International air travel has contributed to the spread of several recent disease epidemics. For example, travelers infected with severe acute respiratory syndrome (or SARS) in 2003 carried the disease around globe. One infected air traveler can carry a disease to a new continent: in 2014, a man infected with Ebola in West Africa flew to the United States and infected two healthcare workers in Dallas during treatment.

Efforts to prevent the spread of SARS, Ebola and other disease outbreaks have included screening air passengers for infection prior to boarding, or immediately after arrival. In these situations, infrared thermometers are often used to check for symptoms of fever and passengers may be asked to fill out questionnaires to assess their risk of exposure to the disease.

However, the effectiveness of these airport screenings is questionable. 1000s of air travelers have been screened during several recent disease outbreaks, but few disease cases were detected. There are many reasons why an infected individual may be missed in airport screens. Passengers who have recently been infected may not yet display any symptoms and some passengers may be able to hide a fever or other symptoms by taking medication. Even if an individual has a fever, infrared thermometers will only detect it about 70% of the time. Also, screening questionnaires may miss passengers who are infected if they lie about any possible exposure to the disease.

Gostic et al. created a mathematical model to help assess how useful airport screening is for detecting cases of disease caused by the SARS coronavirus, Ebola, influenza H1N1 and several other viruses. The model reveals that the effectiveness of airport screening depends on several factors including: how long it takes for symptoms to develop after infection (the incubation period), how much is known about the virus and how it spreads, and whether the epidemic is still growing in size or is starting to slow down.

For influenza H1N1 and other viruses with short incubation periods, fever screening is the most successful method to detect cases throughout the epidemic. However, for viruses with long incubation periods—such as Ebola—questionnaires are more useful in the early stages of an epidemic when the number of cases is rapidly rising. Fever screening becomes more useful later in the epidemic when new cases start to fall because the people who are infected are more likely to be displaying symptoms.

Even so, Gostic et al. point out that in all of these scenarios airport screening will still miss many infected passengers. Thus, a challenge for future outbreaks will be to identify situations in which screening is worthwhile, and obtain better measurements of the factors that influence detection rates.

International air travel drove the spread of SARS in 2003 and influenza A/H1N1p in 2009 (Brockmann and Helbing, 2013), and has since led to imported cases of influenza A/H7N9 (William et al., 2015), MERS-CoV (Cauchemez et al., 2014) and Ebola virus infection (McCarthy, 2014). Traveller screening policies, including fever screening and/or questionnaires at point of departure and/or arrival, have been proposed to limit the geographic spread of infection (Malone et al., 2009; World Health Organization, 2009; Cowling et al., 2010; Khan et al., 2013; Bogoch et al., 2015; Centers for Disease Control and Prevention, 2014a). Fever screening at point of arrival has been criticized, however, because long incubation periods and imperfect efficacy of fever screening devices reduce the probability of detecting symptoms in infected arriving passengers (Pitman et al., 2005; Bitar et al., 2009; Mabey et al., 2014). As the effectiveness of integrated screening programs will depend both on the pathogen-specific natural history of infection and epidemiological knowledge of exposure risk, as well as travel time and efficacy of screening methods, it is important to understand how these different factors contribute to screening effectiveness at departure and arrival.

During screening initiatives for influenza A/H1N1p, MERS-CoV, SARS-CoV and Ebola virus, large numbers of travellers were detained for in-depth assessment, but few or no cases were ultimately detected (Table 1). Although fever is the symptom most commonly measured during screening, it might not be detected in all infected individuals for several reasons. First, those with recent exposure may not yet have progressed to a symptomatic stage (Pitman et al., 2005; Mabey et al., 2014). Second, travellers might be symptomatic but not febrile; the probability a symptomatic patient will have a fever varies by pathogen (Donnelly et al., 2004; Cao et al., 2009; Louie et al., 2009; Assiri et al., 2013; Cowling et al., 2013; Gao et al., 2013; Gong et al., 2014; Sun et al., 2014; WHO Ebola Response Team, 2014). Third, the sensitivity of non-contact infrared thermometers (the devices most often used for airport fever screening) is limited, so passengers with fever may pass through symptom screening undetected (Hausfater et al., 2008; Bitar et al., 2009; Nishiura and Kamiya, 2011). Fourth, passengers may conceal fever and other symptoms during screening using antipyretic drugs (Nishiura and Kamiya, 2011). At the same time, fever is notoriously non-specific as a symptom, leading to high opportunity costs from detaining travellers with non-target illnesses (Anderson et al., 2004; Gunaratnam et al., 2014; Mabey et al., 2014).

Self-reporting of symptoms or potential recent exposure to infection via mandatory questionnaires is also a common component of traveller screening programs (St John et al., 2005; Nishiura and Kamiya, 2011; Hale et al., 2012; Centers for Disease Control and Prevention, 2014a; Cho and Yoon, 2014; Gunaratnam et al., 2014). Because information about risk factors does not depend on the presence of detectable symptoms at the time of screening, there is potential to identify a broader set of exposed travellers. However, epidemiological knowledge on factors linked to risk of infection is limited for some pathogens—particularly for novel emerging pathogens that are often the focus of screening programs. Even for pathogens with well-characterized routes of transmission, not all cases will necessarily have a known source of exposure (Lau et al., 2004; Cao et al., 2009; Tuite et al., 2010; Cowling et al., 2013; Gao et al., 2013; Gong et al., 2014; Sun et al., 2014; WHO Ebola Response Team, 2014). Thus, the contribution of questionnaires to the overall effectiveness of traveller screening programs is unclear.

