Physical health interventions on adolescent mental health inpatient units: A systematic review and call to action

Physical health inequalities experienced by people with mental health conditions are labelled an international scandal; due to the 15 to 30‐year gap in life expectancy, driven mostly by physical health conditions. Lifestyle interventions are recommended to prevent the onset of poor physical health in people with mental illness. Yet, there is less high‐quality evidence for adolescents, particularly those in inpatient settings. We aimed to assess existing literature reporting physical health or lifestyle interventions conducted on adolescent mental health inpatient units.


| INTRODUCTION
People with mental health conditions experience significant physical health inequalities compared with the general population (Correll et al., 2017;De Hert, Schreurs, Vancampfort, & Van Winkel, 2009;Firth et al., 2019). They are more likely to suffer comorbid physical health conditions such as obesity, cardiovascular disease, type II diabetes and some cancers; resulting in a loss of life up to 30 years, (Correll et al., 2017;De Hert et al., 2009). The physical health disparities have been labelled a 'human rights scandal' (Shiers, Bradshaw, & Campion, 2015;Thornicroft, 2011) resulting in multiple national and international health bodies publishing new guidelines for reducing the incidence and impact of physical comorbidities in people with mental illness (NHS England, 2018;Public Health England, 2018;WHO, 2018). In addition, a recent Lancet Psychiatry Commission identified that addressing physical health in young people from the very onset of mental illness is critical, in order to prevent chronic physical health conditions from arising .
Several factors contribute to these physical health inequalities, including side-effects of medication, poor monitoring of physical health (such as fewer physical health assessments, inability to access adequate healthcare and prioritizing mental health care over physical health), and unhealthy lifestyles (Correll et al., 2017;Firth et al., 2019;Shiers et al., 2015). Those with serious mental illness (SMI) are significantly more likely to be inactive and consume obesogenic diets (Stubbs et al., 2016;Vancampfort et al., 2017). This increased risk for poor physical health occurs at an early stage, often during adolescence (Carney, Cotter, Bradshaw, Firth, & Yung, 2016;Carney, Cotter, Bradshaw, & Yung, 2017;Cordes et al., 2017;Lederman, Rosenbaum, Maloney, Curtis, & Ward, 2017;Preyde, Tran, Parekh, & Heintzman, 2018). People who require inpatient care experience additional risks of the 'obesogenic environment' (Faulkner, Gorczynski, & Cohn, 2009;Gorczynski, Faulkner, & Cohn, 2013). Restricted living space, containment, limited access to fresh, healthy food choices and easy access to nutritionally deficient foods, restrictions on cooking facilities and fewer opportunities to be active all pose significant challenges to living a healthy lifestyle while receiving inpatient care and treatment (Faulkner et al., 2009;Gorczynski et al., 2013).
Physical health interventions which increase activity levels and improve diet have shown profound benefits for mental health in both clinical (inpatient and outpatient) and non-clinical populations Vancampfort et al., 2017). Lifestyle interventions can improve wellbeing, mood, anxiety, cognition and overall functioning, as well as improving physical health, cardiovascular fitness and ameliorating antipsychotic-induced weight gain (Curtis et al., 2016;Firth et al., 2019;Vancampfort et al., 2017). To date, there have been a plethora of studies highlighting the benefits of promoting a healthy lifestyle for children and young people (Biddle, Ciaccioni, Thomas, & Vergeer, 2018;Carter, Morres, Repper, & Callaghan, 2016;Curtis et al., 2016;Lubans, Plotnikoff, & Lubans, 2012;Parker & Bailey, 2018).
However, the majority of this work has focussed on young people receiving community care, via specialist outpatient services or on general population studies conducted in schools. In addition, many of the interventions that have been implemented on adult inpatient units may not be appropriate for child and adolescent mental health units, due to environmental factors and different needs of young people. Despite the benefits of exercise for adolescents and young people (Spruit, Assink, van Vugt, van der Put, & Stams, 2016), less attention has been given to those individuals in mental health inpatient units and there have been no reviews summarizing the evidence for physical health interventions for adolescent inpatients on mental health units.

| Aim
The aim was to assess existing literature reporting physical health or lifestyle interventions conducted on adolescent mental health inpatient units. We aim to synthesis the types of interventions that have previously been conducted and summarize the outcomes of these interventions for people's mental health, physical health and behavioural outcomes.

