Protective Practices and Respiratory Illness Among US Travelers to the 2009 Hajj

Victor Balaban PhD, William M. Stauffer MD, Adnan Hammad PhD, Mohamud Afgarshe MD, Mohamed Abd‐Alla MD, Qanta Ahmed MD, Ziad A. Memish MD, Janan Saba MPH, Elizabeth Harton MPH, Gabriel Palumbo MPH, Nina Marano DVM
DOI: 163-168 First published online: 1 May 2012


Background All mass gatherings can place travelers at risk for infectious diseases, but the size and density of the annual Hajj pilgrimage to the Kingdom of Saudi Arabia (KSA) present important public health and infection control challenges. This survey of protective practices and respiratory illness among US travelers to the 2009 Hajj was designed to evaluate whether recommended behavioral interventions (hand hygiene, wearing a face mask, cough etiquette, social distancing, and contact avoidance) were effective at mitigating illness among travelers during the 2009 Hajj.

Methods US residents from Minnesota and Michigan completed anonymous surveys prior to and following travel to the 2009 Hajj. Surveys assessed demographics, knowledge, attitudes, and practices (KAP) related to influenza A(H1N1), vaccination, health‐seeking behaviors, sources of health information, protective behaviors during the Hajj, and respiratory illness during and immediately after the Hajj.

Results Pre‐ and post‐travel surveys were completed by 186 participants. Respiratory illness was reported by 76 (41.3%) respondents; 144 (77.4%) reported engaging in recommended protective behaviors during the Hajj. Reduced risk of respiratory illness was associated with practicing social distancing, hand hygiene, and contact avoidance. Pilgrims who reported practicing more recommended protective measures during the Hajj reported either less occurrence or shorter duration of respiratory illness. Noticing influenza A(H1N1) health messages during the Hajj was associated with more protective measures and with shorter duration of respiratory illness.

Conclusions Recommended protective behaviors were associated with less respiratory illness among US travelers to the 2009 Hajj. Influenza A(H1N1) communication and education in KSA during the Hajj may also have been an effective component of efforts to mitigate illness. Evaluations of communication efforts and preventive measures are important in developing evidence‐based public health plans to prevent and mitigate disease outbreaks at the Hajj and other mass gatherings.

Every year, millions of Muslims, including thousands in the United States, make a pilgrimage called the Hajj to the cities of Mecca and Medina in the Kingdom of Saudi Arabia (KSA). An estimated 2,521,000 pilgrims attended the 2009 Hajj during November 25–29; of these, 1,613,000 were international pilgrims from 163 countries, including 11,066 US Hajj travelers.1,2

While all mass gatherings have the potential to place travelers at risk for infectious and noninfectious hazards, the Hajj presents some of the world's most important public health and infection control challenges.3 A variety of risk factors makes the Hajj an environment where emerging infectious diseases can quickly spread and even evolve into epidemics, including extended stays at Hajj sites, crowded accommodations with other Hajj pilgrims, many of whom are from developing nations, and long periods of time spent in densely packed crowds (crowd densities at Hajj have been estimated to be as high as seven people per square meter).4 Any disease outbreak at the Hajj could potentially be spread globally by returning travelers though major airline hubs, which could become the settings for further dissemination of disease.5

The 2009 Hajj took place during the influenza A(H1N1) pandemic, which led to increased emphasis on understanding ways to mitigate the potential spread of respiratory disease.6 In order to address these concerns KSA, with guidance from national and international public health agencies such as the US Centers for Disease Control and Prevention (CDC), and the World Health Organization (WHO), issued recommendations on measures to mitigate the impact of influenza A(H1N1) among pilgrims performing the Hajj. The recommended behaviors included washing hands often (hand hygiene), use of hand sanitizers, wearing a face mask, covering one's cough or sneeze (cough etiquette), staying away from sick people (social distancing), and not touching objects touched by sick people (contact avoidance).7,8 At the time the survey was developed, CDC recommendations for high‐risk people in crowded settings where influenza A(H1N1) was circulating were to avoid the setting, but if that was not possible, to consider wearing a face mask.9 The 2009 Hajj presented an opportunity to evaluate behavioral interventions for community mitigation of respiratory disease in the context of an extremely large and crowded mass gathering.

