Abstracts for presentation

 

PLENARY

Deborah Caldwell
University of Bristol

Are school-based interventions for prevention of anxiety and depression effective? A network meta-analysis.  

Deborah M Caldwell1, Sarah R Davies2, Paola Caro3, Jennifer Palmer4, Clare French5, Jose Lopez-Lopez6, Judi Kidger7, David Gunnell8, Rona Campbell9, Sarah E Hetrick10, James Thomas11 and Nicky J Welton12.  

1. Senior Lecturer in Public Health Research, Department of Population Health Science, University of Bristol 
2. Deputy Managing Editor, Cochrane Developmental, Psychosocial and Learning Problems Group, School of Policy Studies, University of Bristol 
3. PhD candidate, School of Policy Studies, University of Bristol 
4.  Research Associate in Evidence Synthesis, Department of Population Health Science, University of Bristol 
5. Senior Research Associate in Research Synthesis, Department of Population Health Science, University of Bristol 
6. Assistant Professor of Methodology, Departamento de Psicología Básica y Metodología, Facultad de Psicología, Universidad de Murcia, Spain. 
7. Lecturer in Public Health, Department of Population Health Science, University of Bristol
8. Professor of Epidemiology, Department of Population Health Science, University of Bristol 
9. Professor of Public Health Research, Department of Population Health Science, University of Bristol 
10. Associate Professor of Youth Mental Health, Faculty of Medical and Health Sciences, University of Auckland, Auckland, New Zealand 
11. Professor of Social Research and Policy, EPPI-Centre, University College, London 
12. Professor in Statistical and Health Economic Modelling, Department of Population Health Science, University of Bristol 

Aim: To assess the comparative effectiveness of school-based interventions for prevention of anxiety and depression.  

Background: The prevalence of anxiety and depression is increasing among children and young people (CYP). There is growing interest in the primary prevention of anxiety and depression in CYP.  

Methodology: We conducted a systematic review and network meta-analysis (NMA) of school-based interventions to prevent anxiety and depression in CYP aged 4-18. Searches were conducted in April 2018 and only randomised controlled trials were eligible. NMA enables the simultaneous comparison of multiple interventions in a single evidence synthesis, retaining the distinct identity of each intervention. NMA enables interventions to be ranked according to the probability that each is best for a given outcome. Random effects NMA were conducted in OpenBUGS, using standardised mean differences (SMD), for post-intervention self-report anxiety and depression. Separate analyses were run for universal and targeted populations, in primary and secondary school settings. 

Results: 137 studies met our inclusion criteria, of which 20 were at low risk of bias. In primary settings there was no evidence that universal or targeted interventions were effective for reducing depression or anxiety. In universal secondary settings, there was weak evidence for CBT (SMD −0·07, [95% Credible Interval −0·23 to 0·05]) and strong evidence that mindfulness-based interventions (−0·65, −1·14 to −0·19) reduced anxiety. This reduction was not observed for depression symptoms. In targeted secondary settings, exercise appears effective for reducing anxiety (−0·47, [−0·86 to −0·09]) and third-wave interventions (−3·74, [−4·90 to −2·59]) appear effective for depression. Funnel plots suggested evidence of publication bias for the anxiety outcome.  

Conclusions: Considering unclear/high risk of bias and likely small study effects for anxiety, we conclude there is little evidence that educational setting-based interventions are effective at preventing anxiety or depression. We discuss how these results contrast with recent standard, pairwise meta-analyses.  

 

Professor Isabel Oliver
Public Health England

The impact of flooding on mental health and wellbeing 

Isabel Oliver, Director of Research, Translation and Innovation, Public Health England (PHE), Deputy Director, National Infection Service, PHE, co-Director, National Institute of Health Research (NIHR) Health Protection Research Unit on Evaluation of Interventions at the University of Bristol 
Charles Beck, Consultant Epidemiologist, Field Service, PHE and NIHR Health Protection Research Unit on Evaluation of Interventions at the University of Bristol 
Thomas Waite, Consultant in Public Health, Extreme Events Team, Centre for Radiation, Chemicals and the Environment, Public Health England  
Charlotte Robin, Epidemiology and Information Scientist, Field Service, PHE and NIHR Health Protection Research Unit on Evaluation of Interventions at the University of Bristol 
Ranya Mulchandani, Field Epidemiology Training Fellow, Field Service, PHE 
Clare French, Senior Research Associate in Research Synthesis, NIHR Health Protection Research Unit on Evaluation of Interventions at the University of Bristol 
Flood Study Group, Public Health England with the NIHR Protection Research Units in Evaluation of Interventions, Emergency Preparedness and Response and Environmental Change and Health 

Introduction:  Flooding is common, but little is known about the longer-term effects on health and wellbeing. Public Health England, working with academic partners, established the National Study of Flooding and Health to determine the impact of flooding on mental health to inform public health action.  

Methods: We setup a longitudinal study and invited 8761 households in areas affected by flooding in 2013/2014 to participate. Participants were categorised according to exposure as flooded, disrupted by flooding or unaffected and followed up annually for three years using a questionnaire including validated instruments to screen for anxiety, depression and post-traumatic stress disorder (PTSD) and assess health-related quality of life (HRQoL). We calculated prevalence and adjusted odds ratios (aORs) for each outcome by exposure group relative to unaffected participants. 

Results: At year one, 2126 people responded (household response rate of 23%). The prevalence of psychological morbidity was elevated amongst flooded participants ([n = 622] depression 20.1%, anxiety 28.3%, PTSD 36.2%). Flooding was associated with higher odds of all outcomes (depression aOR 5.91, 95% confidence interval [CI] 3.91-10.99; anxiety aOR 6.50, CI 3.77-11.24; PTSD aOR 7.19, CI 4.33-11.93). The adverse impact persisted for three years in the flooded group compared with the unaffected (depression aOR 8.48, CI 1.04-68.97; anxiety aOR 2.68, CI 0.88-8.20; PTSD aOR 7.74, CI 2.24-26.79). People whose homes were not flooded but their lives were disrupted by flooding also had higher psychological morbidity. Similarly, median HRQoL scores were lower in flooded and disrupted groups (both 80; IQR 70-90), compared with the unaffected group (89; IQR 80-90). Factors associated with worse outcomes included flood-related factors (floodwater depth >100cm or flooding duration >24 hours), displacement, difficulties accessing services and secondary stressors (e.g. insurance issues). 

Conclusion: Interventions to prevent the impact of flooding on people’s mental health need to be developed and evaluated. 