Screening initiatives have also been implemented both at points of departure and arrival. It has been suggested that departure screening is more efficient than entry screening because it needs to be implemented in only a few airports rather than globally (Khan et al., 2013; Bogoch et al., 2015), but there is often local political pressure for arrival screening as well. To understand how departure and arrival screening combine with pathogen natural history, epidemiological knowledge, efficacy of screening methodology, and human behavioral factors to determine overall screening outcomes, we developed a general modelling framework (Figure 1) for the screening process. We used this framework to assess outcomes for six pathogens of current or recent concern: influenza A/H7N9, influenza A/H1N1p, SARS-CoV, MERS-CoV, Ebola virus, and Marburg virus. By separating the contribution of different factors to the probability of detecting infectious travellers, we evaluated pathogen-specific strengths and weaknesses of different screening strategies. We considered scenarios in which the source epidemic is growing or stable, as epidemic phase influences the distribution of times since exposure in potential travellers. We also identified factors that could improve the effectiveness of screening programs for future emerging pathogens.

Figure 1 with 1 supplement see all

Download asset Open asset

For each pathogen, the natural history of infection and state of epidemiological knowledge determined the potential for successful screening at various points in the process. For all six emerging pathogens considered here, the majority of identified cases exhibited a fever (Figure 2A). However, the proportion of confirmed cases who were aware of their exposure risk varied greatly. Influenza A/H1N1p, which can have generic symptoms and can be transmitted via the airborne route, had the lowest reported proportion; Ebola virus, which requires close contact with infected individuals who have conspicuous symptoms, had the highest. Influenza A/H7N9, Marburg virus and SARS-CoV had similar proportions of cases that present with fever, and that had knowledge of exposure risk. We excluded MERS-CoV from the natural history space in Figure 2A because there are no established risk factors for exposure. Moreover, there was limited information available for the fever parameter for MERS-CoV: in a hospital outbreak of MERS-CoV, 20 out of 23 cases presented with fever at onset (Assiri et al., 2013); the small size of this sample means there is greater uncertainty surrounding the estimate for proportion of cases that exhibit fever.

Figure 2

Download asset Open asset

The mean and variance of the incubation period have been recognized as key drivers of the effectiveness of fever screening at arrival, since shorter incubation periods mean a greater likelihood that travellers will progress to symptoms during travel (Pitman et al., 2005; Al-Tawfiq et al., 2014; Mabey et al., 2014). There was considerable variability in incubation period among different pathogens (Figure 2B). Influenza A/H7N9 and A/H1N1p have the shortest incubation periods, while Ebola virus and Marburg virus have the longest. For some pathogens, the estimated variance in incubation period could increase with the addition of more data, which would improve characterization of the tails of the distribution. For instance, the incubation period distribution for Marburg virus was estimated from just five cases with a single exposure opportunity (Martini, 1973); observing more cases might give rise to a right-skewed distribution as seen for Ebola virus. Similarly, the incubation period distribution for MERS-CoV is determined using data from only 23 confirmed cases, and its variance might also expand with the addition of more data. The possibility of a lengthy incubation period presents challenges for symptom screening.

Our focus on the natural history and epidemiology of infection revealed the crucial influence of the time between exposure and the departing flight. When we included the natural history parameters in the model, (see ‘Materials and methods’) we found that the contributions of each component of a screening program depend strongly on the time between exposure to infection and intended departure from the airport (Figure 3). Individuals with more recent exposure were less likely to display symptoms at the time of screening, and hence less likely to be identified by fever screening at departure. For pathogens with long incubation periods, the marginal value of fever screening at arrival was also lower. Note, however, that the 70% efficacy of non-contact infrared thermoscanners means that arrival screening can contribute by catching symptomatic cases missed at departure. Thus, the bulk of the contribution of arrival fever screening is mediated by equipment efficacy rather than natural history.

Figure 3 with 3 supplements see all

Download asset Open asset

Shortly after exposure, we found that detection was typically possible only by risk factor questionnaire screening, as most cases had not yet progressed to symptoms and were undetectable by fever screening (Figure 3). (Again, questionnaire screening at arrival contributes by catching some individuals who did not disclose their exposure risk at departure.) The duration of this phase depended on the incubation period, which is shortest for influenza A/H7N9 and A/H1N1p, and longest for Ebola and Marburg viruses; for MERS-CoV, despite a mid-length incubation period, questionnaire screening contributes nothing due to our ignorance of risk factors. As time since exposure elapsed, fever screening made a greater contribution to case detection, with pathogen natural history factors (i.e., incubation period, and fraction presenting with fever) becoming the primary determinants of screening effectiveness. We found similar qualitative patterns when we assumed reduced efficacy for fever screening devices (Figure 3—figure supplement 1), questionnaire reporting (Figure 3—figure supplement 2), or both tests (Figure 3—figure supplement 3).

The striking patterns in Figure 3 highlight the important role of the ‘infection age structure’ (i.e., the distribution of times from exposure to departure) of the traveller population. As a basic consideration, cases are more likely to have progressed to severe disease or death as time since exposure increases, so the population of infected individuals able to attempt air travel will be skewed toward more recent exposures (i.e., younger infections). The distribution of time since exposure will be influenced by the epidemic phase in the source population (Figure 4, Figure 4—figure supplement 1).

Figure 4 with 1 supplement see all

Download asset Open asset

Overall screening effectiveness was greater in stable than growing epidemics (Figure 4A–B). These gains were driven by increased potential for fever detection in stable epidemics, where cases are less likely to be recently exposed and asymptomatic. In contrast, exposure risk detection does not vary with epidemic phase because exposure risk awareness does not depend on the infection age distribution. Regardless of epidemic phase, the full screening program fails to detect at least 25% of infected travellers, despite our optimistic assumptions. Focusing on the contribution made by screening at point of arrival, our projections suggest that arrival screening will still miss half to three-quarters of infected travellers that manage to complete their flights (Figure 4C–D). For pathogens with short incubation periods (i.e., influenza virus) fever detection was responsible for the majority of case identification in all epidemic phases. However, for pathogens with longer incubation periods (i.e., Ebola, Marburg, and SARS-CoV), exposure risk screening was responsible for half or more of case detection in growing epidemics. For these pathogens, fever detection was dominant only in stable epidemics (Figure 4).

