It’s Not the Combat – Maybe it’s the Drinking in Vulnerable Young Men

JAMA 2013;310(5):496-506

A cohort study published in JAMA tried to answer the question what are the risk factors for suicide in the US military. This is a hot topic as the rate of suicide has increased in US military personnel from about 11/100,000 people in 2005 to about 18/100,000 so that now deaths from suicide outnumber deaths from combat.

What the authors did was to gather together three cohorts in 2000, 2003 and 2006. Importantly they included reservists as well as full time members of the military (reservists make up about 30% of the US army which is relatively high compared to other countries). Participants were asked to complete a baseline questionnaire and then a questionnaire every three years. About a third of people approached agreed to take part in the study and they were more likely to be women, younger and college educated than the typical US military population.

Death, the main outcome, was assessed from the National Death Index and the Department of Defense Medical Mortality Registry. The authors were particularly interested in whether deployment and combat experience were linked to suicide, having gathered information about this from records and surveys.  The questionnaires completed by the participants assessed various other factors such as stressful life events; post-traumatic stress disorder; depression and drinking (although they didn’t ask about other drug use which seems a significant omission).

What they found in the 151,000 participants is that 83 people died as a result of suicide between 2001 and 2008 which gives a crude rate of 11.73 (95% confidence interval 9.21-14.26) per 100,000. What is more important is not the crude rate (as the authors didn’t set out to establish what the “true” rate was in the US military) but the risk factors associated with suicides. They found that suicide was highest in those with bipolar disorder, depression and alcohol problems. There was no link with deployment or combat exposure – in fact those who were not deployed had higher crude rates of suicide than those deployed once.

One way of expressing risk is to report the population attributable risk which tells you how much the rate of a disorder would reduce if you got rid of the risk factor.  (This is useful as it takes into account how common the risk factor is and how much it increases the risk – if a risk factor is rare then from a population perspective how much it increases your risk is relatively unimportant).  In this study the population attributable risk for alcohol related problems was 18%, depression 11% and bipolar disorder 5%.

So in this population, as in most risk factor studies, it was the usual suspects of depression and substance use that was associated with suicide. There are some caveats here though. In this study there were only 83 suicides so there were wide confidence interval intervals in the hazard ratios. There is also the possibility of the healthy deployment issue – those personnel who were unwell or had substance abuse problems may have been less likely to have been deployed or see combat.  The study also only covers only 2005 to 2008 when the increase in suicides was beginning. It will be interesting to see if later data supports these initial conclusions.

However, the study does beg the question of whether there was something “depressogenic” about being in the US military. The population of the US military is generally younger than other high income country armies and deployment when it occurs is longer than other countries with shorter breaks “at home”. Personal testimony from serving and recently retired soldiers also contributes to a view that invading and occupying small defenceless countries far away from the US can affect morale. However, this argument also applies to armies from other countries such as Canada and the UK where there has not been such an increase in military suicides.

Another argument to explain these findings is that the response to mental illness in the military is problematic. We know from the civilian population that identifying and responding to depression in primary care is one of the best ways of reducing suicides. As the authors note mental health problems in the military increase risk of suicide – just as in civilian life- but the response is often different with a much greater reluctance to divulge mental health problems.  This might explain why alcohol abuse – often used to self-medicate psychiatric disorders - was the most important risk factor in this study.

Comparative Efficacy and Tolerability of Antipsychotic Drugs in Schizophrenia

The Lancet, 27 June 2013 (In Press - DOI: 10.1016/S0140-6736(13)60733-3

Recently available on-line in The Lancet is a paper which describes the comparative efficacy and tolerability of 15 antipsychotics in people with schizophrenia written by a group in Germany with a track record of producing such studies. The difficulty with the evidence when comparing drug treatments is that where you have lots of pharmaceutical options there are rarely any head to head comparisons of the different treatments. One way to get a sense of comparative effectiveness is to do a meta-analysis of different studies.