Our survey collected self‐report data on protective practices and respiratory illness among US travelers to the 2009 Hajj. We hypothesized that travelers who engaged in the recommended protective behaviors would be more likely to have lower rates of respiratory illness, as well as less severe illness and shorter duration of illness, compared with pilgrims who did not engage in protective behaviors.


Study Population

The sample in our survey represents approximately 2.0% of US pilgrims to the 2009 Hajj. US Hajj pilgrims in Michigan and Minnesota were administered pre‐travel surveys from October 21 to November 18, 2009; post‐Hajj surveys were administered within 14 days of pilgrims' return, from December 3, 2009, to February 8, 2010. Participants in Minnesota were recruited at a weekly clinic for Hajj travelers conducted by HealthPartners, a Minnesota‐based not‐for‐profit health maintenance organization (HMO). Participants in Michigan were recruited by the Arab Community Center for Economic and Social Services (ACCESS) at multiple settings, including mosques, community health clinics, and the Detroit Wayne County International Airport, and telephone surveys were conducted by health care workers in the language the participant requested (English, Arabic, or Somali). All pre‐Hajj surveys and 129 of the post‐Hajj surveys were conducted in person by health educators; the remaining 35 post‐Hajj surveys were conducted by telephone by health educators when in‐person interviews could not be arranged. All interviews were conducted whenever possible by medically trained persons from the same culture. To ensure anonymity, no identifying information was included on survey forms. Surveys were coded with a survey identification number to allow pre‐ and post‐travel surveys to be linked. This study was reviewed and approved by the ethics review boards of all participating institutions.

Survey Questionnaires

Surveys were developed and piloted by investigators at the Travelers' Health Branch of CDC, in conjunction with investigators at the participating institutions. They were vetted by health professionals from multiple cultures and nationalities, including Somali, Egyptian, Saudi, Palestinian, Lebanese, and Pakistani. The pre‐travel survey consisted of 60 items that assessed demographics, travel itinerary and activities, previous international travel, perceived health risks, health status, sources of health information, seasonal and influenza A(H1N1) immunization status, and knowledge of influenza A(H1N1) symptoms, transmission and prevention. The post‐travel survey consisted of 36 items that assessed the (1) occurrence, (2) severity, and (3) duration of any respiratory illness experienced during Hajj and/or during the first 7 days after return home from travel; protective behaviors during Hajj; and exposure to health messages in KSA during Hajj.

An expanded definition of respiratory illness was used for this study. Respiratory illness was defined as an illness with the presence of one or more of the following localizing signs or symptoms: cough, congestion, sore throat, sneezing, or breathing problems. Two travelers who reported “bronchitis” as a symptom were also included. Severity of respiratory illness was calculated by using a Likert scale from 0 (“did not need to see a doctor or nurse”), to 1 (“ill enough to see a doctor or nurse”), to 2 (“needed to be hospitalized”).

The protective behaviors assessed were five of the community mitigation practices recommended by CDC and WHO: hand hygiene, wearing a face mask, cough etiquette, social distancing, and contact avoidance.7–9 Protective behaviors during Hajj were analyzed both as categorical variables (whether the respondent reported engaging in the behavior) and as continuous variables (the number of behaviors reported by the respondent).


Data were recorded by interviewers and then entered in an Excel spreadsheet. Pearson correlation coefficients, ANOVAs, and chi‐square tests were used to assess variables and determine associations and correlations. Univariate factors with p values <0.2 were entered into multivariable regression analyses. Two‐tailed p values <0.05 were considered statistically significant in multivariable models.