We assessed the influence of pathogen natural history, knowledge of exposure risk, efficacy of screening techniques, and epidemic phase on the ability to detect infected passengers using integrated traveller screening programs. By incorporating pathogen natural history and epidemiology into a mathematical model, we compared screening effectiveness for different pathogens, and showed that detection is driven by screening for exposure risk in travellers with recent exposure and by screening for fever in travellers with older infections. We found that natural history, epidemiological knowledge and epidemic phase combined to determine overall screening outcomes, as well as the relative contribution of each screening method to case detection. Exposure risk screening made a greater contribution to case detection during growing epidemics and for pathogens with longer incubation periods, and when exposure risk factors were well-characterized epidemiologically.

Our results highlight distinct taxonomic patterns in the effectiveness of screening measures. For influenza viruses, which have the shortest incubation periods, our findings suggest that fever screening would be responsible for the majority of case detection in any epidemic context. The picture is more subtle for filoviruses, which have longer incubation periods and distinctive symptoms. When the source epidemic is still growing and many travellers are recently exposed, our model suggests half or more of case detection for filoviruses will be driven by exposure risk screening; when the epidemic has stabilized, the shift in the age structure of infections means fever is likely to become the dominant mechanism of case detection (Figure 4).

Our analysis of coronaviruses illustrates two important features of the screening process. First, the variance of incubation period distributions can strongly modulate screening outcomes: the median incubation period for SARS-CoV is similar to influenza viruses, but because its distribution has a long right tail, the expected incubation time is longer. Thus, despite a short, influenza-like median incubation time, screening outcomes for SARS-CoV are filovirus-like and less favorable. Second, epidemiological knowledge is required to implement risk factor screening. Our results show that for pathogens with long incubation periods, early characterization of exposure risk factors is a powerful screening tool at the beginning of an outbreak; for these pathogens, robust characterization of and screening for specific exposure risk factors can contribute more to case detection than rapid implementation of fever screening (Figures 3–4). However, risk factor screening cannot improve screening effectiveness if factors that increase risk of exposure are not well characterized. For MERS-CoV, recent findings have strengthened the evidence that dromedary camels play a role in primary exposure of some cases, but it remains unclear whether the majority of transmission is driven by contact with infected humans or repeated zoonotic spillovers (Azhar et al., 2014; Cauchemez et al., 2014).

When novel pathogens emerge, it is important to prioritize case-control and other epidemiological studies to establish which factors contribute to exposure risk. As well as enhancing potential case detection using risk factor screening, such knowledge could have the additional benefit of promoting general awareness of exposure risk factors, and hence contribute to reductions in risky behavior in affected regions. Further, when risk factors are known there is potential to conduct extended follow-up with travellers who have exposure risks, but no symptoms at the time of flight. For example, patients with known risks for Ebola exposure should be monitored by local health authorities near their travel destination until the maximum plausible incubation period has elapsed (Brown et al., 2014).

Our results also illustrate that well-characterized exposure risks are not always sufficient: other factors of pathogen natural history can limit the potential effectiveness of risk detection. For example, exposure risks for influenza A/H7N9 are well defined and typically identifiable (exposure to live poultry or infected human contacts [Cowling et al., 2013; Gao et al., 2013]). However, because influenza A/H7N9 infection has a short incubation period and a high probability of generating febrile symptoms, fever screening becomes a potent tool within just a few days of exposure and eclipses the potential contribution of risk factor-driven detection. Thus, even during a growing epidemic, where the proportion of asymptomatic travellers is highest, fever detection is likely to remain the dominant mechanism of detection for pathogens with short incubation periods, such as the influenza viruses (Figure 4A).

We found that once the source epidemic stabilizes or begins to decline, screening has greater overall potential effectiveness: the infection age structure in a stable epidemic allows more fever-driven detection, which has much higher efficacy than risk screening (Figure 4B). This result reinforces other studies that emphasize the need to control outbreaks at the source during the growth phase (Khan et al., 2013; Bogoch et al., 2015; Mabey et al., 2014). Moreover, it suggests that stabilizing the growth of an epidemic could have the added benefit of making passenger screening more effective. Even in this stable phase, however, our results suggest that screening at point of arrival would still miss more than half of incoming infected passengers.

As well as being influenced by the natural history and epidemiological factors described above, the overall effectiveness of traveller screening depends on the efficacy of particular screening techniques. This efficacy in turn depends on a combination of instrument and human factors. We assumed that fever and risk factor screening are implemented with 70% and 25% efficacy, respectively, which we consider upper bounds (see ‘Materials and methods’). Although our results are qualitatively insensitive to these assumptions (Figure 3—figure supplements 1–3), the quantitative results of our main analysis likely represent best-case scenarios. The estimated efficacy of fever screening reflects the sensitivity of non-contact infrared thermal scanner equipment, but human factors may further reduce the efficacy of screening techniques. For example, data from influenza A/H1N1p screening in Tokyo in 2009 suggested antipyretic drug use could have been widespread among febrile travellers (Nishiura and Kamiya, 2011). Additionally, outbreak-affected countries with heavily burdened public health systems may have limited resources to invest in departure screening, while limited preparation, awareness or focus may lower efficacy during arrival screening in countries outside the epidemic region.

