Our study suggests that viruses are a major cause of significant respiratory infections in two large refugee camps in Kenya; at least one virus was detected in specimens from 50% of the participants. Although data from UNHCR's Health Information System suggest that ARI is the highest cause of morbidity and mortality in the refugee camps, this is the first study that we are aware of to demonstrate the viral aetiologies of these infections. AdV and RSV were the leading pathogens identified, accounting for approximately two-thirds of viruses detected. As expected, rates associated with viral infection were highest for children < 12 months of age [4
]. The crude rates of hospitalisation per 1000 for SARI were more than four times higher for children < 1 compared to children 1 to < 5 years old.
Our findings for these two sites were similar to those reported from non-refugee populations in Mozambique, Gambia, Nigeria, South Africa, Indonesia, and India [37
]. Rates per 1000 for severe infections due to RSV for children < 1 year and < 5 years of age were reported to be 15 and 9 (South Africa), 15 and 5 (Mozambique), and 16 and 10 (Indonesia) [39
]. In a 3-year cohort study in rural India, severe lower respiratory tract infection (LRTI) rates per 1000 due to RSV for children aged < 1, 1-2 years, and 2-3 years were found to be 14, 7, and 0 respectively [37
]. In a meta-analysis of incidence of RSV associated severe LRTI in developing countries, Nair et al. estimated incidence in the < 1 year age group per 1000 per year as 22 (95% CI 9-53) [40
]. Rates of respiratory virus infection in children < 5-years-old in the two camps were ~2 times those reported for the United States or Europe but were similar to rates found in special populations groups within these countries [4
]. For instance, in the United States, rates of hospitalisation for respiratory infections were found to be up to 5 times higher among Alaskan Native people than the US average [43
The surveillance period overlapped with the 2009 pandemic influenza A (H1N1). While 2009 pH1N1 virus became the predominant circulating influenza A virus subtype, it was not associated with unusual rates in hospitalisation among the refugee population.
This study adds to our knowledge of the seasonal variations of respiratory viral infections in Africa. Even though both sites have very similar climatic conditions (arid and dry), we found remarkable differences in seasonal variation of ARI and specific virus activity, which might indicate varied transmission patterns in different subregions of the continent. There was distinct seasonality in Dadaab, with peak transmission occurring in November and December. Kakuma, on the other hand, despite having up to 60% more hospitalisation due to SARI, had no such seasonality. These differences may be due to the locations of the two camps. Kakuma, in the northwest of the country near the Ugandan and Sudanese borders, may reflect trends of viral transmission in the broader equatorial regions of central Africa, while Dadaab, located in the east near the Somali border, may reflect trends of transmission in the horn of Africa. Similar differences in seasonality between neighbouring tropical countries in Africa and Asia have been reported previously [39
]. Additional surveillance points are required to further evaluate temporal variations of ARI due to viral illness.
Refugee populations are especially at risk for severe illness from respiratory infections [12
]. Unique challenges make individuals residing in camp settings especially vulnerable to exposure to airborne diseases. The camp system of registration, food, water, and firewood distribution encourages crowding of large groups in small, confined spaces. In addition, malnutrition, high population density and poor shelter conditions may contribute to the elevated rates seen in this population. The availability of new, sensitive, reliable, and cost-effective technologies that can test multiple pathogens simultaneously from one easily obtainable specimen means that surveillance systems can now be set up for previously inaccessible populations [45
Our study has several limitations: 1) Not all eligible individuals meeting the case definitions were identified, and not all those who were identified agreed to participate in the study. Thus the actual rates of SARI and virus infection associated SARI is likely higher than reported. 2) We did not attempt to collect data on all cases of ILI in the camp; therefore, we were not able to estimate rates of ILI. 3) Although the two hospitals are the only ones in the camps, refugees may have sought health care elsewhere; we did not capture information on health utilization among the refugees. If refugees did seek treatment elsewhere, our rates would again underestimate the actual rates. 4) The results of this study should also be interpreted with caution especially for pathogens e.g. AdV with known background carriage in healthy, asymptomatic individuals. Finally, the surveillance system was not designed to measure all indicators of ARI disease; due to lack of data, we could not estimate key outcomes, including mortality and duration of hospitalisation.
Our study found that viruses are a major cause of respiratory infection in the two refugee camps. To decrease the burden of respiratory illnesses requires a multipronged interventional plan. Physical interventions like handwashing with soap have been found to decrease the odds of respiratory infection by as much 55%, and nosocomial transmission has been shown to decrease by 66% when cohort nursing and wearing of gloves and gowns were introduced [46
]. Measures could be put in place to minimise crowding, which is very common in refugee camps and has been associated with increased transmission of respiratory infections, and to target public health education messages during peak transmission months [49
]. All SARI cases in this study were hospitalised in a ward shared by severely sick children presenting with other conditions. While we have not explored hospital acquired infection in this study, measures could also be taken to minimise potential nosocomial transmission. Because of the high rates of morbidity associated with these viruses in the refugee population, in addition to the implementation of innovative and effective approaches to achieve sustained compliance with hand hygiene promotions, refugees should be prioritised for vaccines when they become available [10
] ARI prevention and control in refugee populations should be a key priority area for UNHCR, its partner agencies, and the international community.