I’ve written a couple of articles in the last couple of months on the general theme of the abject state of science/health journalism in the mainstream press – see Why is the Indy Shilling for Big Pharma? and Dumbing Down Dementia– the second of which includes this observation:

“…if one wishes to find good quality source material for a good debunking, then I can heartily recommend that you open any newspaper and scour the news sections for the words ‘health correspondent’.”

Today’s proof of that statement comes from the Guardian, and specifically from the current health correspondent, Polly Curtis, who, so far as one can tell from information in the public domain, appears to possess solid credentials as a generic journalist – runner-up in the feature writer of the year category in the 2000 Guardian Student Media Awards, followed by a steady progression through the ranks at the Guardian from ‘contributor’ and journalist to, first, education correspondent and, now, health correspondent, but exhibits precious little to suggest that she is adequately qualified to correctly evaluate and report accurately on the content of specialist research/journal papers in anything other than a superficial manner.

Now I may, conceivably, be being a little harsh in my evaluation of young Polly’s apparent background where, in truth, she is actually packing a raft of GCSEs, A Levels and a degree in the natural and/or social sciences, some of all of which are relevant to the subject at hand (health) but when set against a benchmark of what science/health journalism looks like when written by a real scientist, of the likes of Ben Goldacre, then it would be fair to say that on the evidence of her efforts today, I have my doubts…

Cannabis use linked to 40% rise in risk of schizophrenia

You knew that was coming, didn’t you? Even if you didn’t see this being trailed on Breakfast TV this morning, you just knew that with cannabis/mental health scare stories being the current flavour of the month, there was going to be a very good chance that this would be the subject of Polly’s article.

Smoking cannabis increases the risk of schizophrenia by at least 40% according to research which indicates that there are at least 800 people suffering serious psychosis in the UK after smoking the drug.

Increases the risk of schizophrenia by 40% from what?

What is the risk of developing schizophrenia if you don’t smoke cannabis, and does that risk apply evenly across the whole population or is contingent on other causal or contributory factors that modify the level of base risk according to individual circumstances? Is this 40% figure a measure of the increase in annual risk or lifetime risk?

Without any of that additional information, the assertion that smoking cannabis increases the risk of developing schizophrenia by 40% is entirely meaningless, even if the quoted figure is true. You simply cannot make an informed evaluation of personal risk from such a statement unless you can assess the percentage increase in risk against a known baseline figure.

Mental health groups called on the government last night to issue fresh health warnings and launch an education campaign to advise teenagers that even light consumption of the drug could trigger long-term mental health problems.

Without wishing to sound cynical – well they would, wouldn’t they?

The fact of the matter is that however one views the motives the of the voluntary sector in general, such an education campaign amounts to, at the very least, as sizeable raft of free publicity for these groups to piggyback on – and that’s if the government decides to develop and operate such a campaign in-house. If, on the other hand, its decides to commission in resource materials or outsource delivery then there’s also the prospect of a few nice fat government grants as well.

The findings came after a rush of ministers declared their cannabis-smoking pasts and an order from the prime minister for officials to consider whether the drug should be reclassified amid fears about its more potent “skunk” form. Last night the Home Office said the research would be considered in that review.

And? I suppose that’s a useful bit of filler that describes the political context behind the current moral panic about cannabis use, but it still tells us nothing of real substance.

The study, an analysis published in the Lancet medical journal of previous research into the effects of the drug on tens of thousands of people, provides the most persuasive evidence to date that smoking cannabis can cause mental illness years after people have stopped using it.

Does it?

Let’s see what the Lancet actually has to say, at least in summary (to get the full article requires a subscription or payment of a day rate of $30).

Background

Whether cannabis can cause psychotic or affective symptoms that persist beyond transient intoxication is unclear. We systematically reviewed the evidence pertaining to cannabis use and occurrence of psychotic or affective mental health outcomes.

Fair enough.

The starting point for this research is ‘we’re not sure if there is evidence to support a causal link between cannabis and long-term mental health problems, so lets review the current evidence that’s available’.