Empirical evidence suggests that the majority of travellers with known exposure would not self-report (Hale et al., 2012; Gunaratnam et al., 2014), so the absolute effectiveness of risk factor screening in our model was lower than the effectiveness of symptom screening once cases progressed to onset, even if many exposed travellers were not yet symptomatic and symptom screening was only 70% effective (Bitar et al., 2009). For questionnaire-based screening, an essential unknown is the probability that travellers will divulge their exposure history if it puts them at risk of detainment or delay. We arrived at a rough, upper-bound estimate of 25% probability of honest reporting for the 2009 influenza A/H1N1p pandemic (see ‘Materials and methods’), but we emphasize that this is a topic in need of further study. Even more valuable would be effective ways to motivate travellers to honestly report their risky exposures. Increasing honest exposure reporting not only has the potential to enhance detection of infected travellers, but is essential for implementation of follow-up monitoring of travellers who may have been exposed but have not yet developed symptoms.

Data from past and ongoing screening initiatives support our suggestion that outcomes predicted in this study should be interpreted as plausible best-case scenarios. For example, during the 2009 influenza A/H1N1 pandemic, arrival screening in Sydney, Australia detected 3 of an estimated 48 infected travellers, giving an empirical sensitivity of 7% (95% CI, 1–18%). This initiative used a combination of risk factor and fever detection in a growing epidemic, and yielded less favorable results than the effectiveness of 32% (95% CI 20–46%) predicted by our model. Also, questionnaire-based arrival screening for influenza A/H1N1 in Auckland, New Zealand detected 4 of 69 infected individuals, for a sensitivity of 6% (95% CI 2–14%). In this case, our model's predicted sensitivity (6%, 95% CI 2–16%) matches the observed pattern well.

Between August and January 2015, screening for Ebola in United States detected neither of two case importations and screening in the United Kingdom did not detect the single known case importation (Department of Health, 2015). Therefore observed sensitivity for Ebola screening has been 0%, which is lower than our model-predicted value of around 50%. While the comparison is not statistically significant with only three data points, these outcomes underscore that screening is inherently imperfect and can be expected to reduce—but not to prevent—disease importations.

These comparisons illustrate that actual screening efficacies and honest reporting fractions may vary considerably, and in some cases appear to be quite poor. Even under best-case scenario assumptions, our model suggests arrival screening will miss half or more of infected travellers. Thus, for screening to be implemented with reasonable effectiveness there is a need to identify behavioral incentives that encourage much better self-reporting and efficacy than the current data indicate.

There are some additional limitations to our framework. Because our model is structured so that fever screening precedes exposure risk screening, case detection through fever screening increasingly overlaps with potential detection through questionnaires as time since exposure increases. The model results appear to show that the effectiveness of exposure risk screening decreases with time since exposure (Figure 3), but in fact this shows that risk factor screening becomes increasingly redundant when passengers are subject to fever screening first. We have treated fever screening as the primary means of detection, as it is much easier to conclusively diagnose infection when symptoms are present (Towner et al., 2004; Centers for Disease Control and Prevention, 2013; Centers for Disease Control and Prevention, 2014b). The overall effectiveness of traveller screening, and the total proportion of cases detected before and after a flight, are independent of screening order. We did not consider the potential for case detection using symptoms other than fever. While other symptoms may aid in case detection, many (coughing, sneezing, etc) are also non-specific and more easily concealed than fever in the early stages of infection (Donnelly et al., 2004; Cao et al., 2009; Louie et al., 2009; Assiri et al., 2013; Gao et al., 2013; Gong et al., 2014; Sun et al., 2014; WHO Ebola Response Team, 2014). Episodic symptoms such as vomiting, and internal symptoms such as gastrointestinal distress, would be difficult to detect via point-screening but could be incorporated into questionnaires.

We also concentrated on the sensitivity rather than specificity of screening measures. This is another respect in which our results should be considered a best-case scenario projection of detection outcomes, assuming that the financial and opportunity costs of imposing additional clinical assessment on a large number of uninfected individuals can be neglected. Past screening programs have been implemented at large financial cost and have delayed large numbers of travellers, while detecting only a few cases (Table 1), reflecting the fact that fever and many risk factors (e.g., contact with live poultry) have low positive predictive value for infection with rare pathogens. Though the high cost and low effectiveness of screening have been noted (St John et al., 2005; World Health Organization, 2009; Cowling et al., 2010; Gunaratnam et al., 2014; Mabey et al., 2014), to the best of our knowledge no formal cost analysis of traveller screening policies has ever been conducted. Such information would greatly aid future policy decisions about screening measures.

Screening at departure rather than arrival has been suggested as a more cost-effective and logistically feasible policy, as departure screening need be implemented only in affected regions, rather than globally (Khan et al., 2013; Bogoch et al., 2015). Our analysis suggests that arrival screening has the potential to make a non-negligible contribution to overall case detection, not only by detecting travellers who develop symptoms in flight, but also by detecting travellers who were missed by imperfect screening at departure. Hence, the additional benefit of arrival screening is greatest when efficacy of departure screening is relatively low, for example if potentially infected travellers primarily depart regions with limited public health re-sources and arrive in regions where public health resources are more abundant. Yet even costly policies that combine exit and arrival screening lack the potential to prevent all case importations. Our analysis suggests that in any context screening would miss a substantial proportion of infected travellers; this result is consistent with other analyses that highlight the limited effectiveness of screening (Pitman et al., 2005; Bitar et al., 2009; Mabey et al., 2014) and with previous or ongoing screening outcomes (Table 1). Policy makers should carefully consider whether resources are better spent on arrival screening, which will reduce but not eliminate the importation of cases, or instead on tracing and containing cases that inevitably do arrive.