Funding: Public Health England and the National Institute for Health Research Health Protection Research Unit in Evaluation of Interventions, Emergency Preparedness and response and Environmental Change and Health  

 

Abstracts for parallel presentation

A1
Natalia Lewis
University of Bristol

Identification and Referral to Improve Safety in community pharmacy: a theory-based intervention adaptation 

Natalia V Lewis, Research Fellow in Primary Care, Centre for Academic Primary Care, Bristol Medical School (PHS), University of Bristol 
Gene Feder, Professor of Primary Care, Centre for Academic Primary Care, Bristol Medical School (PHS), University of Bristol 
Jenny Scott, Senior Lecturer in Pharmacy Practice, Department of Pharmacy and Pharmacology, University of Bath 
Richard Brown, Chief Officer, Avon Pharmaceutical Committee 
Lucy Downes, National Implémentation Manager, Social Franchising, IRIS Interventions 
Jeremy Horwood, Associate Professor in Social Sciences and Health, Bristol Medical School (PHS), University of Bristol 

Aim: to adapt an existing evidence-based domestic violence (DV) intervention for general practice to community pharmacy setting. 

Background: Domestic violence (DV) is a global public health problem rooted in gender and social inequalities. Identification and Referral to Improve Safety (IRIS) is an evidence-based general practice programme for victims of DV, also modified for sexual health settings. IRIS integrates training and support for primary care teams and specialist DV advocacy to improve safety and health of patients affected by DV. The local Domestic Homicide Review and literature review support the case for extending IRIS to community pharmacies. 

Methodology: We applied the Medical Research Council Framework, mixed methods, and the Capability, Opportunity, Motivation – Behaviour framework. First, we synthesised evidence from systematic reviews, case-control study in Clinical Practice Research Datalink, and qualitative interviews with pharmacists (n=20). Second, we produced a prototype IRIS Pharmacy intervention. Third, we refined the prototype with pharmacy and DV stakeholders (n=13). During second stage, we established an intervention adaptation group consisting of members of the research team, pharmacy stakeholders and the IRIS delivery team. 

Findings: Pharmacists felt confident in providing public health services but lacked skills and confidence in identifying and responding to DV. Pharmacists welcomed training on DV, alongside organisation and system-level support, and raising public awareness on the role of pharmacies in multisector response to DV. Our interview study identified predictors of behaviour (asking and responding to DV). We identified key intervention functions grounded in established behaviour change techniques (education, persuasion, training, environmental restructuring, modelling and enablement) and developed a programme theory and logic model for the IRIS pharmacy intervention. IRIS pharmacy will be tested in a feasibility trial. 

Conclusions: Our systematic, theory driven adaptation of IRIS to a new setting and professional group has the potential to increase intervention acceptability, feasibility, and behaviour change. 

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

A2
Georgie MacArthur
Bristol City Council

'Rethink Alcohol': Development and testing of a digital intervention to reduce excessive alcohol use in young people aged 14-15 years. 

GJ MacArthurSpecialty Registrar in Public Health, Bristol City Council
Laura Tinner, PhD student, University of Bristol 
Siobhan Mitchell, Postdoctoral Research Associate in Child Health, University of Exeter 
Claire Garnett, Research Associate, UCL 
Eileen Kaner, Professor of Public Health and Primary Care Research, Newcastle University 
Matt Hickman, Professor of Public Health and Epidemiology, University of Bristol 
Rona Campbell, Professor Public Health Research, University of Bristol 

Background: Despite downward trends among young people, a substantial proportion report excessive alcohol use, which is associated with multiple harms to health. Digital interventions to prevent alcohol use among young people present an appealing and cost-effective mode of intervention. However, relatively few school-based digital alcohol-focused interventions have been developed and tested in the UK.  

Aim: This study sought to: (1) develop a novel digital web-based intervention, ‘Rethink Alcohol’, to reduce excessive alcohol use and related harm among young people aged 14-15, and (2) explore the views of young people, teachers and youth workers in relation to the content, design and usability of the intervention. 

Methods: Intervention development followed the person-based approach, using theories of social norms and social influence. Qualitative ‘Think-Aloud’ one-to-one or paired interviews were conducted with eighteen young people (10F: 8M), four youth workers (3F: 1M), three teachers (2F: 1M) and one clinical professional (1M), recruited via youth groups and professional networks. Interviews were audio-recorded, transcribed verbatim, and analysed thematically using NVivo 12.  

Results: The prototype web intervention, which included normative feedback, information, a quiz, interactive activities and scenarios, was well-received, with average scores of 3.6/5 (72%). Content was reported to be interesting, credible and informative and to embody a neutral and non-judgemental tone. Specifically, participants highlighted a need for stronger messaging regarding social pressures and short-term risks; the importance of credibility and trustworthiness of information; and the need to bring messages to life through design, interactive components and videos. Data also illustrated the importance and challenge of communicating nuanced alcohol-focused public health messages to young people in succinct, clear, non-judgemental and appealing ways. 

Conclusions: Young people report interest in clear, credible, neutral and interactive messages regarding social pressures and short-term risks of alcohol use. There is scope for optimisation and feasibility testing of the Rethink Alcohol intervention.  

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

A3
Sharea Ijaz
University of Bristol

Preventing childhood obesity in the UK primary schools: a realist review  

Sharea Ijaz,1 James Nobles, 1 Laura Johnson,2 Theresa Moore, 3 Jelena Savović,4 Russ Jago5  

1 Senior research associate NIHR ARC West, University of Bristol 
2 Senior Lecturer in Public Health Nutrition, University of Bristol 
3 Methodological Editor, Cochrane Methods Support Unit, University of Bristol 
4 Senior Research Fellow and Evidence Team Lead at NIHR ARC West, University of Bristol 
5 Professor of Paediatric Physical Activity & Public Health, University of Bristol 

Aim: To identify the contextual and mechanistic characteristics of obesity prevention programmes that are implementable in UK primary schools. 

Background:  Childhood obesity is increasing in the UK. A key setting for action is primary school where children spend approximately 25% of their waking hours. A recent Cochrane review[1] found that school-based preventative interventions, can achieve small desirable changes in BMIz scores over a school year. However, as approaches to intervention varied widely in design and degree of success, we still do not know what features work best, for whom, and in what contexts. Our review aimed to identify the features of obesity prevention programmes that could be successfully implemented in UK primary schools. 

Methodology: We undertook a realist review following RAMESES guidance [2] to identify key context-mechanism-outcome (CMO) configurations associated with obesity prevention programmes. We first developed a working programme theory (an iteratively-developed conceptual diagram, which describes the various mechanistic and contextual pathways thought to prevent obesity) based on existing literature. We obtained UK-based school-staff’s input on important features of preventive programmes to inform our programme theory. We identified relevant studies from the recent Cochrane review[1]. Evidence from these studies on CMOs were added to evolve the program theory.  

Results: 24 studies (79 documents) were included. Of 30 potential contextual factors in the program theory, only 11 were reported across all studies and only 3 (baseline BMI status, gender and parental education) impacted intervention effect in 2 or more studies. Mechanisms hypothesized most often were alignment with curriculum (15 studies), education (13), staff upskilling (10), reinforcements (7). However, the mechanism most implicated in changing BMIz was intervention dose (10).  

