Risk Analysis, Vol. 30, No. 12, 2010 DOI: 10.1111/j.1539-6924.2010.01463.x

Prioritizing Environmental Health Risks in the UAE

Henry H. Willis,1∗ Jacqueline MacDonald Gibson,2 Regina A. Shih,3 Sandra Geschwind,4

Sarah Olmstead,4 Jianhui Hu,4 Aimee E. Curtright,1 Gary Cecchine,3 and Melinda Moore3

This article presents the results of a comparative environmental risk-ranking exercise thatwas conducted in the United Arab Emirates (UAE) to inform a strategic planning processled by the Environment Agency-Abu Dhabi (EAD). It represents the first national-level ap-plication of a deliberative method for comparative risk ranking first published in this journal.The deliberative method involves a five-stage process that includes quantitative risk assess-ment by experts and deliberations by groups of stakeholders. The project reported in this ar-ticle considered 14 categories of environmental risks to health identified through discussionswith EAD staff: ambient and indoor air pollution; drinking water contamination; coastal wa-ter pollution; soil and groundwater contamination; contamination of fruits, vegetables, andseafood; ambient noise; stratospheric ozone depletion; electromagnetic fields from powerlines; health impacts from climate change; and exposure to hazardous substances in indus-trial, construction, and agricultural work environments. Results from workshops involving73 stakeholders who met in five separate groups to rank these risks individually and collab-oratively indicated strong consensus that outdoor and indoor air pollution are the highestpriorities in the UAE. Each of the five groups rated these as being among the highest risks.All groups rated soil and groundwater contamination as being among the lowest risks. Insurveys administered after the ranking exercises, participants indicated that the results of theprocess represented their concerns and approved of using the ranking results to inform policydecisions. The results ultimately shaped a strategic plan that is now being implemented.

KEY WORDS: Environmental health; priority setting; risk ranking

1. INTRODUCTION

In recent years, dozens of nations around theworld have developed national environmental healthaction plans.(1) These documents are intended toidentify priority environmental risks to human healthin each country and an agenda of actions for reducing

1 RAND Corporation, Pittsburgh, PA, USA.2 Department of Environmental Sciences and Engineering,

Gillings School of Global Public Health, University of North Car-olina, Chapel Hill, NC, USA.

3 RAND Corporation, Arlington, VA, USA.4 RAND Corporation, Santa Monica, CA, USA.∗Address correspondence to Henry Willis, RAND Corporation,

4570 Fifth Avenue, Suite 600, Pittsburgh, PA 15213, USA; tel:+1-412-683-2300; fax: +1-412-683-2800; hwillis@rand.org.

those risks. The impetus for preparation of these ac-tion plans was the 1992 United Nations Conferenceon Environment and Development, held in Rio deJaneiro on the 20th anniversary of the first globalenvironmental conference. Nations participating inthe Earth Summit, as the Rio conference was known,committed to developing national-level action plansto reduce human impacts on the environment and en-vironmental impacts on human health.(2)

Consistent with this international mandate, in2007 the United Arab Emirates (UAE) initiated aproject to develop a national environmental healthaction plan. The UAE, founded in 1971, is ayoung nation, and its environmental institutions andpolicies are still emerging, with the first national

1842 0272-4332/10/0100-1842$22.00/1 C© 2010 Society for Risk Analysis

Prioritizing Environmental Health Risks in the UAE 1843

environmental law enacted in 1999. Hence, at the na-tional level, the UAE has not previously conducted acomprehensive environmental assessment and prior-ity setting exercise.

This article presents the results of a compara-tive risk-ranking exercise that was conducted in theUAE to inform the national environmental healthstrategic planning process. The work described hereultimately shaped the strategic plan that is now be-ing implemented.(3) To help the UAE identify itspriority environmental health risks, the authors ofthis article used the “deliberative method for rankingrisks,” a method for which the initial pilot test resultswere published in this journal.(4−7) The deliberativemethod provides a systematic process for combiningquantitative risk assessment with stakeholder partic-ipation to evaluate the relative importance of risks tohealth and safety, from the perspective of differentstakeholders. Thus, this exercise represents the firstnational-level application of this deliberative methodfor comparative risk ranking.

1.1. Emerging Environmental Risks to PublicHealth in the UAE

The UAE, a confederation of seven emirates onthe Arabian Gulf, has undergone industrial develop-ment at a rate that is unprecedented in the history ofthe world. Until oil was discovered in 1958 and ex-ports began in 1962, the UAE was a sparsely inhab-ited region populated primarily by Bedouin nomadsand fishermen. The main industry was pearl diving.In 1968, the first year for which official census figuresare available, the UAE population was 180,000.(8)

The population at the end of 2009 was estimatedat 6 million,5 a growth of more than 3,000% in just40 years.

Population growth in the UAE has been accom-panied by an extraordinary increase in per-capita in-come and prosperity. The per-capita gross domesticproduct in the UAE in 2009 was estimated at $41,800,the 17th highest in the world.(9) This increased pros-perity has also brought increased life expectancy,with average male and female life expectancies in thecountry now at 74 and 79 years, respectively. Wa-terborne and other environmentally related diseasesassociated with lack of sanitation and hygiene havebeen virtually eliminated.(10)

5 http://uaeinteract.com/docs/UAE population to rise in 2010 to7.5million/39347.htm.

Although the UAE’s dramatic increase in pros-perity has brought substantial benefits, it also has cre-ated new environmental risks to public health thatare gaining increasing attention in the country. As anexample, Li et al. recently estimated that 545 prema-ture deaths per year in the UAE can be attributedto elevated levels of particulate matter in ambientair.(11) Sources of the increased load of particulatematter include petroleum refineries, heavy industry,power plants, and traffic—all familiar contributors todegraded ambient air quality in developed nations.Other examples of environmental risks that have ac-companied development include increased loads ofhazardous and solid wastes and sewage. In addition,because of the warm climate residents spend signif-icant time inside poorly ventilated, air-conditionedbuildings, in which indoor air pollutants may accu-mulate.

1.2. Overview of This Analysis

The remainder of the article is organized as fol-lows. Section 2 describes the motivation for, devel-opment of, and steps in the deliberative method forranking risks and how these steps were implementedin the UAE. Section 3 presents the results of the risk-ranking exercise in the UAE, an evaluation of theprocess used to elicit the rankings, and a compari-son of the results to evaluations published during thedevelopment of the deliberative method for rankingrisks. Section 4 describes how these results were laterincorporated into the environmental health strategicplanning process in the UAE.

2. THE DELIBERATIVE METHODFOR RANKING RISKS

The first step in guiding the UAE towards anenvironmental health strategic plan was to iden-tify a method for prioritizing environmental risksto health that would be both scientifically defensi-ble and would include stakeholder participation. Themethod would need to consider the quantitative as-pects of risks while also incorporating other, non-quantitative characteristics (such as controllabilityand the catastrophic nature of some risks with verylow probabilities of occurrence). Approaches to com-parative risk assessment, such as the WHO burden ofdisease methodology,(10) address the quantitative as-pects of risks. However, few address the other char-acteristics that affect concerns about risks. In fact, athorough review of the literature on comparative risk

1844 Willis et al.

assessment revealed that the deliberative method forranking risks is the only method that meets these cri-teria and that has been empirically tested in large pi-lot studies.

2.1. Development of the Deliberative Methodfor Ranking Risks

The deliberative method for ranking risksemerged from projects in the United States in the1980s and 1990s to set environmental priorities, be-ginning with U.S. Environmental Protection Agency(EPA) 1987 report Unfinished Business: A Compar-ative Assessment of Environmental Priorities.(12,13)

These early comparative risk projects in the UnitedStates were criticized for being ad hoc and lack-ing sufficient stakeholder participation.(14) The com-parative risk projects undertaken in Europe in the1990s had similar limitations, leading participants in aNATO- and WHO-sponsored workshop focused onthese assessments to recommend the developmentof systematic approaches for incorporating quantita-tive and qualitative risk information into such exer-cises.(1)

As a result of the limitations identified in U.S.risk-ranking exercises, the U.S. Office of Science andTechnology Policy (part of the Executive Office ofthe President) convened a workshop involving ex-perts in fields related to risk assessment to discusswhether a more systematic process for setting priori-ties could be developed. In preparation for this work-shop, Morgan et al.(15) developed a proposal for arisk-ranking method that includes both quantitativeand qualitative metrics and a systematic process forstakeholder participation. The resulting publicationsreferred to the process as the “deliberative methodfor ranking risks.”

The researchers who developed the deliberativemethod sought to design a ranking approach thatwould

1. use existing scientific knowledge (includingrecent findings from research in decision the-ory, risk analysis, and the psychology of riskcommunication);

2. help those doing the ranking “to systemat-ically consider all relevant information,” in-cluding available quantitative risk informa-tion;

3. help participants construct rankings consis-tent with their own values;

4. ensure that participants understand and aresatisfied with the procedure; and

5. determine the level of agreement and sourcesof disagreement among participants in therisk-ranking process.

This deliberative method for ranking risks wastested in a set of experiments involving 218 profes-sional risk managers(5,6) and expanded and tested forcomparing risks that affect both ecological and hu-man health endpoints.(7)

2.2. Steps in the Deliberative Methodfor Ranking Risks

The deliberative method for ranking risks in-volves five steps, shown in Fig. 1. The first step, defin-ing and categorizing the risks to be ranked, involvessorting risks into categories that can be meaningfullycompared. The second step, identifying the risk at-tributes, involves enumerating the metrics (such asnumbers of deaths and illnesses of various kinds,controllability, level of scientific understanding, andothers) that will be used to compare the risks. Thethird step, describing the risks, requires conductingsystematic, quantitative assessments of each risk andsummarizing that information, along with qualitativerisk attributes, in four-page “summary sheets” pre-pared in a standard format. The fourth step, perform-ing the risk ranking, involves holding focus groupsessions with the full range of stakeholders con-cerned about environmental risks to review the risksummary sheets and rank the risks. The last stepis the quantitative analysis of results of the rank-ing exercise. The next section provides more detailson each of these steps and explains how we imple-mented the method in the UAE.

2.3. Implementing the Deliberative Methodin the UAE

2.3.1. Step A: Defining and Categorizing Risks inthe UAE

The Environment Agency-Abu Dhabi (EAD)and the World Health Organization Centre for En-vironmental Health Activities (WHO-CEHA) initi-ated the strategic planning process that our assess-ment was designed to inform. These organizationsprovided our research team with a listing of poten-tial environmental health concerns that served asthe starting point for defining which risks to rank.The third column in Table I shows the concerns

Prioritizing Environmental Health Risks in the UAE 1845

Step CDescribe the risks in risk summary sheetsin terms of the attributes.

Step DSelect participants and perform the risk rankings.

Step EDescribe the issues identified and the resulting rankings.

Step ADefine and categorize the risks to be ranked.

Step BIdentify the risk attributes that should be considered.

Step ADefine and categorize the risks to be ranked.

Step BIdentify the risk attributes that should be considered.

