Making medicines safer for all of us

This post follows on Being Black, I can’t breathe and I can’t breathe II and will be followed by Drawing Breath (I and maybe II).  Patsy Stephenson has managed to wriggle her way in again, see comments on I can’t breathe II.  Whether she managed to stage this photo or not, as a red-head Patsy has had to live with a terrible stigma – although perhaps not as bad for a woman.

Another Game in Town

I can’t breathe II features women and antidepressants.  Why the mass consumption of numbing agents?  With benzos, it was men handing them out.  With SSRIs its other women handing them out.  (While more likely to get headline abuse from their fathers, daughters are more likely to be trapped in an ongoing struggle with their mothers).

Women and ADHD is another game in town.  This used to be a disorder of young boys until Lilly started to market Adult ADHD.  Adult ADHD is not the same condition as hyperactivity in young boys.  People with childhood ADHD do not become Adult ADHD. See NZ ADHD.

Over 80% of childhood cases are boys, easily recognisable by their bee in a jam-jar activity level.  But over 60% of cases in adults are women.  Adult ADHD is not a disease – there is no disease in medicine that changes sex ratio like this, never mind so dramatically.

There is no disease in medicine where it was possible to run a simple back of an envelope experiment in 2010 in North Wales, a part of the world that had no Adult ADHD then, where doctors responded saying they didn’t believe there was such a thing, but at the same time indicating they expected that in five years from then they would be diagnosing this condition and prescribing for it. And they are.  See Adult ADHD.

There are other ways to frame things like ADHD. It can be seen as extraversion, something to cherish, an indicator of creativity.  Why would women not want to be creative?  Success in life is about finding the right relationship and work niches to make the gifts we have work for us and others rather than Stepfordizing ourselves.

It seems many of us view being damaged as the way to go. Especially women. Granted there is the possibility of weight loss on stimulants and society puts tremendous pressure on women to be svelte.

Whether extraverted or introverted, women are supposed to be able to multi-task.  Having a stimulant induced increase in focus, an increase in stereotyped looping, sounds like giving up on a birth-right.  Unless being a woman has lost its appeal.

Rapid Onset Gender Dysphoria

One of the most striking phenomena of recent times is Rapid Onset Gender Dysphoria.

Transgender states have been around for a long time. Men, mostly older men, have figured that being a woman, or something like a woman, was very appealing. And for some men, the switch seems to work well.

There are profound poorly understood issues here.

What is novel and unprecedented is a surge in the last decade of teenage girls who want to be…  It’s not even want to be boys. It seems more just don’t want to be girls.  Teenage girls are now transitioning in ever increasing numbers, unchecked by anyone – their parents, the culture, medicine, anyone.

See Transgender Meds and Snakes in a Love Drug 2.

They are also detransitioning in ever increasing numbers, leaving both themselves and the transgender community pretty confused.

Rapid Onset Death Wish

The most confusing of all is the young women seeking medically assisted dying in Belgium and the Netherlands supposedly for Treatment Resistant Depression (TRD) and soon in Canada and elsewhere.  TRD is not a real illness – See MAiD in Canada and M.A.D. Treatment Resistant Depression.  TRD is primarily caused by the psychotropic treatments these young women have been previously given.

Antidepressants are now the second most commonly taken drugs by adolescent women, who migrate into their early twenties on them – with heading toward 25% of women in their twenties on them.

The clinical trials of antidepressant drugs for ‘depressed’ adolescents are the greatest concentration of negative trials for anything ever in human history.  Essentially 30 out of 30 trials done are negative, including the two trials that got Prozac (fluoxetine) licensed.  Prozac has more negative trials in this age group than any other drug.  See Adolescent Antidepressant Trials.

In the Treatment of Adolescent Depression Study (TADS), a supposedly independent study of fluoxetine compared to placebo, there were 34 suicidal events on Prozac compared to 3 on placebo in the trial.  All written out of the script – saved in an obscure hiding place by Goran Hogberg. It appears that the data for this trial, sold as an independent, supposed National Institute of Health trial (conducted on Lilly forms and paper) has now been destroyed.  A euphemistic way of saying….

Young women given antidepressants for the last twenty years in ever increasing amounts stood very little chance of being helped.  Not being helped these days is not a recipe for your drug being stopped at some point.  Instead, mood stabilisers and stimulants get added into the mix.  The upshot of this is Treatment Resistant Depression and Medically Assisted Death.

Young White women go along with this while Black and Muslim women have so far been more cautious but are now coming under pressure, from members of their own communities sent to medical school, to drop their stone-age attitudes. Black and Muslim women are now doing it to Black and Muslim women.

Consumerism?

We have moved into a world of Medical Consumerism. MAiD, medical assistance in dying, is an ultimate expression of this.

Gender dysphoria seems like another. Medical advances appear to make it possible for the health industries to do anything to us and our bodies.  Change sex?  Sure, step this way.  Become invulnerable like Achilles?  No problem

There has always been some degree of medical consumerism, albeit against a background of things that were more readily distinguishable from current fashions. The Treatment Resistant Depression that now leads to MAiD, which according to industry is close to epidemic, is not a fashion.  It is caused by meds and before MAiD used to lead to ECT and still does.  Some services turned requests for ECT for these patients down on the basis they were unlikely to respond but this becomes more difficult if both doctor and patient are signed up to the idea. And its women mainly. If we are willing to fry the brains of people unlikely to benefit from ECT, if they insist on it, are we going to say no to MAiD?  Maybe a few times, until we get used to it.

This scenario portrays women as helpless – consumed by Vampires. Is this all there is to it?

Young women have also been ignored visionaries from Cassandra in Troy, to the young women in Fatima who put sexual abuse on the radar for the Catholic Church and didn’t back down where Freud did, and the young women who seem to have prophesied the Rwandan massacres.

Greta Thunberg now looks more like a leader for our times than pretty well anyone in political office at the moment.

While some young women lead, others are left choking on the air they breathe, or perhaps today its their pills, inhaled on the cultural air they breathe.  What would have to happen for them all – Black, Muslim, Asian, and White – to breathe more freely?

This post continues a sequence that began with Being Black and I can’t breathe. and will be continued with I can’t breathe III and Drawing Breath. 

The picture features Patsy Stephenson, one of a well-heeled group of women, including Kate Middleton – likely there for the photo-op – protesting the death of Sarah Everard in London, killed for walking down a suburban street at the wrong time. Whether they are black or white, young or old, Muslim, Christian, Jew or nothing, women like Sarah Everard are in a similar position to George Floyd.  Patsy isn’t in this position but the photo is eye-catching. 

If someone in a group said they had AIDs, others would likely get nervous.  If they add, they are on Triple Therapy everyone relaxes.

If someone said they have epilepsy but manage it without anticonvulsants, everyone would get nervous. If they say ‘only joking, I’m taking an anticonvulsant’, everyone relaxes.  (Even though lots of epilepsy can be managed without meds and the meds double the risk of suicidal or homicidal outbursts).

If a woman says she is taking an SSRI, everyone relaxes.

As though this demonstrates her competence and reduces the threat to everyone else – the men anyway, perhaps other women. In the case of other women, the reassurance may lie in ‘she’s one of us, she’s on the pills.’

It was a woman explained this to me, a psychiatrist, who was taking meds.

Forty years ago, saying you were on Valium was an admission of weakness. Feminists fought against the subjugation of women by drugs.  Ironically the benzodiazepines were more likely to lead women to rebel against the patriarchy than any SSRI does.

SSRIs are Stepford drugs blunting awkward emotions – all emotions that is.

But saying you take an SSRI, if you are a woman, is a statement of competence.

Among a younger generation of white women taking an SSRI has moved beyond being a statement of competence.  Good for you if along with eating the right foods, spending hours in the gym, you not just admit but declare you are taking Zoloft, Lexapro or whatever.  Their 800,000 followers cheer Influencers on Instagram who make declarations like this.

The word depression may slip in somewhere, but SSRIs don’t treat depressive illness. As Joe Davis in Chemically Imbalanced explains, really what is being said is I am treating a glitch – just like I wear glasses – that is holding me back from seeing and being seen properly. Holding me back from being my true self.

SSRIs numb us, ease the pain. They make us more authentically us in the way alcohol does.  At the moment, non-white and non-Christian groups are less likely to buy into this, just as they are less likely to buy into alcohol.  You are supposed to have a community for times like this when you need support.

SSRIs though give lots of people the impression they treat an illness.  They do treat an illness – it’s called dependence.  Miss your pills for a few days, and we feel desperately ill, which getting back on our pills magically relieves.  The extreme form of this illness is called Treatment Resistant Depression.

Because so many figure the pills are saving their lives (by relieving withdrawal), if anyone questions these pills, without any bidding from pharma or their doctors, millions of defenders will march out to say these pills saved their lives.  Close to 15% of the population at the latest estimate.