It’s a meta-analysis study, which a perfectly valid research technique but one not without its potential problems and pitfalls, which is something I, unfortunately, cannot explore in detail without access to the full article, so for the purposes of this article we’ll take it as read that the study has been competently conducted and provides a valid set of outcomes and conclusions.

Methods

We searched Medline, Embase, CINAHL, PsycINFO, ISI Web of Knowledge, ISI Proceedings, ZETOC, BIOSIS, LILACS, and MEDCARIB from their inception to September, 2006, searched reference lists of studies selected for inclusion, and contacted experts. Studies were included if longitudinal and population based. 35 studies from 4804 references were included. Data extraction and quality assessment were done independently and in duplicate.

Okay, that’s also very useful is as much as it tells use two very important things.

First, given the scope of the publication databases listed as having been searched for suitable studies, it seems highly likely that the data used in the study in not UK specific. Nothing wrong with that, but it does mean that we may find some statistical discrepancies when comparing the output data in this study with actual data from the UK.

Second, this is research is based on data drawn only from longitudinal population studies. Again, there is nothing wrong with that, but what it does tell us is that it’s highly unlikely that the research will be able to provide anything more substantial than indications of any observed correlations between cannabis use and mental health problems – which means that any claims of causation deriving from or attributed to this research need to be regarded sceptically.

Findings

There was an increased risk of any psychotic outcome in individuals who had ever used cannabis (pooled adjusted odds ratio=1·41, 95% CI 1·20–1·65). Findings were consistent with a dose-response effect, with greater risk in people who used cannabis most frequently (2·09, 1·54–2·84). Results of analyses restricted to studies of more clinically relevant psychotic disorders were similar. Depression, suicidal thoughts, and anxiety outcomes were examined separately. Findings for these outcomes were less consistent, and fewer attempts were made to address non-causal explanations, than for psychosis. A substantial confounding effect was present for both psychotic and affective outcomes.

Now that is very interesting, because what the study is reporting – if we assume that baseline for no increased risk is a pool adjusted odd ratio of 1 – is that the 40% increase in risk cited by Curtis applies to ‘any psychotic outcome’ and not just schizophrenia.

There are actually a wide of conditions, diseases and even situations that can give rise to psychotic episodes and outcomes, ranging from the psychological (schizophrenia, bi-polar disorders, severe clinical depression and/or stress) to the organic (brain tumours, multiple sclerosis, Alzheimer’s disease and other dementias, syphilis – in rare cases even influenza and mumps can induce psychotic episodes) to the broadly self-inflicted, which includes cannabis but also covers a wide range of other drugs that can, and do, have psychoactive effects ranging from dear old alcohol through a range of prescription drugs (barbiturates, benzodiazepines, some anti-depressants and anti-epileptics) to the classic range of ‘street drugs’ which encompassed pretty much anything that’s cocaine or amphetamine-based plus all the usual hallucinogenics.

It also cites a doubling of the risk of psychosis (a 100% increase) in data relating to long-term use of cannabis and indicates some of the potential shortcomings of the research that the study reviewed. There’s both consistency in outcomes and in efforts to eliminate external factors in studies looks specifically at the incidence of psychosis, much less consistency and effort to account for external factors is studies focussing on depression/anxiety and a substantial confounding effect across all studies that needs to allowed for, statistically, before reaching any valid conclusions.

On the whole, that looks pretty good as research methodologies go, certainly good enough to suggest that this is, indeed, a competent and exacting piece of research and give the study’s conclusions some serious consideration.

Interpretation

The evidence is consistent with the view that cannabis increases risk of psychotic outcomes independently of confounding and transient intoxication effects, although evidence for affective outcomes is less strong. The uncertainty about whether cannabis causes psychosis is unlikely to be resolved by further longitudinal studies such as those reviewed here. However, we conclude that there is now sufficient evidence to warn young people that using cannabis could increase their risk of developing a psychotic illness later in life.

So, what the study actually concludes is that there is sufficient statistical evidence to take the issue of risk seriously although not enough to prove a clear causative relationship or accurately define the parameters of that relationship. Indeed the study goes on to indicate that, so far as statistical population studies go, we’ve pretty much hit the wall in terms of what they can actually tell us about the possible relationship between cannabis and mental health, which means that any further answers are going to have to come from those strands of research currently examining this relationship in terms of genetics and biochemistry.