Screening policies have been implemented during several recent epidemics (Samaan et al., 2004; Pitman et al., 2005; St John et al., 2005; Nishiura and Kamiya, 2011; Hale et al., 2012; Khan et al., 2013; Centers for Disease Control and Prevention, 2014a; Gunaratnam et al., 2014), and will likely continue to be discussed in response to future disease outbreaks. Certain aspects of screening, particularly fever screening at arrival, have been criticized as having little scientific justification (Pitman et al., 2005; Bitar et al., 2009; Mabey et al., 2014), but political leaders and health policy makers are likely to consider implementing screening programs when public pressure becomes intense. Thus there is a need to characterize the potential contributions of screening programs when implemented at different times, in different combinations, and for different pathogens; ultimately a quantitative understanding will be needed, to factor into cost-benefit calculations. In this study we begin to address these issues by demonstrating that screening outcomes depend strongly on pathogen natural history and epidemiological features, as well as human factors in implementation and compliance. Our results emphasize the need to characterize basic properties of emerging pathogens, as this knowledge can enhance disease control measures.

1. 1. Al-Tawfiq JA
   2. Hinedi K
   3. Ghandour J
   4. Khairalla H
   5. Musleh S
   6. Ujayli A
   7. Memish ZA

   (2014) Middle East respiratory syndrome coronavirus: a case-control study of hospitalized patients

   Clinical Infectious Diseases 59:160–165.

   https://doi.org/10.1093/cid/ciu226

   • Google Scholar
2. 1. Anderson RM
   2. Fraser C
   3. Ghani AC
   4. Donnelly CA
   5. Riley S
   6. Ferguson NM
   7. Leung GM
   8. Lam TH
   9. Hedley AJ

   (2004) Epidemiology, transmission dynamics and control of SARS: the 2002-2003 epidemic

   Philosophical Transactions of the Royal Society B: Biological Sciences 359:1091–1105.

   https://doi.org/10.1098/rstb.2004.1490

   • Google Scholar
3. 1. Apuzzo M
   2. Fernandez M

   (2014)

   5 U.S. Airports set for travelers from 3 west African Nations

   New York Times.

   • Google Scholar
4. 1. Assiri A
   2. McGeer A
   3. Perl TM
   4. Price CS
   5. Al Rabeeah AA
   6. Cummings DA
   7. Alabdullatif ZN
   8. Assad M
   9. Almulhim A
   10. Makhdoom H
   11. Madani H
   12. Alhakeem R
   13. Al-Tawfiq JA
   14. Cotten M
   15. Watson SJ
   16. Kellam P
   17. Zumla AI
   18. Memish ZA
   19. KSA MERS-CoV Investigation Team

   (2013) Hospital outbreak of Middle East respiratory syndrome coronavirus

   New England Journal of Medicine 369:407–416.

   https://doi.org/10.1056/NEJMoa1306742

   • Google Scholar
5. 1. Azhar EI
   2. El-Kafrawy SA
   3. Farraj SA
   4. Hassan AM
   5. Al-Saeed MS
   6. Hashem AM
   7. Madani TA

   (2014) Evidence for camel-to-human transmission of MERS coronavirus

   New England Journal of Medicine 370:2499–2505.

   https://doi.org/10.1056/NEJMoa1401505

   • Google Scholar
6. 1. Bausch DG
   2. Borchert M
   3. Grein T
   4. Roth C
   5. Swanepoel R
   6. Libande ML
   7. Talarmin A
   8. Bertherat E
   9. Muyembe-Tamfum JJ
   10. Tugume B
   11. Colebunders R
   12. Kondé KM
   13. Pirad P
   14. Olinda LL
   15. Rodier GR
   16. Campbell P
   17. Tomori O
   18. Ksiazek TG
   19. Rollin PE

   (2003) Risk factors for Marburg hemorrhagic fever, Democratic Republic of the Congo

   Emerging Infectious Diseases 9:1531–1537.

   https://doi.org/10.3201/eid0912.030355

   • Google Scholar
7. 1. Bausch DG
   2. Nichol ST
   3. Muyembe-Tamfum JJ
   4. Borchert M
   5. Rollin PE
   6. Sleurs H
   7. Campbell P
   8. Tshioko FK
   9. Roth C
   10. Colebunders R
   11. Pirard P
   12. Mardel S
   13. Olinda LA
   14. Zeller H
   15. Tshomba A
   16. Kulidri A
   17. Libande ML
   18. Mulangu S
   19. Formenty P
   20. Grein T
   21. Leirs H
   22. Braack L
   23. Ksiazek T
   24. Zaki S
   25. Bowen MD
   26. Smit SB
   27. Leman PA
   28. Burt FJ
   29. Kemp A
   30. Swanepoel R
   31. International Scientific and Technical Committee for Marburg Hemorrhagic Fever Control in the Democratic Republic of the Congo

   (2006) Marburg hemorrhagic fever associated with multiple genetic lineages of virus

   New England Journal of Medicine 355:909–919.

   https://doi.org/10.1056/NEJMoa051465

   • Google Scholar
8. 1. Bitar D
   2. Goubar A
   3. Desenclos JC

   (2009)

   International travels and fever screening during epidemics: a literature review on the effectiveness and potential use of non-contact infrared thermometers

   Euro Surveillance 14:.

   • Google Scholar
9. 1. Bogoch II
   2. Creatore MI
   3. Cetron MS
   4. Brownstein JS
   5. Pesik N
   6. Miniota J
   7. Tam T
   8. Hu W
   9. Nicolucci A
   10. Ahmed S
   11. Yoon JW
   12. Berry I
   13. Hay SI
   14. Anema A
   15. Tatem AJ
   16. MacFadden D
   17. German M
   18. Khan K

   (2015) Assessment of the potential for international dissemination of Ebola virus via commercial air travel during the 2014 west African outbreak

   Lancet 385:29–35.

   https://doi.org/10.1016/S0140-6736(14)61828-6

   • Google Scholar
10. 1. Brockmann D
    2. Helbing D

    (2013) The hidden geometry of complex, network-driven contagion phenomena

    Science 342:1337–1342.

    https://doi.org/10.1126/science.1245200−

    • Google Scholar
11. 1. Brown CM
    2. Aranas AE
    3. Benenson GA
    4. Brunette G
    5. Cetron M
    6. Chen TH
    7. Cohen NJ
    8. Diaz P
    9. Haber Y
    10. Hale CR
    11. Holton K
    12. Kohl K
    13. Le AW
    14. Palumbo GJ
    15. Pearson K
    16. Phares CR
    17. Alvarado-Ramy F
    18. Roohi S
    19. Rotz LD
    20. Tappero J
    21. Washburn FM
    22. Watkins J
    23. Pesik N

    (2014)

    Airport exit and entry screening for ebola — august–november 10, 2014

    Morbidity and Mortality Weekly Report 63:1163–1167.