The authors here combined the results of 212 blinded randomised controlled trials that included 43049 patients who had schizophrenia. They excluded studies where people had mainly negative symptoms, those who were treatment resistant and those done in stable patients (mainly relapse prevention studies).  Of the people in the trials the mean duration of illness was about 12 years and they had an average age of 38 years.

Using a sophisticated form of metal analysis they then produced hierarchies of effect sizes for overall efficacy, discontinuation, weight gain, extrapyramidal side effects, prolactin increase, QTc prolongation and sedation. What they found for overall efficacy is that clozapine (by some margin) was the most effective followed by amisulpiride, olanzapine and risperidone.  Haloperidol came in at seventh in the list. When looking at overall acceptability the top three least likely to be discontinued compared to placebo were amisulpiride, olanzapine and clozapine.

The true interest in this study is that it challenges two ideas. The first is that all antipsychotics have the same effectiveness. Whilst the difference in effectiveness between the drugs was small (and smaller than for the side effects) it was a robust finding that didn’t alter much when tested in the sensitivity analysis. The second challenge is that thinking about antipsychotics as first generation or second generation isn’t very helpful as it obscures differences in adverse effects between all antipsychotics. For example aripiprazole is less sedating than quetiapine but haloperidol is somewhere between; similarly olanzapine causes more weight gain than risperidone but chlorpromazine is in between them.

As the authors state, “Antipsychotic drugs differ in many properties and can therefore not be categorised in first generation and second generation groupings. The suggested hierarchies in seven major domains should help clinicians to adapt choice of antipsychotic drug to the needs of individual patients”.

The DSM-5 and the Complexities and Capitalizing of Classification

Well it’s not actually a journal article but as everyone and their dog has an opinion on the launch of DSM-5 next week I thought I would pitch in as well.

First why bother with classification at all? We have classification because it is useful for communication and ultimately inevitable. The reason that it is inevitable is that once you recognize that some people share something in common that other people do not have – low mood or forgetfulness for example – you are creating a classificatory system. The other choices are everyone is unique (not that useful because it means you can’t apply lessons from one person to another) or that everyone is the same (again not that useful).

So given that we have to have a system of classification in medicine what should it look like? There are three choices – classification by symptoms, by course of the disorder or by aetiology.

Until the mid-19^th century most disorders in medicine were classified by symptoms – so a reading of medical textbooks from the 1700s would have several chapters on different types of fever. This changed as more knowledge was gained about how the body worked and links were made between pathology and symptoms in life. For most medical disciplines the classification changed from symptoms to aetiology so that physicians today don’t diagnosis central crushing chest pain disorder (a symptom), they diagnosis a myocardial infarction (an aetiology). Classification by course of a disorder has been tried in medicine but is never really that successful as it can only be done retrospectively.

The reason that classification never shifted from symptoms to aetiologies in psychiatry is that the brain is the most complex organ in the human body (whose workings we still don’t fully understand), which is enclosed in a bony box (the skull), making it hard to study (unlike the heart or pancreas for example). So in psychiatry, with a few exceptions, we are still stuck with a symptomatic classification of disorders. The trouble with symptomatic classifications is that they are not really that powerful in helping decisions about treatment or prognosis – hence the focus on formulations in training psychiatrists which attempt to take a wider view of the aetiology and impact of disorders.

Which brings us to DSM-5  This is a classification by symptoms from a particularly U.S. point of view. The two big criticisms of DSM-5 are that it medicalizes what should be normal and that, while it pretends to be biomedical, the evidence for the biological basis for most disorders is lacking.

In my opinion there is considerable merit in the argument about DSM-5 being an attempt to medicalize the normal – a sort of “psychiatric mission creep”. The suspicion here is of the influence of pharmaceutical companies on the designers of DSM-5 to create new markets.

There are many examples of undeclared conflicts of interest, particularly in U.S. academic psychiatry, influencing the research agenda and the interpretation of research. The other financial conflict concerns the American Psychiatric Association who publishes the DSM. The drafting of DSM-5 has missed most deadlines except the final publication and launch date, leading to the suspicion that the APA is in poor financial straits and needs the DSM to come out now in order to collects the money it makes from its sales. 