| Method
This review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for reporting systematic reviews , and informed by the guidance for reporting scoping reviews (Tricco et al., 2018). Due to the scoping nature of this review, the search strategy and inclusion criteria were kept intentionally broad.

| Eligibility criteria
Eligible studies were original research articles published in peerreviewed English language journals. Eligible populations were from child/adolescent generic or secure mental health inpatient services. This included any inpatient or residential services where young people aged between 12 and 25 were receiving inpatient treatment for any mental health disorder. The upper age limit was identified in line with literature highlighting the growing evidence for young people's mental health services to extend up to age 25, and definitions of adolescence based on young people's cognitive development (Malla et al., 2016;Rickwood et al., 2019;Sawyer et al., 2012;Wilson et al., 2018). Studies which reported physical health interventions, defined as any form of activity designed to increase activity levels or improve diet quality, for example, sports, exercise, physical yoga, were all included. Trials of all design were eligible including non-randomized and uncontrolled interventions, and interventions delivered through routine care. Studies that reported interventions which had a psychological/therapeutic basis, without any focus on physical health were excluded, as were those containing only older populations (18 years plus), studies which included mixed samples including those under 25, but who could not be extrapolated from the remainder of the sample, or those which were purely qualitative studies.

| Search strategy
An electronic database search was conducted on 13th June 2019 using Ovid MEDLINE, PsycINFO, Embase, Cochrane Central Register of Controlled Trials and AMED. Search terms synonymous with 'inpatient settings', 'mental health', 'children and adolescents' and 'physical health interventions' were used (strategy available on request). A manual search of Google Scholar was conducted and reference lists were scanned to identify any additional papers. The search strategy was intentionally broad and included all publications from database conception until the date of the search.

| Study selection and data extraction
Reviewers independently screened articles for eligibility. Disagreements were resolved through discussion. An excel tool was developed to record: (a) study characteristics (author, year of publication, country of origin, study design); (b) sample demographics (sample size, gender, age); (c) clinical demographics (diagnoses, description of service, interventions offered); (d) physical health interventions (type, delivery, duration, content, control group, adherence); (e) prepost measurements (physical measurements, for example, weight; psychological measurements; behavioural outcomes); (f) summary of findings.

| Analyses
A narrative synthesis was conducted on the data. Systematically extracted data were combined to identify similarities and differences in the types of intervention, for example, duration, delivery of intervention and content. Tabulated data were created to summarize the key study characteristics. Quantifiable data were combined, and key outcomes were described for each paper, summarized according to measurement type, for example, physical outcomes.

| Search results
Study selection and exclusion is summarized in Figure 1 if they reported an ineligible population (eg, no mental health condition; n = 12), if they did not report any physical health interventions (n = 8), were conference abstracts which did not lead to a full text (n = 4) or individual case studies (n = 1). One further paper was identified through a manual search of Google Scholar. Studies were conducted in Australia (n = 1), the United States (n = 1) and New Zealand (n = 1). A summary of the studies can be found in Table 1.

| Clinical settings
The clinical setting varied, however, all individuals were receiving inpatient treatment for a diagnosed mental health condition. This included a specialized early intervention service providing inpatient treatment and crisis care (Woodhead et al., 2018), tertiary mental health facilities for young people with a severe psychiatric disorder (Bobier et al., 2009) and a locked residential unit for young people

| Summary of interventions
All studies employed a pre-post design. None of the studies had a control group, and all were made available to all of the individuals within the service. Two studies looked at physical health interventions, which were integrated as part of usual routine activities (Bobier et al., 2009;McIlvain et al., 2015), whereas one reported a new pilot intervention to be offered within the service as an adjunct to usual care (Woodhead et al., 2018). The content of the interventions varied (see Table 2). This included a multi-faceted lifestyle intervention delivered by Accredited Exercise Physiologists, a yoga programme focusing on movement and activity (rather than spirituality), and a routinely administered sporting intervention. However, the type of sport used in Bobier et al. (2009) could not be determined and was described as a routinely offered sport session making up generic ward-based activity. The duration of interventions ranged from 8 weeks to 6 months.
Little information was available on how each of the interventions was developed and none of the studies reported using models of behaviour change or theoretical underpinning.

| Outcomes
Eligible studies reported a range of different outcomes.

