When reactor No 4 at the Chernobyl nuclear power plant exploded 20 years ago on Wednesday, the world had no experience of where such a disaster might lead.

Though the aftermath of the Hiroshima and Nagasaki atomic bombs had been studied extensively, there was no precedent for a civil nuclear catastrophe. The plume of fallout that spread across Europe was widely predicted to cause hundreds of thousands of deaths.

In the years since 1986 it has become clear that the consequences have not been this severe. The official death toll stands at 62, including the 31 plant workers who died in the fire. Chernboyl has caused almost 5,000 cases of thyroid cancer, but only 15 of these have been fatal.

The most serious long-term effects, it is true, have yet to be felt: tumours take time to grow and the World Health Organisation (WHO) estimates that another 9,000 people will die. Chernobyl remains by far the worst civil nuclear incident in history but it has a way to go to catch up with the Bhopal chemical leak of 1984, which killed 15,000.

This picture is not above challenge. Particularly telling is the charge, made in Nature this week by two scientists, that the WHO has presented its figures with more certainty than is justified. Dillwyn Williams and Keith Baverstock, both respected, independent experts, pointed out that even 20 years is too short an interval to be sure of the long-term effects of radiation. Many health impacts of the Japanese atomic bombs did not come to light for up to45 years. By suggesting a “certainty unwarranted by the underlying statistics”, the WHO has opened itself up to critics with an axe to grind.

This is fair comment, as is their call for a rigorous study of the populations exposed to Chernobyl’s fallout, along the lines of research conducted by Japan since 1945. Without it “speculation about Chernobyl’s human cost will be unconstrained by hard evidence and interested parties will be able to exaggerate or underplay the consequences”. A prime example of this is another criticism of the WHO’s figures published this week: a Greenpeace report claiming that Chernobyl has contributed to 200,000 deaths and stands to cause 93,000 more.

If official estimates are too conservative, this one strays in the other direction. As campaigning NGOs are wont to do, Greenpeace has exploited uncertainty to quote results at one extreme end of the error bar, the end that suits its anti-nuclear agenda.

An accurate account of Chernobyl’s impact is very important, for targeting aid as well as refuting conjectures that serve political goals.

It is not necessarily relevant, however, to decisions about nuclear safety that have to be made today. Chernobyl might be our only experience of a nuclear catastrophe but its impact on health is unlikely to be duplicated by the failure of a modern reactor.

You may have noticed that I have not described Chernobyl as an accident: that is because it was not. It was a badly designed reactor, tested to destruction during an ill-considered safety check. A comparable explosion could not happen in modern plants, which are built to fail-safe. Once the fire had started it was managed with a staggering incompetence that amplified its effects. An exclusion zone was drawn arbitrarily in a circle, taking no account of wind direction. Evacuations were delayed so that May Day parades could proceed as normal.

The victims of Chernobyl are the victims of poor design and crisis management just as much as they are the victims of radiation. Whatever the ultimate toll turns out to be — and it’s too soon to judge — it is important to recognise it as that of a worst-case scenario. Only then can balanced debate about the safety of nuclear power begin.

Mark Henderson is the Times science correspondent