Fig. 1. Steps in the deliberative methodfor ranking risks. (Adapted from Floriget al.(5))

Table I. Framework for CategorizingEnvironmental Risks to Health

in the UAE

Exposure Route Risks Evaluated for this ProjectIssues Specified by the

UAE and WHO

Air (breathing) 1. Ambient (outdoor) air pollution Ambient air pollution2. Indoor air pollution in residential

environmentsIndoor air pollution

Water (drinking, 3. Drinking water contamination Waterbathing, inhalingdroplets)

4. Coastal water pollution

Soil (dermal contactfollowed byingestion)

5. Soil and associated groundwaterpollution due to solid andhazardous waste disposal

Waste management

Food (eating) 6. Seafood contamination Food7. Produce (fruits and vegetables)

contaminationSound and 8. Ambient noise above healthful Noise pollution

electromagnetic levelsradiation (contactingin ambient

9. Ultraviolet radiation abovenatural levels as a result of

Stratospheric ozonedepletion

environment) stratospheric ozone depletion10. Electromagnetic fields from power

linesElectromagnetic fields

Occupational 11. Industries Chemical safetyenvironments 12. Construction

13. Agriculture

Global climate change 14. Global climate change Global climate change

identified by the EAD and the WHO-CEHA. To theextent possible, we categorized these risks accordingto the traditional exposure routes (air, water, soil,and food) typically used in environmental risk assess-ments. The first two columns of Table I show how weorganized the list of risks received from the EAD andWHO-CEHA to correspond approximately to expo-sure route. We also added categories for exposureto excess noise and electromagnetic radiation frompowerlines. Finally, to encompass all of the concernsexpressed by the EAD and the WHO, we includedcategories for exposure risks in occupational envi-ronments (which may occur via multiple exposureroutes) and climate change.

Two risks we were asked to consider are ex-cluded from Table I: vector-borne disease andionizing radiation. We surveyed infectious diseaseinformation from the UAE and found no evidenceof diseases that would be expected if animal vectorswere a sanitation concern. Similarly, we found no ev-idence other than radon, which we consider as anindoor air pollutant, to warrant inclusion of ioniz-ing radiation. Data in the scientific literature indicatethat background radiation levels across the UAE arelower than global averages. Although risks of expo-sure to ionizing radiation may exist from future acci-dents related to nuclear energy or nuclear materialsprocessing, this exposure route was not considered

1846 Willis et al.

because it is more related to emergency responsethan to environmental health risk management andbecause the UAE currently does not have nuclearpower plants (although in 2009 the UAE govern-ment entered into an agreement with a South Koreanconsortia led by Korea Electric Power Corporationto construct the UAE’s first nuclear power plant by2017).

For the purposes of this project, the category per-taining to hazardous substances in occupational en-vironments was divided into three distinct subsets,corresponding to exposures in (1) industry, (2) con-struction, and (3) agriculture. Risks in each of theseoccupational subgroups were evaluated separately.The expansion of that single category into three cat-egories resulted in a final list of 14 hazards to beassessed in the risk-ranking exercise, shown in thesecond column of Table I.

2.3.2. Step B: Identify the Risk Attributes

As is well known in the field of risk analysis, alarge number of characteristics may influence peo-ple’s perceptions of health, safety, and environmen-tal hazards.(16−18) Examples of these characteristicsinclude number of deaths or illnesses, age of thosemost affected, latency of the illness, nature of theillness, and ability to control exposure to the risk.The second step of the deliberative method involvesidentifying the risk attributes that should be con-sidered when comparing the risks. Empirical studieson risk perception have demonstrated that while awide variety of characteristics are important in peo-ple’s perceptions of risk, the attributes can be sortedinto three categories: number, knowledge, and dread.Within each category, the attributes are highly corre-lated, so that if a risk-ranking exercise considers just afew attributes from each category, the ranking resultswill not depend significantly on the specific attributesthat are assessed.(16,15) Table II shows examples ofattributes in each of these three categories. For theUAE risk-ranking exercise, we relied on attributesidentified as most suitable in previous pilot tests ofthe deliberative method.(5)

2.3.3. Step C: Describe the Risks in the UAE inTerms of the Attributes

To assess each risk according to the attributes inTable II, we reviewed more than 400 technical re-ports and scientific publications that described expo-sures and health consequences to environmental haz-

ards. Where possible, the literature review focusedon studies specific to the UAE. When UAE-specificliterature did not exist, the team next turned to stud-ies of environmental health risks in other countriesin the Arabian Gulf based on an assumption thatsimilarities in economic, sociopolitical, and climaticconditions would result in comparable outcomes. Fi-nally, in some cases, the review had to rely on infor-mation about exposures and health effects in otherregions (e.g., Europe or the United States) if a casecould be made that the exposures could be expectedto be comparable between the two regions.

The results of this literature review were sum-marized in four-page risk summary sheets that de-scribe the 14 risks. The risk summary sheets weredesigned based on lessons of effective risk commu-nication learned while developing the deliberativerisk-ranking process, and the design was tested in pi-lot evaluations of the method.(5−7) The first page ofeach sheet contains a summary table showing howthe risk measures along each of the 12 attributes (e.g.,see Fig. 2). The subsequent interior pages summarizethe scope of the risk, describe what is known aboutthe risks from the exposure generally and specifi-cally within the UAE, and provide an overview ofwhat has already been done in the UAE to managethe risk of the exposure. The risk summary sheetsdeveloped for this effort can be found online athttp://www.sph.unc.edu/uae.

2.3.4. Step D: Perform the Risk Rankingsfor the UAE

In January 2009, we held five workshops in AbuDhabi and Dubai with stakeholders from throughoutthe UAE to elicit their concerns about environmen-tal health risks in the UAE. Invitations to participatein these workshops were sent to organizations se-lected by the EAD and an interagency steering com-mittee appointed to participate in leadership of theproject. (The EAD and WHO-CEHA selected mem-bers of the steering committee based on the role theywould ultimately need to play in implementing theNational Strategy and Action Plan for EnvironmentalHealth.)