| Psychological outcomes
The HONOSCA (Health of the Nation Outcome Scale for Children

| Physical outcomes
Physical health measurements improved in relation to the 'Bod Squad' intervention (Woodhead et al., 2018). Body mass index (BMI) and waist, hip and chest circumference all decreased after the intervention. However, the sample of individuals who had pre-post outcomes was limited, therefore, conclusions were not based on any significant statistical tests nor any longitudinal follow ups. Physical activity levels or other lifestyle behaviours were not reported in any of the studies, despite the focus of each intervention being to improve physical activity and physical health.

| Feasibility of studies
The feasibility of the studies was assessed by looking at levels of adherence, which varied considerably. Older individuals and those with more severe mental health difficulties on admission were more likely to engage in weekly sporting sessions than younger individuals (Bobier et al., 2009). In addition, individuals with mood disorders (eg, depression) had lower levels of adherence than those with psychosis and mixed diagnoses (Bobier et al., 2009 frequently (Woodhead et al., 2018). This related to the observation that young people either engaged well with the intervention or they only attended a few sessions.

| DISCUSSION
The aim of this review was to identify studies that reported lifestyle interventions (physical activity and diet) administered within child and adolescent inpatient units. There is a paucity of evidence for the implementation of lifestyle interventions in this setting. However, preliminary work suggests lifestyle interventions may be feasible within routine inpatient care. Therefore, this could potentially benefit young people's mental and physical health, as observed in other populations and care settings (Bobier et al., 2009;Firth et al., 2019;Gordon & Lubitz, 2009;McIlvain et al., 2015;Noetel et al., 2016;Van der Baan-Slootweg et al., 2014;Vancampfort et al., 2017;Woodhead et al., 2018). For example, increased social support, higher activity levels, improved behavioural and psychosocial functioning and physical health improvements, have all been observed in the existing studies (Bobier et al., 2009;McIlvain et al., 2015;Woodhead et al., 2018). However, relatively few studies have been conducted to date and there is a lack of high-quality research in this area, meaning we are unable to make any conclusions on the efficacy of existing interventions or specific variables which could improve efficacy and outcomes.
Despite the dearth of current studies, the few that were identified suggest lifestyle interventions may be beneficial for young people. For example, individuals who took part in the 'Bod Squad' intervention found having a personalized approach to looking after their mental and physical health was useful (Woodhead et al., 2018).
Many also valued the social element of the intervention, and claimed it had empowered them to engage in more physical activity in their daily life. The remaining studies suggest activity-based interventions may be acceptable and feasible when offered routinely within child and adolescent services. Rather than treating physical health independently to mental health, implementing activity-based sessions as part of routine ward management has the potential to benefit outcomes for young people.

| Clinical implications
There is a clear evidence and implementation gap, which neglects the impact lifestyle interventions may have for young adolescents on inpatient units. Given the plethora of existing evidence within clinical populations Vancampfort et al., 2017), it is surprising that this knowledge has not yet extended into child and adolescent research. There are many reasons to suggest lifestyle interventions will benefit young people in inpatient services. For instance, multiple meta-analyses in adult patients have already demonstrated that exercise, diet and broader lifestyle interventions can reduce weight gain and improve a broad spectrum of metabolic health outcomes in people with severe mental illness Vancampfort et al., 2019). In recognition of this, lifestyle interventions are recommended by National Institute for Health Care and Excellence (NICE), for use within mental health services (National Collaborating Centre for Mental Health, 2014). However, a recent review found they are not consistently offered in mental health trusts in the UK (Swaby et al., 2017). None of the studies included in this review had a focus on nutrition or looked at nutritional outcomes or knowl- edge. Yet, nutritional needs are critical in inpatient environments where young people are often given little choice over the foods they are provided, or receive little psychoeducation on the foods they consume both on the unit and upon discharge (Faulkner et al., 2009;Gorczynski et al., 2013). Therefore, ensuring appropriate interventions are implemented in inpatient units, in line with evidence-based practice is an area of clinical importance which would minimize the potential for iatrogenic harm.
Adolescent inpatients are a particularly vulnerable group. They often present with risk factors for poor physical health which precede the onset of illness. For example, many have high BMI values on admission, which rapidly increase with duration of stay (Carney, Imran, Law, Folstad, & Parker, 2019). They also display pre-cursors to developing metabolic disorders and preventable physical health conditions, such as dysregulated blood metabolites and hyperprolactinemia (Carney et al., 2016;Carney et al., 2019;Cordes et al., 2017;Eapen et al., 2012;Vancampfort et al., 2017). This affords the opportunity to intervene to prevent the development of comorbid conditions. Adolescents with mental health conditions are also less likely to receive high-quality physical health care across both inpatient and community mental health services than the general population (Carney et al., 2016;Carney, Bradshaw, & Yung, 2018;Cordes et al., 2017;Eapen et al., 2012;Vancampfort et al., 2017). Previous evaluations have shown young people who are prescribed antipsychotic medication do not receive physical health monitoring according to NICE guidelines (Gnanavel & Hussain, 2018;Pasha, Saeed, & Drewek, 2015). There is evidence that metabolic health is also poorly monitored in adolescent inpatient settings, and many difficulties often go unrecognized and untreated (Eapen et al., 2012).
Those admitted to adolescent inpatient units experience the additional risk of the 'obesogenic environment' (Faulkner et al., 2009;Gorczynski et al., 2013). This refers to the additional restrictions and barriers to living a healthy lifestyle; including reduced opportunities to be active and increased access to high-calorie foods (Faulkner et al., 2009;Gorczynski et al., 2013). Many existing interventions have been conducted in community settings or specialist services for outpatient treatment (Biddle et al., 2018(Biddle et al., 2018Carter et al., 2016;Curtis et al., 2016;Lubans et al., 2012;Parker & Bailey, 2018;Vancampfort et al., 2017). Therefore, tailoring interventions for adolescent inpatient units is necessary to optimize the care young people receive and may even improve long-term mental and physical health outcomes.

