were in place while they wereliving in the community.

13% of the study population, consumed waterthat was not bottled or boiled.

We defined “need to know” about the boilwater orders as not having evacuated or hav-ing returned home while boil water orderswere still in place. The determination ofwhether evacuees returned while orderswere active was made by comparing the self-reported duration of evacuation to the number

of days that boil water orders were in placefor the respective communities according tothe Louisiana Department of Health and Hos-pitals. A total of 97 respondents (49%) hadnot evacuated or had returned to their homeswhile boil water orders were still active fortheir community, thus needing to know aboutthe orders; 99 respondents (51%) returnedonly after boil water orders were lifted and,

thus, did not need to know about the orders(Figure 3) Awareness of boil water orders waslower among those who needed to know thanamong those who did not need to know(P < .05). A similar proportion of both groupsbelieved that they were living in the commu-nity while boil water orders were active and,among those, a similar proportion reportedcomplying with the orders.

Among the total of 27 persons whoboiled water, 70% used a kitchen stove;22% used another stove, such as a hotplateor crawfish boiler; 7% of respondents couldnot remember which stove was used; and1% did not answer the question. Whenasked how long water was supposed to beboiled, 52% answered more than 5 min;11% specified a duration between 1 and 5 min;4% specified a duration less than 1 min;7% answered 1 min; and 26% reported notknowing the proper duration for boiling.Boiled water was used for various purposesincluding cooking (46%), hand washing orbathing (42%), drinking (39%), brushingteeth (23%), or preparing drinks (12%).Other uses were reported by 26%, andthese included washing dishes, preparingbaby formula, and shaving.

Overall, 163 respondents (83%) reportedknowing at least 1 method of water disinfec-tion at the household level. The followingtechniques were specified by these respon-dents: boiling (78%), chlorination (27%), fil-tration (25%), and iodination (4%). Amongthe 44 respondents who suggested chlorina-tion as a household disinfection technique,only 1 (2%) correctly named the dose ofchlorine recommended by the LouisianaDepartment of Health and Hospitals (one-eighth teaspoon of commercial bleach pergallon of water); 21 (50%) stated a dosethat was higher than recommended, 6(14%) stated a dose that was lower than rec-ommended, and 13 (31%) reported notknowing the recommended chlorine dosefor drinking water disinfection. Ninety-twopercent of respondents reported having acontainer of bleach at home before Hurri-cane Rita, and 87% said that they hadbleach after Hurricane Rita. Seventeen per-cent of respondents reported having tinctureof iodine or iodine drops at home beforeand after the hurricane.



aNeed to know: Residents who had not evacuated or had returned while boil orders were still in placebDid not need to know: Residents who had returned after boil orders were withdrawn

FIGURE 3—Awareness of boil water orders among hurricane-affected respondents:Louisiana, 2005.

DISCUSSION

This study highlights the challenges of com-municating to the public about drinking waterissues during and after large-scale disasters.Awareness of boil water orders was lowamong hurricane-affected respondents andparticularly low among those who needed toknow because they had not evacuated or hadreturned before the orders were lifted. Thisfinding reflects the need to identify effectiveand efficient means of mass communicationduring the early phase of a water emergency.Many who returned after boil water orderswere lifted believed that the orders remainedactive, indicating the equal failure in ade-quately communicating the termination of thewater emergency. Our findings also under-score the need to enhance the public’s knowl-edge of the consequences of drinking contami-nated water and alternative household waterdisinfection techniques given that prolongedand widespread disruptions in electricity andnatural gas services may prevent boiling.

Previous studies have documented the diffi-culty of communicating to the public regarding

boil water orders in the context of known con-tamination or an outbreak of disease amongconsumers of municipal water supplies.3,4 Thelack of accurate knowledge regarding boilwater orders among our survey respondentshas several possible explanations. Interruptionsin electricity service may have prevented per-sons in hurricane-affected areas from accessingthis information through usual media channels,such as television. Although not statistically sig-nificant, perhaps because of limited sample size,we found that, among persons aware of boilwater orders, those who boiled water usedbottled water for drinking less frequently thanthose who did not boil water. Residents maynot have heeded announcements to boil waterif they had purchased bottled water in prepara-tion for the hurricane or if they received bot-tled water from emergency response teamsafter the hurricane. In disasters that occur with-out warning, residents may not be able to stockup on bottled water in advance and may beforced to use water from the tap or othersources, irrespective of its microbiologic quality.