The Dark is for Mushrooms

Close to 25% of the population of women of child-bearing years are defenders. The last group you’d expect to insist on access to meds they might not be able to get off. Meds that double the rate of miscarriages, and double the rate of birth defects. They also drastically lower libido and make love-making a chore, so maybe miscarriage and birth defects are not an issue.

The defenders of the meds, decent progressive women, portray comments like this is an attack on a woman’s right, duty, to look after herself.  Women are too quick to sacrifice themselves for others and this has to stop.  Calls for women to think twice before taking drugs that will make them feel better are now a touch twentieth century, second millennium even, we are told.  Time to move on.

This is today’s equivalent of Edward Bernays organizing for women to march down the street smoking – as a sign of liberation, equality and access to opportunity.  See

The Dark is for Mushrooms

Preventing Precaution 

Mumsnet

Herding Women

Looking after themselves conflicts with the instincts of many women (instincts put there just by the patriarchy?) to take themselves off drugs when they find out they are pregnant, even when male ObGyns are telling them its fine to stay on treatment and an untreated illness can harm their baby more than the meds.

The ObGyns are offering quintessential neo-Medicalism (aka neo-Liberalism) as outlined in Shipwreck of the Singular. There cannot be any clearer divide between how we once thought and the neo-Medical/Liberal way of thinking than the question of taking drugs in pregnancy.  It runs smack into Margaret Thatcher’s ‘there is no such thing as society’.

A pregnancy is either an embryonic society or a weird and inexplicable weight gain that happens to some chromosomally compromised individuals.

That progressive women in particular should buy into the no such thing as society mantra shows how much the Left or progressivism has lost its way.

There is a lot going on in between the lines here.  Sure there are women who have to insist they are diseased and need treatment, maybe never taking the drugs, in order to get benefits.  Telling them they are not depressed is like whipping a lifebelt away from them.  And it’s not sin to trumpet a disorder which you are now treating in order to maintain some self-respect.  And there are times when the termination of a pregnancy may be the best option.

But there is very little sign of the women’s movement saying to women – sure play the game but remember it is a game. The women’s movement now buys into Evidence Based Medicine and believes the Fake literature (entirely ghost-written literature on drugs with zero access to clinical trial data) is the best evidence we have. Our Bodies Ourselves is now an outsourced marketing arm of Pharma. The movement is petrified of being branded Anti-Vaxx or Anti-Science. Settle for anything other than the full scientific deal – Us? No Way!

A Seat at the Table

The medical model emerged around the same time, or just before capitalism – a movement that figured on making money out of producing goods. Capitalism was and is different to mercantilism or Free Trade and has nothing to do with neo-Liberalism. See Shipwreck of the Singular.

Capitalism was novel and successful – leading to great increases in wealth and goods so much so that, while some of us rejected capitalists, most of us who weren’t capitalists figured it was worth having a seat at this table.  All could benefit from this engine which did seem to driving us ‘forward in an acceptable direction’.

Socialism (including trade unionism) was all about having a seat at the table. The communists in contrast figured we needed a completely new set up.  Socialists figured there was absolutely no point in not being at the table – so much so they executed communists like Rosa Luxembourg.

The medical model was the Table within health.  It produced solid advances in our understanding putting us in a position to correct things and able to choose whether to take a poison or undergo a mutilation in order to correct something that would otherwise kill and disable us.  It gave us new abilities to investigate things and intervene so that for most of us having a seat at this table was a no brainer.

Through to 1980 – 1990, medicine and capitalism increased health and wealth.  Within medicine the AIDs pandemic is the best example of what could be done.  An intense mobilization by activists (lay people with the disease who mostly believed in the medical model) led in remarkably short time to a life-saver if not a complete cure. The activists split into two groups – those looking for a seat at the table (mostly male) who went into FDA and got clinical trial rules and FDA culture changed and those who rejected the entire way medicine was going and called for a totally new set-up (many more women in the mix).

AIDs activism was the final flourish of medicine as was.  The vast majority of people now taking poisons have nothing wrong with them – something inconceivable in medical model terms.

Medicine has given way to a neo-Medicalism just as Capitalism has given way to neo-Liberalism.  Now financialization leads to money chasing money while poverty rises, and pharmaceuticalization has led to drugs chasing drugs while life expectancies fall.

Socialists hung around at the table too long and embraced the operationalism that made neo-Liberalism, figuring as Maggie said ‘there is no alternative’.

The Vanished

Now that the table has vanished – neither wealth nor health is increasing – socialists have no idea what do.  Their ongoing defence of the once very reasonable idea that it is important to have a seat at the table now validates a system that is causing increasing harm.

The idea of a medical model that underpins the giving of a poison or performance of a mutilation within a relationship that is at least semi-genuine still offers a glimmer of what a table we could all get around might look like.

As things stand women seem among the most likely to get cut by the cutting edge of the operationalism that underpins both neo-Medicalism and neo-Liberalism. Their music, art and science vanishes as though written on water. They too can vanish and no-one much is bothered.

Women though may be better placed than men to work out what needs to happen next. For centuries, they have been forced to figure its better have a seat at the table the men are seated around than not. Even when the cost is getting abused, or vanishing while walking down suburban streets in daylight

We can’t go back to the 1980s.  We have to go forward.  We need to redefine what a seat at the table should look like and unless the table and seating works for women, black, white and brown, rich and poor, old and young, its not likely to be a table worth sitting at.

Drawing Breath in two posts time will offer what seems to me a crazily idealistic proposal about what the table might look like – but the only viable proposal.

The Fault lies in our Stars not in Ourselves:
Randomized Controlled Trials & Clinical Knowledge

In the Beginning

In 1947, a trial of streptomycin introduced RCTs to medicine.  From then, through to their incorporation into the 1962 amendments to the Food, Drugs and Cosmetics Act, occasioned by the thalidomide tragedy, there were questions about the epistemological link between RCTs and clinical reality. Since 1962, there have been disputes about the best statistical approach to take to RCT data – whether confidence intervals are preferable to significance testing, for instance. There have also been efforts to account for a heterogeneity of treatment effects (HTE) within the wider Evidence Based Medicine (EBM) movement, which touch on the issues raised here, but this questioning assumes RCTs connect with clinical reality and the only task is one of smoothing some statistical edges.

Repeated characterizations of RCTs as offering gold-standard evidence likely leave many clinicians thinking these trials have a solid epistemological foundation, even as clinicians recognize difficulties in translating from population or average effects to individual patients. In legal settings, RCTs are pitched as generating evidence that is generalizable and knowledge that lies within confidence limits in contrast to the views of clinicians and case reports.

Pre 1962: Hill, Fisher and Randomization

A Medical Research Council trial of streptomycin in 1947 demonstrated the feasibility of randomization as a control of the subtle biases involved in evaluating a medicine. Tony Hill, the MRC trial lead, got the idea of randomization from a horticultural thought experiment about fertilizers outlined two decades previously in which Ronald Fisher proposed that randomization could control for unknown confounders. Hill thought that randomization might control for the difficult to detect ways in which clinicians steer patients likely to respond well into an active treatment arm.  Hill’s randomization was a method for fair allocation, not a means of controlling for the unknowns linked to doctors not knowing what they were doing (Healy 2020).

Hills’ trial missed the tolerance that develops to streptomycin and the deafness and other problems it causes – information evident in a prior trial of streptomycin that controlled for confounders in the then standard way and depended on clinical judgement (Healy 2020).

RCTs brought statistical significance in their wake because Fisher argued that the only things that can interfere with expert judgement not being correct every time are unknown confounders and chance. Significance testing could control for chance and randomization for unknown confounders. Fisher’s model had an anchor in the real world – an expert whose judgements were invariably correct – such as offering a view that wearing a parachute if you jump from a plane at 5,000 feet will save your life. For Fisher, experiments were a way of demonstrating that we knew what we were doing rather than a leap into the unknown. They should get the same result every time.

The more doctors know what they are doing, the more they approach Fisher’s expert, but no one runs RCTs in situations where we are likely to get the predicted result every time.

In the case of breast cancer, on the basis of advances in physiology, it was hoped that giving Herceptin to Her 2+ receptor breast cancers might produce better responses than cisplatin, a more indiscriminate toxin, which nevertheless extends longevity compared to placebo. Trials confirm this but also reveal that even using Herceptin in Her 2+ breast cancers, we do not get the same result every time – there is a lot we don’t know.

In contrast, in trials comparing stents to other cardiac procedures, doing what seems physiologically obvious does not produce the expected results.  The issue is not whether stents work but whether we know what we are doing, which we mostly don’t. While recent stent trials demonstrate the power of RCTs to stall a therapeutic bandwagon, the view that clearing blocked arteries might not produce a good outcome had been accepted clinical wisdom in vascular leg surgery and for stents in some quarters prior to any RCT.

