Purpose: Mass childhood vaccination against Hepatitis A virus (HAV) in areas of high or intermediate endemicity has been highly effective. In some low endemic areas, disease transmission has fallen dramatically in the absence of vaccination programs. It is unclear whether this is due to secular disease trends, periodic variation, or both. Lack of ongoing disease transmission might suggest that evaluating cost-effectiveness of vaccination programs is unnecessary.

Methods: Monthly incidence of HAV reported cases were obtained from the Canadian Disease Notification Reporting for 1980-1999. A power spectral analysis (PSA) of the monthly incidence time series was conducted to decompose the series into pure component frequencies and to determine their relative strength. A dominant peak at frequency F corresponds to periodic outbreaks every 1/F years. Less-dominant frequencies can also be identified.

An inter-epidemic interval can also be predicted by a Susceptible(S)-Infected(I)-Recovered(R) dynamic model. An SIR model captures the time evolution of S(t), I(t), and R(t) using coupled equations of time-dependent rates of change. The inter-epidemic interval in an SIR model is estimated by T=2Π√AD where ‘A' and ‘D' are the mean age of infection and duration of infectiousness, respectively. ‘A' and ‘D' were derived using age-specific incidence data and a literature search.

Results: For Canada, the average HAV incidence rate for 1980-1999 was 6.5/100,000, with three peak periods in 1984-1985, 1990-1992, and 1995-1997.

The power spectrum analysis revealed a dominant peak at 0.15/year, corresponding to outbreaks every 7 years. The SIR model predicted outbreaks every 7.5 years (i.e., A=25 years, D=3 weeks). The existence of this peak suggests that much HAV in Canada comes from transmission in the general population, beyond sporadic outbreaks in high-risk groups (1990-1992, 1995-1997).

The PSA also exhibited a spectral peak every year in December (i.e., a 28% increase from 112 to 142 cases for other months), suggesting a role for seasonality in HA transmission.

Conclusions: Results from the PSA indicate ongoing transmission in the general population, and the SIR model suggests that another incidence peak in the coming years is possible. Cost-effectiveness studies using a dynamic model of different primary prevention options are indicated. Evaluation of periodicity can elucidate transmission patterns, and aid in planning cost-effectiveness studies by suggesting whether they should be performed, and if so, which strategies should be considered.