Conclusions: Intervention effects are impacted by child’s BMI, gender and parental education level. Appropriate intervention dose is a feature of the intervention that often makes a difference to the outcome. 

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

A4
Laura Tinner (presented by Rona Campbell)
University of Bristol

Perceptions of adolescent health risk behaviour and socioeconomic position: A grounded theory study of young adults in Bristol. 

Laura Tinner1, Deborah Caldwell2, Matthew Hickman3, Rona Campbell4
1PhD Student, Population Health Sciences, University of Bristol, UK 
2 Senior Lecturer in Public Health Research, Population Health Sciences, University of Bristol, UK 
3 Professor of Public Health and Epidemiology, Population Health Sciences, University of Bristol, UK 
4 Professor of Public Health Research, Population Health Sciences, University of Bristol, UK 

Aim: To explore young adults’ perceptions of adolescent health risk behaviour and socioeconomic position.  

Background: Health risk behaviours such as tobacco smoking, alcohol consumption, drug misuse, unhealthy diet and unprotected sexual intercourse are global health issues, often initiated in adolescence. There is contrasting evidence on the association between health risk behaviours and socioeconomic position in adolescence and young adulthood, with little qualitative evidence to illuminate the relationship. 

Methods: Nested qualitative study within a cohort study using semi-structured interviews of 25 young adults aged 26-28 in born in Bristol. A grounded theory approach was adopted. Theoretical sampling was conducted based on participants’ socioeconomic background and engagement in health risk behaviours. Data collection and analysis were undertaken iteratively to aid constant comparison.  

Findings: Categories of peer influence, family influence, “drive and determination” and personal responsibility were identified. These data underline the importance of family and peers in relation to both health risk behaviours and socioeconomic life chances. Engaging in some level of health risk behaviour allowed young people to gain experiences and bond with peers but avoiding certain risk behaviours helped construct their identity and self-security. Some participants adopted a moralistic or neoliberal perspective in labelling peers who engaged in certain risk behaviours, such as drug use, as irresponsible or immature. Adolescent alcohol use was seen as normative and therefore less tied to socioeconomic position. The study highlighted class stigma, with some young people from low socioeconomic backgrounds avoiding health risk behaviours as a strategy to evade further stigmatisation. 

Conclusions: Family and peers are central to how young people perceive health risk behaviour and socioeconomic position. Class stigma is connected to health risk behaviours for some young people. Therefore, strategies to reduce health risk behaviours should adopt measures to avoid further stigmatising young people from low socioeconomic backgrounds. 

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

B1
Rebecca Abbott
University of Exeter

Effectiveness of pharmacist home visits for individuals at risk of medication-related problems: a systematic review and meta-analysis of randomised controlled trials.

Rebecca Abbott, Senior Research Fellow, University of Exeter Medical School
Darren Moore, Lecturer, Graduate School of Education, University of Exeter
Morwenna Rogers, Information Specialist, University of Exeter Medical School
Alison Bethel, Information Specialist, University of Exeter Medical School
Ken Stein, Professor of Public Health, University of Exeter Medical School
Jo Thompson Coon, Professor of Evidence Synthesis and Health Policy, University of Exeter Medical School

Aim: To determine the effectiveness of pharmacist home visits for individuals at risk of medication-related problems.

Background: Medication mismanagement is a major cause of both hospital admission and nursing home placement of frail older adults. Medication reviews undertaken in the home may facilitate identification of medicine issues and may provide a more receptive environment in which to provide medication advice and education.

Methods: We carried out a systematic review and meta-analysis of randomised controlled trials (RCTs) assessing home visits by community pharmacists for populations at risk of medication-related problems. Thirteen databases were searched from inception to December 2018 and citation chasing of included studies was also performed. Articles were screened for inclusion independently by two reviewers. Data extraction and quality appraisal were performed by one reviewer and checked by a second. Random-effects meta-analyses were performed where sufficient data allowed and narrative synthesis summarised all remaining data.

Results: Twelve RCTs (reported in 15 articles) were eligible for inclusion. The data from eight trials were suitable for meta-analysis of the effects on hospital admissions and mortality, and from three trials for the effects on quality of life. Overall there was no evidence of reduction in hospital admissions (risk ratio (RR) of 1.01 (95%CI 0.86 to 1.20, I2 = 69.0%, p=0.89; 8 studies, 2314 participants)), or mortality (RR of 1.01 (95%CI 0.81 to 1.26, I2 = 0%, p=0.94; 8 studies, 2314 participants)). There was no consistent evidence of an effect on quality of life, medication adherence or knowledge.

Conclusion: This systematic review of twelve RCTs assessing the impact of pharmacist home visits for individuals at risk of medication-related problems found no evidence of effect on hospital admission or mortality rates, and limited evidence of effect on quality of life. Future studies should focus on using more robust methods to assess relevant outcomes.

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

B2
Frank de Vocht
University of Bristol

Evaluating place-based natural experiments yourself: with examples from alcohol licensing decisions, health and local crime

Frank de Vocht1,2,3, Cheryl McQuire1

1 Population Health Sciences, The University of Bristol, Bristol, UK
2 NIHR School for Public Health Research
3 NIHR Applied Research Collaboration West (ARC West)

Background: Natural experiments offer the opportunity to evaluate the impact of interventions using routinely collected data in a wide variety of public health settings. However, it is not always straightforward to recognise when natural experiments occur, how to analyse them, and what inferences can be made. Bringing in external expertise can help, but academically rigorous evaluations cost time and money. In many situations a relatively precise estimate of impact may be sufficient to guide policy.

Aim: This presentation aims to discuss natural experiments and provide an easy-to-implement statistical method to evaluate place-based natural experiments using freely available software; using case studies of alcohol licensing decisions as examples.

Methodology: The evaluation of the effect of the closure of a nightclub and the closure of a restaurant/nightclub following reviews on local health and crime using timeseries data of emergency department admissions, ambulance call-outs, and crimes will be discussed. Subsequently, it will be illustrated how, using the R statistical software environment and the CausalImpact package, these can be evaluated in a straightforward manner using programme defaults.

Results: The case studies show how closure of the nightclub was associated with temporary reductions in anti-social behaviour (-18%) and no effects on other outcomes, while closure of the restaurant/nightclub was not associated with measurable changes in outcomes. This will be discussed in light of local knowledge from practitioners. A short “guide” will be provided for attendees to facilitate evaluation of such natural experiments of sufficient precision in their own local environments and areas of specialty.

Conclusion: It is possible to evaluate the impact of small-scale local public health interventions using the provided framework. Using programme defaults, estimates of impacts of these place-based interventions of sufficient precision to guide policy can be obtained. A short guide for this will be provided.

