planning and delivery of liaison psychiatry services requires many decisions regarding, for example, staffing profile; the type of work to be undertaken and by which members of the team; the relation between the team’s work and the work of other parts of the local mental health services; and how outcomes will be measured. Putting all these decisions together with the rationale for them amounts to what is sometimes called a programme theory or logic model for the service – outlining what is expected to work for whom and under what circumstances. These decisions are not made in isolation: key influences on the final shape of a service will depend upon the views (and financial support!) of service commissioners, which are in turn influenced by national policy. Even a carefully worked-out configuration for a liaison service is unlikely to be realised exactly as planned – implementation (delivery) will be influenced by the practical limitations experienced by staff in day-to-day work, and by the views of patients and of the health professionals who do (or do not) refer to the service.

Lifetime prevalence of self-harm (self-poisoning or self-injury) is 5–6% in the UK adult population. Rates are increasing especially among young women, where lifetime prevalence is now 25%. The commonest immediate causes are social and interpersonal problems. The most commonly associated mental health problems are mood disorders and alcohol misuse. Assessment involves careful characterisation of the act of self-harm, the person who self-harms and their social circumstances, but should not entirely be focussed on risk. It should identify needs in mental health or physical health, social needs and psychological needs, and should be a therapeutic experience for the person who has self-harmed. The assessment itself may be helpful for some. Research into cost-effective treatment is limited, but even so it is reasonable to offer brief psychological therapies to everybody and longer-term therapies to those with severe repeated self-harm. Quality improvement activities should aim to improve the experience for people making contact with mental health and to increase the availability and accessibility of therapeutic interventions for self-harm.

outcome measurement in healthcare is to establish the effect of a clinical or organisational activity. A useful initial step is therefore to clarify not just what outcome is being assessed but also what activity is being evaluated by measuring that outcome. In terms of describing the activity of liaison psychiatry, the service can be understood as a complex intervention in healthcare. That is, it has multiple components, each with its own mechanism of action and possible effects. This raises a fundamental question about outcome measurement – what part of this complex system is being evaluated? Before considering the particulars of outcome measurement, we therefore need to have a way of characterising liaison services and how they might work.

Mental health’ is a catch-all term used to describe a wide range of conditions – from emotional states such as worry about academic progress or distress due to relationship problems or social challenges such as poverty, through to psychosis and other severe mental illness. To start therefore, this chapter will outline the definitions of some common terms used to describe those disturbances of mental health that are usually called emotional or mood disorders, and what might define mood disorders that are severe enough to be considered an illness. It will review what can be said with some certainty about social media use and disorders of mood, in a field where there is in truth a great deal of uncertainty. It will consider the common causes of mood disorders in general and what features of social media use might therefore lead us to be concerned that in any association between social media use and mood disorder, the former explains the latter. It will conclude with pointers to the research that needs to be done if the situation is to be clarified.

The main quantitative research designs are reviewed, to indicate their strengths and limitations. Even the best designed and conducted studies run the risk of being influenced by the ever-present threats of chance, bias, and confounding, terms that are explained and illustrated with examples, and of sampling and measurement error. The informed reader should therefore look beyond the results of a study, using critical appraisal skills to ask whether the study’s design, sample, and measures are robust enough to allow the findings to be taken as likely to be accurate.

In this chapter we review the research evidence on the relationship between social media use and self-harm. There is much discussion about the possible role of social media in rising rates of self-harm, seen especially in young girls, a discussion that is accompanied by a worry that there may be an associated increased risk of death by suicide. It is beyond the scope of this chapter to go into detail about current understanding of why some people self-harm and some die by suicide - suffice to say that reasons for both are complex and multifactorial. Here we focus specifically on what we know about any potential role of social media and discuss possible mechanisms of action towards harm or benefit in relation to self-harm content.