All, so far, are epidemiological observational studies of populations one of the weakest types of investigations because you cannot isolate all the variables. A 1998 Finnish study of 3 million children could find no association between the MMR jab and autism. At first glance, the evidence is fairly compelling (Lancet, 1998; 351: 1327-8). The Finnish National Board of Health and National Public Health Institute had launched a vaccination campaign in 1982 to give the MMR vaccine to all children at 12-15 months and then at age 6. By 1996, 3 million doses of vaccine had been given. All adverse effects were supposed to be reported. During that time, the study says, 31 children developed gastrointestinal symptoms, with 21 admitted to hospital.There are important differences between this study and Wakefield's (see box, p 3). For one, the Wakefield study began its investigation with children diagnosed with an autisticspectrum disorder and then went on to make the link with the bowel problems. The Finnish study, on the other hand, only noted children with gastrointestinal problems, among whom none had been diagnosed as autistic. In other words, nobody was actually investigating autism as a side effect.

The most important difference is that this was a passive study, which largely relied upon doctors to report side effects. As in this country, many doctors may not have connected autism with a seemingly unrelated factor like a vaccine, or indeed bothered to record it.

As for the English studies (Lancet, 1999; 353: 2026-9), these also represent massaging of epidemiological data. The authors examined 498 children diagnosed as autistic in eight North Thames health districts and compared clinical data with immunisation records to see if there was a temporal relationship between the onset of the condition and receipt of the MMR jab. They also looked at the overall incidence of autism to determine whether an increase had indeed occurred after the introduction of MMR in Britain.

According to this study, there was no link found in the 498 children studied. Overall, autism hadn't increased within a timeframe that could pin it to the jab.

The authors themselves admit that the study design and this kind of broad brush epidemiological sweep is weak.

Autism has increased dramatically in the North Thames health districts by 25 per cent every year. Although the study says that this increase occurred in children born before the MMR was released in 1998, that ignores the PHLS's 'catch up' policy of vaccinating children born between 1985-6 with MMR when they were age two or three, as soon as the triple jab was introduced.

If you take these children into account, there would be a significant increase in autism cases in children born in 1986 if the MMR jab was responsible. Indeed, the figures show just such a peak. Furthermore, there is also a large peak of parents who first noticed autistic symptoms in their children six months after the children had the jab.