    • Google Scholar
12. 1. Cao B
    2. Li XW
    3. Mao Y
    4. Wang J
    5. Lu HZ
    6. Chen YS
    7. Liang ZA
    8. Liang L
    9. Zhang SJ
    10. Zhang B
    11. Gu L
    12. Lu LH
    13. Wang DY
    14. Wang C
    15. National Influenza A Pandemic (H1N1) 2009 Clinical Investigation Group of China

    (2009) Clinical features of the Initial cases of 2009 pandemic influenza a (H1N1) virus infection in China

    New England Journal of Medicine 361:2507–2517.

    https://doi.org/10.1056/NEJMoa0906612

    • Google Scholar
13. 1. Cauchemez S
    2. Fraser C
    3. Van Kerkhove MD
    4. Donnelly CA
    5. Riley S
    6. Rambaut A
    7. Enouf V
    8. van der Werf S
    9. Ferguson NM

    (2014) Middle East respiratory syndrome coronavirus: quantification of the extent of the epidemic, surveillance biases, and transmissibility

    The Lancet Infectious Diseases 14:50–56.

    https://doi.org/10.1016/S1473-3099(13)70304-9

    • Google Scholar
14. 1. CBS

    (2014)

    Plane passengers hospitalized due to Ebola concerns

    CBS News.

    • Google Scholar
15. 1. Centers for Disease Control and Prevention

    (2013)

    CDC SARS fact sheet for clinicians-interpreting SARS-CoV test results

    CDC Online.

    • Google Scholar
16. 1. Centers for Disease Control and Prevention

    (2014a)

    Enhanced Ebola screening to start at five U.S. airports and new tracking program for all people entering U.S. from Ebola-affected countries

    CDC Online.

    • Google Scholar
17. 1. Centers for Disease Control and Prevention

    (2014b)

    Ebola virus disease – diagnosis

    CDC Online.

    • Google Scholar
18. 1. Cho KS
    2. Yoon J

    (2014) Fever screening and detection of febrile arrivals at an International Airport in Korea: association among self-reported fever, infrared thermal camera scanning, and tympanic temperature

    Epidemiology and Health 36:e2014004.

    https://doi.org/10.4178/epih/e2014004

    • Google Scholar
19. 1. Cowling BJ
    2. Jin L
    3. Lau EH
    4. Liao Q
    5. Wu P
    6. Jiang H
    7. Tsang TK
    8. Zheng J
    9. Fang VJ
    10. Chang Z
    11. Ni MY
    12. Zhang Q
    13. Ip DK
    14. Yu J
    15. Li Y
    16. Wang L
    17. Tu W
    18. Meng L
    19. Wu JT
    20. Luo H
    21. Li Q
    22. Shu Y
    23. Li Z
    24. Feng Z
    25. Yang W
    26. Wang Y
    27. Leung GM
    28. Yu H

    (2013) Comparative epidemiology of human infections with avian influenza A H7N9 and H5N1 viruses in China: a population-based study of laboratory-confirmed cases

    Lancet 382:129–137.

    https://doi.org/10.1016/S0140-6736(13)61171-X

    • Google Scholar
20. 1. Cowling BJ
    2. Lau LL
    3. Wu P
    4. Wong HW
    5. Fang VJ
    6. Riley S
    7. Nishiura H

    (2010) Entry screening to delay local trans-mission of 2009 pandemic influenza A (H1N1)

    BMC Infectious Diseases 10:82.

    https://doi.org/10.1186/1471-2334-10-82

    • Google Scholar
21. 1. Department of Health

    (2015)

    Health minister updates parliament on UK’s preparedness for Ebola

    UK Government Online.

    • Google Scholar
22. 1. Donnelly CA
    2. Fisher MC
    3. Fraser C
    4. Ghani AC
    5. Riley S
    6. Ferguson NM
    7. Anderson RM

    (2004)

    Epidemiological and genetic analysis of severe acute respiratory syndrome

    The Lancet Infectious Diseases 4:672–683.

    • Google Scholar
23. 1. Donnelly CA
    2. Ghani AC
    3. Leung GM
    4. Hedley AJ
    5. Fraser C
    6. Riley S
    7. Abu-Raddad LJ
    8. Ho LM
    9. Thach TQ
    10. Chau P
    11. Chan KP
    12. Lam TH
    13. Tse LY
    14. Tsang T
    15. Liu SH
    16. Kong JH
    17. Lau EM
    18. Ferguson NM
    19. Anderson RM

    (2003) Epidemiological determinants of spread of causal agent of severe acute respiratory syndrome in Hong Kong

    Lancet 361:1761–1766.

    https://doi.org/10.1016/S0140-6736(03)13410-1

    • Google Scholar
24. 1. Gao HN
    2. Lu HZ
    3. Cao B
    4. Du B
    5. Shang H
    6. Gan JH
    7. Lu SH
    8. Yang YD
    9. Fang Q
    10. Shen YZ
    11. Xi XM
    12. Gu Q
    13. Zhou XM
    14. Qu HP
    15. Yan Z
    16. Li FM
    17. Zhao W
    18. Gao ZC
    19. Wang GF
    20. Ruan LX
    21. Wang WH
    22. Ye J
    23. Cao HF
    24. Li XW
    25. Zhang WH
    26. Fang XC
    27. He J
    28. Liang WF
    29. Xie J
    30. Zeng M
    31. Wu XZ
    32. Li J
    33. Xia Q
    34. Jin ZC
    35. Chen Q
    36. Tang C
    37. Zhang ZY
    38. Hou BM
    39. Feng ZX
    40. Sheng JF
    41. Zhong NS
    42. Li LJ