To analyze the effects of protective behaviors during Hajj on respiratory illness, additional factors that have been shown to influence compliance with relevant health behaviors were also included in multivariable models.10 The variables included in multivariable models were (1) demographic and health factors: age, gender, education, whether respondent was US‐born, health risk factors, seasonal influenza vaccination in the previous 12 months, influenza A(H1N1) vaccination prior to Hajj, and taking medication for respiratory illness during or post‐Hajj; (2) travel‐related factors: length of trip, international travel in the previous 12 months, and whether respondent had made a previous Hajj; and (3) influenza A(H1N1) knowledge and attitudes: if respondent received pre‐travel health information, level of influenza A(H1N1) knowledge, perceived severity of influenza A(H1N1), and noticing influenza A(H1N1)‐related health messages during the Hajj.

Influenza A(H1N1) knowledge was calculated as the number of correct influenza A(H1N1) symptoms, modes of transmission, and methods of prevention that the respondent provided when asked. Perceived severity of influenza A(H1N1) was calculated by asking respondents how serious a disease they felt influenza A(H1N1) was on a Likert scale from “not serious” (defined as “like a cold”) to “very serious” (defined as “it can kill you”).


Pre‐travel surveys were completed by 221 participants; 186 (84.2%) completed the post‐Hajj survey after their return (Table 1). Reasons for not completing post‐Hajj surveys included travelers not receiving a visa for Hajj (which forced trip cancellation), travelers receiving a visa but choosing not to go to Hajj, travelers making extended visits to other countries lasting past the time‐frame for the survey, and being lost to follow‐up. Analyses were conducted among the 186 participants who completed both pre‐ and post‐travel surveys.

View this table:
Table 1

Demographic characteristics of 2009 Hajj pilgrims from Michigan and Minnesota

Factorn (%)
Mean = 48.9 y (range 16–89)
Gender (female)
94 (50.5)
Country of birth
 Lebanon64 (34.6)
 Iraq40 (21.6)
 Somalia21 (11.4)
 Pakistan17 (9.2)
 USA13 (7.0)
 Yemen6 (3.2)
 Egypt5 (2.7)
 Bangladesh4 (2.2)
 Palestine4 (2.2)
 Jordan3 (1.6)
 India2 (1.1)
 Kuwait2 (1.1)
 Ghana1 (0.5)
 Iran1 (0.5)
 Morocco,1 (0.5)
 United Arab Emirates1 (0.5)
Years lived in US (if foreign‐born)
Mean = 16.3 y (range 0.5–44)
 Elementary or less44 (23.8)
 High school62 (33.5)
 University35 (18.9)
 Graduate or professional44 (23.8)
Health risk factors
 Low risk (age < 65 and no chronic condition) (Chronic conditions in the pilgrim population were: diabetes, hypertension, and asthma)153 (82.3)
 Age > 65, no chronic condition20 (10.8)
 Chronic condition, age < 6513 (7.0)
 Age > 65 and chronic condition18 (9.7)

The mean length of stay at Hajj was 24.1 days. Protective behaviors during the Hajj were reported by 144 (77.4%) of the 186 respondents. Hand hygiene was reported by 125 (67.2%), wearing a face mask by 91 (48.9%), cough etiquette by 86 (46.2%), social distancing by 64 (34.4%), and contact avoidance by 45 (24.2%). Seasonal influenza vaccination in the previous 12 months was reported by 138 (63.0%) respondents. Influenza A(H1N1) vaccinations were reported by 72 (38.7%) respondents.

Respiratory illness during the Hajj and/or in the first 7 days post‐Hajj was reported by 76 (41.3%) respondents (respiratory illness during Hajj = 32 (17.3%) respondents and post‐Hajj =53 (29.0%) respondents). Among the 76 respondents who reported respiratory symptoms, coughing was reported by 56 (73.7%), sneezing by 48 (63.2%), sore throat by 29 (38.2%), fever by 25 (31.1%), congestion by 16 (32.9%), breathing problems by 4 (5.3%), and “bronchitis” by 2 (2.6%). Of the 76 respondents who reported respiratory illness, 18 (23.7%) met criteria for self‐reported influenza‐like illness (ILI), defined as fever plus sore throat and/or coughing.11

Three protective behaviors were associated with reduced risk of respiratory illness: social distancing, hand hygiene, and contact avoidance (Table 2). When the number of protective practices was analyzed as a continuous variable, reduced risk of respiratory illness was associated with engaging in more protective behaviors during the Hajj (F = 3.13,p = 0.03) (Figure 1). Engaging in more protective measures was associated with noticing influenza A(H1N1) health messages during the Hajj (F = 6.93,p = 0.01).