Seventy-three people from the federal govern-ment, emirate governments, municipal governments,private-sector organizations, and nongovernmentalorganizations participated in the workshops, rep-resenting five of the seven emirates in the UAEand a diverse range of expertise and perspectiveson managing environmental health risks. Table III

Prioritizing Environmental Health Risks in the UAE 1847

Table II. Examples of Attributes AlongWhich Risks Can Be Compared(Adapted from Morgan et al.(15))

AttributeCategory Example Attributes Units

Number Number of deathsa Deaths/yearNumber of more serious

long-term illnessesaCases/year

Number of less serious long-termillnessesa

Cases/year

Number of more seriousshort-term illnessesa

Cases/year

Number of less serious short-termillnessesa

Cases/year

Expected number of annualperson-years lost

Person-years

Expected number of annualperson-years lost due tononfatal illnesses

Person-years

Total expected number of annualperson-years lost

Person-years

Chance in a million of death peryear for the average individuala

Probability

Chance in a million of death peryear for the individual athighest riska

Probability

Knowledge Time between exposure andhealth effects (degree to whichimpacts are delayed)a

Constructed scale (e.g., <1 year,1–10 years, >10 years)

Degree to which risk is observable Constructed scale (e.g.,unobservable, with difficulty,with ease)

Uncertainty in number of deaths,illnesses (degree to which risk isknown)a

Constructed scale (e.g., low,medium, high)

Degree to which risk is reversible Constructed scale (e.g., yes,frequently, no)

Degree of scientific understandingof riska

Constructed scale (e.g., low,medium, high)

Dread Individual controllabilitya Constructed scale (e.g., low,medium, high)

Catastrophic potential (e.g.,greatest number of deaths in asingle event)a

Number (or other appropriatemeasure)

Outcome equity (number whoreceive benefits from riskdivided by half the sum ofnumber who receive benefitsand number at risk)

Constructed scale (e.g., high =1–3, medium = 4–10, low >10)

Intergenerational risk Constructed scale (e.g., negligible,modest, large)

aAttributes included in the UAE risk-ranking exercise.

summarizes the characteristics of participants, ob-tained though surveys that participants completed atthe end of the workshops.

The workshops were conducted in groups rang-ing in size from 8 to 20 people. Although the goal wasfor each group to include about 10 participants, thegroup size had to be adjusted to accommodate the

availability of participants. In an attempt to ensurethat diverse views were represented in each group,we separated participants from the same organiza-tion into different groups when possible.

The majority of the workshop participants werefluent in English, as English is emphasized inschools in the UAE, although Arabic is the primary

1848 Willis et al.

DDRINKING WATER

SUMMARY

Access to clean drinking water is essential for health. However, drinking water

quality is a serious problem worldwide. Problems with drinking water are seen more

often in developing countries than in industrialized countries like the United Arab

Emirates. Drinking water hazards include microbial contamination that can lead to

diarrheal diseases, compounds that may cause acute toxicity, compounds that may

cause cancer, and radiological contamination. The UAE has established drinking water

quality guidelines that are comparable to international guidelines, and a World Health

Organization estimate of risks associated with water, sanitation, and hygiene indicates

that these risks in the UAE are comparable to those in other industrialized nations.

Risk Characteristic Low

Estimate Best

Estimate High

Estimate Fatalities Number of deaths per year

0 Not zero but low

147

Chance in a million of death per year for the average citizen

0 Not zero but low

33

Chance in a million of death per year for the citizen at highest risk

0 Not zero but low

33

Greatest number of deaths in a single event 1 Illness or Injury More serious long-term cases per year

0 0 Not

Reported Less serious long-term cases per year

0 0 Not

Reported More serious short-term cases per year

0 Not zero but low

Not Reported

Less serious short-term cases per year Not Reported

Not zero but low

Not Reported

Other Factors Time between exposure and health effects Immediate to 10 - 30 years Quality of scientific understanding High Combined uncertainty in death, illness, and injury

High

Ability of resident to control exposure to hazard

Moderate

Fig. 2. Example of the first page of a risksummary sheet.

language. Thus, the workshop materials were writ-ten in English, and discussions primarily occurred inEnglish. However, a few members of the study teamwere fluent in Arabic and able to answer specificquestions or clarify technical aspects of the materi-als in those cases where participants had lower levelsof English fluency.

The steps used throughout the workshops, de-picted in Fig. 3, were designed to elicit informed andreliable measures of participants’ concerns. Through-out each six- to eight-hour workshop, participants

were asked to provide several rankings of the 14risks that reflected how concerned they were aboutthe risks. The workshops provided opportunities forpeople to familiarize themselves with scientific infor-mation about the risks by reading the risk summarysheets before providing a first ranking. Then, theywere assisted in developing individual rankings basedon their levels of concern about the risk attributes.Later, participants had the opportunity to learn fromeach other about the risks through group discussionsand group ranking of the risks.

Prioritizing Environmental Health Risks in the UAE 1849

Table III. Characteristics of Participants in theRisk-Ranking Workshops

Sector Emirate Expertise

• Emirate government (39) • Abu Dhabi (18) • Environment (18)• Federal government (11) • Ajman (1) • Health (12)• Private sector (3) • Dubai (11) • Food (2)• Nongovernmental • Fujairah (1) • Education (1)

organizations (2) • Sharjah (2) • Petroleum (1)• Other (2) • Other industry (2)

Note: The number of responses in this table does not sum to thetotal number of participants, 73. A number of participants chosenot to answer these questions on the exit survey, and others leftthe workshops before the forms were distributed.

To avoid problems associated with group pro-cess leading to forced consensus, the workshops con-cluded by allowing participants to provide a final in-dividual ranking. This allowed participants to adjusttheir first ranking to reflect new insights they gainedabout the risks or their preferences through the struc-tured ranking or group discussion processes. Finally,to aid in interpretation of results, participants wereasked to answer several questions about their viewson the process and the rankings that resulted.