Pre 1962: Neyman and Confidence Intervals

Jerzy Neyman and Egon Pearson took issue with Fisher’s real-world anchor – a semi-infallible expert. They borrowed from Carl Friedrich Gauss’ use of confidence intervals to manage the error in astronomical measurements of stars.  Gauss’ ideas were picked up by Pierre-Simon Laplace and their combined input (1809-1827) to the central limit theorem, least-squares optimization and efficient linear algebra provided celebrated benefits for the physical sciences, engineering, astronomy, and geodesy.

Applied to imprecise measuring instruments and invariant entities like stars, confidence intervals have an anchor in the real world, helping us to decide if our varying measures reflect the presence of one or two stars. Taking successive measurements of a pulse in an individual is similar to determining the precise location of a star – the tighter the confidence interval bounding our measurements the more apparent we can do things reliably.

Confidence intervals could be used in a manner consistent with their use in astronomy to distinguish between a repeated set of pulse measurements before and during (but not after) administration of a drug – to one individual. The current use of confidence intervals in RCTs seems predicated on the idea that a cohort of patients in standard parallel group trials can be regarded as a single object like a galaxy. But pulses can increase in response to a drug in one individual and decrease in a second in response to the same drug. This is not measurement error.

In cases like this, claiming the true effect of the drug likely lies near some mean of the effects in a group of individuals, potentially giving us a best estimate of no effect, is wrong. A mechanism to decide whether there are one or two stars present should not turn up the answer there are none.  If the gap between Average Treatment Effects (ATE) and Heterogenous Treatment Effects (THE), despite trial designs to mitigate the problem, is too great, there is some recognition that the notion of ATE falls apart (Kravitz et al 2004).  In the case of stars, we knew enough about what we were doing to make reasonable inferences from varying measurements.  We need to know as much to make comparable inferences when giving medicines – and we rarely do.

1962: Lasagna, Hill & Primary Endpoints

The 1938 U.S. Food, Drugs and Cosmetics Act required pharmaceutical companies to establish the safety of their products. The birth defects thalidomide caused produced a political crisis in which something needed to be seen to be done.  Louis Lasagna, through Estes Kefauver, proposed that companies should also be required to demonstrate treatment effectiveness – an ineffective treatment cannot be safe. The 1962 Amendments to the 1938 Act paired the word Effectiveness with Safety throughout, with two placebo controlled RCTs later proposed by Lasagna as the means of demonstrating effectiveness.

These provisions were put in place before it was realized that demonstrating effectiveness rather than a treatment effect was not a realistic gateway to the market. In 1962, it was also assumed that RCTs offered generalizable knowledge and a positive result would invariably be replicated but this has not been borne out.

Before 1962, RCTs were not seen as offering gold standard knowledge about what drugs do. As Tony Hill put it in a 1965 lecture, RCTs have a place in the study of therapeutic efficacy, but they are only one way of doing so and any belief to the contrary is mad (Hill 1965). Hill’s lecture ties RCTs to the investigation of one effect and places the information they yield within the framework of clinical judgement.

Fisher’s significance testing, and Gauss’ confidence intervals, require a focus on one effect. In medical RCTs, a focus on a primary endpoint is key to ensuring that only chance or measurement error will get in way of the correct result.  Ipso facto, this means RCTs are not a good way to evaluate a medicine.

A horticultural expert focused on whether a fertilizer improved corn yield would likely have no more accurate a view of its effects on worms in the ground or insects in the air than a non-horticulturalist – in respect of whose views significance testing by Fisher’s definition would not be appropriate.  Similarly, Gauss’ confidence intervals applied to measurements of the location of a star are of little use when it comes to pinpointing the trajectories of satellites crossing the path of the observations.

It is often assumed that the primary endpoint in an RCT is the commonest effect of a drug. Treatment heterogeneity leading to wider confidence intervals than are ideal can be accommodated against this background as can missing other effects assumed to be rare or not appearing within the duration of the trial. But the RCT measuring process is often not trained on the commonest thing a drug does. The commonest effect of an SSRI is genital anesthesia, which appears almost universally and within 30 minutes of taking a first pill. It should not be possible to miss it, but this effect has been missed in all RCTs of these drugs for nervous conditions because of an RCT required focus on a primary endpoint.

The measuring attention given to a primary endpoint essentially creates an act of hypnosis in which common treatment effects can be missed entirely or given an incidental status. Casting RCTs as offering gold standard evidence about a drug, rather than one effect of the drug, creates an ignorance about the ignorance they generate.

RCT evidence should never trump an evident safety effect that appears after treatment.  If a person becomes suicidal after taking an antidepressant, the issue of what is happening in that case is a matter of assessing the effects of their condition, circumstances, prior exposure to similar drugs, dose changes on the medication and whether there are other evident effects of treatment consistent with a link between suicidality and treatment. Unless RCTs have been designed specifically to look at the effects of treatment on a possible emergence of an effect like suicidality (and there have been none), RCT evidence is irrelevant and it is pernicious to pitch irrelevant RCTs as science that should count for more than clinical “anecdotes” containing CDR, dose response, and other evidence.

The transformation of RCTs from a hurdle industry had to surmount to make gold into gold standard knowledge has made RCTs a gold standard way to hide a drug’s 99 other effects.

Post 1962: Confounding & Causality

Discussions of the results of epidemiological studies apparently linking drugs to treatment effects often caution that confounding by indication undercuts any easy assumption of a link.  RCTs, which are essentially epidemiological studies, rarely come with this rider.  Many clinicians likely think that randomization takes care of confounding by all unknown unknowns, including by indication, with many saying RCTs demonstrate cause and effect where other epidemiological studies produce correlations.

Consider scenarios involving the antidepressants imipramine and paroxetine.  Imipramine was discovered in 1957 and launched in 1958 without any RCT input. Among other actions, it is a serotonin reuptake inhibitor. In later RCTs, it (and other older antidepressants all discovered and marketed without RCTs) “beat” SSRI antidepressants in trials involving patients with melancholia (severe depression). Melancholic patients are 80 times more likely to commit suicide than mildly depressed patients.

By 1959, clinicians praising imipramine’s benefits also noted it could cause agitation and suicidality in some patients that cleared when the drug was stopped and reappeared when restarted. This Challenge-Dechallenge-Rechallenge (CDR) evidence, especially as it was replicated by several clinicians with different patients, offers close to Fisherian expert like certainty that imipramine causes suicide in certain individuals.

Despite being able to cause suicide, in an RCT of melancholic patients, imipramine seems likely to protect against suicide on average by reducing the risk from melancholia to a greater extent than placebo. In contrast, in the RCTs that brought SSRIs to the market, these drugs doubled the rate of suicidal acts. This was because, weaker than imipramine, SSRIs had to be tested in people with mild depression at little risk of suicide. The low placebo suicidal act rate revealed the risk from the SSRI – as it does for imipramine when put into trials of mild depression. RCTs can, in other words, mislead as regards cause and effect – potentially getting results all the way along a spectrum from “causes”, to possible risk, likely protective and “cannot cause”.

In any trial where both condition and treatment cause superficially similar problems, as when antidepressants and depression cause suicidality or bisphosphonates and osteoporosis both lead to fractures, a dependence on RCT data rather than clinical judgement risks misleading. This is likely the case for a majority of RCTs in clinical conditions, which are Treatment Trials rather than Drug Trials.

Drug Trials are done on healthy volunteers, and ordinarily do not have a primary endpoint.  In these, treatment effects stand out more clearly. SSRI Drug Trials in the 1980s demonstrated sexual effects were common, often debilitating, and might endure after treatment stopped, that agitation up to suicidality was common and that dependence commonly occurred after exposures of two weeks. The correct choice of primary endpoint in subsequent Treatment Trials could eliminate these effects. The non-confidential Drug Trial data remain unpublished.

Paroxetine was later put into Treatment Trials of patients with Major Depressive Disorder (MDD) and patients with Intermittent Brief Depressive Disorders (IBDD). IBDD patients (borderline personality disorder) are repeated self-harmers. The depressive features IBDD patients have mean that they can readily meet criteria for MDD.

In April 2006, GlaxoSmithKline (GSK) released RCT data showing a worrying increase in suicidal events in MDD patients on paroxetine (Table). The data from IBDD RCTs in the GSK release were better.  We can add 16 suicidal events to the paroxetine IBDD column and still get an apparently protective rather than problematic result for paroxetine when MDD and IBDD data are added together.

Table: Suicidal Events in MDD & IBDD Trials

This effect has been noted as a hazard of meta-analyses but it must apply to some extent in every trial that recruits patients who have a superficially similar but in fact heterogenous conditions such as depression, pain, breast cancer, Parkinson’s disease, diabetes or almost any medical disorder. Every time there is a mixture of more than one patient group in a trial, randomization will ensure some patients hide some treatment effects – good and bad. Trials of standard treatments for back pain, for instance, mask the beneficial treatment effects of an antibiotic on back pains linked to infections (up to 10% of back pains).

This is Heterogeneity of Treatment Conditions (HTC) rather than HTE. In epidemiological studies confounding by indication is commonly taken to mean that we should not for example interpret results apparently associating a treatment like an antidepressant with suicidality given the possibility that depression can cause suicidality, but in fact this effect likely more commonly hides the adverse effects of treatment. It is even possible to design Treatment Trials to hide adverse effects – as above.