In short, it doesn’t really tell us anything new and it certainly doesn’t establish a clear causal relationship between cannabis use and long term mental health problems, but it does suggest that there’s enough statistical evidence of risk to justify warning young people of the existence of such risks.

All very reasonable and nothing that’s really the stuff of scare stories and moral panics if understood (and presented) properly.

Getting back to Curtis’ article, the good news is that she’s at least taken the time and trouble to actually include some of the research outcomes in a form that can be evaluated.

The overall additional risk to cannabis smokers is small, but measurable. One in 100 of the general population have a chance of developing severe schizophrenia; that rises to 1.4 in 100 for people who have smoked cannabis.

The bad news is that she’s made a complete hash of presenting this information.

Notice the shift in tone – ‘the overall additional risk is small, but measurable’ – which is certainly true, but not what was implied at the start of the article when she was banging on about a 40% increased risk.

She also claims that this increased risk relates specifically to the possibility of developing ‘severe schizophrenia’, which is certainly not what the summary given by the Lancet suggests – remember that refers to the increase risk of ‘any psychotic outcome’, not just schizophrenia and certainly not just ‘severe schizophrenia’ – unless there is such a specific reference in the full article then it looks very much as if Curtis is over-egging the pudding and misrepresenting the conclusions of the study in terms that suggest that the qualitative elements of risk (i.e. the severity of the condition linked to cannabis use) are significant greater than those indicated by the actual study.

But the risk of developing other psychotic symptoms among people who smoke large quantities or are already prone to mental illness is significant, the researchers say.

People who smoke cannabis daily have a 200% increased risk of psychosis.

Well, yes, the study does indicate an increased risk with regular use and higher dosage, but that risk is a little over double the baseline risk (actually 109% greater) not the 200% that Curtis cites – a basic mathematical error, one would presume, but one that still gives a distorted picture of the actual outcomes of the research.

Moreover, while we’re talking about baseline risks, the 1 in 100 figure cited is the generally quoted figure for schizophrenia in the general population of the US, the actual figure given by the National Statistical Office for the UK is around 1 in 200 for all psychotic disorders, not just schizophrenia. In terms of assessing risk, this is actually rather important as what it does suggest is that even allowing for the degree of increased risk noted in the research study, the actual risk of a regular cannabis user developing a psychotic disorder in UK is only about the same as base risk of schizophrenia in the US.

In fact, if one narrows the scope down just to schizophrenia, then the generally cited rate of prevalence in the UK is only 3 cases per 1000 population with an annual incidence rate of between 0.1 and 0.2 cases per thousand, which makes accounting for the differences between US and UK rates even more important.

What this may well indicate, more than anything else, is marked differences in diagnostic practices between the UK and US that need to be considered and factored into our calculations, and which certainly needs to be accounted or allowed for in assessing the value of this research in terms of public policy.

They estimate that 14% of 15- to 34-year-olds currently suffering schizophrenia are ill because they smoked cannabis, a figure previously thought to be between 8% and 10%.

Because?

How the hell do you get ‘because’ from a study that concludes:-

The uncertainty about whether cannabis causes psychosis is unlikely to be resolved by further longitudinal studies such as those reviewed here.

The answer is, ‘you don’t’, because, as yet, there is no definitive proof of causation and the most promising line of inquiry at the present time suggests that the risks of developing psychosis as a consequence of cannabis use are likely to be confined to a segment of population that has a genetic predisposition towards mental health problems to begin with.

Where, exactly have those figures come from?

According to the current diagnosis rates about 800 people would have been spared schizophrenia if they had not smoked cannabis.

Ah, now I see – they come from mixing and matching incidence rates (i.e. numbers diagnosed each year) with estimates of risk derived from prevalence rates.

Having done a bit of scouting around to see if I could find any reported studies where a comparison of outcomes between them and this new study might give the figures cited by Curtis, arriving eventually at this report, which appears to fit the bill.

If cannabis causes schizophrenia – and that remains in question – then by 2010 up to 25 per cent of new cases of schizophrenia in the UK may be due to cannabis, according to a new study by Dr Matthew Hickman of the University of Bristol and colleagues, published in Addiction journal.