The second argument about the DSM-5 is that it does not reflect biological reality and the comparison is often made with the rest of medicine. However, many disorders in medicine are equally subjective (pain for example) or their cause is obscure (headaches or migraines anyone?). Also disorders in other areas of medicine regularly undergo reclassification (epilepsy or acute coronary events come to mind).

Critics often condemn the “medical model” when what they really are referring to is a reductionist biological model that equates all disease with biological pathology. However, anyone who spends any time on a medical ward round or out-patient clinic will quickly discover that the “medical model” actually means integrating biology, psychology and sociology in a complete package.

So in psychiatry we are still left with a predominantly symptomatic classification to understand people who present with distress. Our focus should be on improving and defending services for such people and, as good clinicians, integrating psychology and biology to do something helpful.

The role of classificatory symptoms is often exaggerated – the best quote I have heard about them is that “they are like lines of longitude and latitude – nothing like them exists in the real world – but they are helpful in finding your way around”.

Screening to Assess who is at Risk for Suicide Falls Short - Screening to Assess Depression Likely More Constructive in Managing the Risk of Suicide

Annals of Internal Medicine; Published first online April 23, 2013

Suicide and how to prevent it is a hot topic. From the evidence that we have, investing in primary care to improve the detection and treatment of depression would appear to be the place where you get the biggest bang for your buck. Depression is clearly related to suicide, so you would think that screening for suicide in primary care might be helpful in suicide prevention.

However, a recent systematic review of “Screening for and Treatment of Suicide Risk Relevant to Primary Care” in the Annals of Internal Medicine concluded that screening in primary care was of limited usefulness. The authors searched for English language studies only (again!) up to December 2012 which looked at two questions:

“What are the benefits and accuracy of screening instruments in primary care?”

“What is the effectiveness of suicide prevention interventions in primary care or mental health settings?”

Addressing the first question, the authors found five studies which showed no clear short term benefits (within two weeks) of screening and that the accuracy of screening instruments was poor. It should be noted that while the authors talk about screening, this is not screening as it would be applied to other disorders in medicine. In other disorders screening refers to the time between biological onset (say the presence of cancerous cells) and developing symptoms (for example breast lumps).

For screening to be effective, treatment given before symptoms develop needs to be more effective than treatment given after symptoms appear. Clearly if people respond positively to questions about suicide then “symptoms” have already developed, so what the authors are really talking about here is case finding rather than screening. Perhaps more effort should be put into screening people for depression rather than suicide specifically.

The second part of this review looked at interventions in both primary care and mental health settings that might reduce suicide risk. As one of the authors of a study included in this review I was interested to see their conclusions. What the authors found were 49 trials looking at reducing suicidal risk.

Psychotherapies reduced suicide attempts in adults but had little effect on suicidal ideas whereas interventions which tried to enhance usual care had little effect on suicide risk. Nearly all the studies were not done in primary care and recruited patients at high risk of suicide from general hospitals (usually people who presented with self-harm to the emergency department) so the results cannot really inform what to do with people detected as being suicidal in primary care.

So what next? Risk assessment tools do not predict who will commit suicide or repeat self-harm. What are needed are better risk management systems rather than yet another risk assessment form. Screening for depression and offering brief effective treatments – possibly computerized therapies – that can be used in primary care may prove to be a more useful strategy than simply screening to find “suicidal” people.

Childhood Trauma and Abuse is the Smoking of Psychiatry

Annals of Internal Medicine 2013; 158(3): 179-190

Childhood trauma and abuse is the smoking of psychiatry. As a risk factor for mental illness it is comparable to how smoking a pack of cigarettes per day increases the risk of lung cancer and heart disease.

As an adult psychiatrist I see the consequences of poor starts to life and do my best to manage the consequences. However, doing my best often isn’t that effective or proves to be only a temporary solution. So the real question at hand is - what if we could do something to prevent child abuse from happening?

