Wednesday, 21 April 2010


Drugs work. They do. They can be very, very helpful indeed.

But as Richard Ashcroft of The Verve penned, regarding the effect of drugs on his father dying of cancer when he was 11 years old, "and I hope you’re thinking of me, as you lay down inside, now the drugs don’t work, they just make you worse, but I know I’ll see your face again."

We think drugs work well. But for an 11 year old boy, reality hit hard, seeing his dad die and drugs fail to change things.

Drugs work, but drugs aren't miraculous.

Having a discussion on this is always contentious. There are two wholly valid reasons for this. Firstly the statistical evidence can be challenged/critiqued/appraised. Secondly experiential learning (of having drugs and being cured/being worse) gives valid and wholly accurate evidence that they're brilliant/harmful. These two sources of information, aggregate trial data of large numbers and single patient therapeutic trials with an n of 1 both give different perspectives and facts, both supporting enthusiastic use/reticence to use antidepressant drugs.

DeeDee's comment stirred my thoughts on this, again.

Like any meaningful, complicated, multifactorial issue, in or out of health care, a dichotomous "this is good" or "this is bad" doesn't really work. Much as it's cozy and comfortable for a newspaper to publish that antidepressants are great and patients are missing out through not being properly treated, or antidepressants aren't wonderful and patients have shabby care through doctors dishing them out inappropriately, the truth is more complicated. Newspaper headlines can't be complicated. Many newspaper articles need to be timely, snappy and simple so can't be complicated. Complicated health issues (with personal and socioeconomic consequences) aren't easily discussed or debated in mainstream media. The issues are left to wither. So it goes.

If I was to be horribly reductionist and come down with a quick and easy message my thoughts'd be that antidepressants can work well as part of a package of care for some people some of the time so the message would be, "Use antidepressants appropriately and get it right!"

Which isn't massively helpful.

The crux of it is that it's incredibly useful for individual patients, as DeeDee describes. Or it's unhelpful/harmful for individual patients, as others describe.

Beyond patient numbers, we get the same pattern. Drug companies have had to evidence efficacy (that the drugs work) to get a marketting authorisation to sell their antidepressants. Clinical trial data shows that the drugs work in clinical trials. Out of clinical trials, in the really real world, a study this year published in the Journal of the American Medical Association showed that antidepressants work no better than placebo in mild, moderate and severe depression (with benefit emerging just through very severe depression).

This means if you're a GP seeing someone with mild, moderate or severe depression you're informed that, statistically, prescribing an antidepressant for the patient sat in front of you will have as much effect as prescribing placebo. Yet, clinically, some patients respond brilliantly.

It's a flaw of evidence based medicine that effects which are uncommon but highly significant for a small number of people get diluted/lost in the trial data. Trials aren't usually sufficiently powered to evidence statistical significance through rare but highly meaningful events. This is even more true in looking at clinical effectiveness of treatments (i.e. how it works in real clinical practice) rather than trials of efficacy (rigid clinical trials with strict patient inclusion/exclusion criteria).

In the really real world, outside clinical trials, people have low mood. A lot. Most people with low mood do not meet ICD-10 diagnostic criteria for clinical depression. Most people with low mood do not have a somatic syndrome, common in chemical (functional, endogenous) mood disorder. For most people, chemical solutions (of antidepressant medication) therefore has little benefit. Which is what patient and trial evidence, and clinical experience, shows us.

Yet, for people with chemical mood disorders, psychosocial interventions have some but modest utility and chemical treatments (or treatments effecting chemical changes in the brain, like ECT) can work brilliantly.

Antidepressants therefore have a very important and very valuable role to play, but in a very defined subgroup of people who have a mood problem. Of all those with mood difficulties, those with endogenous chemical depression do well. Those with reactive depression, feeling depressed because of events, responding (as most of us would) with depressed mood to a depressing situation, are low in mood through their situation not their brain chemistry, so plying medication 'pon them unsurprisingly effects little benefit.

Gets you thinking.

If appropriate and successful drug treatment hinges on accurate diagnosis and subtyping of diagnosis (it's pretty robustly evidenced and understood that chemical functional mental illnessness of ICD-10 recurrent mood disorders and bipolar disorders merit antidepressants) then is it fair to ask GPs to do this?

In older adults it's even more complicated. Loss of health, role, mobility, opportunity, income, friends and family is common in older adults we see. There's often been a lot of adversity. If life's not peachy, should folk be feeling peachy? Then, as well as loss events, there're cognitive changes. In neurodegenerative dementia like Alzheimer's Disease the limbic system, controlling mood, is always affected before memory is. Everyone with Alzheimer's Disease has brain damage affecting their mood area of the brain so frustration, irritability, low mood and changeable mood is common, before even accounting for the changes in their life that dementia causes. Teasing out if older adults have a mood disorder that's sufficient to attract an ICD-10 diagnosis of clinical depression isn't quick and easy. Determining if medication has a role to play is complex.

Our APC and PCT's been wrestling with this. Clinical care is one consideration. Cost of the drugs is another. So the question last month was, "Is it fair to ask GPs to initiate antidepressant medication, or should this always be undertaken within specialist care?"

I'm sure nobody will be bold enough to answer it.


Socrates said...

This post should be tattooed on the back of every GP's hand.

Thanks :)

DeeDee said...

I started writing a comment but it got really long so I wrote it as a post on my blog: