Three hours from London, Jacomo's free, taking his woes down to the sea

This month's Kaleidoscope (pp. 64–65) opens with talk of biopsychosocial, but it is really sociopsychobio in terms of emphasising what has an impact on the wider determinants of health and well-being. Senior et al (pp. 4–5) describe the toxicity of debt and poverty to mental health, and conversely how mental illness can lead to impoverishment. I was fascinated to learn that in the USA, suicide rates in men fall sharply when they reach the eligibility age for social security; I was less surprised to read how complex universal credit and work capability assessment forms have been linked to mental illness and suicide in the UK. The authors propose guiding principles on how a welfare system might improve to both avoid harm and indeed promote mental well-being through reducing bureaucracy, training assessors in mental illness and having unconditional components to payments. Their call for governmental ‘therapeutic jurisprudence’ – where legislation is assessed for the potential impact on mental health – felt especially innovative.

‘One in four’ is the ubiquitous prevalence descriptor on mental illness, so it is important to check it is, well, correct. Bebbington & McManus (pp. 55–57) do this via the British Adult Psychiatric Morbidity surveys of 2000, 2007 and 2014. They find a better description is ‘one in six’, and this has been very stable over time. This inevitably opens wider debate on what constitutes ‘illness,’ and indeed the figure varies considerably depending upon inclusion or exclusion of personality disorder, attention-deficit hyperactivity disorder and substance use disorder. The original expression is well understood and used by the general population and media, and has a social as well as scientific value: I suspect it is unlikely to shift.

In search of a lifetime to tell when he's home, in search of a story that's never been known

Speaking of the media, they seem to me to be curiously preoccupied by environmental causes (and miracle cures) for dementia in a manner that does not occur for other disorders. Well, we have all bases covered in this month's BJPsych. Russ et al note (pp. 29–34) a previously reported link with aluminium and fluorine in drinking water, and update with the largest longitudinal study on the topic. Their data, from a Scottish cohort, support the association, and interestingly even low levels of either of these elements was associated with greater risk of developing dementia. Any causative mechanism remains unclear, although aluminium has been shown to influence β-amyloid oligomerisation and oxidative stress, and fluorine increases its absorption from drinking water. Peters et al (pp. 16–28) systemically reviewed the longitudinal trajectory of blood pressure, body mass index and cholesterol in those subsequently diagnosed with dementia. They found that body mass index first rises, followed by blood pressure, with both then dropping back considerably in the 5–10 years before diagnosis; data on cholesterol were less clear. The authors propose that it is the trajectory of change, rather than a current value of these measurements, that is of most utility in identifying those at higher risk.

Vik-Mo et al (pp. 43–48) take a different tack, longitudinally measuring neuropsychiatric symptoms annually over 12 years in those in primary care with suspected mild dementia. Neuropsychiatric symptoms include issues such as aggression and agitation, delusions, irritability and aberrant motor behaviour. These have a major impact on quality of life for individuals and their carers, and are associated with higher rates of institutionalisation, but most work to date has been cross-sectional, and any change over time has been less understood. In this work, very severe neuropsychiatric symptoms were generally single episodes, but 8% had such symptoms in a steady manner across time. More broadly, outcomes were strikingly heterogeneous and a stable course of neuropsychiatric symptoms was uncommon. Nagina Khan explores further in this month's Mental Elf blog: https://elfi.sh/bjp-me20

George Savulich et al’s editorial (pp. 1–3) notes how anxiety and depression are commonly found in dementia but can be difficult to detect and disentangle from the pathology of neurodegeneration. This matters, as correct discrimination allows targeted interventions, and indeed these may be modifiable risk factors for cognitive decline. Gill Livingston's team report (pp. 35–42) on the START (STrAtegies for RelaTives) psychological intervention for family carers of those with dementia. Earlier work had shown START effective at reducing symptoms of depression and anxiety at 2 years, and this follow-up lengthens this to a 6-year time point. Compared with treatment as usual, the intervention group continued to show significant clinical benefit at this extended time. Importantly in a constrained healthcare system, there were no differences in costs over treatment as usual, demonstrating the programme to also be economically effective.

Three hours is needed to leave from them all, three hours to wonder and three hours to fall

Wrapping up this month's BJPsych are a couple of papers on non-dementia aspects of older persons’ mental health. Fornaro et al (pp. 6–15) systematically reviewed the prevalence and correlates of major depressive disorder (MDD), bipolar affective disorder and schizophrenia among nursing home residents who did not have dementia. This is a cohort that is growing significantly each year, both in total number and as a percentage of the population: these data matter, but frustratingly such individuals are frequently not included in research. This work pooled 32 observational studies that covered over 13 000 such individuals and found a mean prevalence rate for MDD of just under 19%. A paucity of data meant that reliable figures could not be obtained for other conditions, and a problematic gap in our knowledge remains for those important areas. Arthur et al (pp. 49–54) continue the theme, looking for changes in prevalence and treatment of MDD in older people more generally. Data from two English population cohort studies were compared: the Cognitive Function and Aging Studies I and II that took measurements in 1990–1993 and 2008–2011, respectively. They found a relative, but not statistically significant, decrease in the prevalence of depression in those ≥65 years, with a case-level prevalence of 6.8% in the latter sample. However, the proportion taking antidepressant medication showed about a threefold increase across the two-or-so decades, going from 4.0% to 10.7% of this population. Of course, with antidepressant ‘number’ debates, it is the underpinning issue of appropriateness that matters: are the ‘right’ people getting medication. Unfortunately, this work replicated what we have seen elsewhere: many with depression were still not getting any treatment, and many on treatment did not have depression.

Finally, Kaleidoscope (pp. 64–65) discovers the differences between men and women in how they view their ex-partners. Tut-tut, I am disappointed by your instant anticipation of what this showed, which is a purely stereotyped guess (I am a psychiatrist – as people at parties note, I must be able to read your mind). Turn to the back and have your prejudices confirmed or refuted.