The A(H1N1)pdm09 outbreak occurred in Rwanda at the beginning of the normal influenza season period and three months after the detection of the first cases in neighboring countries (Kenya, Tanzania, Uganda). Prior to the outbreak, any increased influenza activity in the surveillance system could have identified the outbreak. Instead, the index case was detected in a non-sentinel hospital, suggesting that sentinel surveillance may not be the best system to promptly detect the first case. The trigger of outbreak detection was the guidance provided by the Ministry of Health to the index case to seek care, and have nasopharyngeal and oropharyngeal swabs taken and tested for A(H1N1)pdm09 virus infection.
Following the initial pH1N1 detection, control measures such as airport screening of incoming passengers, promotion of hand hygiene, use of facial mask, and self-isolation of laboratory-confirmed cases were implemented at the beginning of the outbreak. However, these measures did not prevent the spread of pH1N1 in the general population. The containment phase investigation documented at least two additional likely introductions of pH1N1 to Rwanda; these laboratory-confirmed cases likely acquired infection from external settings. Moreover, the virus spread among children was related to school and social contact of the index case’s children with classmates and other students in Kigali. Having children at home and contact with a patient with a similar illness were the exposures more frequently identified among laboratory-confirmed pH1N1 cases compared to seasonal influenza patients (p<0.0001).
As mentioned earlier, during the pH1N1 outbreak the age group<15 years old had the highest proportion (53.2%) of influenza positive cases compared to the inter-pandemic period (44.6%). This is probably due to the demographic structure of Rwandan where more than 45.9% are <15 years 
. The pandemic strain has co-circulated with seasonal A(H3) and influenza B; but seasonal A(H1) has no longer circulated since the detection of A(H1N1)pdm09 ().
During the shift from containment to mitigation phase, laboratory sampling methods were changed to selectively target most at risk and severely ill cases. Thus, a higher proportion of laboratory-confirmed influenza was expected during the mitigation phase. Instead, a slightly lower proportion −23.4% (218/931) - was reported compared with 24.7% (276/1114) –during the containment phase. The mitigation strategies implemented during the containment phase such as school closure and systematic distribution of oseltamivir to all suspect and laboratory confirmed cases may have helped to reduce viral transmission. It is also possible that the mitigation phase took place when the most susceptible population had already been exposed, and the rates in the general population had begun to decline.
Overall, disease severity was relatively mild with 70% of cases classified as ILI, 12% hospitalized, and no deaths. These findings are consistent with other reports from the region and internationally 
. Our findings demonstrated that clinical symptoms and duration for pH1N1 virus infection were similar to what has been described for normal seasonal influenza illness 
. We identified lower proportions of diarrhea (8.3% vs. 37%) and vomiting (16.7% vs.41%) compared to hospitalized pH1N1 patients in Kenya 
. Symptoms significantly associated with A(H1N1)pdm09 positive cases versus negative cases were fever, cough, sore throat, nausea and muscle pain (p<0.0001).
Although Rwanda received 15,000 treatment courses of oseltamivir from WHO, this drug was used sparingly during the pandemic period to avoid drug resistance as the outbreak was considered to be a mild disease. Our attack rates were limited by the fact that they were only calculated within circumscribed communities (i.e.: schools, prisons) where it was possible to obtain denominator data. These attack rates are below the cumulative attack rate of 7.7% among all age groups in Lima, Peru 
, of 14.7% in Chicago among children aged 5–14 years 
and of 43.4% in a community cohort study involving persons aged 5–14 years in Hong Kong 
Districts with influenza sentinel sites accounted for 69% of all positive confirmed cases and the share from the sentinel sites increased substantially from containment to mitigation phase (). These numbers demonstrated that the sentinel surveillance system served as the driving force behind the detection and sample collection of new cases of pH1N1. However, the lack of reporting from non-sentinel sites and districts also showed that the decentralized district hospital-driven strategy envisioned during the tabletop exercise was not fully realized. Although an initial attempt was made during the transition from containment to mitigation phases to train district hospitals on outbreak surveillance and management, the limited reporting reinforced the need to improve the functioning of influenza rapid response teams at the district level.
Our study has some limitations. We used different methods during containment (contact tracing and systematic testing) and mitigation phases (targeted testing) to identify cases and this may have introduced biases that reduce the representativeness of the findings with respect to the general population. The reported cases were limited to the hospitals where influenza surveillance is conducted and in hospitals where clinicians conducted systematic surveillance for pH1N1. Thus, additional cases of pH1N1 may have been missed at other health facilities and thus were not reflected in this report. Lastly, sentinel surveillance of influenza began too recently to provide robust data on seasonal influenza epidemiology and burden of disease to allow appropriate comparison with pandemic data ().
This study describes the epidemiology, clinical features, and the response to the outbreak of Influenza A(H1N1) pdm09 in Rwanda from October 2009 to May 2010. The outbreak occurred and peaked during the influenza season in Rwanda. Unlike seasonal influenza, children<15 years were the most affected. Our findings demonstrate that clinical symptoms and duration for A(H1N1pdm09 virus infection were similar to what has been described in the region and internationally. The outbreak helped to identify gaps in the incipient national influenza surveillance system. The lessons learned from the outbreak response also included the need to expand laboratory capacity to manage increased demand for specimen testing during epidemics; and to strengthen technical capacity at the district level to ensure the successful decentralization of outbreak management. Additionally, the quality of the integrated disease surveillance and response (IDSR) system must be improved in order to provide reliable surveillance data on current and future influenza outbreaks. Given the importance of ISS as a surveillance backbone during the pH1N1outbreak response, the system’s quality should be evaluated and improved to ensure timely detection of novel influenza strains with pandemic potential, and for better understanding of the epidemiology of seasonal influenza.