Table 3.

Influenza-associated excess mortality rates per 100 000 person-years in southern Brazil, Singapore, Hong Kong, USA and South Africa

Country	Period of study	Statistical method	Underlying pneumonia and influenza deaths (/100 000 person-years)	Underlying respiratory and circulatory deaths (/100 000 person-years)
Southern Brazil	1980–2008	Serfling regression model  using monthly number of deaths	All ages: 1·4 ⩾60 yr: 10·0	All ages: 9·2 ⩾60 yr: 86·6
Singapore [13]	1996–2003	Negative binominal regression  model using monthly number of  deaths and monthly proportion  of positive influenza test results	All ages: 2·9 ⩾65 yr: 46·9	All ages: 11·9 ⩾65 yr: 155·4
Hong Kong [14]	1996–1999	Poisson regression model using  weekly number of deaths and  weekly proportion of positive  influenza test results	All ages: 4·1 ⩾65 yr: 39·3	All ages: 12·4 ⩾65 yr: 102·0
USA [15]	1976–2007	Poisson regression model using  weekly number of deaths and  weekly proportion of positive  influenza test results	All ages: 2·4 ⩾65 yr: 17·0	All ages: 9·0 ⩾65 yr: 66·1
South Africa [28]	1998–2005	Serfling regression model using  monthly number of deaths	⩾65 yr: 42·0