    (2013) Clinical findings in 111 cases of influenza A (H7N9) virus infection

    New England Journal of Medicine 368:2277–2285.

    https://doi.org/10.1056/NEJMoa1305584

    • Google Scholar
25. 1. Ghani A
    2. Baguelin M
    3. Griffin J
    4. Flasche S
    5. van Hoek AJ
    6. Cauchemez S
    7. Donnelly C
    8. Robertson C
    9. White M
    10. Truscott J
    11. Fraser C
    12. Garske T
    13. White P
    14. Leach S
    15. Hall I
    16. Jenkins H
    17. Ferguson N
    18. Cooper B

    (2009)

    The early transmission dynamics of H1N1pdm influenza in the United Kingdom

    PLOS Currents 1:RRN1130.

    • Google Scholar
26. 1. Gong Z
    2. Lv H
    3. Ding H
    4. Han J
    5. Sun J
    6. Chai C
    7. Cai J
    8. Yu Z
    9. Chen E

    (2014) Epidemiology of the avian influenza A (H7N9) outbreak in Zhejiang Province, China

    BMC Infectious Diseases 14:244.

    https://doi.org/10.1186/1471-2334-14-244

    • Google Scholar
27. 1. Gunaratnam PJ
    2. Tobin S
    3. Seale H
    4. Marich A
    5. McAnulty J

    (2014) Airport arrivals screening during pandemic (H1N1) 2009 influenza in New South Wales, Australia

    The Medical Journal of Australia 200:290–292.

    https://doi.org/10.5694/mja13.10832

    • Google Scholar
28. 1. Hale M
    2. Hoskins R
    3. Baker M

    (2012) Screening for influenza A(H1N1)pdm09, Auckland International Airport, New Zeland

    Emerging Infectious Diseases 18:866–868.

    https://doi.org/10.3201/eid1805.111080

    • Google Scholar
29. 1. Hausfater P
    2. Zhao Y
    3. Defrenne S
    4. Bonnet P
    5. Riou B

    (2008) Cutaneous infrared thermometry for detecting febrile patients

    Emerging Infectious Diseases 14:1255–1258.

    https://doi.org/10.3201/eid1408.080059

    • Google Scholar
30. 1. Khan K
    2. Eckhardt R
    3. Brownstein JS
    4. Naqvi R
    5. Hu W
    6. Kossowsky D
    7. Scales D
    8. Arino J
    9. MacDonald M
    10. Wang J
    11. Sears J
    12. Cetron MS

    (2013) Entry and exit screening of airline travellers during the A(H1N1) 2009 pandemic: a retrospective evaluation

    Bulletin of the World Health Organization 91:368–376.

    https://doi.org/10.2471/BLT.12.114777

    • Google Scholar
31. 1. Lau J
    2. Tsui H
    3. Lau M
    4. Yang X

    (2004) SARS transmission, risk factors, and prevention in Hong Kong

    Emerging Infectious Diseases 10:587–592.

    https://doi.org/10.3201/eid1004.030628

    • Google Scholar
32. 1. Louie J
    2. Acosta M
    3. Winter K
    4. Jean C

    (2009) Factors associated with death or hospitalization due to pandemic 2009 influenza A (H1N1) infection in California

    JAMA 302:1896–1902.

    https://doi.org/10.1001/jama.2009.1583

    • Google Scholar
33. 1. Mabey D
    2. Flasche S
    3. Edmunds WJ

    (2014) Airport screening for Ebola

    BMJ 349:g6202.

    https://doi.org/10.1136/bmj.g6202

    • Google Scholar
34. 1. Malone JD
    2. Brigantic R
    3. Muller GA
    4. Gadgil A
    5. Delp W
    6. McMahon BH
    7. Lee R
    8. Kulesz J
    9. Mihelic FM

    (2009) U.S. airport entry screening in response to pandemic influenza: modeling and analysis

    Travel Medicine and Infectious Disease 7:181–191.

    https://doi.org/10.1016/j.tmaid.2009.02.006

    • Google Scholar
35. 1. Martini GA

    (1973) Marburg virus disease

    Postgraduate Medical Journal 49:542–546.

    https://doi.org/10.1136/pgmj.49.574.542

    • Google Scholar
36. 1. McCarthy M

    (2014) Liberian man being treated for Ebola in Texas dies

    BMJ 349:g6145.

    https://doi.org/10.1136/bmj.g6145

    • Google Scholar
37. 1. Nishiura H
    2. Kamiya K

    (2011) Fever screening during the influenza (H1N1-2009) pandemic at Narita International Airport, Japan

    BMC Infectious Diseases 11:111.

    https://doi.org/10.1186/1471-2334-11-111

    • Google Scholar
38. 1. Pattyn S

    (1978)

    International Colloquium on Ebola virus infection and other Haemorrhagic fevers, organized in Antwerp

    International Colloquium on Ebola virus infection and other Haemorrhagic fevers, organized in Antwerp, Belgium, Elsevier.