Funding: This work was funded by the National Institute for Health Research School for Public Health Research (NIHR SPHR). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

B3
Russ Jago
University of Bristol

Change in children’s physical activity from age 6 to 11: implications for prevention 

Jago R1, Salway R2, Emm-Collison L3, Sebire SJ4, Janice L. Thompsonand Deborah A. Lawlor 3, 4 

Professor of Peadiatric Physical Activity & Public Health, Centre for Exercise, Nutrition & Health Sciences, School for Policy Studies, University of Bristol
2Senior Research Associate (Statistics), Centre for Exercise, Nutrition & Health Sciences, School for Policy Studies, University of Bristol 
3Senior Research Associate (Physical Activity), Centre for Exercise, Nutrition & Health Sciences, School for Policy Studies, University of Bristol 
4Senior Lecturer in Physical Activity & Public Health, Centre for Exercise, Nutrition & Health Sciences, School for Policy Studies, University of Bristol 
5Professor of Public Health Nutrition, School of Sport, Exercise and Rehabilitation Sciences, University of Birmingham 
Professor of Epidemiology, Population Health Sciences, Bristol Medical School 

Aim: To examine the association of body mass index (BMI) with change in children’s physical activity and sedentary time between ages 6 and 11.  

Participants: A total of 2132 children participated from 57 schools in Southwest England, from the B-PROACT1V study. 

Methods: Mean minutes of moderate-to-vigorous-intensity physical activity (MVPA) and sedentary time per day were derived from accelerometer-based measurements at ages 6, 9 and 11. Linear multilevel models examined the association of BMI categories with MVPA and sedentary time between 6 and 11, adjusting for seasonality, wear time, gender and household education. Differences in change over time were examined using interaction terms.   

Results:  Average weekday MVPA decreased between ages 6 and 11 by 2.2 min/day/year (95% CI: 1.9 to 2.5), with a steeper decline at weekends. Average sedentary time increased at a rate of 12.9 min/day/year (95% CI: 12.2 to 13.6). There were no differences in mean levels of MVPA by BMI categories at age 6, but differences emerged as children aged, with the gap between children who were healthy weight and overweight increasing by 1.7 min/day (95% CI: 0.8 to 2.6) every year, and between healthy and obese by 2.0 min/day (95% CI: 0.9 to 3.1) each year. Children who were overweight/obese engaged in less average weekday sedentary time at age 6 than those of healthy weight, but the gap closed by age 11. 

Conclusion: MVPA declines and sedentary time increases on average for all children between ages 6 and 11. While there are no differences in activity levels by BMI category at age 6, differences in MVPA emerge over time for those who are overweight and obese. Developing interventions that support children to retain activity levels as they approach older childhood, particularly those who are overweight/obese could improve public health is critical for prevention efforts.  

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

B4
David Troy
University of Bristol

An exploration of whether childhood psychopathology mediate associations between adverse childhood experiences and multiple risk behaviours in adolescence: analysis using the ALSPAC birth cohort 

Dr David Troy1, Dr Abigail Russell2, Dr Caroline Wright2 
1Research Associate, Bristol Medical School, University of Bristol, UK. 
2Senior Research Associate, Bristol Medical School, University of Bristol, UK. 

Aim: To explore whether childhood psychopathology (at 14 years old) mediates associations between adverse childhood experiences (ACE) before 9 years old and multiple risk behaviours (MRB) at age 16 years. 

Background: Childhood adversities and adolescent MRBs are both associated with several of the leading causes of death in adults including chronic health conditions, morbidity and premature mortality. It is therefore important to understand the relationship between ACEs and MRBs and what could mediate this association.  

Methodology: Data from the Avon Longitudinal Study of Parents and Children (ALSPAC), were used to explore pathways from childhood to adolescence (N=1348). Using structural equation models, we explored the associations between ACEs before 9 years old and MRBs at age 16 years. We also explored potential mediating pathways through dimensional psychopathology measured by the Strength and Difficulties Questionnaire subscales at age 14 years.  

Results: There were strong positive associations between ACEs experienced before 9 years old (beta=0.19, 95%CI=[0.11,0.28], p<0.001) and MRB at age 16 years suggesting that each additional ACE is associated with a 0.19 increase in MRBs. We also found evidence of mediating pathways from ACEs through conduct problems (beta=0.03, 95%CI=[0.01,0.05], p=0.007), hyperactivity/inattention (beta=0.03, 95%CI=[0.01,0.05], p=0.002) and peer relationship problems (beta=-0.03, 95%CI=[-0.05,-0.01], p=0.002) and MRBs. There was no evidence that emotional symptoms (beta=-0.01, 95%CI=[-0.02,0.01], p=0.315] or prosocial behaviour (beta=0.001, 95%CI=[-0.005,0.008], p=0.764] mediated the association between number of ACEs and MRBs. There was a strong positive association of ACEs and all five mediators (beta=0.21, 95%CI=[0.12,0.30], p<0.001) and MRBs. 

Conclusions: There were strong positive associations between ACEs and adolescent MRBs. There were distinct mediating pathways through conduct problems, hyperactivity/inattention and peer relationship problems in mid-adolescence. These results suggest that interventions aimed at reducing adolescent MRBs should target conduct problems and symptoms of hyperactivity and inattention in those exposed to multiple childhood adversities.  

Key words: Adverse childhood experiences, multiple risk behaviours, adolescence, psychosocial attributes, mediation analysis, structural equation modelling, ALSPAC, UK birth cohort study. 

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

C1
Charlotte Bigland
University of West of England

Systems leadership in Practice: Case Studies from UK Public Health

Bigland, C
Specialty Registrar in Public Health

Aim: This study mapped and detailed systems leadership case-studies in UK public health. Specific objectives were to: (i) provide empirical evidence on the nature of systems leadership in UK public health; (ii) provide evidence-informed guidance for the upcoming FPH specialty training curriculum refresh; and (iii) signpost public health registrars in seeking systems leadership learning opportunities.

Background: There is a recognition of the need to develop systems leadership in UK public health. However, despite interest in and discussion of systems leadership in public health, there is a lack of UK-focused empirical evidence documenting and analysing systems leadership in real-world public health practice. This study builds on a recent scoping report by Bolden et al. (2019)1 on developing systems leadership for public health registrars and new consultants.

Methodology: Using a case-study approach, three cases of systems leadership were identified in the three domains of public health respectively. A total of 27 semi-structured interviews were conducted across the three case-studies. These included a mix of public health professionals and professionals from partner organisations in the case study systems. Data were thematically analysed to identify the components of systems leadership in each case and its impact.

Results:The case studies provided empirical public health examples of what good systems leadership looks like in each of the domains of public health. Thematic analysis provided insight into the ways people, cultures and structures interact in systems, and explored the paradoxes around power, system disruption, conflict, uncertainty and temporal outlooks inherent in all systems leadership experiences. The analysis also explored how these themes change during the different phases of systems working, and identified indicators of successful system leadership.