The World Health Organization’s Guidelinesfor Drinking-Water Quality recommend that

boil water orders be issued only if the respon-sible public health authority is convinced ofan ongoing risk to the public’s health that out-weighs the risks associated with boiling, whichinclude scalds and burns.5 The decision toissue boil water orders in Calcasieu andCameron Parishes after Hurricane Rita wasbased on loss of electric power to the watertreatment plants or loss of water pressure inthe distribution system and not necessarily onthe basis of confirmed microbiologic contami-nation. Perceptions of risk regarding the com-munity water systems may have been low, be-cause affected communities were not told ofconfirmed microbial contamination or of po-tential adverse health effects that might resultfrom drinking the municipal water supply.Moreover, general awareness of the healthconsequences of drinking contaminated waterwas poor among survey respondents, whichmay have contributed to low risk perception.In an assessment of adherence to boil waterorders in the context of cryptosporidial con-tamination of the municipal water supply in anEnglish community, participants recommendedthat describing the health effects of drinkingunboiled water would have been useful.4

Hundreds of community water systemswere affected by the 2005 hurricane season.Although the water system companies wereprovided with sample text to address the boilwater orders (Figure 1), they were not man-dated to use the provided text. Because sev-eral weeks had elapsed since the time of thehurricane, we were unable to assess theexact messages delivered by each water sys-tem company supplying the mobile homecommunities in this study. Thus, we cannotknow whether awareness of the boil waterorders was hampered by a failure to deliverthe messages or by use of unclear messages.Leaflets and word of mouth were the mostcommonly reported sources of informationregarding boil water orders among hurri-cane-affected respondents. Such techniquesmay be most effective when usual media,such as television and radio, are renderedunavailable during emergency situations.However, these strategies may be difficultto scale up in the midst of very large disas-ters, in which hundreds of thousands ofpersons may be affected. Moreover, partic-ularly when word of mouth is used as a



TABLE 2—Recommendations for Household Water Disinfection From Federal and Hurricane-Affected State Agencies: United States, 2005

Recommended Disinfection With Liquid Chlorine

Duration of Type of chlorine Chlorine Tincture of Agency or State Boiling recommended Clear water Cloudy water Tablets Iodine Iodine Tablets

Environmental Protection Agency: 1 min 1% (nonscented, 10 drops per quart; Double the 1 tablet per 5 drops 1 tablet per

www.epa.gov/safewater/faq/emerg.html household 40 drops per amount gallon gallon

chlorine gallon

bleach)

4—6% 2 drops per quart; Double the

8 drops per amount

gallon

(1/8 teaspoon)

7—10% 1 drop per quart; Double the

4 drops per amount

gallon

Centers for Disease Control and Prevention: 1 min Household 1/8 tsp per gallon 1/4 tsp per Follow . . . Follow

http://www.bt.cdc.gov/disasters/foodwater.asp chlorine (approximately gallon directions directions

bleach 0.75 mL) (approximately on label

1.50 mL) on label

Alabamaa: http://www.adph.org/environmental . . . . . . . . . . . . . . . . . .

Louisiana: http://www.dhh.louisiana.gov/offices/faq.asp?ID=145 1 min Unscented, 1/8 tsp per gallon 1/4 tsp per . . . 5 drops . . .

liquid gallon

chlorine

bleach

Mississippi: www.health.ms.gov/msdhsite/index.cfm/23,0,148,html 1 min Unscented, 1/8 tsp per gallon . . . . . . . . . . . .

ordinary

household

chlorine

bleach

Texas: http://www.dshs.state.tx.us/news/releases/20050926.shtm 1 min Ordinary 1/8 tsp (approximately . . . . . . Follow . . .

household 8 drops) directions

bleach per gallon on label

aThe Alabama Department of Health Web site has a link to the Centers for Disease Control and Prevention emergency hurricane health information.

communication tool or when each watersystem is left to produce its own communica-tion materials, there may be a risk of passingalong incomplete or inaccurate information.Managers of water systems would benefitfrom receiving communications materials,such as ready-to-post signs and letters, frompublic health officials regarding boil waterorders. The text in such messages should beaccessible to audiences of wide literacy skillsand should contain information regardingalternatives to boiling for household waterdisinfection. Such pre-prepared materialsmay minimize the dissemination of misinfor-mation about water safety issues.