This process represents an abbreviated versionof the ranking process used in the work to developthe deliberative method.(5−7) Specifically, this formatomitted steps to allow the participants to comparetheir first ranking with one constructed from theirranking of the risk attributes, to allow the group torank the risk attributes and to compare their group

ranking to one constructed from the group’s rankingof the attributes, and finally to allow the individualsto again rank the attributes at the end of the pro-cess. These steps were omitted because it was neces-sary to reduce the overall time required of the partic-ipants, and experimental work with the deliberativerisk-ranking process suggested that these steps addedthe least marginal value.

2.3.5. Step E: Describe the Resulting Rankingsin the UAE

The final step in the process is to analyze andcommunicate the results of the ranking exercise. Theresults presented in the next section were dissemi-nated in the UAE during the process of developingthe environmental health strategic plan. As describedlater in this article, the ranking exercises revealed sig-nificant consensus about which risks are of greatestand least concern. This work also identified risk cat-egories for which consensus on importance is lack-ing.

3. RESULTS FROM THE RISK-RANKINGWORKSHOPS

The rankings obtained from the workshops pro-vide insights into which environmental health risksresidents of the UAE are most concerned about andwhy. Fig. 4 summarizes these results by displayingthe average results of the 56 participants who submit-ted final rankings. The horizontal bars show the 25thto 75th percentiles of these rankings. Because some

Fig. 3. Overview of process used duringthe risk-ranking workshops.

1850 Willis et al.

Ambient Air PollutionIndoor Air Pollution

Occupational Exposure in IndustryOccupational Exposure in Construction

Occupational Exposure in AgricultureEating Contaminated Seafood

Global Climate ChangeEating Contaminated Fruits and Vegetables

Drinking WaterStratospheric Ozone Depletion

Electromagnetic FieldsCoastal Recreational Water

Ambient NoiseResidential Soil

Ambient Air PollutionIndoor Air Pollution

Occupational Exposure in IndustryOccupational Exposure in Construction

Occupational Exposure in AgricultureEating Contaminated Seafood

Global Climate ChangeEating Contaminated Fruits and Vegetables

Drinking WaterStratospheric Ozone Depletion

Electromagnetic FieldsCoastal Recreational Water

Ambient NoiseResidential Soil

25th and 75th percentiles

Average ranking among individuals 1

Final Ranking

2 3 4 5 6 7 8 9 10 11 12 13 14

Fig. 4. Participants’ final rankings ofenvironmental health risks in the UAE.

individuals were unable to participate in the entireeight-hour workshop and the recorded rankings ofothers erroneously omitted some risks, this analysiscould not be conducted on rankings from 17 of theoriginal 73 participants.

These results indicate that participants weremost concerned with risks related to ambient air pol-lution and were generally also concerned with risksfrom indoor air pollution and occupational expo-sures. In contrast, participants were generally lessconcerned about risks associated with electromag-netic fields, exposures to coastal recreational wa-ter, ambient noise, and contaminated residential soil.Furthermore, neither the group nor average individ-ual rankings correspond perfectly to the rankings ofrisks by expected mortality, suggesting that partici-pants’ judgments were influenced by other risk at-tributes (Table IV).

3.1. Agreement and Disagreement AboutEnvironmental Health Risks

The workshop results demonstrate substantialagreement about which environmental health risksare of most concern in the UAE. Both the aver-age individual rankings and group rankings illustratethat participants consistently ranked outdoor air pol-lution, indoor air pollution, and occupational expo-sures as being of greater concern than other risks(Table IV).

Producing consensus was not the goal of the risk-ranking workshops. The different values and inter-ests of participants can lead to valid disagreementsabout which risks are of greatest concern. How-ever, disagreements could also be the result of confu-

sion or misunderstandings about definitions of envi-ronmental health risks or their consequences. Thus,effective risk communication can lead to greateragreement by reducing confusion and misunder-standing. One goal of the workshops was to increaseparticipants’ knowledge of the risks and eliminatemisunderstanding or misconceptions. The averagecorrelations among participants’ rankings within agroup provide a measure of agreement among partic-ipants. These data suggest that agreement generallyincreased through the ranking workshops (Table V),a trend that also was observed during the pilot eval-uations of the deliberative risk-ranking method.(6,7)

Comments made during the group discussionphase of the workshops revealed reasons for agree-ment about the level of concern about several ofthe risks. In discussions, participants consistentlysaid that outdoor and indoor air pollution are high-priority risks because all residents in the UAEare exposed, individuals can do little to avoid ex-posure (especially for outdoor air), and estimates ofthese risks placed them among the leading contrib-utors to the environmental burden of disease. Someparticipants noted that indoor air pollution risks canbe avoided by not smoking tobacco, whereas oth-ers mentioned that the uncertainty about indoor airpollution risks is high because of the lack of UAE-specific data on indoor air quality.

Discussions of occupational risks suggested thatparticipants’ concerns were motivated by a differentset of characteristics. In these cases, concerns wereless motivated by the estimates of the overall burdenof disease and more motivated by the higher levelsof risk to which each segment of the work force isexposed. Compounding this issue, participants notedthat while mechanisms exist for workers to protect

Prioritizing Environmental Health Risks in the UAE 1851

Table IV. Average Individual and Group Rankings of Risks

Rank by Average Group RankingsExpected Ranking by

Risk Mortality Individuals Group 1 Group 2 Group 3 Group 4 Group 5

Ambient (outdoor) air pollution 1 1 1 1 1 1 1Indoor air pollution in residential

environments2 2 2 2 3 3 4

Occupational exposures (agriculture) 3 5 8 5 5 4 12Ultraviolet radiation above natural levels

as a result of stratospheric ozonedepletion

4 10 12 8 12 7 10

Global climate change 4 7 13 6 11 6 5Occupational exposures (construction) 6 4 6 4 4 5 2Occupational exposures (industries) 7 3 5 3 2 2 2Seafood contamination 8 6 4 9 8 8 7Electromagnetic fields from power lines 9 11 9 11 9 10 11Drinking water contamination 10 9 3 14 6 12 13Ambient noise above healthful levels 10 13 10 12 10 11 8Coastal water pollution 12 12 11 7 13 13 9Produce (fruits and vegetables)

contamination12 8 7 10 7 9 6

Soil and associated groundwaterpollution due to solid and hazardouswaste disposal