Conclusions:This study provides an empirically derived systems leadership resource which public health professionals of all levels can use to optimise systems leadership practices in all phases of systems work.

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

C2
Ruth Sampson
Bath and North East Somerset Council

Evaluating Offer of E-cigarettes for Smoking Cessation in Pregnancy 

Ruth Sampson & Janani Arulrajah 
Health Improvement Officer, Public Health Department, Bath & North East Somerset Council Public Health Registrar 

Aim: To explore the experiences of pregnant women receiving free e-cigarettes as an addition to their usual smoking cessation treatment using a qualitative approach in order to inform evidence-based practice in this novel area of health improvement. 

Background: 10.6% of women in England are smokers at the time of delivery. Searches of existing literature reveal a lack of published evidence concerning the role of e-cigarettes in improved engagement of pregnant women with smoking cessation. The Health in Pregnancy service in Bath and North East Somerset (B&NES) piloted the provision of e-cigarettes to pregnant women interested in quitting smoking from October 2018-October 2019.  

Methodology: A qualitative study exploring the experience of service users provided with e-cigarettes during a quit attempt with the B&NES Smoking in Pregnancy Service. 15 women who had a recorded, completed quit attempt during the pilot period were selected at random and invited to participate in a research interview. An interview guide was formulated in collaboration with the University of Bath. Transcripts are analysed thematically by two members of the research team using methods of coding and constant comparison to identify and inter-relate emerging themes.  Approval was gained from the NHS Ethics Committee. 

Results: Preliminary results show that the offer of an e-cigarette has helped motivate women to better engage. Emerging themes also include the impact of e-cigarettes on household smoking behaviours, the pros and cons of the device used and the benefits of receiving the device directly from the HIPs team.  The full data analysis will be complete by Spring 2020. 

Conclusions: This study provides valuable qualitative data from an initial cohort in an e-cigarette provision pilot giving insight into the value and acceptability of e-cigarettes in health improvement for pregnant women. Such evidence is necessary to inform policy formulation, service development and strategic commissioning across smoking cessation services. 

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

C3
Kaiseree Dias
University of Bristol

International Comparison of the Levels and Potential Correlates of Objectively Measured Sedentary Time and Physical Activity among Three-to-Four-Year-Old Children 

Dias K,  PhD Student, University of Bristol 
White, J Reader, Deputy Director in Population Health Trials, Cardiff University 
Jago, R Professor of Paediatric Physical Activity and Public Health, University of Bristol 
Cardon, G  Professor, Ghent University 
Davey, R  Professor of Health Research, University of Canberra 
Janz, K  Professor Emeritus, University of Iowa 
Pate, R  Professor, Director of Children's Physical Activity Research Group, University of South Carolina 
Puder, J  Senior Physician, Lausanne University Hospital 
Reilly, J Professor of Physical Activity and Public Health Science, University of Strathclyde 
Kipping, R Consultant Senior Lecturer in Public Health, Programme Director for the MSc in Public Health, Honorary Consultant in Public Health, Specialty Tutor for Academic Public Health Specialty Registrar Training, University of Bristol 

Aim: The study aimed to determine the levels and potential correlates of sedentary time (ST), total physical activity (TPA), and moderate-to-vigorous physical activity (MVPA) in preschool-aged children. 

Background: Physical activity patterns track from childhood through to adulthood, making preschoolers a key population to target for public health interventions. Few studies have been conducted looking at the number of under-fives who are meeting current UK and WHO physical activity guidelines. It has not previously been possible to draw international comparisons of ST and PA levels due to differences in study accelerometry data processing methods. Little is known about the associations between potential correlates with ST and PA.  

Methodology: Cross-sectional analyses of 1052 children aged three-to-four-years-old from six studies included in the International Children’s Accelerometry Database. Multilevel linear regression models adjusting for age, gender, season, minutes of wear time and study clustering effects were used to estimate associations between age, gender, country, season, ethnicity, parental education, day of the week, time of sunrise, time of sunset and hours of daylight and the daily minutes spent in ST, TPA and MVPA. 

Results: Across the UK, Switzerland, Belgium and the USA, children in our analysis sample spent 490 minutes in ST per day and 30.0% and 21.2% of children did not engage in recommended daily TPA (≥180 min) and MVPA (≥60 min) guidelines respectively. There was evidence for an association between all 10 potential correlates analysed and at least one of the outcome variables; average daily minutes spent in ST, TPA and/or MVPA. 

Conclusions: There is a need for public health interventions in high income countries to decrease ST and increase PA in preschool-aged children. The correlates identified in this study can be explored further to help inform the design of such interventions.  

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

C4
Laura Tinner
University of Bristol

Is adolescent multiple risk behaviour associated with socioeconomic status in young adulthood and is it moderated by early life socioeconomic status? Findings from two UK birth cohorts.  

Laura Tinner1, Caroline Wright2, Jon Heron3, Deborah Caldwell4, Matthew Hickman5, Rona Campbell6.  

1PhD Student, Population Health Sciences, University of Bristol, UK 
2 Senior Research Associate, Population Health Sciences, University of Bristol, UK 
3 Senior Research Fellow, Population Health Sciences, University of Bristol, UK 
4 Senior Lecturer in Public Health Research, Population Health Sciences, University of Bristol, UK 
5 Professor of Professor in Public Health and Epidemiology, Population Health Sciences, University of Bristol, UK 
6 Professor of Public Health Research, Population Health Sciences, University of Bristol, UK

Aim: To examine the association between adolescent multiple risk behaviour (MRB) and young adult socioeconomic status (SES) and to assess whether young people from low SES backgrounds suffer greater negative impact.  

Background: Multiple risk behaviour (MRB) refers to the occurrence of two or more risk behaviours. MRB is associated with multiple negative outcomes such as police arrests, unemployment and premature mortality and morbidity. What is unknown is whether MRB is associated with SES in young adulthood. 

Methods: Prospective cohort studies; British Cohort Study 1970 (BCS70) and Avon Longitudinal Study of Parents and Children (ALSPAC), born in 1991-1992, were used with comparable MRB and SES variables. Logistic regression was used to determine the strength of association between MRB and young adult SES. The moderating effect of early life SES variables was assessed using logistic regression models with and without interaction parameters. Evidence to support moderation was determined by likelihood ratio tests ≤p=0·05. Multiple imputation was used to account for missing data. 

Findings: There was a negative association between adolescent MRB and university degree attainment in young adulthood in both BCS70 (Adjusted OR 0·81, 95% CI: 0·76, 0·86) and ALSPAC (Adjusted OR 0·85, 95% CI: 0·82, 0·88), and occupational status at age 34 in BCS70 (Adjusted OR 0·86 95% CI: 0·82, 0·90). Each additional risk behaviour resulted in reduced odds of attaining a university degree. In BCS70, there was evidence that maternal education (p=0·03), parental occupational status (p=0·009) and household income (p=0·03) moderated the effect of adolescent MRB on young adult SES. No evidence of moderation was found in the ALSPAC cohort. 