A majority of survey respondents drank bot-tled water to meet their drinking water needsand, because of its availability, many who wereaware of the boil water orders may not havefelt the need to comply with those orders. Bot-tled water was commonly purchased by respon-dents in anticipation of Hurricane Rita, andmany relief agencies carried out distribution ef-forts in the aftermath of the hurricane. Al-though bottled water distribution may have metthe need for potable water in these communi-ties, it is unlikely to have served residents veryearly in the posthurricane period. First, distribu-tion centers must be established and stockedwith necessary supplies. Next, the location of

centers must be communicated widely; as withthe communication of boil water orders, the dis-semination of such information might be signifi-cantly hampered when electricity services aredisrupted and residents cannot access usualsources of information, such as television andradio. Because bottled water and other suppliesare generally distributed at centralized locations,affected residents would have to be able to walkto the distribution center or have access to afunctioning vehicle to collect necessary supplies.All of these potential barriers to procuring bot-tled water suggest that residents of disaster-prone areas should be aware of when and howto disinfect water at the household level.



Boiling water was the most widely recog-nized technique for household water disinfec-tion among survey respondents. With electric-ity and natural gas services interrupted, itwould have been nearly impossible for resi-dents to comply with boil water orders in theimmediate posthurricane period. Small-scalewater purifiers, such as those available fromsporting or camping stores, may be useful inthe emergency context but are likely to beprohibitively expensive for some. The publicshould be aware of inexpensive emergencywater disinfection techniques, such as chlori-nation or iodination. Only a minority of re-spondents was aware of these alternatives toboiling. Chlorination of water at the house-hold level has been shown repeatedly to im-prove the microbiological quality of drinkingwater and to decrease the risk of diarrhea inthe context of poorly functioning municipalwater supply systems.6,7

Among respondents who were aware thatchlorine could be used to disinfect water,knowledge of the recommended dose of chlo-rine per gallon of water was poor. Guidancefrom federal and state public health and watersafety authorities may be confusing to UnitedStates residents who are not in the habit ofcollecting water in containers and addingbleach to make it safe to drink (Table 2). TheUS Environmental Protection Agency providesguidance regarding 3 different concentrationsof bleach, and other federal and state agenciesrefer to “household bleach.” Many US resi-dents may not be aware that bleach can in-deed be purchased in varying concentrations.Moreover, the common recommendation touse one-eighth teaspoon of bleach per gallonof water may be challenged by the absence ofone-eighth teaspoons in sets of baking mea-sures typically available in American homes.Notably, the overwhelming majority of respon-dents had a container of bleach at home and,thus, had the capacity to disinfect water bychlorination before and after the hurricane.Emergency preparedness plans should priori-tize enhancing the public’s awareness of chlo-rination as a household disinfection techniqueand preparing user-friendly guidelines on ap-propriate doses of bleach for such disinfection.

This study faced some key limitations. Thesurvey population was restricted to thosewhose homes were inhabitable and who had

returned to the area; hence, a potentially morevulnerable population of persons with greaterproperty damage would not have been in-cluded in the survey. The survey was con-ducted 6 weeks after Hurricane Rita, whichmay have limited respondents’ ability to recallcompliance with the boil water orders accu-rately. All of the practices were self-reportedand, thus, the practice of boiling water mayhave been overestimated by respondents. Fi-nally, the sample size within each mobilehome community was small and preventedcomparison of the boiling water practice be-tween communities. Despite these importantlimitations, this study documented importantdeficiencies in communication regarding theinitiation and cancellation of boil water ordersand alternative techniques to boiling.