    • Google Scholar
39. 1. Pitman RJ
    2. Cooper BS
    3. Trotter CL
    4. Gay NJ
    5. Edmunds WJ

    (2005) Entry screening for severe acute respiratory syndrome (SARS) or influenza: policy evaluation

    BMJ 331:1242–1243.

    https://doi.org/10.1136/bmj.38573.696100.3A

    • Google Scholar
40. 1. Roddy P
    2. Thomas SL
    3. Jeffs B
    4. Nascimento Folo P
    5. Pablo Palma P
    6. Moco Henrique B
    7. Villa L
    8. Damiao Machado FP
    9. Bernal O
    10. Jones SM
    11. Strong JE
    12. Feldmann H
    13. Borchert M

    (2010) Factors associated with Marburg hemorrhagic fever: analysis of patient data from Uige, Angola

    Journal of Infectious Diseases 201:1909–1918.

    https://doi.org/10.1086/652748

    • Google Scholar
41. 1. Samaan G
    2. Patel M
    3. Spencer J
    4. Roberts L

    (2004)

    Border screening for SARS in Australia: what has been learnt?

    Medical journal of Australia 180:220–224.

    • Google Scholar
42. 1. St John RK
    2. King A
    3. de Jong D
    4. Jong DD
    5. Bodie-Collins M
    6. Squires SG
    7. Tam TW

    (2005) Border screening for SARS

    Emerging Infectious Diseases 11:6–10.

    https://doi.org/10.3201/eid1101.040835

    • Google Scholar
43. 1. Sun J
    2. Gong Z
    3. Lv H
    4. Chen Z
    5. Chai C
    6. Liu S
    7. Ling F
    8. Lu Y
    9. Cai J
    10. Yu Z
    11. Miao Z
    12. Ren J
    13. Chen E

    (2014) Comparison of characteristics between patients with H7N9 living in rural and urban areas of Zhejiang Province, China: a preliminary report

    PLOS ONE 9:e93775.

    https://doi.org/10.1371/journal.pone.0093775

    • Google Scholar
44. 1. Thomas HL
    2. Zhao H
    3. Green HK
    4. Boddington NL
    5. Carvalho CF
    6. Osman HK
    7. Sadler C
    8. Zambon M
    9. Bermingham A
    10. Pebody RG

    (2014) Enhanced MERS coronavirus surveillance of travelers from the Middle East to England

    Emerging Infectious Diseases 20:1562–1564.

    https://doi.org/10.3201/eid2009.140817

    • Google Scholar
45. 1. Towner JS
    2. Rollin PE
    3. Bausch DG
    4. Sanchez A
    5. Crary SM
    6. Vincent M
    7. Lee WF
    8. Spiropoulou CF
    9. Ksiazek TG
    10. Lukwiya M
    11. Kaducu F
    12. Downing R
    13. Nichol ST

    (2004) Rapid diagnosis of Ebola hemorrhagic fever by reverse transcription-PCR in an outbreak setting and assessment of patient viral load as a predictor of outcome

    Journal of Virology 78:4330–4341.

    https://doi.org/10.1128/JVI.78.8.4330-4341.2004

    • Google Scholar
46. 1. Tuite AR
    2. Greer AL
    3. Whelan M
    4. Winter AL
    5. Lee B
    6. Yan P
    7. Wu J
    8. Moghadas S
    9. Buckeridge D
    10. Pourbohloul B
    11. Fisman DN

    (2010) Estimated epidemiologic parameters and morbidity associated with pandemic H1N1 influenza

    CMAJ 182:131–136.

    https://doi.org/10.1503/cmaj.091807

    • Google Scholar
47. 1. WHO Ebola Response Team

    (2014) Ebola virus disease in West Africa—the first 9 months of the epidemic and forward projections

    New England Journal of Medicine 371:1481–1495.

    https://doi.org/10.1056/NEJMoa1411100

    • Google Scholar
48. 1. Wilder-Smith A
    2. Goh Kee T
    3. Paton NI

    (2003) Experience of severe acute respiratory syndrome in singapore: importation of cases, and defense strategies at the airport

    Journal of Travel Medicine 10:259–262.

    https://doi.org/10.2310/7060.2003.2676

    • Google Scholar
49. 1. William T
    2. Thevarajah B
    3. Lee SF
    4. Suleiman M
    5. Jeffree MS
    6. Menon J
    7. Saat Z
    8. Thayan R
    9. Tambyah PA
    10. Yeo TW

    (2015) Avian influenza (H7N9) virus infection in Chinese Tourist in Malaysia, 2014

    Emerging Infectious Diseases 21:142–145.

    https://doi.org/10.3201/eid2101.141092

    • Google Scholar
50. 1. World Health Organization

    (2009)

    New influenza A(H1N1) virus: WHO guidance on public health measures 11 June 2009

    Weekly Epidemiological Report 84:261–268.

    • Google Scholar
51. 1. Zumla A
    2. Mwaba P
    3. Bates M
    4. AlTawfiq JA
    5. Maeurer M
    6. Memish ZA

    (2014) The Hajj pilgrimage and surveillance for Middle East Respiratory syndrome coronavirus in pilgrims from African countries

    Tropical Medicine & International Health 19:838–840.

    https://doi.org/10.1111/tmi.12318

    • Google Scholar

Author details

1. Katelyn M Gostic

   Department of Ecology and Evolutionary Biology, University of California, Los Angeles, Los Angeles, United States

   Contribution

   KMG, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Contributed equally with

   Adam J Kucharski

   For correspondence

   kgostic@ucla.edu

   Competing interests

   The authors declare that no competing interests exist.

2. Adam J Kucharski

   1. Centre for the Mathematical Modelling of Infectious Diseases, Department of Infectious Disease Epidemiology, London School of Tropical Hygiene and Medicine, London, United Kingdom
   2. Fogarty International Center, National Institutes of Health, Bethesda, United States

   Contribution

   AJK, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Contributed equally with

   Katelyn M Gostic

   Competing interests

   The authors declare that no competing interests exist.

3. James O Lloyd-Smith

   1. Department of Ecology and Evolutionary Biology, University of California, Los Angeles, Los Angeles, United States
   2. Fogarty International Center, National Institutes of Health, Bethesda, United States

   Contribution

   JOL-S, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

• 4,765

  views

• 482

  downloads

• 43

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.