Conclusions: Intervening on adolescent MRB could improve later SES status and thus affect health outcomes and inequalities later in life. Proportionate universalism that addresses the whole population with disproportionate attention on disadvantaged groups will need to underpin any such interventions. 

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

Mini-Oral presentations – sessions D1-D4 – please see Mini-Oral list

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

E1
Elizabeth Smout
South Gloucestershire Council

Estimating blood-borne virus prevalence and linkage to care through routine, opt-out emergency department testing 

Smout E, Parker L, Rezai R, Phyu K, Evans A, McLaren J, Kelly S, Ruf M, Aldersley M, Page EE, Hughes GJ. 

Elizabeth Smout, Public Health Registrar, South Gloucestershire Council; UK Field Epidemiology Training Programme, Public Health England 
Lee Parker, Service Manager, Leeds Teaching Hospitals NHS Trust 
Roozbeh Rezai, Leeds Teaching Hospitals NHS Trust 
Khine Phyu, Specialty Trainee in Genitourinary Medicine, Leeds Teaching Hospitals NHS Trust 
Amy Evans, Consultant in Genitourinary Medicine, Leeds Teaching Hospitals NHS Trust 
Joscelyne McLaren, BBV Testing Nurse Specialist, Leeds Teaching Hospitals NHS Trust 
Sarah Kelly, Hepatitis Nurse Specialist, Leeds Teaching Hospitals NHS Trust 
Murad Ruf, Director of Medical Affairs, Gilead Sciences Ltd. 
Mark Aldersley, Consultant Hepatologist, Leeds Teaching Hospitals NHS Trust 
Emma Page, Consultant in Sexual Health & Virology, Leeds Teaching Hospitals NHS Trust 
Gareth Hughes, Consultant Epidemiologist, Public Health England 

Aim: To evaluate the introduction of integrated, routine, opt-out testing for HIV, hepatitis B virus (HBV) and hepatitis C virus (HCV) infections in hospital emergency departments (EDs) at Leeds Teaching Hospitals NHS Trust, UK. 

Background: Routine testing for blood-borne viruses (BBVs) within EDs has the potential to provide access to testing for individuals who do not routinely access health services. This could enable earlier diagnosis and access to treatment and care, supporting progress towards elimination targets and improving health outcomes. 

Methodology: From October 2018, HIV antibody, HBV surface antigen (HBsAg) and HCV (HCV antibody [anti-HCV] and HCV-RNA) testing were offered to all ED attendees aged 16–65 years who had a blood test for urea and electrolytes as part of their routine care. Testing uptake, BBV seropositivity and linkage to care were calculated for a 9-month testing period. 

Results: Testing uptake was 57.0% (16,053/28,178). We identified 70 HIV, 345 HCV-Ab, 157 HCV-RNA and 72 HBsAg-positive ED attendees. Overall HIV seroprevalence was 0.4%; HCV-Ab 2.1%; HCV-RNA 1.0% and HBsAg 0.4%. 95.5% (150/157) of HCV-RNA positive, 24.3% (17/70) of HIV positive, and 52.8% (38/72) of HBsAg-positive individuals required linkage to care (newly diagnosed or previously diagnosed but lost to follow-up). Of those eligible for linkage, 82% (9/11) HIV-positive, 53% (36/68) of HCV-RNA positive and 81% (13/16) HBsAg-positive individuals had been linked to care at the time of analysis.  

Conclusions: Routine BBV testing within the ED was feasible and effective, with BBV seroprevalence higher than population estimates for Leeds. We recommend implementing BBV testing as part of routine care within ED settings in areas with a high BBV prevalence as a cost-effective method of detecting undiagnosed infections. Increased focus is required to ensure those testing positive are supported with community-based treatment services to improve engagement with care and health outcomes.  

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

E2
Helen Fay
University of Bristol

Qualitative insights into the Papageno Effect: Can internet content deter the uptake of suicide methods? 

Fay H.1, Derges J.2, Goldsmith, C. 3, Gunnell D. 1, Biddle L. 1 

1 National Institute of Health Research Bristol Biomedical Research Centre, University Hospitals Bristol NHS Foundation Trust and University of Bristol, Bristol, UK; Bristol Medical School, University of Bristol, Bristol, UK; Samaritans, Surrey, UK  
 
Aim: To explore instances in which online depictions of suicide methods can deter their uptake. 

Introduction: Choice of method is a key determinant in the lethality of a suicide attempt. It has been established that the media can increase method specific suicide rates. In an era of increasing internet usage, there are concerns that online information on suicide methods can encourage suicide. Contrary to this narrative, we explored instances in which internet content discouraged individuals from employing suicide methods. 

Methodology: We conducted a thematic analysis of interviews undertaken between 2014 and 2016 exploring suicide-related internet use in three United Kingdom samples: young persons (n=13); adults in the community (n=20); and self-harm patients (n=20). Transcripts were coded in accordance with the interpretive tradition to identify the key themes associated with the media deterring the uptake of specific suicide methods. 

Results: Drawing upon dual process models, two decision making pathways – rational and emotional – were identified by which internet content could discourage method uptake.  Rationally, online information created negative perceptions of characteristics associated with suicide methods. Emotionally, graphic depictions could instigate a negative emotional response (e.g., disgust) resulting in an immediate change in suicidal cognitions (e.g., method rejection). Negative characterisations of suicide methods could counter positive misperceptions previously held of the method. 

Conclusions/Limitations:  Whilst our findings are preliminary, they provide novel theoretical perspectives on the Papageno effect – the suicide protective influence of the media – and the potential of the internet within suicide prevention. If corroborated by further experimental research, future suicide prevention strategies could utilise the media to reduce the cognitive availability of high-lethality suicide methods.  

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

E3
Noreen Orr
University of Exeter

The impact of visiting or resident animals on the health and wellbeing of residents in care homes: a systematic review of the qualitative and quantitative evidence (PETROC)  

Noreen Orr, Research Fellow, University of Exeter Medical School 
Rebecca Abbott, Senior Research Fellow, University of Exeter Medical School 
Rebecca Whear, Research Fellow, University of Exeter Medical School 
Alison Bethel, Information Specialist, University of Exeter Medical School 
Sarah Paviour, NIHR Pre-doctoral Fellow, University of Exeter Medical School 
Ruth Garside, Associate Professor of Evidence Synthesis, European Centre of Environment and Human Health & University of Exeter Medical School 
Jo Thompson Coon, Professor of Evidence Synthesis and Health Policy, University of Exeter Medical School 

Aim: To understand the short- and long-term impacts of visiting or resident animals on the physical and mental health, social wellbeing, and quality of life of older people living in residential/nursing care 

Background: Moving from independent living to residential/nursing care is a key transition in an older person’s life and can significantly affect an individual’s quality of life.  A homelike environment is important to residents and the presence of animals or pets may contribute to a feeling of less institutionalised living.   