View this table:
Table 2

Protective behaviors and occurrence of respiratory illness among 2009 Hajj pilgrims from Michigan and Minnesota

Respiratory illness
Protective behavior% reporting illnessOdds ratio (OR), 95% CIp
Contact avoidance
 Practiced27.9% (12/43)OR = 0.51, 95% CI = 0.24–1.110.06
 Not practiced43.0% (43/100)
Cough etiquette
 Practiced37.6% (32/85)OR = 0.92, 95% CI = 0.46–1.820.47
 Not practiced39.7% (23/58)
Face mask
 Practiced41.6% (37/89)OR = 1.42, 95% CI = 0.70–2.880.21
 Not practiced33.3% (18/54)
Hand hygiene
 Practiced35.0% (43/123)OR = 0.36, 95% CI = 0.14–0.940.03
 Not practiced60.0% (12/20)
Social distancing
 Practiced28.1% (18/64)OR = 0.44, 95% CI = 0.22–0.900.02
 Not practiced46.8% (37/79)
Figure 1

Protective behaviors reported during Hajj and occurrence of respiratory illness.

Respiratory illness mild enough that the respondents did not need to see a doctor or nurse was reported by 47 (65.3%) respondents, 23 (31.9%) were ill enough to see a doctor or nurse, and 2 (2.8%) needed to be hospitalized. No protective behaviors during Hajj were associated with less severe respiratory illness. Reduced severity of respiratory illness during Hajj was associated with fewer years lived in the United States (F = 4.72,p = 0.01).

The mean duration of respiratory illness reported during Hajj was 7 days (range = 1–21d). Practicing contact avoidance during Hajj was associated with shorter duration of respiratory illness (F = 3.54,p = 0.06). Shorter duration of respiratory illness during Hajj was also associated with younger age (r2 = 0.361,p = 0.002), fewer health risks (F = 3.99,p = 0.02), and higher levels of perceived influenza A(H1N1) severity (F = 8.02,p < 0.001). A multivariable model contained two significant predictors of reduced duration of respiratory illness: practicing contact avoidance (β = −0.38,p = 0.01) and noticing influenza A(H1N1) health messages during Hajj (β = 0.25,p = 0.06). These factors also explained a significant proportion of variance in the duration of respiratory illness (r2 = 0.13,F6,45 = 2.29,p = 0.05). When the number of protective practices was analyzed as a continuous variable, engaging in more protective measures during Hajj was correlated with shorter duration of respiratory illness (r2 = −0.307,p = 0.02 ) (Figure 2).

Figure 2

Protective behaviors reported during Hajj and duration of respiratory illness.


Engaging in CDC‐ and WHO‐recommended protective behaviors during the Hajj was a predictor of reduced occurrence and duration of respiratory illness during and after the Hajj. Pilgrims who practiced contact avoidance, social distancing, and hand hygiene during the Hajj reported less respiratory illness. Practicing contact avoidance was also associated with shorter duration of respiratory illness.

The number of protective practices carried out by pilgrims was also a predictor of Hajj‐related respiratory illness. Pilgrims who reported carrying out more protective practices during Hajj reported less illness and shorter duration of illness (Figures 1 and 2). Although engaging in multiple protective behaviors may have a cumulative protective effect, it is likely that travelers who engaged in more behaviors might have been better informed before and/or during travel and thus more conscientious in practicing recommended behaviors. This hypothesis is consistent with the finding that noticing influenza A(H1N1) health messages during the Hajj was a predictor of the number of protective behaviors engaged in by pilgrims, and was also associated with reduced occurrence and duration of respiratory illness. These findings suggest that the influenza A(H1N1) communications and education carried out by the KSA during the 2009 Hajj may have been an important component of efforts to mitigate illness among travelers to this mass gathering. Future evaluations of health communications conducted during Hajj, combined with objective observations of protective behaviors and confirmation of respiratory disease would help to delineate the role played by health messages during the Hajj.