14 14 14 13 14 14 14

Table V. Agreement Among Individuals’ First and Final Rankings as Measured Through Mean Pairwise Correlations of Results

Mean Pairwise Correlation Among Individuals’ Rankings

Current Study Previous Studies

Ranking Step Group A Group B Group C Group D Group E Willis et al. (2004) Morgan et al. (2001)

Agreement among first rankings 0.45 0.52 0.35 0.45 0.17 0.39 0.59Agreement among final rankings 0.70 0.87 0.45 0.93 0.79 0.87 0.86p-values for test that mean correlation <0.0001 <0.0001 0.22 <0.0001 <0.0001 0.0102 <0.0001

among final rankings is larger thanamong first

themselves from occupational exposures, workers inthe UAE frequently lack awareness of the risks andproper use of personal protective technologies andwork practices for avoiding hazardous exposures.

The risk-ranking results reveal similar levels ofagreement about risks that were consistently viewedas low priority. These risks included stratosphericozone depletion, electromagnetic fields, coastal wa-ter pollution, ambient noise, and residential soil con-tamination. In each of these cases, participants notedin discussions that the scientific evidence about theexistence of the risks is weak (e.g., risks from electro-magnetic fields), that exposures to the risks are easilyavoidable (e.g., most Emiratis experience low UV ex-posures because traditional dress covers most of the

skin), or that there is little evidence of exposure (e.g.,residential soil contamination).

Finally, discussions also provided insight intoreasons for agreement and disagreement about therelative ranking of risks from environmental expo-sures. Participants disagreed most about the relativerisks associated with drinking water. Individuals whoviewed drinking-water risks as being of great con-cern frequently noted the importance of clean waterin the UAE, where water resources are scarce andmore than 95% of the potable water supply comesfrom desalination plants. They also pointed out thatmany UAE residents drink bottled water rather thantreated water from the tap, in part because of a lackof awareness about the quality of treated drinking

1852 Willis et al.

water and belief that reports of the high quality ofthe drinking water may not be consistently reliable.Others viewed risks from drinking water to be of lowconcern, frequently citing both the high level of treat-ment that drinking water receives at desalinationplants and the relatively low incidence in the UAEof diseases that are frequently associated with poordrinking water quality, such as cholera and dysen-tery.

3.2. Assessing the Quality and Level of Supportfor the Ranking Results

When considering whether and how to use theresults of the risk-ranking workshops to shape pub-lic policy, it is important to ask whether the work-shops achieved their objective of eliciting informed,reliable judgments of participants’ concerns. At thesame time, it is important to ask whether participantsviewed the workshops as achieving this objective andwould support using the results of the workshops infurther planning efforts. Answers to these questionscan be found in responses that participants providedto evaluation surveys and inferences drawn from theranking results.

3.2.1. Assessing Whether Rankings RepresentInformed Judgments

The workshop process was designed to supple-ment participants’ knowledge about environmentalhealth risks in several ways. The participants wereprovided with concise summaries about the risks,guided through a structured ranking process to helpthem better understand the characteristics of therisks, and provided an opportunity to discuss the riskswith other participants. Responses to the evaluationsurvey confirmed that each of these opportunities tolearn contributed to the participants’ current knowl-edge of the risks (Fig. 5).

The workshops incorporated opportunities forparticipants to express their personal judgments andlearn about the risks by participating in a group exer-cise. To the extent that participants viewed the groupprocess as instructive, one would expect that eachperson’s final rankings were influenced by the grouprankings. This is a desirable outcome when the influ-ence indicates participants reached a common under-standing of the scope and the expected consequencesof the risks. It is an undesirable outcome if the influ-ence represents a forced consensus generated by thegroup discussion process.

Responses to evaluation questions suggest thatparticipants’ final individual risk rankings were influ-enced both by the participants’ first ranking and thegroup ranking to which they contributed (Fig. 6). Bytrying to predict individuals’ final rankings using eachparticipant’s first and group ranking, it is possible toassess whether the rankings themselves confirm thesurvey responses presented in Fig. 6.

A regression model that describes this relation-ship while controlling for the effect of group mem-bership confirms that both participants’ individualranking and group ranking were found to be predic-tive of their final individual ranking. The R2 value forthe model regressing first and group rankings on thefinal individual rankings and including group mem-bership as a random effect variable was 0.72, with aWald chi-squared statistic of 101 (df = 2; p-value =0). Regression coefficients for the first ranking andgroup ranking were 0.39 and 0.65, respectively, withassociated standard errors of 0.08 and p-values of lessthan 0.01. Again these results corresponded to previ-ously published pilot tests of the deliberative methodfor ranking risks.6

As further evidence that the participants learnedfrom group discussions but did not feel that they wereforced to adopt one view or another, they generallyjudged the group ranking portion of the workshop tobe open and engaging (Fig. 7).

3.2.2. Assessing Participants’ Satisfactionwith the Results

Possibly the most important measure of whetherthe risk-ranking workshops captured informed andreliable judgments of participants’ concerns aboutenvironmental health risks is the extent to which par-ticipants support using the results of the workshop toinform further risk management planning and policy.To answer this question, participants were asked howsatisfied they were with the group rankings, whetherthe group rankings were representative of their con-cerns, and whether they would support submittingthe results from the ranking workshops to the EADfor use in decision making. Responses to these ques-tions indicated that participants were generally satis-fied with the group rankings and that they strongly

6 The authors of Ref. 6 reported average regression coefficients forindividual rankings and group rankings of 0.29 and 0.65, respec-tively, for studies involving health and safety risks. The authorsof Ref. 7 reported regression coefficients for individual rankingsand group rankings of 0.28 and 0.73, respectively, for studies in-volving ecological and human health risks.