The assumption is that in Treatment Trials placebos simply control for natural variation.  But placebos can have potent treatment effects, making them another treatment like an antibiotic in a backpain RCT. We do not know enough about placebo responses to know the extent to which, in the context of randomization, they might confound the data.

Every medicine that gets on the market, by definition, beats placebo (often inconsistently).  As a result, it has become unethical to use placebos in clinical practice, when for those for whom it works a placebo may be preferable to therapeutic poisoning.

A quantitative approach to data generated by algorithm rather than an approach based on judgement also increases the risk that minor events in a placebo arm will be offset against significant events in an “active” treatment arm creating an opportunity to claim that nothing specific has happened, when it has.

Finally, the suicidality, sexual dysfunction, agitation, and insomnia antidepressants cause in clinical trials are commonly folded into a primary endpoint, the Hamilton Depression Rating Scale (HDRS), which includes questions on suicidality, sexuality, sleep and agitation. These changes render confidence intervals around scores on these items meaningless, compromise the use of the scale more generally, and risk hiding a benefit.

Post 1980: From Therapeutic Poisoning to Sacraments

In 1947, treatment with medicines was viewed as therapeutic poisoning.  As of 1951, FDA made most new medicines prescription-only on the basis that they are unavoidably risky.  But from the mid-1990s, regulators have licensed drugs on the basis of a favorable risk-benefit profile. This implies a balance in which benefits and risks can be weighed, but there is no balance. One statistically significant effect is taken to count for more than all other effects, even serious effects that occur more frequently and can include death, but which by design are not significant, transforming poisons into sacraments (hyper-real agents from which only good can come).

In 1959, clinicians could readily distinguish between treatment emergent suicidality and suicidality caused by melancholia.  In 1961, Frank Ayd, the discoverer of amitriptyline a year before could distinguish the sexual dysfunction it causes from the sexual dysfunction melancholia causes. Through to 1991, clinical knowledge of the range of effects drugs can cause derived primarily from clinical experience, embodied in case reports and published in clinical journals. A steady rise of mechanical evaluations, however, allied to a sequestration of trial data, has relegated clinical evaluations that drug X causes effect Y, even when buttressed by evidence of CDR, to the status of anecdotes. From 1991, leading journals stopped taking anecdotes about “side” effects that almost by definition must be rare compared to the treatment effect.

As a result, where in the 1960s the harms of treatments took at most a few years to establish after a drug came on the market, by 1990 it could take decades for significant harms such as with impulse control disorders on dopamine agonists, persistent sexual dysfunction on isotretinoin, antibiotics, finasteride and other drugs, mental disorders on fluoroquinolones or leukotriene antagonists, or dependence on psychotropic drugs, to be accepted.

This growing delay underpins a perception that pharmacovigilance is in crisis. Proposed solutions mention the need for systems to detect rare treatment effects not found in RCTs. There is a turn to a mining of electronic medical records or other observational approaches. New signal detection methods and investigative approaches are always welcome, but these are not the answer to the problems we face, which lie not in a failure to detect rare effects but in a systematic failure to acknowledge common effects.

Through to 1991, clinical knowledge also derived from Drug Trials on healthy volunteers and this is almost self-evidently a better approach than relying on signal detection methods.

The ability of RCTs to focus on one effect suits Regulatory Trials but this focus does not suit an evaluation of treatments, the intention of which is to poison or mutilate in the hope of producing an overall benefit. Studies run on a primary endpoint chosen for commercial reasons cannot be expected to produce the kind of information that might inform therapeutic poisoning. Nor can we know a priori if data-handling methods developed for fertilizers and stars can encompass the complexity of therapeutic poisoning.

The question of whether the suicidality the patient in front of me is experiencing comes from their illness or their treatment is not a matter of deciding if there are 1 or 2 stars. In this case, we already know there are two stars and a lot about them, and one patient may have both kinds.  Instruments (checklists) specifically designed with the characteristics of each star in mind may facilitate the distinction between the two, but in practice it’s a case of pattern recognition and a judgement call as to whether increasing or reducing the dose of treatment is more appropriate. The high stakes may make the option of falling back on an operational approach appealing – but it is not good science or good medicine.

If registered on adverse event forms, treatment emergent suicidality or sexual dysfunction should almost de facto be causally linked to treatment. Without clinical context, and the opportunity to dechallenge and rechallenge, faced with a requirement to tick boxes as to the likelihood of a link, the ethos of RCTs, which replaces clinical judgement with decisions based on analytic processes rather than an interrogation of people, steers investigators toward designating the effect as possibly unrelated.

Facing claims in 1983 that spontaneous reporting of adverse events was unsophisticated and not scientifically rigorous, and the only proper method of establishing effects was through trials, Lasagna, once a leading advocate for RCTs, responded that “this was only the case in the dictionary sense of sophisticated meaning “adulterated” and spontaneous reporting was in fact more worldly-wise, knowing, subtle and intellectually appealing than [trials]” (Lasagna 1983).

Implications: Objectivity

A few years later, Lasagna offered the view that:

“In contrast to my role in the 1950s which was trying to convince people to do controlled trials, now I find myself telling people that it’s not the only way to truth… Evidence Based Medicine has become synonymous with RCTs even though such trials invariably fail to tell the physician what he or she wants to know which is which drug is best for Mr Jones or Ms Smith – not what happens to a non-existent average person”.

Concerns about what is often termed the population effects of RCTs, or Average Treatment Effects (ATE) and the mismatch between these and the responses of individual patients has been framed in terms of HTE and recognized in EBM as needing an incorporation of RCT evidence into the judgement of clinicians and the values and preferences of patients.  Designating RCTs as offering gold standard evidence, however, effectively side-lines the judgements of clinicians and patients.

The view that RCTs give population or average treatment effects assumes a valid population with individual outliers. In the case of antidepressants, however, there is no knowing how any individual will respond. Fisher expected us to get the same result in every individual case, and within limits confidence intervals offer the same guarantee. Neither Fisher nor Gauss would recognize a problem in translating from a population to an individual level. Diagnostic imprecision, and individual heterogeneity, mean we do not have Gaussian populations. To adapt Shakespeare, the fault lies in our stars not in ourselves.

In recent years, there has been sophisticated consideration of the statistical techniques employed in epidemiological studies, including RCTs (Greenland et al 2017), and of the merits of RCTs applied to complex situations in the social sciences (Deaton and Cartwright 2019).  Both considerations have stressed the role of judgement in deciding what populations and experimental design are appropriate, and how results should be interpreted. Both view RCTs, and related designs using statistics, as assay systems yielding results specific to the system, rather than experiments that generate the ‘knowledge from nowhere’ that means we don’t have to worry whether the laws of gravity will apply to the next patient.

These positions are compatible with the argument here, which is that rather than assay systems that might in the right circumstances offer applicable information, RCTs have become algorithmic or operational procedures.  DSM criteria in mental health, and the metrics for blood pressure, peak flow rates and bone densities are similarly operational. The creators of the DSM criteria claimed that of course just ostensibly meeting criteria for an illness doesn’t mean the person has the illness, clinical judgements are needed to establish what is really going on, just as they are in the case blood pressure, peak flow or bone density readings. In practice, however, operational exercises like RCTs, DSM criteria, and many medical metrics nudge us toward a suspension of judgement and put a third party, like the pharmaceutical industry, in a strong position to contest any introduction of judgement by a doctor or patient on the basis that the figures are supposedly more objective than any clinician or patient judgement can be.

Even facing strong epidemiological evidence that a drug causes birth defects or strokes, many clinicians will dismiss these as observational data and be unwilling to adjust practice until an RCT has demonstrated the effect.  Industry openly play on clinical difficulties in identifying RCTs as producing observational data.

Science traditionally generates data and challenges us to interpret them. New techniques (like a new drug) can throw up new observations (data) that challenge prior judgements. The application of statistical techniques to data yields outputs, not observations. While these techniques and their outputs can be useful, the mission of science has not been to replace judgement by technical outputs.

Individual judgement of course is suspect. This argument does not advocate replacing collective evaluation by a reliance on individuals or doctors; the argument is for collective evaluation rather than its replacement by algorithmic processes. Collective evaluation has a clear footing in the real world, as the Mayo Clinic streptomycin trial demonstrated.  The idea that clinical RCTs as happen now have as clear a footing is assumed not established.

Arguments favoring RCTs to point to a small series of treatments, such as internal mammary ligation, that RCTs demonstrated did not work, with the implication that clinical judgement can get things wrong. The internal mammary ligation trial only happened because the dominant clinical judgement was that this treatment didn’t work – an article in the Reader’s Digest notwithstanding. And randomization didn’t work in this 17-patient trial.