| Strengths and limitations
To the best our knowledge, this is the first review of physical health interventions on adolescent inpatient units. We have identified a gap in the evidence base and highlighted the need for more research in this area. The scope of this review was intentionally broad, and few eligible studies were identified. The heterogeneity of the studies, and various outcomes and methodologies used did not allow direct comparison or meta-analyses. In addition, outcome measurements were often not standardized (McIlvain et al., 2015), and were devised by the researchers for the purpose of the study, introducing further bias.
Formal quality assessments were not conducted due to the limited amount of studies identified; two of which looked at routinely administered physical health interventions, rather than controlled trials. One further consideration is that many child and adolescent services might indeed be conducting physical health interventions routinely but have not been formally evaluated. Therefore, we cannot draw firm conclusions on the efficacy of existing interventions compared with treatment as usual. However, we can make the call for further research in this area.

| Future work
Future work should explore lifestyle interventions for this vulnerable group. High-quality studies are needed in this area to identify effective health promotion interventions for adolescent inpatients. Many existing interventions are not appropriate for young people, or do not take into account the additional considerations of an inpatient unit such as limited opportunities to exercise, staffing requirements, patient safety, risk concerns and high levels of psychological distress.
It may also be that an individualized approach is needed, adapted to enhance engagement across the varied diagnoses observed on child and adolescent inpatient units. This is reflected in the studies included in this review showing people with low mood had lower levels of adherence than those with schizophrenia spectrum disorders (Bobier et al., 2009). Therefore, future research should address this and identify feasible ways to increase physical activity levels and improve dietary intake on adolescent inpatient units. Interventions should be developed using the MRC Framework for Developing Complex Interventions (Craig et al., 2008); ensuring there is a solid theoretical underpinning for behaviour change, and that any approaches taken are appropriate for the target population. It will also afford the opportunity to explore mechanisms of action and change (Michie, Van Stralen, & West, 2011). Adopting a youth-friendly approach is important; one which is developed in collaboration with young people, and affords the opportunity for continued engagement after transition from inpatient mental health settings. Given the young age of this population, there is potential to explore wider interventions that continue outside of the ward environment, by including families or community-based health services.

| CONCLUSION
Physical health interventions may be useful, feasible and acceptable within routine inpatient care for adolescents. However, there is a paucity of high quality, robust evidence for the use of lifestyle interventions on adolescent inpatient units. Significantly more research is required in this area to identify the best ways to improve outcomes for young people. It is important that we seek to reduce the physical health inequalities experienced by people with serious mental illness, and do so at the earliest possible stage. Adolescent inpatient units represent a potential target to intervene early and alleviate the physical health disparities of this group.
Therefore, ensuring appropriate lifestyle interventions are implemented in inpatient units, according to evidence-based practice is an area of clinical importance, to minimize the potential for iatrogenic harm.