Volume 12, Number 4—April 2006
Research
1951 Influenza Epidemic, England and Wales, Canada, and the United States
, Theresa Tam†, Douglas Fleming‡, Mark A Miller*, and Lone Simonsen*
Table 3
Test for a pandemic signature in age-specific, influenza-related death rates of 1951 influenza epidemic, Canada*
| Influenza seasons | Age coefficient | Intercept | p value† |
|---|---|---|---|
| Major epidemic seasons, 1950–1999‡ | |||
| Average (SD)§ | 0.15 (0.02) | –7.9 (1.9) | Reference |
| Minimum-maximum | (0.11–0.18) | (–9.9; –4.1) | Reference |
| 1951 epidemic (A/H1) | 0.14 | –5.2 | 0.30; 0.13 |
| 1957 pandemic (A/H2) | 0.10 | –2.7 | <0.001¶ |
| 1968 pandemic (A/H3) | 0.08 | –2.5 | <0.001¶ |
*Coefficients of exponential models explaining the increase in pneumonia and influenza (P&I) excess death rate with age for the 1951 epidemic, 1957 and 1968 pandemics, and other major epidemics in the interpandemic periods, 1950–1999. Exponential models use 5-year age groups starting at age 55 years and ending at age >90 years (see Figure 4). All R2 between 0.85 and 0.99 (R2 of linear model applied to log-transformed P&I excess death rate, equivalent to an exponential model for untransformed P&I).
†p value for comparison of respectively age and intercept coefficients with the "null" distribution of major epidemic seasons. Based on 1,000 bootstrap samples.
‡"Null" distribution of major influenza seasons in the interpandemic periods, 1950–1999. Based on 17 seasons that had seasonal P&I excess mortality rates above the median of all seasons (including seasons in the A/H1N1, A/H2N2, and A/H3N2 era but excluding the 1951 epidemics and the 2 pandemics).
§SD, standard deviation based on bootstrap resampling of the null distribution of observed epidemic patterns.
¶Evidence of pandemic signature (not consistent with an epidemic mortality age pattern).