Compared with other protective behaviors, wearing face masks during Hajj seemed to have little protective effect. Wearing a face mask was actually associated with greater likelihood of respiratory illness. This finding is consistent with previous findings that face masks either offered no significant protection or were associated with sore throat and with longer duration of sore throat and fever symptoms among Hajj pilgrims,12–15 but in contrast to other studies that have found protective effects of face masks at Hajj.16 Evidence for the efficacy of face masks for preventing the transmission of influenza is limited.17 In addition, a recent study of influenza transmission suggests that poor face mask compliance decreases their utility in mitigating the spread of disease, and there is anecdotal evidence that many pilgrims at the 2009 Hajj may not have worn masks correctly (eg, mistakenly positioning the top of the mask below the nose) 18 (S. Ebrahim, personal communication). Since our survey asked only if respondents had worn face masks during Hajj, but did not ask whether masks had been worn correctly or consistently, or what types of masks were worn, it is not possible to determine the effectiveness of face masks from our data.

The occurrence, symptoms, and severity of illness in this population are consistent with previous studies of Hajj travelers.3,12,13,19–22 The rates of ILI are consistent with a study of French Hajj pilgrims and with previous studies that have found 8.0–9.8% of Hajj pilgrims with acute respiratory infection to have influenza.13,14,22 Pilgrims who reported respiratory illness during the Hajj and those who reported post‐Hajj illness were not the same travelers: only 17% of travelers with respiratory illness reported illness both during and after the Hajj. This finding suggests that surveys that only assess respiratory illness during or after mass gatherings might risk underreporting the burden of respiratory disease associated with mass gatherings.

The present study has several limitations. The study population might not be representative of the Muslim population in the United States. Compared with the US‐Arab population, the study population had a higher proportion of people of Iraqi (32 vs 4%) ancestry and a lower proportion of Egyptian (3 vs 11%), and Syrian ancestry (0 vs 10%).23 Nor could we systematically evaluate the effects of pre‐Hajj health information, since there was no consistent communication or education outreach for Hajj travelers. Many respondents were contacted during pre‐Hajj clinic visits, leading to confusion over whether the visit itself was also a source of pre‐Hajj health information. Finally, all health information was collected by self‐report and so could not be independently corroborated, although self‐reported symptoms of respiratory illnesses have shown close congruence with physician documentation.11 It is also unclear whether self‐reported duration of illness corresponds to actual severity of respiratory infection (ie, greater viral load). This association likely represents a subjective measure of respondents' perceived severity of their illness.

Our findings highlight the role that both protective behaviors and health communications can play in mitigating respiratory illness, even during extremely large and densely crowded mass gatherings such as the Hajj. Our study also demonstrates the value of conducting enhanced surveillance of international travelers both during and immediately after large mass gatherings. The fact that more than 40% of pilgrims reported respiratory illness during or after the Hajj illustrates the potential for Hajj pilgrims to be a major contributor in the international transmission of respiratory disease. The possible role of mass gatherings in the worldwide spread of respiratory disease is highlighted by a recent study speculating that a large Easter mass gathering of two million people in Iztapalapa, Mexico City may have been a key contributing factor in the rapid spread of influenza A(H1N1) throughout Mexico at the beginning of the 2009 pandemic.24

Mass gatherings such as the Hajj pilgrimage provide an opportunity to conduct large trials to evaluate the role of communication campaigns and protective behaviors in mitigating respiratory illness. Future research should focus on prospective studies of predictors of protective behaviors that also include objective confirmation of respiratory disease. If KAP were assessed using reliable, consistent instruments prior to and after travel to mass gatherings, and observational and behavioral studies of actual protective behaviors were conducted during gatherings, it would be possible to better determine the effectiveness of protective behaviors, and which factors predict protective behaviors during travel. The results from these types of studies could then be used to develop evidence‐based interventions to help prepare for future pandemics.

Declaration of Interests

The authors state they have no conflicts of interest to declare.


View Abstract