We welcome the review by Gaskin et al ^{Reference Gaskin, Elsom and Happell1} of interventions to reduce use of seclusion. Although studies reviewed were conducted in adult and child settings, the authors did not differentiate the developmental needs between these patient populations.

The determinants of emotional distress and aggression may differ between children and adults. In adult psychiatric units, aggression is frequently associated with psychosis. Seclusion may reduce staff injury but increases patient distress. ^{Reference Steinert, Bergbauer, Schmid and Gebhardt2} In contrast, aggression is typically the most common reason for referral to child psychiatric units. Underlying diagnoses include disruptive behavioural and developmental disorders, and are complicated by high rates of abuse and neglect. ^{Reference Dean, Duke, George and Scott3} Admission goals may include learning prosocial behaviour, necessitating use of behavioural management. Community-based studies indicate that parent management training, using contingency reinforcement and consequences such as ‘closed time-out’, are effective in reducing aggressive behaviours. ^{Reference Sanders4} Seclusion may function similarly to time-out, in that it can take the child away from a situation reinforcing negative behaviour and it encourages the child to self-regulate.

We agree with Gaskin et al that more evidence is needed to guide use of such interventions. We draw readers' attention to a recent study reporting reductions in aggression in a child and adolescent in-patient unit, following the introduction of a behavioural management programme. ^{Reference Dean, Duke, George and Scott3} The intervention incorporated staff training, contingency management and promoted use of less restrictive interventions. In keeping with current practice parameters, ^{Reference Masters, Bellonci, Bernet, Arnold, Beitchman, Benson, Bukstein, Kinlan, McClellan, Rue, Shaw and Stock5} if a restrictive intervention was required the preferred intervention was a form of seclusion. This intervention led to a significant reduction in aggressive incidents and injuries to staff and patients. Although the number of episodes of locked interventions did not decrease, there was a significant reduction in the duration of time patients spent in seclusion and a reduction in physical restraint. These outcomes were achieved without reducing admission numbers, changing the types of admissions, increasing staff costs, or increasing utilisation of medication as needed.

We concur with Gaskin et al that seclusion may exert counter-therapeutic effects, and that effective alternatives should be identified. ^{Reference Gaskin, Elsom and Happell1} However, we remain open to the possibility that predictable, time-limited locked interventions may have therapeutic effects when used within a broader behavioural management programme in young patient populations. In addition, the ultimate goal of interventions in this area should emphasise reducing the demand for seclusion, rather than just the use of seclusion per se. We need to acknowledge that some aspects of the in-patient environment can contribute to patient distress and seek to optimise the therapeutic effects of the in-patient milieu. Protocols for use of seclusion and for reduction in demand for seclusion need to be incorporated into the developmental needs of the specific patient group.