Methods: We carried out a systematic review of the qualitative and quantitative evidence exploring the impact of resident or visiting animals (often termed animal assisted interventions) in care homes. We restricted quantitative evidence to randomised controlled trials (RCTs). Thirteen electronic databases were searched in March 2019 and this was supplemented with forward and backward citation chasing. All study selection, data extraction and quality appraisal were conducted independently by two reviewers. The qualitative and quantitative literature were brought together in the synthesis. 

Results: 14 qualitative studies (reported in 16 papers) and 17 RCTs (reported in 19 papers) met inclusion criteria.  Interactions with animals were described as having a positive impact on aspects of resident well-being including helping to maintain identity, offering companionship and being something to care for. The impact of animals inspiring storytelling and reminiscence were also key themes across the studies. The RCT evidence of effects on psychological well-being, behaviour, quality of life and engagement was mixed: the strongest evidence being for depression. Meta-analysis showed evidence of a reduction in depression in residents receiving animal assisted therapy compared to those receiving usual care (-0.34 [95%CI -0.73 to 0.04], p=0.08). 

Conclusion: While not necessarily for everyone, visiting or resident animals appear to have a positive impact on the wellbeing of older adults resident in care homes.   

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

E4
Samuel Coleborn
University of West of England /South Gloucestershire Council

Inequalities in health risk behaviours among sexual minority adolescents and the role of protective/risk factors: A secondary analysis of existing survey data across three West of England Local Authorities

Samuel Coleborn, Public Health Analyst, South Gloucestershire Council, Masters Student, University of West of England

Aim: Compare the prevalence of self-reported health risk behaviours (HRB) between sexual minority (SM; i.e. non-heterosexual) and non-SM adolescents in South Gloucestershire, Gloucestershire and Wiltshire (SGGW), and understand the role of protective/risk factors.

Background: Local survey data provides an opportunity to investigate HRBs and the role of protective/risk factors among SM adolescents.

Methodology: A secondary analysis of survey data (n=20,766) using Chi-square analysis and adjusted logistic regression to investigate the interaction between protective/risk factors and SM status on the likelihood of reporting higher levels of HRBs in SM (n=2452) and non-SM adolescents (11-18 years).

Findings: We found differences for all HRBs between SM and non-SM groups. Adjusted logistic regression found that SMs showed increased odds of smoking, regular alcohol consumption, self-harm, having tried illegal drugs, frequency of substance misuse and having had sexual intercourse (AORs 1.19-2.00, CIs 1.04-2.33), and lower odds of contraception use at last intercourse (AOR 0.69, CI 0.49-0.9999). SMs were found to have lower odds of reporting feeling safe at home, feeling safe in school, school enjoyment, friendships, physical activity and perceiving their school deals with bullying well (AORs 0.39-0.80, CIs 0.35-0.88) and increased odds of having been bullied weekly or more in the last year (AOR 2.2, CI 1.92-2.52).

We found protective associations between physical activity and smoking in the total sample and non-SMs (AORs 0.81-0.84, CIs 0.71-0.94) compared to SMs (AOR 1.32; CI 1.0-1.75). Similarly, the association between bullying and regular alcohol consumption was greater for SMs (AOR 2.19; CI 1.55-3.08) compared to non-SMs and the total sample (AORs 1.15-1.41, CIs 0.92-1.69).

Conclusions: There are inequalities in HRBs between SM and non-SM adolescents in SGGW. As SMs lack supportive resources, this should be addressed through interventions. Research using varied ontological perspectives is required to better understand how SMs experience protective/risk factors and the mechanisms underlying their associations with HRBs.

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

F1
Hannah Christensen
University of Bristol

The relationships between respiratory viral infections and meningococcal carriage in healthy adolescents 

1Elizabeth Oliver, 2Begonia Morales-Aza, 3Rosy Reynolds, 4Jenny Oliver, 5Berenice Armbruster, 6Harry Chappell, 7Paulina Sikora-Liszka, 8Barry Vipond, 9Peter Muir, 10Hannah Christensen11James Stuart, 12Adam Finn 

1 Research Technician, School of Cellular and Molecular Medicine, University of Bristol, Bristol, UK 
2 Senior Research Associate, School of Cellular and Molecular Medicine, University of Bristol, Bristol, UK 
3 Senior Research Associate in Medical Statistics, School of Population Health Sciences, University of Bristol, Bristol, UK 
4 Senior Research Associate (Clinical Research Manager), School of Population Health Sciences, University of Bristol, Bristol, UK 
5 Student, School of Cellular and Molecular Medicine, University of Bristol, Bristol, UK 
6 Student, School of Cellular and Molecular Medicine, University of Bristol, Bristol, UK 
7 Research Technician, School of Cellular and Molecular Medicine, University of Bristol, Bristol, UK 
8 Clinical Scientist (Head of Molecular), Public Health England Virology Laboratory, Southmead Hospital, Bristol, UK 
9 Consultant Clinical Scientist, Public Health England Virology Laboratory, Southmead Hospital, Bristol, UK 
10 Lecturer in Infectious Disease Mathematical Modelling, School of Population Health Sciences, University of Bristol, Bristol, UK 
11 Honorary Professor, School of Population Health Sciences, University of Bristol, Bristol, UK 
12 Professor of Paediatrics, School of Cellular and Molecular Medicine and School of Population Health Sciences, University of Bristol, Bristol, UK 

Aim: We explored the associations between respiratory viral infection and presence and density of meningococcus. 

Background: Investigating the relationships between viral infections and bacterial colonisation in healthy adults is challenging because the former are relatively unusual and few studies have considered the role of bacterial density. 

Methodology: Of 1813 Bristol students recruited to UKMenCar4 in 2014-2015 who had throat swabs taken, 917, including all those willing whose initial swab was meningococcus PCR+ were enrolled for up to 5 further swabs taken over the winter at monthly intervals into STGG broth and frozen until analysis. Nucleic acid extracts were subjected to quantitative PCR analysis for meningococcus (porA) and a respiratory virus panel (rhinovirus (RhV), influenza (flu) A H1 and H3 and B, parainfluenza viruses 1,2 and 3 (PF), respiratory syncytial virus (RSV), adenovirus, human metapneumovirus and enterovirus). Data were analysed using logistic regression, including exact and mixed effects models as appropriate to take clustering and meningococcal carriage in a previous sample into account. 

Results: 433/5448 samples tested positive for at least one virus (323 RhV, 46 flu, 37 RSV, 15 PF, 11 AdV) but most individual viral species were too rare for analysis. Only 5% of students had more than one positive test and only RhV was detected at more than one visit in the same student (n=22). In contrast multiple porA carriage-positive visits were seen in 67 (7.3%) of these 917 students. The co-incident detection of RhV corresponded to an approximately 3 (95%CI 1.05–9.0) fold increase in density of meningococcus (p<0.043) but no association between presence of RhV or any virus in the previous sample and meningococcal density in the following sample was found.  