This is the question addressed in a systematic review published in the Annals of Internal Medicine in February this year. The authors update the US preventive services task force recommendations from 2004 on the prevention of child abuse and neglect. They reviewed the literature up until the middle of 2012 and found 11 randomised controlled trials which tested different interventions to prevent child abuse.

However, they only included English language studies and didn’t search outside MEDLINE, PsycINFO or the Cochrane Central Register of Controlled Trials which means non-English language studies were likely to be missed. (It seems strange that systematic reviewers ignore non-English language studies – is the implication that they are somehow less “good” than English language papers or that we can’t learn from “foreign” countries?).

The authors found one study in a paediatric clinic (although in other countries this would more likely be a family physician practice) which sought to identify and then refer those children up to age 5 at high risk of abuse and neglect. The program found that the intervention produced a significant reduction in child protective service reports up to four years after the child was seen. The evidence was less clear cut for various forms of home visits by nurses or “trained laypersons” up to three years after birth.

The conclusion of the review is that risk assessment and behavioural interventions in paediatric clinics reduced abuse and neglect for young children, but the evidence is inconsistent for home visitation programs. Plus, of course, it’s clear that additional research is needed.

This seems to be one of those areas where there is the choice between focusing on those at high risk or trying to address abuse at a population level – the classic Geoffrey Rose problem first articulated over salt and hypertension. Providing leadership through the development of infant mental health services may be one way forward.

Antipsychotics Not the Answer for Treatment Resistant Depression

PLOS Med 2013; 10(3): e1001403

The so called atypical antipsychotics are no more effective than the older ones but the neuromuscular adverse effects have been replaced by metabolic ones.

One exception to this broad generalization is their use as an adjunctive treatment for treating depression, although there is a suspicion that this is a sign of “indication creep” for drugs which may be coming off patent. To address this, a systematic review raises its head in the March PLOS Medicine on-line Journal.

The authors identified 14 trials (all funded by drug manufacturers) where patients with depression who were “treatment resistant” were randomized to receive either an antipsycotic or placebo. The trials were short term studies ranging from 4 to 12 weeks and included aripriprazole, an olanzapine/fluoxetine combination, quetiapine and risperidone.

All four drugs had significant effects on remission and response (except the olanzapine/fluoxetine combination which did not affect response rates). However the effect was small with a mean difference of about 2.5 on the Montgomery-Asberg Depression Rating Scale. Additionally, on measures of quality of life the drugs had no or very small effect (with the exception of risperidone which had a small to moderate effect). Numbers needed to treat for remission compared to placebo were 9 for aripiprazole, quetiapine and risperidone and 19 for the olanzapine/fluoxetine combination.

Treatment was associated with weight gain, akathisia, sedation and abnormal metabolic laboratory results.

So would I have an antipsychotic if I had treatment resistant depression? Probably not and if I did I'd want to stop it pretty quickly.

Association between Depression and Hospital Outcomes among Older Men

Canadian Medical Association Journal; 185.2 (Feb. 5, 2013) p117. 

It is a little known fact that the Canadian Medical Association Journal ranks at number 8 in the list of the world’s top medical journals as judged by impact factor. So to emphasize yet again that there is no health without mental health and the global nature of healthcare, in February the Journal published a cohort study of over 5400 Australian men.

These men were over 65 years old and were enrolled in the “Health in Men study” where they were assessed for depression at baseline on the Geriatric Depression Scale.  Two years later of the 339 men who scored more than 7 on the rating scale 152 (44.8%) had at least one non-psychiatric emergency hospital admission compared to 1164 of 5072 (22.9%) non depressed men.

The depressive symptoms also predicted whether these men were admitted to hospital (for non-psychiatric conditions), the number of admissions and the total length of stay. The system of separating mental health care from other aspects of health care is anachronistic and can no longer be supported by the evidence. Once funders recognize that taking into account the whole person improves quality and reduces costs this separation will increasingly be seen to be out of date and a sign of organizational stigma.