Hurricanes often present with several days ofadvance notice, allowing residents and govern-ment officials to prepare by stocking up on nec-essary supplies, disseminating important healthinformation, or evacuating. Other natural disas-ters, such as earthquakes, do not afford this lux-ury of advance warning. With several seismicallyactive regions within the country and forecastsof increasingly intense hurricane seasons tocome, US residents and government agenciesshould be prepared to cope with disasters andwith the consequences for drinking water sup-ply. Enhancing awareness of household waterdisinfection strategies, including boiling and chlo-rination, may protect the public’s health in theaftermath of large-scale disasters.

About the AuthorsPavani K. Ram, Elizabeth Blanton, and Eric D. Mintz arewith the Diarrheal Diseases Epidemiology Team, Centersfor Disease Control and Prevention, Atlanta, Ga. Pavani K.Ram, Mary Platek, and Matthew R. Bonner are also withthe Department of Social and Preventive Medicine, Univer-sity at Buffalo, Buffalo, NY. Debra Klinghoffer, Janet Piper,and Susanne Straif-Bourgeois are with the Office of PublicHealth, Louisiana Department of Health and Hospitals,Lake Charles and New Orleans, La.

Requests for reprints should be sent to Pavani K. Ram,Department of Social and Preventive Medicine, School ofPublic Health and Health Professions, University at Buf-falo, 3435 Main St, Room 270 Farber Hall, Buffalo,NY 14214 (e-mail: [email protected]).

This research article was accepted October 21, 2006.

ContributorsP. K. Ram, M. R. Bonner, and E. D. Mintz originatedthe study. P.K. Ram developed the protocol, oversawdata collection and data analysis, and drafted the article. E. Blanton, D. Klinghoffer, and M. Platek

contributed to data collection, management, and analy-sis and reviewed the article. J. Piper and S. Straif-Bour-geois contributed to data analysis and article review.M.R. Bonner and E.D. Mintz contributed to interpreta-tion of results and article revision.

AcknowledgmentsThe National Foundation for the Centers for DiseaseControl and Prevention graciously provided funding forthis project but did not provide technical input to studydesign, data collection or analysis, or preparation or ap-proval of this article.

The authors thank the Louisiana residents whotook time to speak with them during very difficult days.They appreciate greatly the fieldwork and humor of thefollowing University at Buffalo PhD students: TheodoreBrasky, Gabrielle Foley, Lisa Leuchner, SreenivasaSarikonda, and Michael Yeh.

Human Participant ProtectionThis study protocol was deemed exempt from institu-tional review under CFR 46.101 (b)(2) by the socialand behavioral sciences institutional review board atthe University of Buffalo and the Human ResearchProtection Office of the Centers for Disease Controland Prevention.

References1. Centers for Disease Control and Prevention. As-sessment of health-related needs after Hurricanes Kat-rina and Rita–Orleans and Jefferson Parishes, New Or-leans area, Louisiana, October 17–22, 2005. MMWRMorbid Mortal Wkly Rep. 2006;55:38–41.

2. Louisiana Department of Health and Hospitals.Boil orders related to Hurricanes Katrina & Rita. BatonRouge, LA: Louisiana Department of Health and Hos-pitals; 2005.

3. Angulo FJ, Tippen S, Sharp DJ, et al. A commu-nity waterborne outbreak of salmonellosis and the ef-fectiveness of a boil water order. Am J Public Health1997;87:580–584.

4. O’Donnell M, Platt C, Aston R. Effect of a boilwater notice on behaviour in the management of awater contamination incident. Commun Dis PublicHealth 2000;3:56–59.

5. World Health Organization. Guidelines for Drinking-Water Quality, Third Edition. Geneva, Switzerland:World Health Organization; 2004.

6. Luby SP, Agboatwalla M, Hoekstra RM, RahbarMH, Billhimer W, Keswick BH. Delayed effectivenessof home-based interventions in reducing childhood di-arrhea, Karachi, Pakistan. Am J Trop Med Hygiene2004;71:420–427.

7. Semenza JC, Roberts L, Henderson A, Bogan J,Rubin CH. Water distribution system and diarrhealdisease transmission: a case study in Uzbekistan. Am JTrop Med Hygiene 1998;59:941–946.

household water disinfection in hurricane-affected...

Documents