Andy Ho

A cautionary tale for science reporters

Early last month, some readers sent me a story from a Canadian newspaper, The Toronto Star, headlined A Wonder Drug's Dark Side. This was about Gardasil, a vaccine given to females aged 10 to 25 to prevent infection with the human papilloma virus (HPV).

As this shot is available in Singapore, these readers who were parents wanted to know what to make of the paper's claim to have found a link between the vaccine and illness in young women getting the shot. They asked if they should give their daughters the vaccine.

In 10 per cent of cases of HPV, which is a sexually transmitted infection, the virus causes cervical cancer. Gardasil prevents that cancer, but the 123-year-old Toronto broadsheet claimed it had found that "since 2008, at least 60 Canadians have experienced debilitating illnesses after inoculation" with Gardasil.

It focused on several young women who became ill after getting the shot. Because the vaccine's safety and efficacy are well established, many experts took The Star to task.

It initially responded by attacking its critics, calling one "a rural doctor" and others "loony American right(wingers)".

But after two weeks of responding combatively to criticism, it finally admitted that readers were likely to have concluded from its report that it "had uncovered a direct connection between a large variety of ailments and the vaccine".

Conceding that the report was alarmist, it said: "There is no scientific medical evidence of any 'dark side' to this vaccine." It had not looked at the many studies attesting to the vaccine's safety. There was "confusion between anecdote and evidence", it said.

It told of young women racked with chronic debilitating pain, some even dying suddenly after getting the shot. This "litany of horror stories... gave the incorrect impression that the vaccine caused the harm", it confessed.

Eating humble pie, the paper agreed that its report had not met "the standards of responsibility expected in evidence-based science journalism". It took the rare step of unpublishing the report from its website.

The parents who had e-mailed me the original report wondered what to make of its removal from the website: If Gardasil is safe, how did a reputable paper get it so wrong?

To answer these questions, one must identify how the paper was led astray.

The first error was mistaking association for cause-and-effect. Noting that the women developed health problems some time after the shots, it inferred that the vaccine was the cause. Of one patient, it said: "After the first Gardasil shot, the pain started in her back. After the second and third shot, it spread to other joints, her knees and hips."

Just because A occurred before X, The Star inferred that A caused X. But other things may have occurred in the intervening weeks.

Who helped it make the leap? A naturopath whom one of the patients saw asked the woman if she had been vaccinated recently. Preferring "natural" cures, naturopaths are dead set against drugs and vaccines. This one told the patient's family that aluminium in Gardasil was causing the health problems.

While Gardasil does contain aluminium salts, these have been used safely in vaccines for 90 years now. (The only other approved HPV vaccine called Cervarix also contains aluminium salts.) The aluminium helps to boost the body's immune response to the weakened virus proteins found in a vaccine.

The naturopath prescribed a course of "natural" treatment. Serendipitously, the patient's health did improve afterwards. The Star mentioned in passing that the patient was later diagnosed with a condition called fibromyalgia.

But wait a minute. Patients with fibromyalgia experience unending bodily pains over several months with no obvious cause. These pains lead to sleep problems, tiredness and depression.

It is likely caused by mistakes in how the nervous system processes stimuli: Those that are normally not painful are interpreted as painful while stimuli that are somewhat painful are interpreted as being very painful.

Much commoner in women than men, the pain waxes and wanes in fibromyalgia. So when there is symptom improvement with any treatment, natural or otherwise, it might well be just the disorder entering its waning phase. Thus getting better is no proof that aluminium in Gardasil was the problem.

The newspaper's second error was its misuse of databases that report adverse side effects of drugs and vaccines. These are the Vaccine Adverse Event Reporting System (VAERS) in the US and the Vigilance Adverse Reaction Online Database in Canada.

The Star said it "found thousands of suspected cases, including more than 100 deaths" with Garda-sil on VAERS and more than 50 "serious" incidents on the Canadian database. This, it thought, established Gardasil's guilt.

As early warning systems that pick up side effects of drugs on the market that were not detected in clinical trials done before they were approved, these databases have to be open, so anyone can submit a report. However, critically, no one verifies any report. So someone reporting a side effect of a drug by no means proves that the drug caused it.

The paper's third error was to accept uncritically what some doctors had to say although they were known to be either anti-vaccine activists or else those with funding from the maker of Cervarix, the HPV vaccine used in Europe. This is a lucrative market: A full regime of Gardasil requires three doses - about $500 in all - plus boosters later, so money may skew some expert opinions.

The upshot is that science journalists must know the experts they rely on, and study the research literature in order to question the experts closely.

The Toronto Star provided a reminder to science journalists everywhere about the need to provide readers with evidence-based information when reporting on public health issues.