Fear can drive us to make hasty decisions that are not necessarily based on good and robust data. This can include rushing into a mass population vaccination program before we have adequate safety and efficacy data. Also we might try to deliver the vaccine in ways that put people at increased risk for other infections (e.g. by using multi-dose vials).

With the current H1N1 (swine flu) pandemic, a fundamental issue we need to resolve before we rush into any mass vaccination program is what is its virulence and who are the most susceptible to complications and death.

One group with a disproportionate risk are pregnant women. Their relatively frequent and serious complications have been a surprising element of this influenza outbreak. In the US their rate of hospital admissions is about four times higher than the general population and 100 times higher than for other women of their age.

If we have a virus that spreads readily, and is associated with a 1 per cent or more mortality, that needs to be approached differently to a virus that has similar (or less) virulence than many seasonal influenza strains and may only be associated with case fatality rate of about one in 10,000 people.

Surprising, readily available and detailed data on the age of those who have died, their underlying diseases and even admissions to ICU has been difficult to find so far to answer these important questions. NSW Health has, however, recent and excellent data on its website that helps us greatly. Hopefully similar detailed data for all of Australia will become available soon on the web.

In NSW up to August 11 2009, there have been 4677 lab confirmed cases, 1043  admitted to hospital, 154 admitted to ICU and 32 deaths associated with H1N1 (31 of whom had underlying chronic conditions). Reported rates per 100,000 population for ICU admissions were two and for hospital admissions were 15. My estimation of the death rate was 0.4 per 100,000 population.

What we don’t know is what percentage of people in NSW has already been infected by swine flu (and this will affect any calculations we make of virulence). It is likely that at least 10 per cent of the population have already been infected. If we assume a 10 per cent infection rate, then the rates per 10,000 people infected for hospital admissions, ICU admissions and deaths are 15, 2 and 0.4 respectively. This gives an overall case fatality rate of 0.004 per cent.

The other surprising element of the NSW data is that most ICU admissions and deaths were in those older than 40, with the peak age group at risk for more serious complications being the 50-60-year-olds. Only nine of the 32 deaths were in people younger than 50, and 55 per cent of ICU admissions with H1N1 were in those older than 40.

What does this all mean?

More than 99 per cent of people with H1N1 will have a relatively mild or moderate illness. They will fully recover and not need hospital admission. Obviously, however, some people do get seriously ill with H1N1 influenza but it appears to be mainly those with recognisable risk factors (and for this infection this includes pregnancy and obesity). Thus our national “Protect” policy needs to be followed as this is designed to identify and promptly assess and treat those most at risk.

It also means we need to be very careful with any vaccination program, given the low morbidly and mortality associated with this infection in the vast majority of those infected so far. Some argue that the virus may mutate and become more virulent. This, however, does not appear to have happened previously. The increased deaths in the second and third waves in 1918/19 Spanish flu were due to bacteria and pneumonia and not any more virulent virus. In addition, even if the virus did mutate, this is likely to mean that antigenic drift occurred and hence any vaccine that was rolled out early based on the original virus will probably have less efficacy.

In nearly all states we seem to have passed the peak of new swine flu cases with overall influenza-like illness rates not higher than what has occurred in previous years (e.g. 2007) and less than in 1997. We are not seeing second peaks occurring despite children returning to school after their winter holidays. Thus we are not likely to see new waves of swine flu until winter next year. This means we have time to look at the best way of rolling out any national vaccine program. We do not need to unduly rush its introduction. We have time to learn from whatever experience occurs in the northern hemisphere’s winter season and see what effects the vaccines have there when introduced.

If a rushed mass vaccine program is envisaged, then there may the temptation to use vaccine in multi-dose vials. There have been many examples in the past of blood-borne virus infections been transmitted by multi-dose vials when some vials become contaminated after initial use. Bacterial infections have also occurred. This is why nearly all International Infection Control Guidelines recommend against using multi-dose vials.

A rapid vaccine roll-out may also mean that adequate safety evaluations may not have been done and/or be available for examination by our regulators. In the 1970s when a new swine flu epidemic was a concern in the US, more than 40 million people were vaccinated. However, about one per 100,000 of vaccine recipients developed an ascending a paralysis (the Guillain-Barré syndrome). We need to be sure that there has been adequate safety evaluation done in many thousands of people before we roll out any vaccine, especially on any vaccines that are different to how we make or formulate seasonal flu vaccines. After vaccination starts, we need an adequate and ongoing surveillance system so that any rare side effects, such as ascending paralysis can be rapidly identified.

We also need data on vaccine efficacy. Inactivated flu vaccines may only have a 50% efficacy in young adults and children. We not only need to know not only that they work based on in-vitro antibody and neutralisation experiments but also follow up large numbers of people who come into contact with H1N1 next winter to ensure we have some clinical efficacy data following exposure to the virus. We should also consider having study groups where we know there is a very low risk of serious complications (e.g. healthy young men) we should consider the merits of doing studies with comparative arms where they do not receive vaccine or are given placebo. Such studies in smaller groups are often the only way we will ever know how effective a vaccine was and the results will has important implication then for the future use of similar vaccines.

In summary, for most in the population the recent NSW Health data and those from others, suggest that to date swine flu has been no worse than what we have seen in many years with seasonal influenza. Even the age distribution of those with complications appears to becoming closer to that seen with seasonal influenza in comparison to what was seen early on in this epidemic. Thus, it is important that we do not rush into any mass vaccine program before we have reasonable safety and efficacy data. We also have the ability to wait until early next year and see what have been the effects or problems with vaccine when used in the northern hemisphere and their upcoming swine flu winter season. It is also important that vaccines are administered with the lowest other associated infective risks — which means multi-use vials should be avoided.

Peter Collignon is a professor at the Canberra Clinical School. Australian National University.