These arguments fail to note that most of the current treatment classes we have were introduced in the 1950s without RCTs. That the treatments introduced then from anti-hypertensives and hypoglycemics to psychotropic drugs are more effective than treatments introduced since.  That RCTs facilitate the introduction of treatments with lesser effects.

Our most important failure is our complicity in a sequestration of trial data, fooled perhaps in some instances into thinking that analytic outputs are data. Data means the people entering into a study, who lie behind any table of figures or the outputs of any analytic process applied to those figures. At present, with the exception of a very few RCTs, case reports with names attached are the only form of controlled clinical investigation that offer the possibility of interrogating the data and an opportunity to ground any conclusions in the real world.

Drug interventions (therapeutic poisoning) invariably harm; the hope is that some good can also be brought from their use.  Evaluations of a medicine by RCT harm (generate ignorance), but if used judiciously some good can be brought out of the ignorance they necessarily generate. It is less likely that good will be brought out of ignorance, if we rely solely on a data handling formula. Analytic methods can describe data but whether good comes from their use requires the kind of judgement calls that statistical approaches ordinarily make a virtue of side-lining. A recent study looking at 29 ways to analyze a dataset, generated from referees giving red cards to dark and light-skinned soccer players, demonstrated that different techniques can lead to a wide variation in results with none able to guarantee what is happening in the real world (Silberzahn et al. 2018).

Clinical practice is essentially a judicial rather than an algorithmic exercise.  The view offered here is that our best evidence as to what happens or is likely to happen on treatment lies in the ability to examine and cross-examine the persons (interrogate the data) given that treatment. What holds true at the individual level must be true at the population level also.  The evaluation of a treatment cannot be algorithmic.

An endorsement of clinical judgement does not suit health service managers or the pharmaceutical industry, for whom the supposed generalizability of RCT knowledge and confidence intervals that can be offered for such knowledge are legally appealing.

Implications: The Place for Randomized Controlled Trials

Randomization, placebo controls, confidence intervals and primary endpoints all have a place in the evaluation of treatments.  Confidence intervals are clearly appropriate in instances where measurement error is likely to play a part.  Randomization is an extra control on clinical bias. There is a place for it, unhooked from primary endpoints and statistical significance, as happens in large pragmatic trials – but here the word pragmatic concedes our limited understanding of what we are doing.

An increasing use of RCTs in social science, economic and political settings makes it clear that complexity is not a necessary bar to their use in trials with an appropriate focus on a primary endpoint. In medicine, the multidimensional nature of therapeutic poisoning adds an extra layer of complexity and makes a focus on a primary endpoint problematic, other than when a claimed benefit is contested.

RCTs may be better suited to evaluate time-limited surgical interventions as opposed to chronic therapeutic poisoning, as well as in studies to evaluate programs, and treatment studies that have an endpoint like all-cause mortality, but even here we risk being misled by findings of no change in mortality into missing a switch from cardiac events to cancers when many patients might prefer to die by heart attack (Mangin et al 2007).

RCTs also have a merit as a gateway to the market; randomization means that trials require less patients and can be run quickly. A positive result in commercial trials may indicate a compound has an “effect”. Trials aimed at establishing effectiveness, in contrast, require hard outcomes and time. This is not a realistic gateway to the market. Demonstration of an effect, as with SSRIs for depression, means it is not correct to say this drug does nothing and on this basis entry to the market could be permitted, although strictly speaking this is inconsistent with current statutes.

After 1962, RCTs became the standard through which industry would make gold, As they proliferated, the mantra that they provide gold standard medical evidence took hold. The ignorance of ignorance in claims that the only valid information on medicines comes from RCTs compounds a series of other factors that make RCTs a gold standard way to hide adverse events and encourage over-use of treatments.

The launch of a drug licensed on the basis of a treatment effect should be the point when more comprehensive clinical evaluations start, aimed at generating consensus as to the place of the drug in practice. As a general tool to evaluate the effects of a drug, Regulatory Trials should take second place to both the observations of a group of experienced clinicians, unconstrained by checklists and an investigation tailored to one effect, as well as to the values of patients who increasingly need to reduce their medication burden to achieve optimal benefits.

In addition, seasoned clinicians, allied to increasingly health-literate patients, are better placed than RCTs to determine cause in the case of the 99 other effects every drug has, especially effects such as sexual or suicidal effects of antidepressants, which need to be distinguished from superficially similar condition effects.

The fact that pharmaceutical companies run “RCTs” for regulatory and marketing purposes may have generated a belief that any problems with RCTs stem from a link to commerce.

The difficulty in recognizing adverse effects has for instance been compounded by company sequestration of trial data and ghostwriting of the clinical literature that hypes the benefits and hides the harms of treatments, compounded by a regulatory willingness to avoid deterring patients from treatment benefits by placing warnings on drugs.

Clinical practice is also compromised by licensing indications and by guidelines.  There are no drugs licensed to treat adverse effects. When a person becomes suicidal on an SSRI, there is no treatment licensed to treat this toxicity. Clinicians wanting to help feel compelled to diagnose depression rather than toxicity but a depression diagnosis inevitably leads to a further treatment with an antidepressant rather than something more appropriate like a benzodiazepine, a beta-blocker, or red wine.

The incorporation of RCTs into the regulatory apparatus has introduced surrogate markers, which mean that in real life treatments may not show effectiveness consistent with RCT demonstrations of a treatment effect. Trials showing antidepressants work, for instance, have more deaths and both suicidal and homicidal events in their treatment arms compared to placebo.

Commercial trials have given rise to the idea of an abstract Risk-Benefit ratio which along with treatment effect sizes, the Number Needed to Treat (NNT) or to Harm (NNH) are based on the outputs from analytic processes rather than in clinical reality.

Possible answers to these problems lie with medical journals who should insist on the publication of data from Drug and Treatment Trials. Our hierarchies of evidence should come clean on whether they regard a ghostwritten article without access to trial data as better than or inferior to a Case Report that embodies dose responsiveness and CDR elements. And those deploying an analytic process should clarify how the resulting outputs might translate into the real world, rather than assuming they do.

It is not unreasonable to want to discard the industry bathwater but save the RCT baby. But doing so requires an explicit recognition that industry activities avail of an epistemological gap between the conduct of RCT assays and a consideration of the implication of their results rather than constitute the gap.

Coda

Evaluating treatment effects properly is hugely important. When drugs work, they can like parachutes save lives. Given the importance of the task, the notion of a hierarchy of evidence topped by mechanisms that do the deciding for us has a potent allure.

Relegating judgement to the bottom of the evidence hierarchy in medicine brings out our discomfort with judgement. Succumbing to an operational solution, however, is at least as dangerous as depending on judgement.

RCTs have led many to view drug treatments as comparable in effectiveness to parachutes.  As a result, by the age of 50, close to 50% of us are now on three or more drugs and by the age of 65 on 5 or more drugs. For the past five years, our life expectancies have been falling and admissions to hospital for treatment-induced morbidity rising, an outcome that contrasts with the added safety of having parachutes and other gadgets in planes (Healy 2020). Adding parachutes and gadgets that are effective (rather than just have an effect) enhances aviation safety, although recent Boeing crashes point to the perils of too great a reliance on automatic decision tools. Combining five pluripotent drug gadgets almost certainly brings risks of interactions that airplane gadgets don’t bring and current data indicates that reducing medication burden from 10 or more drugs to 5 or less reduces hospitalization, increases life expectancy and improves quality of life (Garfinkel and Mangin 2010). But if RCTs of medicines essentially produce evidence that it is not correct to say this drug has no possible benefit, rather than that they are effective, our methods of evaluation rather than just the chemicals we prescribe may be contributing to increasing levels of mortality and morbidity.

Recent data on life expectancies and treatment linked morbidities call for an evaluation of the role of RCTs in the evaluation of drug treatments (Healy 2020).  So does data indicating antidepressants are now the second most commonly used drugs by young women in the face of 30 out of 30 trials negative on the primary outcome, which advocates of RCTs, with no links to industry, claim to be able meta-analyze and extract positive effects from data taken from ghostwritten publications, without access to trial data.

References:

Deaton A, Cartwright N (2018).  Understanding and misunderstanding randomised controlled trials.  Social Science and Medicine 2018, 210, 2-21.

Garfinkel D, Mangin D.  Feasibility study of a systematic approach for discontinuation of multiple medications in older adults.  Arch Intern Med 2010, 170 1648-54.

Greenland S, Senn SJ, Rothman K, Carlin JB, Poole C, Goodman SN, Altman DG. Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations.  Eur J Epidemiol, 2016, 31:337–350.

Healy D, The Shipwreck of the Singular; Healthcare’s Castaways. Samizdat Press, Toronto (Forthcoming 2020).

Hill A.B. Reflections on the Controlled Trial. Annals Rheum Disease 1966; 25, 107-113.

Kravitz RL, Duan N, Braslow J.  Evidence-based medicine, heterogeneity of treatment effects, and the trouble with averages.  Milbank Quarterly 2004; 82; 661-687.