Conclusions: This study provides evidence that intercurrent RhV infection may be associated with increases in meningococcal density, which could result in increased onward transmission.  

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

F2
Lizzy Winstone
University of Bristol

Association between social media screen-time and mental health in adolescents in South West England 

Lizzy Winstone, PhD student in Population Health Sciences (NIHR SPHR), University of Bristol 
Dr Judi Kidger, Lecturer in Public Health, University of Bristol 
Dr Becky Mars, Vice Chancellor Fellow, University of Bristol 

Aim: This study examined whether time spent using social media on a typical weekday and typical weekend-day was associated with depression, anxiety, self-harm and well-being in Year 9 students, using cross-sectional data from a survey in secondary schools in South West England.  

Background: Several studies have demonstrated associations between weekday social media screen-time and depression and well-being. Evidence on weekend social media screen-time and on anxiety and self-harm outcomes is lacking. 

Methodology: 2,645 respondents completed an online survey designed specifically for the study (56% girls). Self-reported social media screen-time and mental health outcomes were measured using items from other published surveys. The data were analysed using mixed effects regression models, taking account of clustering by school.

Results: Greater time spent on social media was associated with poor mental health in almost all models. Compared to 1-3hrs use, high use of social media (3hrs or more) on a weekday was associated with increased risk of depression (OR girls: 2.14 (95% CI: 1.57 to 2.92); OR boys: 2.37 (95% CI: 1.57 to 3.57)); anxiety (OR girls: 1.79 (95% CI: 1.35 to 2.36); OR boys: 1.63 (95% CI: 1.09 to 2.46)); self-harm  (OR girls: 1.87 (95% CI: 1.36 to 2.57); OR boys: 1.56 (95% CI: 0.97 to 2.51)); and with lower levels of well-being in girls (β girls: -2.87 (95% CI: -4.24 to -1.49). Associations were similar for weekend-day use.

Conclusions: The study contributes to the evidence base by extending findings regarding associations between social media use and mental health to anxiety and self-harm and by demonstrating associations for weekend-days as well as weekdays. Contrary to previous research, the study finds evidence of a moderate association between social media screen-time and poor mental health for both girls and boys. Further research should explore mechanisms underpinning these associations and investigate their direction longitudinally. 

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

F3
Judi Kidger
University of Bristol

The WISE Study: A Cluster Randomised Controlled Trial of an intervention to improve mental health training and support for secondary school teachers 

Kidger J1, Bell S2, Copeland L3, Evans R4 Fisher H2, Turner N6 on behalf of the WISE study team 
1Lecturer in Public Health, Population Health Sciences, University of Bristol 
2Senior Research Associate, Population Health Sciences, University of Bristol 
3Research Associate, DECIPHer School of Social Sciences, Cardiff University 
4Senior Lecturer, DECIPHer School of Social Sciences, Cardiff University 
6Senior Research Associate, Bristol Randomised Trials Collaboration, Population Health Sciences, University of Bristol 

Aim: This cluster randomised controlled trial examined whether an intervention that comprised i) mental health first aid (MHFA) training and ii) a confidential peer support service for secondary school teachers led to improved mental health outcomes for teachers and students. 

Background: Teachers are at risk of poor mental health yet no randomised controlled trials have attempted to addressing this problem. Improving teachers’ mental health is likely to lead to better performance at work and may have positive impacts on student mental health and academic learning.  

Methodology: 25 mainstream, non-fee paying secondary schools in SW England and SE/South Central Wales, stratified by geographical area and free school meal entitlement were randomly allocated to intervention or ‘usual practice’ control groups after collection of baseline measures. We analysed data using mixed-effects repeated measures models in the intention-to-treat population, adjusted for stratification variables, sex, and years of experience (teacher outcomes only). The primary outcome was teacher wellbeing (Warwick-Edinburgh Mental Wellbeing Scale). Secondary outcomes were teacher depression, absence, and presenteeism, and student (year 8) wellbeing, mental health difficulties, attendance, and attainment. Follow-up was done at months 12 and 24. A mixed-methods process evaluation examined whether the intervention worked as hypothesised in the logic model. Participants completed peer supporter logs and training evaluation sheets, and semi-structured focus groups and interviews were conducted and analysed thematically. 

Results: Teacher wellbeing did not differ between groups (adjusted mean difference –0·90, 95% CI –2·07 to 0·27), nor did any of the secondary outcomes (adjusted p=0·203–0·964). The intervention had high acceptability, but participants reported a lack of support from senior leadership, and a failure to create meaningful whole-school change.  

Conclusions: Interventions to improve teacher wellbeing and mental health need to ensure genuine senior leadership buy-in and specifically target help-seeking stigma and work-related stress alongside support for individuals. 

----------------------------------------------------------------------------------------------------------------------------------------------------------------------- 

F4
Sarah Webb Phillips
South Gloucestershire Council

Trends in mortality and health outcomes by local area deprivation.  What are the patterns observed in health inequalities over time?  

Webb Phillips, S Senior Public Health Intelligence Analyst, South Gloucestershire Council 

Aim: This study investigated differences in trends over time in several health indicators by local area deprivation and to what extent this has affected the inequalities gap. 

Background: As we strive to reduce the inequalities gap, it is important to know both the current inequalities picture and its historical trajectory. Whilst trends at a local authority or CCG level are available, there is no suite of indicators to show within-area inequality over time. 

Methodology: Data spanning 2007/8-2017/18 were obtained from Hospital Episodes Statistics, Primary Care Mortality and National Child Measurement Programme. Local deprivation quintiles based on the indices of multiple deprivation 2015 were applied to examine indicators including hospital admissions, mortality, life expectancy and childhood excess weight.  Data were presented as three year pooled rates to smooth annual fluctuations, prevent disclosure and improve statistical certainty. 

Results: Significant inequalities existed in 24 of 25 indicators. Life expectancy at birth was found to be lower for the most compared to the least deprived fifth of the South Gloucestershire (SG) population, with a 4.8 years gap. Indicators with the greatest gaps include years of life lost, mortality considered preventable, intentional self-harm and dental extraction due to decay in under-5s. Time-trend analysis revealed that the gap between most and least deprived increased in 17 indicators, remained stable in three, and decreased in five.  While we see reducing gaps in some key overarching indicators, many show signs of stalling progress or increasing inequalities between the most and least deprived. 

Conclusions: These findings are unlikely to be unique to SG as recent research found similar patterns for life expectancy nationally.  Public health professionals and commissioners should be aware of the degree to which inequalities exist in their local area, how they may be evolving and that they are often concealed within broader trend data.