Last night I dreamt I went to Manderley again

To paraphrase Jerry Seinfeld, what's the deal with psychiatrists? Remember the old days when you more confidently prescribed lithium? Well we appear to be forgetting some older but evidence-based aspects of our medication practice. July's BJPsych reported on the considerable underuse of lithium, despite its excellent effectiveness data. This month David Owens writes (pp. 516–518) on the converse issue, the dramatic rise and overuse of the mood stabiliser sodium valproate. He notes that it might be perceived as easier and safer to use, because of a reduced need for blood tests and lack of serious toxicity, but the evidence clearly shows it inferior to lithium in long-term maintenance of bipolar affective disorders. We inadequately understand how it alters some key physiological processes, and, Owens argues, it should really be the preserve of a small number of patients with bipolar affective disorders. Local audit, discussion and – to adopt the à la mode phrase – a ‘QI (quality improvement) methodology’ might be helpful to contrast, understand and potentially change lithium and valproate prescribing where you work.

Both lithium and valproate are avoided where possible in pregnancy. There has been growing evidence that perinatal depression (PND) has important pathophysiological differences to other forms of depressive disorder, with notable recent positive therapeutic trials on the use of the allopregnanolone analogue brexanolone. Mehta et al (pp. 519–527) ask if sensitivity to oestrogen signalling – rather than changes in absolute circulating levels – is a risk factor for developing PND. They tested sex-steroid responsiveness in 60 women exposed to either a gonadotrophin-releasing hormone agonist or placebo. Across the 116 sex-steroid-responsive transcripts tested (which had previously been associated with PND), a large proportion of their dynamic changes were significantly associated with depressive symptoms, oestradiol levels and serotonin transporter binding. Oestrogen signalling has an impact on various brain functions, including hippocampal neurogenesis, and the findings offer promise as potential biomarkers.

Gureje et al (pp. 528–535) compared a low-intensity World Health Organization recommended PND treatment of psychoeducation, addressing psychosocial stressors and reactivating social networks, with a high-intensity variant that added 8-weekly problem-solving therapy sessions. Both interventions were effective in leading to symptom remission in almost 700 women in Nigeria, but there was nothing to differentiate them in terms of the impact on the mother or baby except that high-intensity variant showed some additional gains in the women who were most severely depressed. Optimising resource allocation is a critical part of all healthcare systems, but this is particularly so in low- and middle-income countries, where interventions are often given by non-medical practitioners with exceptionally heavy workloads.

We can never go back again, that much is certain

We are all products of our past and what cannot be undone, and several papers in this month's BJPsych explore this. Deighton et al (pp. 565–567) explore the prevalence of mental health problems in schools. The results from the 28 000 child data-set are shocking, with about 40% scoring above thresholds for emotional, conduct or hyperactivity problems – a figure considerably greater than previous estimates. Deprivation (eligible for free school meals) and being a child in need all increased the odds of experiencing these; interestingly children from minority ethnic groups had reduced rates. Boys were more likely to have behavioural and attentional problems, girls to have emotional symptoms. In the UK, a recent green paper has proposed introducing mental health leads and support teams for all schools, facilitating accessing specialist care: these findings would support that. Clearly not all with adverse childhood experiences go on to develop clinical psychopathology, and Dhont et al (pp. 559–564) explore specific mediators that might predict or prevent such change. Taking the longitudinal Growing Up in Ireland cohort of almost 9000 children, they found childhood adversity reported by just over 28%, and that this was significantly associated with both internalising and externalising problems. Curiously neither a positive parent–child relationship nor hobby participation had an impact on outcomes. However, crucially, parent–child conflict accounted for about half the relationship with persisting externalising problems in adolescence. This presents a prime target for interventions: the authors note how targeted interventions on parental education and training have been shown to reduce antisocial behaviour, and might serve as a model for an expanded repertoire of help.

Kathryn Abel and colleagues reverse this (pp. 513–515), looking at children and adolescents living with parental mental illness. There is a cumulative 53% risk that a child aged 16 will have experienced a mental health problem in their mother (most data-sets ‘ignore’ fathers, so the figures there are less certain). The authors argue that as well as the obvious target of assisting the parent that children are key to identifying solutions in promoting their own resilience. A call is made for developing evidence-based child-centred interventions for the most vulnerable and at-risk in this group; their team at Manchester is at the forefront of this.

And the ashes blew towards us with the salt wind from the sea

We typically have more epidemiological and clinical than neurocognitive and neuroimaged data in adolescent cohorts, but this is a time of considerable cognitive and emotional remodelling, and their capture is important. Barzilay et al (pp. 552–558) characterise the cognitive profile of young people with suicidal ideation. A US community sample of over 6000 was analysed, having undergone a comprehensive battery that tested executive functioning, episodic memory, complex reasoning and social cognitive functioning. Those with suicidal ideation (n = 672) had worse levels of clinical functioning, but fascinatingly did better on neurocognitive testing across multiple domains, an effect that was particularly pronounced in post-pubertal males. There is no obvious explanation for this finding, although the authors note the difference between suicidal thinking and suicidal acts, and longitudinal work might unpick any causal aspects. Gin Malhi et al (pp. 545–551) looked at the emergence of emotional symptoms in adolescent girls. Clinical data collected at baseline and 2 years later were used to divide them into two groups: those with and those without emotional symptoms. Comparing their functional connectivity from neuroimaging taken at baseline, significant differences were seen in the left lateral prefrontal network, with coupling strength between this region and its right-sided counterpart – areas involved in self-referential information processing and approach–avoidance behaviours – predicting subsequent depression and anxiety. Shubhangi Karmakar from Trinity College Dublin writes more on the topic, leading this month's Mental Elf blog: https://elfi.sh/bjp-me18.

Finally, Kaleidoscope (pp. 571–572) invokes Plato, and asks if honesty is, for the most part, less profitable than dishonesty: ἄνθρωπος μέτρον.