Lasagna L. Discovering adverse drug reactions. JAMA 1983; 249: 2224-5

Mangin D, Sweeney K, Heath I.  Preventive health care in elderly people needs rethinking. BMJ 2007, 335 285-287.

Silberzahn S, Uhlmann EL, Martin DP, et al. Many Analysts, One Data Set: Making Transparent How Variations in Analytic Choices Affect Results. Advances in Methods and Practices in Psychological Science. 2018; 1:337-56.

Explanatory Note

An earlier version of this article was sent out to over 20 senior figures with expertise on areas to do with RCTs and Evidence, chosen mostly by Mark Wilson.  Their responses are reproduced in Fawlty Stars. 

Publication of the target piece and responses happens in some journals but throwing the argument open to anyone who wishes to contribute is a unique venture in peer review.

The argument lies at the heart of Chapter 6 of Shipwreck of the Singular which will be published in January 2021. The extra details in Shipwreck may help with getting to grips with what is going on.  

Some weeks ago, following the publication of an article by Andrea Cipriani in their journal, Jim Gottstein and I wrote to the Frontiers in Psychiatry Bigwigs asking for the Cipriani article to be removed – Meta-Analytic SuperSpreader.

They have since replied Vicariously through Elena Vicario.

From: Editorial Office <editorial.office@frontiersin.org>
Date: Tue, 3 Nov 2020 at 04:34
Subject: Re: Letter re Retraction
To: David Healy <david.healy54@googlemail.com>

Dear Dr. Healy,

I am writing with an update on our investigation into the concerns raised about article “Antidepressants in Children and Adolescents: Meta-Review of Efficacy, Tolerability and Suicidality in Acute Treatment”.

Following the assessment of the journal Chief Editor, the methodology and systematic protocol used in the article is in line with PRISMA guidelines and the articles included in the systematic review are clearly outlined and available for consultation by the readers. In addition, we note that the conclusions drawn by the authors from the meta-analysis are cautious and broad.

We recognize the importance of the subject matter and thank you again for bringing this to our attention.

Best regards,

Elena Vicario, PhD

Senior Manager, Research Integrity: Elena Vicario, PhD
Frontiers | London Office
12 Moorgate
EC2R 6DA
London

Most people reviewing the issue of the journal, likely Elena too, figure matters are pretty balanced overall with all other articles skeptical about the use of antidepressants for children and teenagers and a hard hitting editorial from Michael Hengartner to make matters absolutely clear.

This not however how the system works.  Decades back, at academic meetings, in what seem like more innocent times, pharmaceutical companies were good at having someone talk about CBT, or another therapy, for depression on a program mostly dedicated to selling some antidepressant.

It looked like there was a degree of balance you might expect from a scientific meeting but the idea of having a therapist in practice came closer to the vaudeville model of having a comedian come on stage in between striptease acts.  The customers needed a chance to draw breath and relax and perhaps an excuse to tell people where they had been – they weren’t there for the comedian.

Its the same in this case, even though there are eight or nine comedians and only one stripper. The internet and open access makes it possible for industry to distribute the Cipriani article far and wide and for free and even to be able to point to its publication in a virtuous setting – almost nothing but comedians – while doing so.  The eight or nine other pieces will go nowhere fast.

There is probably an appropriate riff on the old line about lies getting half way around the world before the truth gets its boots on.

There is another unsettling angle.  The editors are essentially blaming Michael H.  They’ve deferred to his judgment call as to whether there is an issue here.

It’s what Jose-Mario Bergoglio, Donald Trump, Joe Biden, Emanuel Macron and any other ‘leader’ you care to mention would do.  When asked about ghost-writing combined with a lack of access to trial data, something that is not acceptable in any conceivable scientific universe,  the fall-back line is ‘we’re not scientists, we will check with EMA or FDA or MHRA or whoever’ and ‘we’re sure you will understand we have no option but to do as they advise us’.  See The Perfect Killing Machine.

To be absolutely clear, my problem if I were one of the comedians in this show is not that there is a paper offering another point of view, its that there is a paper masquerading as science, which is just not science.  A cuckoo in the scientific nest.

Its a bit like if in the old Vaudevillian days a Drag Queen had done a striptease.  The audience would likely have seen this as a comic act.  These days apparently Drag Queens do striptease’s in front of children – for equality’s sake.  And, besides what their advisers say, for equality’s sake Jose and Joe and Emanuel likely feel they cannot prevent a Cipriani article (of this type) being passed off as science.

In this case Michael Hengartner has no option but to take the blame. Once the requirements of the algorithm are satisfied, two positive reviews, there is no option but to publish.

There was a notable fuss last week about a Danish Mask Trial, which failed to show a convincing benefit for Masks in the current pandemic – perhaps because it didn’t ask the right questions.  It was nevertheless a trial – the only trial of Masks in a time of Covid.  When several doctors linked to it on Facebook pages, they found their link was blocked on the basis that they were posting misinformation.

A lot of mostly male docs were offended – how dare Facebook do this.  But there is no-one doing this to anyone.  The algorithm is doing it.

It hasn’t fully drawn all power to itself – there is still a missing ring, a precious, that needs retrieving.  Soon now…

Meta-Analytic SuperSpreader outlined a disturbing sequence of events leading up to the publication of an article by Andrea Cipriani and colleagues effectively promoting the benefits of antidepressants.  Andrea and colleagues are part of an Oxford University Precision Psychiatry ‘lab’.  The idea that this article in some way helps move us toward a precision psychiatry is mind-boggling.

Michael Hengartner ended up as the only editor standing for the issue of Frontiers Psychiatry in which the Cipriani article appeared.  He wrote a covering editorial for the collection of articles that mentions the Cipriani article and that most readers will think tries hard to get over a ‘decent’ message – there will be more on being decent in a follow-up post.

Antidepressant Prescriptions in Children and Adolescents

The use of antidepressants in children and adolescents has a troubled history, for almost all principles of good evidence-based medicine were violated or compromised. It is a history characterized by systematically biased research, financial conflicts of interest, and professional recklessness (1–3). In 2004, the Lancet Editors (4), in an article titled “Depressing research” bluntly stated that “The story of research into selective serotonin reuptake inhibitor (SSRI) use in childhood depression is one of confusion, manipulation, and institutional failure” (p. 1335). It is now well-established that most pediatric antidepressant trials were industry-sponsored and had serious methodological limitations; many trials remained unpublished due to unfavorable results, and those published were mostly ghost-written, selectively reported efficacy outcomes and misrepresented the true rate of treatment-emergent suicidal events (5–9). Drug regulators issued a suicidality warning for pediatric antidepressant use in 2003 (MHRA) and 2004 (FDA) and advised to use fluoxetine only. By consequence, some authors argued that SSRI should be reserved as a second-line option for youth with severe and resistant forms of depression (10).

However, in most countries antidepressant use has considerably increased in children and adolescent over the last 10–15 years (11–13), despite suicidality warnings, the serious limitations of the evidence-base (14), and ongoing controversies surrounding risks and benefits (15) as well as the placebo response (16). The aim of this special topic was thus to provide a collection of articles broadly focused on two main issues; first, on the current scientific evidence for the efficacy and safety of antidepressants, with a special emphasis on suicidality and related regulatory warnings, and, second, on recent trends in prescription rates and patterns of utilization, including antidepressant overuse, and the increasingly medicalized approach to mental health.

Safer and Zito reviewed the efficacy of new-generation antidepressants for pediatric depression. They found no meaningful benefits in children and only marginal benefits in adolescents based on placebo-controlled short-term trials. Moreover, they considered the evidence for maintenance treatment based on discontinuation (placebo-substitution) trials problematic and inconclusive due to high dropout rates, potential withdrawal syndromes that mimic relapse, and relapse rates not dissimilar from the natural course of the disorder.

Boaden et al. conducted a meta-review on the efficacy, tolerability, and suicidality-risk of antidepressants for the treatment of various pediatric disorders. The meta-review found that just a few antidepressants were effective and well-tolerated. For instance, only fluoxetine was more effective than placebo in major depression, and only fluvoxamine and paroxetine were effective in anxiety disorders. Venlafaxine (in major depression) and paroxetine (in anxiety disorders) were associated with significantly increased risk of suicidality. However, of the nine meta-analyses included, only one met criteria of high quality; five were rated moderate quality, one was of low quality and two of critically low quality. The authors further state that the quality of the available evidence is inadequate due to short trial duration, selective reporting and publication bias, and they emphasize the paucity of data on suicidal ideation and behavior in antidepressant trials.

The issue of increased risk of suicidality with antidepressants was specifically addressed in two articles. In the first, Spielmans et al. review the scientific evidence and conclude that the FDA black-box suicidality warning was justified and firmly rooted in solid data from placebo-controlled antidepressant trials. They further detail that prominent claims suggesting that the FDA warning has led to decreasing prescription rates and thus increasing suicide rates were based on methodologically weak and potentially misleading ecological studies.