Note, once again, the big ‘if’.

Not only that, but if we look at this study’s reported methodology, we find:

The research study matches historic trends in cannabis use and exposure from a national population survey against estimates of new occurrences of schizophrenia in three English cities (Nottingham, Bristol and the London Borough of Southwark). The researchers assess what might happen to schizophrenia cases if we assume a causal link between cannabis use and onset of psychotic symptoms, an association widely recognised by some psychiatrists and researchers and considered recently by the Advisory Council on the Misuse of Drugs.

Now there is a potential problem with that methodology and that is:-

Ethnic minority groups are at increased risk for all psychotic illnesses but African-Caribbeans and Black Africans appear to be at especially high risk for both schizophrenia and mania. These findings suggest that (a) either additional risk factors are operating in African-Caribbeans and Black Africans or that these factors are particularly prevalent in these groups, and that (b) such factors increase risk for schizophrenia and mania in these groups. (source)

In terms of demographics, the Black population of Southwark amounts to 25.9% of the total population of the borough; for Nottingham its 4.34% and for Bristol its 2.32%.

For the UK as a whole, the Black population amounts to 1% of the total population, so the Bristol study is taking its data from areas in which a ethnic minority population that is known to be at increased risk of developing psychotic disorders is over-represented in comparison to the national population by anything from 2 to 25 times.

Not to mention that the study, itself, works from a premise that assumes that a causal relationship between long-term mental health problems and cannabis use will be established.

And is all that were not enough…

The researchers said the evidence was the strongest yet to show that cannabis caused psychotic mental illnesses, and not just that people who were ill smoked more. Dr Stanley Zammit, of Cardiff University, said: “We think the evidence is such that we need a new official warning about the risk.”

At the risk of repetition, the summary in the Lancet explicitly states:

The uncertainty about whether cannabis causes psychosis is unlikely to be resolved by further longitudinal studies such as those reviewed here.

And you’ll notice that the apparent claim of causality attributed to the researchers is not an actual quotation, in fact the only direct quotation given calls for the a warning about the apparent risks, which is supported by the conclusions of the research study but only the context of statistical correlations and not proof of causation.

Paul Corry, director of public affairs at the mental health charity Rethink, echoed calls for more warnings but said it was not evidence in itself that cannabis should be reclassified. “Rather than focusing its attention on the reclassification debate, the government would do well to crack on with the more important job of informing the public about the health implications,” he said.

Well yes, quite…

But then before doing that, should we not be just a little more exacting in how those are expressed, in terms of giving an accurate assessment of the evidence of risk provided by these studies and by not claiming causation when there is still no definitive evidence to support such claims?

There is rather more to this than just a blogger venting off a bit of steam about yet another piece of shoddy health journalism. In matters of public health education, particularly with young people and especially if what one is trying to achieve amounts to behaviour modification, there is nothing more important than trust – and there is nothing more certain to undermine trust than the misreporting and misrepresentation of evidence.

If you want young people to take on board the evidence of possible risks arising from the use of cannabis you have to be upfront in giving them accurate information, because no soon as you are caught cheating and overplaying your hand then that’s it, you’ve blown it and they won’t listen to you.

Moral panics might be good for selling newspapers, but they’re a lousy way of going about public health education.

UPDATE

Having worked your way through all that, this is the Daily Mail’s take on the story…

A single joint of cannabis raises the risk of schizophrenia by more than 40 per cent, a disturbing study warns.

The Government-commissioned report has also found that taking the drug regularly more than doubles the risk of serious mental illness.

Overall, cannabis could be to blame for one in seven cases of schizophrenia and other life-shattering mental illness, the Lancet reports.

The grim statistics – the latest to link teenage cannabis use with mental illness in later life – come only days after Gordon Brown ordered a review of the decision to downgrade cannabis to class C, the least serious category.

I suppose the only saving grace is that neither of the journalists bylined on the story – Fiona Macrae and Emily Andrews – appears to be billed a ‘health correspondent’, although its questionable as to whether they should even be billed as journalists.