Prioritizing Environmental Health Risks in the UAE 1853

How much is your current knowledge of environmental health risks in the UAE based on...

00 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6

your prior

knowledge?

what you learned by completing

your first ranking?

Average = 4.03 Average = 3.92 Average = 4.14 Average = 4.59

Numberof

responses

5

10

15

20

25

30

0 = Not at all6 = Very much

what you learnedfrom ranking

the risk attributes?

what you learned from the group

discussion?

Fig. 5. Participants’ responses toquestions of how information sourcescontribute to their current knowledge ofenvironmental health risks in the UAE.

Numberof

responses

0 = Not at all6 = Very much

How much was your individual final risk ranking influenced by...

0

5

10

15

20

25

30

0 1 2 3 4 5 6 0 1 2 3 4 5 6

your first risk ranking? your group's risk ranking?

70.4 = egarevA08.3 = egarevA

Fig. 6. Participants’ responses toquestions about whether their finalindividual rankings were influenced bytheir first ranking and their group’sranking.

supported EAD using the results to guide future de-cisions about how to manage environmental healthrisks (Fig. 8).

4. TRANSLATING CONCERNS ABOUT RISKSTO RISK MANAGEMENT PRIORITIES

The combined evidence about how participantsviewed the stages of the ranking workshop and theresults suggests that the workshops achieved the goalof capturing informed judgments about environmen-tal health risks in the UAE. It also suggests that theresults provide a sound basis upon which to set prior-ities for risk management strategies.

However, translating these assessments of con-cern regarding risks into risk management efforts isnot a straightforward task. Agreement that a risk iscurrently of low concern is not necessarily reasonto forego efforts to manage the risk. Low-cost ef-forts to greatly reduce or better understand theserisks may be prudent. Similarly, agreement that arisk is of great concern does not necessarily jus-tify an all-out effort to manage that risk. If effec-tive or affordable alternatives to manage the risksdo not exist, it may be more prudent to invest inresearch to generate better alternatives in the fu-ture than to direct resources to costly, ineffectivesolutions.

1854 Willis et al.

To what extent did the group consider and discuss different points of view and

encourage each member to express his or her opinion?

0

5

10

15

20

25

0 1 2 3 4 5 6

Average = 4.46

Numberof

responses

0 = Not at all6 = Very much

Fig. 7. Participants’ responses toquestions about whether their groupconsidered and openly discusseddifferent points of view during the groupranking process.

0

5

10

15

20

25

0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6

How satisfied or dissatisfied are you with

your group’srisk ranking?

To what extent is the group's risk ranking representative of

your concerns aboutthese risks?

How strongly would youapprove of submitting your group’s rankings to EAD for

use in making decisions?

63.4 = egarevA00.4 = egarevA12.4 = egarevA

Numberof

responses

0 = Very dissatisfied6 = Very satisfied

0 = Not very representative6 = Very representative

0 = Strongly disapprove6 = Strongly approve

Fig. 8. Participants’ responses toquestions about their views of the resultsof the risk-ranking workshop.

For these and other reasons, when presentedwith the results of the ranking exercise, EAD’s man-agers and their advisors from the WHO-CEHA re-tained on the list of priorities some risks that werejudged as having relatively low priority in the rankingexercise. Both coastal water pollution and soil con-tamination were carried forward for further analysisin strategic planning, even though these risks wereranked low. The EAD and WHO-CEHA were con-cerned that although they currently presented lowrisks, both of these categories could increase in im-portance in the future due to the burgeoning pop-ulation. For example, major developments are cur-rently being planned in once-remote areas that wereused for waste disposal. Only electromagnetic fields,stratospheric ozone depletion, and ambient noisewere eliminated as concerns for strategic planningpurposes.

In subsequent phases of the strategic planningprocess, information from this risk-ranking exercisewas integrated with views from across the UAEabout the feasibility, effectiveness, and attractive-ness of strategies for managing environmental healthrisks. Additional workshops involving some of theparticipants in the risk-ranking exercise, as well asothers, were held to identify specific actions to ad-dress each risk area, timelines for these actions, or-ganizations that need to be involved, and ways tomeasure progress. All of this information is com-piled along with the results of the ranking exercisein the National Strategy and Action Plan for Environ-mental Health, United Arab Emirates, 2009 (availableat: http://www.sph.unc.edu/uae/). The results of theranking exercise and the strategic plan have been for-mally presented to some of the UAE’s most promi-nent politicians, who have endorsed the results. Plans

Prioritizing Environmental Health Risks in the UAE 1855

are being put in place to establish an interagency cen-ter that will focus on implementing the plan.

As the process of implementing the strategy andaction plan continues, it may be useful to revisit theconcerns of UAE leaders and residents about envi-ronmental health risks. Doing so would provide anopportunity to judge how risk management effortsare being perceived, identify risks that are emergingas new concerns, and broaden the outreach processto engage all of the emirates and more residents inthe process of managing risks from exposures to haz-ards in the UAE environment.

The positive results from this study suggest thatthe deliberative method for ranking risks could havebroader applicability for public policies. The ap-proach could be useful for revisiting the assessmentof environmental priorities in the United States andother nations. It could also be useful in other dis-ciplines. For example, in the areas of public health,natural disaster preparedness, and terrorism security,policymakers are debating the appropriate prioritieswithin and across threats. Approaches like the onedescribed and demonstrated here provide a mecha-nism for identifying priorities that reflect accuratelyinformed judgments of concern that interested indi-viduals and groups have about the risks with whichthey are confronted.