In the second, Whitely et al. describe how prominent psychiatrists and influential mental health organizations challenged the black-box suicidality warning for adolescents and young adults. The authors argue that various ecological studies were cited misleadingly as evidence that increased antidepressant use reduces youth suicide risk. Contrary to these claims, they further show that, in Australia, both antidepressant use and suicide rates increased substantially from 2008 to 2018.

Another serious safety issue was addressed by Kapra et al.. In their mini review they discuss the evidence for and against a potential effect of antidepressant use during pregnancy on autism spectrum disorders in the offspring. The authors found evidence for an association between prenatal SSRI exposure and an increased risk of autism spectrum disorders based on several observational studies, but caution that causality has not been demonstrated yet due to confounding by indication. The authors conclude that there is a need for more research on this serious safety issue, as accumulating data from animal studies suggest that SSRI exposure may alter normal brain development.

Trends of increasing antidepressant use in young people were addressed in two articles. In the first, Zito et al. analyzed administrative claims of Medicaid-insured youth aged <20 years from 1987 to 2014. During this 28-year period, antidepressant use grew 14-fold. They further show that in 2014, antidepressants were prescribed six times more often for youth in foster care than for their income-eligible Medicaid-counterparts. Off-label prescribing was also very common: a quarter of antidepressant-medicated youth were diagnosed with a behavioral disorder.

In the second, Cosgrove et al. state that antidepressant use in children and adolescents rose substantially over the last 15 years in part due to commercially driven off-label prescriptions, despite ongoing controversy over their effectiveness and safety. From the perspective of institutional corruption, they discuss two drivers of overuse resulting from an increasingly medicalised approach to mental health. The first is the empirically unsupported demand for depression screenings in youth and the second the emphasis on scaling up diagnosis and treatment of mental disorders as part of a renewed Global Mental Health Movement.

Last but not least, Locher et al. make an interesting case for open-label placebos in the treatment of chronic pain conditions in children and adolescents as an alternative to long-term antidepressant use. The authors acknowledge that this approach still lacks empirical evidence, but also stress that open-label placebos constitute a promising avenue for future research as they may help to mitigate the serious adverse effects of antidepressants.

In conclusion, the articles in this special topic demonstrate that pediatric antidepressant use is still controversial. Although antidepressant use in children and adolescents has increased substantially over the last 10–15 years, convincing evidence that the benefits outweigh the risks is lacking and treatment-emergent suicidality remains a major concern. Overuse and off-label prescribing are pressing issues, and there certainly is a need for safer and more effective treatments, both pharmacological and psychological (17). It is hoped that this article collection will spur innovative research and critical discussion.

Author Contributions

The author confirms being the sole contributor of this work and has approved it for publication.

Conflict of Interest

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

I thank David Healy and Irving Kirsch for their help in recruiting authors for this special topic.

References

1. Jureidini JN, McHenry LB. Key opinion leaders and paediatric antidepressant overprescribing. Psychother Psychosom. (2009) 78:197–201. doi: 10.1159/000214440

PubMed Abstract | CrossRef Full Text | Google Scholar

2. Leo J. The SSRI trials in children: disturbing implications for academic medicine. Ethical Hum Psychol Psychiatry. (2006) 8:29–41. doi: 10.1891/ehpp.8.1.29

PubMed Abstract | CrossRef Full Text | Google Scholar

3. Raz A. Perspectives on the efficacy of antidepressants for child and adolescent depression. PLoS Med. (2006) 3:e9. doi: 10.1371/journal.pmed.0030009

PubMed Abstract | CrossRef Full Text | Google Scholar

4. Lancet Editors. Depressing research. Lancet.(2004) 363:1335. doi: 10.1016/S0140-6736(04)16080-7

PubMed Abstract | CrossRef Full Text

5. Jureidini JN, Doecke CJ, Mansfield PR, Haby MM, Menkes DB, Tonkin AL. Efficacy and safety of antidepressants for children and adolescents. BMJ.(2004) 328:879–83. doi: 10.1136/bmj.328.7444.879

PubMed Abstract | CrossRef Full Text | Google Scholar

6. Jureidini JN, McHenry LB, Mansfield PR. Clinical trials and drug promotion: selective reporting of study 329. Int J Risk Saf Med. (2008) 20:73–81. doi: 10.3233/JRS-2008-0426

CrossRef Full Text | Google Scholar

7. Jureidini JN, Amsterdam JD, McHenry LB. The citalopram CIT-MD-18 pediatric depression trial: deconstruction of medical ghostwriting, data mischaracterisation and academic malfeasance. Int J Risk Saf Med. (2016) 28:33–43. doi: 10.3233/JRS-160671

PubMed Abstract | CrossRef Full Text | Google Scholar

8. Le Noury J, Nardo JM, Healy D, Jureidini J, Raven M, Tufanaru C, et al. Restoring Study 329: efficacy and harms of paroxetine and imipramine in treatment of major depression in adolescence. BMJ. (2015) 351:h4320. doi: 10.1136/bmj.h4320

PubMed Abstract | CrossRef Full Text | Google Scholar

9. Whittington CJ, Kendall T, Fonagy P, Cottrell D, Cotgrove A, Boddington E. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet. (2004) 363:1341–5. doi: 10.1016/S0140-6736(04)16043-1

PubMed Abstract | CrossRef Full Text | Google Scholar

10. Cohen D. Should the use of selective serotonin reuptake inhibitors in child and adolescent depression be banned? Psychother Psychosom.(2007) 76:5–14. doi: 10.1159/000096360

PubMed Abstract | CrossRef Full Text | Google Scholar

11. Bachmann CJ, Aagaard L, Burcu M, Glaeske G, Kalverdijk LJ, Petersen I, et al. Trends and patterns of antidepressant use in children and adolescents from five western countries, 2005-2012. Eur Neuropsychopharmacol. (2016) 26:411–9. doi: 10.1016/j.euroneuro.2016.02.001

PubMed Abstract | CrossRef Full Text | Google Scholar

12. John A, Marchant AL, Fone DL, McGregor JI, Dennis MS, Tan JO, et al. Recent trends in primary-care antidepressant prescribing to children and young people: an e-cohort study. Psychol Med. (2016) 46:3315–27. doi: 10.1017/S0033291716002099

PubMed Abstract | CrossRef Full Text | Google Scholar

13. Schroder C, Dorks M, Kollhorst B, Blenk T, Dittmann RW, Garbe E, et al. Outpatient antidepressant drug use in children and adolescents in Germany between 2004 and 2011. Pharmacoepidemiol Drug Saf. (2017) 26:170–9. doi: 10.1002/pds.4138

PubMed Abstract | CrossRef Full Text | Google Scholar

14. Healy D, Le Noury J, Jureidini J. Paediatric antidepressants: benefits and risks. Int J Risk Saf Med.(2019) 30:1–7. doi: 10.3233/JRS-180746

PubMed Abstract | CrossRef Full Text | Google Scholar

15. Cipriani A, Zhou X, Del Giovane C, Hetrick SE, Qin B, Whittington C, et al. Comparative efficacy and tolerability of antidepressants for major depressive disorder in children and adolescents: a network meta-analysis. Lancet.(2016) 388:881–90. doi: 10.1016/S0140-6736(16)30385-3

PubMed Abstract | CrossRef Full Text | Google Scholar

16. Cohen D, Deniau E, Maturana A, Tanguy ML, Bodeau N, Labelle R, et al. Are child and adolescent responses to placebo higher in major depression than in anxiety disorders? A systematic review of placebo-controlled trials. PLoS ONE. (2008) 3:e2632. doi: 10.1371/journal.pone.0002632

PubMed Abstract | CrossRef Full Text | Google Scholar

17. Zhou X, Teng T, Zhang Y, Del Giovane C, Furukawa TA, Weisz JR, et al. Comparative efficacy and acceptability of antidepressants, psychotherapies, and their combination for acute treatment of children and adolescents with depressive disorder: a systematic review and network meta-analysis. Lancet Psychiatry. (2020) 7:581–601. doi: 10.1016/S2215-0366(20)30137-1

PubMed Abstract | CrossRef Full Text | Google Scholar

Keywords: antidepressant, pediatric, prescribing, children, adolescents, efficacy, safety, suicidality

Citation: Hengartner MP (2020) Editorial: Antidepressant Prescriptions in Children and Adolescents. Front. Psychiatry 11:600283. doi: 10.3389/fpsyt.2020.600283

Received: 29 August 2020; Accepted: 30 September 2020;
Published: 30 October 2020.

In a recent post Infection Super-Spreaders, I outlined the disturbing publication of an article by Andrea Cipriani and colleagues that gives a green light to prescribing antidepressants to children.  Prescribing antidepressants to children could even be part of Precision Psychiatry – a new Oxford University brand to complement Evidence Based Medicine.

Concerned about developments, Jim Gottstein and I wrote to the Bosses of Frontiers – normally we would have written to an editor about something like this but editing is outsourced in the Frontiers business model.

