During the study period from September–December 2009, 5.69% of the Korean population was prescribed antiviral drugs and 2.3/1,000 people were admitted as confirmed or suspected cases of infection. The proportion of females was higher among severe infection cases. A dominant prevalence of female cases was also reported in Canada 
. However, a gender-specific infection could not be concluded clearly, because other variables associated with females, such as pregnancy, 
were not included in the present analyses.
Kim et al
. (2010) 
studied the trend of the spread of this novel influenza strain by comparing three monitoring tools used in Korea during the pandemic. The patterns of spread from the three methods were generally similar but details, such as peak time, were different. We found that illness severity was greater among patients who were ≥ 60 yr, who were in a low-income group, and who had comorbidities. This finding persisted in the results for analysis of the confirmed group only.
Most previous studies have reported the characteristics of novel influenza A (H1N1) lab-confirmed cases. However, as novel influenza A (H1N1) became a pandemic, routine testing for the infection was not recommended, and prompt treatment was given instead to mitigate damage from the infection. Therefore, an analysis of only confirmed cases would certainly lead to selection bias in the results. Because the entire population that was given antiviral drugs, including those that were treated during the peak period of the pandemic, were considered in this study, we were able to conclude that there was no difference between the results for all cases and the results for only confirmed cases. This increases the confidence of our findings.
The age-specific immunity of novel influenza A (H1N1) reported in a previous study by Miller et al
. (2010) 
suggested that a great portion of older people had pre-existing immunity to the novel influenza A (H1N1) virus as a result of pre-exposure to antigenically related influenza A earlier in their lifetime. Moreover, the significantly higher incidence of infection in those ≤ 19 yr occurred presumably because these patients were exposed to an increased potential for transmission within their schools, and confirmation tests were performed vigorously 
. In our study, the characteristics of novel influenza A (H1N1) appeared to be age-specific in terms of mortality. Younger aged individuals were susceptible to the influenza A (H1N1) infection and had a higher incidence of severe outcomes than those in other age groups. But, the persons in the elderly group (≥60 yr) were much more likely to suffer a severe case or even more fatal cases once they became infected. Schools in Korea were not closed nationwide during the novel influenza A (H1N1) outbreak, and the decision was left to the discretion of each school.
The incidence and admission rate in the younger age group of outpatients and inpatients were higher than those in the older group. ICU incidence was high in the groups < 10 and ≥ 60 yr. Considering that the prevalence was significantly lower in the older group, the risk of severity was assumed to be higher in those aged ≥ 60 yr. This result is consistent with our supplementary analysis of confirmed cases. The incidences in children, who were confirmed patients classified as outpatients, inpatients, and those admitted to the ICU, were higher than those in the older group, but an exception was found for deaths assumed to be caused by novel influenza A (H1N1), which were higher in the older group than those in children. The mortality rate showed a J-shaped curve with the greatest risk in those aged ≥ 60 yr (1.31/100,000). This finding can be interpreted as most young children recovered from the severe illness, whereas severe illness more often led to death in the older group.
A risk factor associated with obesity was not found in Korea, which differed from results reported in Mexico 
and the United States 
. However, being underweight was one of the risk factors for severity in our study. The reason for the reverse association between BMI and severe outcome is unclear. We used the general cut-off point of BMI as recommended by the WHO. However, Asian populations have different associations between BMI, the percentage of body fat, and health risks compared to Western populations 
. Blumentals et al. 
identified the highest influenza pneumonia rates in underweight individuals in their retrospective cohort study of UK patients and suggested an association between low BMI, malnutrition, and immune function. This was also most likely influenced by the observation that a substantial proportion of patients with a severe underlying disease such as cardiovascular disease or chronic pulmonary disease have a lean body mass in general 
. One of the limitations of this study was that information regarding the severity of the individual underlying diseases did not exist.
We showed various outcomes in the incidence and mortality among regions. Regional variations in illness magnitude may have been caused by the density and composition of the population. Approximately 49% of the Korean population lives in the capital area (Seoul) and around this area, including the city of Incheon and Kyonggi province from among the 16 cities and provinces in Korea. One social issue in Korea is that the average age of the population in rural areas is increasing; thus, it is assumed that age-specific immunity and mortality were the cause of the observed variations in incidence in the regions, together with differences in the transmission potential according to population density. After classifying the region into two groups such as city and province, the incidence of influenza A (H1N1) and the risk of severe outcomes were higher in provinces. The proportion of working people aged in their 20 s to 50 s among residents, the lower risk groups for influenza A (H1N1), was greater in the city.
We found that individual economic status influenced infection severity. Although only two groups were used in this study, consistent results were found throughout the analysis. Patients in the Medical Aid program showed greater disease severity. Accessibility to medical treatment and hygiene could differ according to individual economic conditions. This may have caused a delay in seeking medical care after symptom onset. The length of time from symptom onset to treatment is associated with illness severity 
Underlying medical conditions are a risk factor for severe influenza 
. Echevarria-Zuno et al
. (2009) 
reported that the presence of an underlying disease increased the risk of dying to by 6.1 fold in Mexico. We found that the OR for death cases with underlying disease was 2.218 (95% CI, 1.504–3.271) adjusted by all other variables including the phase of prescription refill.
Our study is an aggregated case report including almost all cases of confirmed and suspected infection during and around the pandemic peak. Individual case reports at an early stage of a pandemic are important to make appropriate policy decisions. However, while such reports at the early stage of a pandemic can explain groups susceptible to transmission; they cannot help identify risk groups in the total population. Moreover, these data cannot predict the degree of severity, because the aim of hospitalization at this time is isolation in general 
. Including the probable cases of novel influenza A (H1N1) in this study would not likely result in overestimating the incidence rate considering that novel influenza A (H1N1) is a relatively mild infection 
or even an asymptomatic infection for which a majority of cases were not captured 
Our study had several limitations. Given that we used data from the ADSS, there was an absence of detailed clinical symptom information. Data related to the type of medication may be limited in its ability to reflect the true conditions of the infection. Another limitation was that prescription information was entered by staff at hundreds of clinics across the county, which may have reduced reliability of the data, but antiviral drugs distributed from national stores were counted and rechecked by the district health center to verify their use and the number of remaining drugs. We were also unable to gather information on underlying disease severity, which precluded a conclusion as to which type of underlying disease most influenced outcomes.