ACKNOWLEDGMENTS

Funding for this project was provided by the En-vironment Agency-Abu Dhabi (EAD). We wish toexpress special thanks to H. E. Majid Al Mansourifor support of this work. Thanks are also due tomany EAD staff members, including Dr. Jaber AlJaberi and Dr. Ahmed Bashir. We appreciate thehard work of Ms. Ayesha Al Suweidi, Ms. RuqayaMohamed, and Ms. Ayesha Abushahab in identi-fying and inviting participants for the risk-rankingworkshops. Thanks are due as well to the many par-ticipants in the workshop, too numerous to name.We also thank Dr. Amir Johri at the WHO Cen-tre for Environmental Health Activities and Dr.M. Z. Ali Khan for technical advice in support ofthis work. Finally, we wish to recognize several col-leagues at RAND Corporation and the Universityof North Carolina who shared suggestions and guid-ance as we conducted this study, including: NicholasBurger, Zainab Farah, Tiina Folley, Ying Li, MichaelMattock, Hanine Salem, Preethi Sama, and JasonWest.

REFERENCES

1. Briggs DJ, Stern R, Tinker TL (eds). Environmental Healthfor All: Risk Assessment and Risk Communication for Na-tional Environmental Health Action Plans. Dordrecht, TheNetherlands: Kluwer Academic Publishers, 1999.

2. United Nations. Earth Summit Agenda 21: The United Na-tions Programme of Action from Rio. New York: UnitedNations Department of Economic and Social Affairs, 1992.Available at: http://www. un.org/ esa/ dsd/ agenda21/ in-dex.shtml, Accessed on March 9, 2010.

3. Environment Agency-Abu Dhabi. National Strategy and Ac-tion Plan for Environmental Health, United Arab Emirates,2009. Abu Dhabi, UAE: Environment Agency-Abu Dhabi.Available at http://www.sph.unc.edu/images/stories/units/uae/documents/ strategy action plan 022310.pdf, Accessedon March 8, 2010.

4. DeKay ML, Florig HK, Fischbeck PS, Morgan MG, MorganKM, Fischhoff B, Jenni KE. The use of public risk ranking inregulatory development. Pp. 208–230, Chapter 10 in FischbeckPS, Farrow RS (eds). Improving Regulation: Cases in Envi-ronment, Health, and Safety. Washington, DC: Resources forthe Future, 2001.

5. Florig HK, Morgan MG, Morgan KM, Jenni KE, Fischhoff B,Fischbeck PS, DeKay ML. A deliberative method for rankingrisks (I): Overview and test bed development. Risk Analysis,2001; 21(5):913–921.

6. Morgan KM, DeKay ML, Fischbeck PS, Morgan MG, Fis-chhoff B, Florig HK. A deliberative method for ranking risks(II): Evaluation of validity and agreement among risk man-agers. Risk Analysis, 2001; 21(5):923–937.

7. Willis HH, DeKay ML, Morgan MG, Florig HK, FischbeckPS. Ecological risk ranking: Development and evaluationof a method for improving public participation in envi-ronmental decision making. Risk Analysis, 2004; 24:363–378.

8. Heard-Bey F. The tribal society of the UAE and its tradi-tional economy. Pp. 98–116 in Vine P, Al Abed I (eds). UnitedArab Emirates: A New Perspective. London: Trident Press,2001.

9. CIA. World Factbook. United Arab Emirates, 2010.Available at: www.cia.gov/library/publications/the-world-factbook/geos/ae.html#Econ, Accessed on February 19,2010.

10. World Health Organization. Country profiles of environmen-tal burden of disease: United Arab Emirates, 2009 Avail-able at http://www.who.int/quantifying ehimpacts/national/countryprofile/unitedarabemirates.pdf, Accessed on March 9,2010.

11. Li Y, MacDonald Gibson JA, Jat P, Puggioni G, Hasan M,West JJ, Vizuete W, Sexton K, Serre M. Burden of diseaseattributed to anthropogenic air pollution in the United ArabEmirates: Estimates based on observed air quality data. Sci-ence of the Total Environment, in review.

12. U.S. Environmental Protection Agency (EPA). UnfinishedBusiness: A Comparative Assessment of Environmental Prob-lems. Alexandria, VA: National Technical Information Ser-vice Report, EPA Number: 230287025a. NTIS PB88–127048,1987

13. U.S. Environmental Protection Agency (EPA) Science Advi-sory Board, Relative Risk Reduction Strategies Committee.Reducing Risk: Setting Priorities and Strategies for Environ-mental Protection. Washington, DC: U.S. EPA Report SAB-EC-90–021, 1990.

14. Hornstein DT. Reclaiming environmental law: A normativecritique of comparative risk analysis. Columbia Law Review,1992; 92(3):562–633.

15. Morgan MG, Fischhoff B, Lave L, Fischbeck P. A proposalfor ranking risk within federal agencies. Chapter 6 in Davies

1856 Willis et al.

JC (ed). Comparing Environmental Risks: Tools for SettingGovernment Priorities. Washington, DC: Resources for theFuture, 1996.

16. Slovic P, Fischhoff B, Lichtenstein S. Facts and fears: Under-standing perceived risk. Pp. 181–214 in Schwing R, AlbersW A Jr. (eds), Societal Risk Assessment: How Safe is SafeEnough? New York: Plenum Press, 1980.

17. McDaniels T, Axelrod LJ, Slovic P. Perceived ecological risksof global change: A psychometric comparison of causes andconsequences. Global Environmental Change, 1996; 6:159–171.

18. Willis HH, DeKay ML, Fischhoff B, Morgan MG. Aggregate-and individual-level analysis of risk perception data. RiskAnalysis, 2005; 25:405–428.