Kamila and Miriam

Kamila Markram C.E.O.
Mirjam Eckert Publishing Director
Frontiers Media SA
Avenue du Tribunal Fédéral 34
1005 Lausanne
Switzerland

editorial.office@frontiersin.org

September 24^th 2020

Dear Drs Markram and Eckert

On September 2^nd you published an article by Andrea Cipriani and colleagues:

K Boaden, A Tomlinson, S Cortese, A Cipriani. Antidepressants in Children and Adolescents: Meta-Review of Efficacy, Tolerability and Suicidality in Acute Treatment.  Frontiers in Psychiatry, doi: 10.3389/fpsyt.2020.00717

For the reputation of Frontiers journals in general, a matter that will be of concern for you, for the safety of anyone taking antidepressants, children in particular, as well as the integrity of the legal system, we think this article should be retracted.

Andrea Cipriani is known personally to one of us and is a very charming man concerned we believe to do his job well. The problem we are raising here is somewhat novel in that we are not criticising the technical methods used in his study – although it can be argued these methods cannot support the conclusions he draws. But there is a more important point.

The general wisdom is that it is a bad idea to introduce a Holocaust comparison into arguments, but we are drawing here on the work of the Jewish sociologist of the Holocaust, Zygmunt Baumann, who was also a leading analyst of technique in general.  Applying his insights on technical matters to this instance, we might get the following:

A Frontiers journal could conceivably publish an article on the technical specifications for a sealed rail carriage that allowing for the elimination of defecation and micturition could be used to transport animals several hundred miles ensuring that most if not all of them got there alive – without paying any heed to the fact that the animals to be transported are human.

Dr Cipriani’s article was accepted because some of your reviewers thought the technical specifications were reasonable but the material to which his techniques are applied are a set of ghost-written articles reporting the results of clinical trials to whose data there is no access. Without any consideration of the moral issue of the likely effects on people of publishing his results, the lack of access to data breaches the canons of empirical method,

The consequences of a breach of this kind, which is relatively widespread across medicine, become very clear in the case of trials done with antidepressants in children, where the entire literature has been essentially ghost or company written and where there is no access to the raw data from the relevant trials.  In some cases, the data has been destroyed. In one case the mismatch between published articles claiming the trials showed the drugs worked well and are free of problems led New York State to take legal action for fraud against GlaxoSmithKline over paroxetine.  In two other cases the US Department of Justice took action against companies for illegal off-label marketing where scientific evidence for their support was lacking, resulting in criminal and civil fines of $3 Billion against GlaxoSmithKline for paroxetine (at the time the largest corporate settlement with the US Dept. of Justice), and  $313 million against Forest Laboratories for citalopram in 2012.

It would be a mistake to think that these actions mean the system has managed to spot the bad apples in the barrel – the articles that led to legal actions were not outliers. These trials and articles in fact were likely more ethical than many of the trials done since on which Dr Cipriani and colleagues have depended.

The upshot in the case of antidepressants used for children who are diagnosed with depression is that in fact as one of us has published, there have been 30 trials done, all negative – the greatest concentration of negative trials for any indication ever in human history –  yet because of ghost-written publications and articles like Dr Cipriani’s which works from those ghost-written publications, antidepressants may now be the second most commonly taken drugs by teenage girls.  The science should make this impossible – it is the ghost-written reporting on this science that has enabled it. Dr Cipriani’s attention was drawn to this, but he chose to ignore the point making him complicit in turning the science inside out, and in so doing he has made Frontiers complicit in this also.

The editor it appears thought there was no room in the Frontiers process for him to draw attention to the issues

There will be deaths as a direct result of this article.  In addition to antidepressant scripts escalating rapidly in teenage girls, their suicide rate in the UK has doubled over the past 5 years.

A further concern is the knock-on effect this article will have on legal processes.  One of us is a lawyer who has represented people who do not want to take medicines, most often neuroleptics, recommended to them by doctors.  These doctors, operating on the basis of a ghost-written literature in which the supposed results of clinical trials are reported which hype the benefit of treatments and hide their harms, have a power they exercise routinely to deprive people of their liberty and force them to take medicines.  It turns out that when the real data comes to light, the lack of benefit and the harms such people complain about correspond with what the actual data shows – and the ghost-written literature denies.  This situation almost certainly contributes to the widely reported 20-25 year reduction in life expectancy for people diagnosed with serious mental illness.

At the moment, Pontius Pilate-like, the courts and inquests have washed their hands of this issue, but it only takes one judge to put some journal and publication in the firing line.

For the above reasons, we would encourage Frontiers to engage with our concerns about your publication of the Cipriani article.

Yours,

David Healy MD                                              James B. (Jim) Gottstein, Esq.
Professor of Psychiatry                                   Pres. Law Project for Psychiatric Rights
McMaster University, Canada                        Alaska, USA.

As we composed this letter, the Lancet changed its retraction policy having been shamed by their publication of an article on hydroxychloroquine – see article from Guardian below. The drugs and conditions differ but the key issue the Lancet identified – access to the raw data – is the issue that concerns us and hopefully you.

The Lancet changes editorial policy after hydroxychloroquine Covid study retraction

Melissa Davey  @MelissaLDavey

Tue 22 Sep 2020 05.08 BST

One of the world’s leading medical journals, the Lancet, has reformed its editorial policies following a shocking case of apparent research misconduct involving the study of hydroxychloroquine as a treatment for Covid-19.

In May, the Lancet published a peer-reviewed study about the controversial drug hydroxychloroquine, which concluded Covid-19 patients who received the drug were dying at higher rates and experiencing more heart-related complications than other virus patients.

The large observational study analysed data purported to be from nearly 15,000 patients with Covid-19 who received the drug alone or in combination with antibiotics, comparing this data with 81,000 controls who did not receive the drug.

Questions raised over hydroxychloroquine study which caused WHO to halt trials for Covid-19

This data was recorded by hospitals around the world in a database by a US data analytics company known as “Surgisphere”, the Lancet paper said. The findings prompted the World Health Organization to halt its clinical trials of the drug, given the paper’s findings that it was linked with deaths and complications.

But days after the paper was published, Guardian Australia revealed issues with the Australian data in the study. Figures on the number of Covid-19 deaths and patients in hospital cited by the authors did not match up with official government and health department data. Senior clinicians involved in Covid-19 research told Guardian Australia they had never heard of the Surgisphere database.

Researchers from other countries identified similar issues with the data from their hospitals, and a further Guardian Australia investigation revealed doubts over whether the database used by the study authors even existed. Sapan Desai was a co-author of the paper and founder of the Surgisphere database. Following the revelations, information about Surgisphere was deleted from the internet.

It was also revealed that none of the co-authors of the paper had seen the Surgisphere data for themselves, and they said that Desai did not give them access to it even after questions about the paper were raised by Guardian Australia and the research community. The paper’s co-authors, which included a highly respected vascular surgeon, supported the retraction of the paper and distanced themselves from the data.

While the latest available data shows hydroxychloroquine does not reduce deaths among severely unwell patients in hospital with Covid-19, or reduce illness in those with moderate disease, the higher death rates among those given the drug outlined in the Surgisphere study have never been replicated.

Unreliable data: how doubt snowballed over Covid-19 drug research that swept the world

The publication of the Surgisphere study by the Lancet meant well-controlled studies to definitely determine the drug’s efficacy in preventing or treating the virus were stopped prematurely. Given the drug has been highly politicised by figures such as US president Donald Trump, who has made numerous false claims about its usefulness against Covid-19, rigorous studies into the drug remain important.

World Health Organization studies into hydroxychloroquine resumed following Guardian Australia’s Surgisphere investigation, and the Lancet retracted the Surgipshere paper and vowed to review its publication policy. Such rapid retractions are rare, and followed pressure from the international research community who questioned how the study passed quality control processes.

The new policy, published three months after the study was retracted, requires that more than one author on a paper must directly access and verify the data reported in the manuscript.

“For research articles that are the result of an academic and commercial partnership, one of the authors named as having accessed and verified data must be from the academic team,” the policy states. “In addition, all authors will be asked to sign the author statements form to confirm they had full access to the data reported in their article and accept responsibility for submitting the article for publication.”

One of the questions raised by the publication of the Surgisphere paper was how the paper passed the peer-review process.

The Lancet has made one of the biggest retractions in modern history. How could this happen?

The Lancet has updated its peer-review policy, stating: “Editors will ensure that at least one peer reviewer is knowledgable about the details of the dataset being reported and can understand and comment on its strengths and limitations in relation to the research question being addressed.”

For studies that use very large datasets, such as the Surgisphere dataset, editors will ensure that in addition to statistical peer review, a review from an expert in data science is obtained.

“Finally, we will explicitly ask reviewers if they have concerns about research integrity or publication ethics regarding the manuscript they are reviewing,” the new policy states. The new policy is effective immediately.

https://amp.theguardian.com/world/2020/sep/22/the-lancet-reforms-editorial-policy-after-hydroxychloroquine-covid-study-retraction