key: cord-0323617-sul5sry0 authors: Kalita, N.; Cooper, K.; Baird, J.; Woods-Townsend, K.; Godfrey, K. M.; Cooper, C.; Inskip, H.; Barker, M.; Lord, J. title: Do dietary and physical activity interventions in adolescents provide a cost-effective use of healthcare resources? Model development and illustration based on the Engaging Adolescents in Changing Behaviour (EACH-B) programme date: 2021-05-08 journal: nan DOI: 10.1101/2021.05.05.21256672 sha: ff6446d26bdd4c3c867ddcf83b0cb90e5e8e463a doc_id: 323617 cord_uid: sul5sry0 Objective: To assess costs, health outcomes and cost-effectiveness of interventions that aim to improve quality of diet and level of physical activity in adolescents. Design: A Markov model was developed to assess four potential benefits of healthy behaviour for adolescents: better mental health, Type 2 diabetes, higher earnings and reduced incidences of adverse pregnancy outcomes. The model parameters were informed by published literature. The analysis took a societal perspective over a 20-year period. One-way and probabilistic sensitivity analyses were conducted. Setting: Secondary schools. Participants: A hypothetical cohort of adolescents aged 12-13 years. Interventions: An exemplar school-based, multi-component intervention that was developed by the Engaging Adolescents for Changing Behaviour programme, compared with usual schooling. Primary and secondary outcome measures: Incremental cost-effectiveness ratio as measured by cost per quality-adjusted life year (QALY) gained. Results The model suggested that an intervention for improving diet and physical activity has the potential to offer a cost-effective use of healthcare resources for adolescents in the UK at a willingness-to-pay threshold of GBP 20,000 per QALY. The key model drivers are the intervention effect on levels of physical activity, quality of life gain for high levels of physical activity, the duration of the intervention effects and the period over which effects wane. Conclusions: The model focused on short to medium-term benefits of healthy eating and physical activity exploiting the strong evidence base that exists for this age group. Other benefits in later life, such as reduced cardiovascular risk, are more sensitive to assumptions about the persistence of behavioural change and discounting. (EACH-B) was designed to develop and test an intervention to encourage UK-based school 128 students, aged 12-13 years, to adopt healthy behaviours such as eating better and exercising 129 more (Trial registration: ISRCTN 74109264). EACH-B involves a cluster randomised 130 controlled trial as a test of intervention effectiveness. Further details of the trial design are 131 given elsewhere. 20 The research programme is funded by the National Institute for Health 132 Research (RP-PG-0216-20004). The "LifeLab Plus" intervention developed as part of this 133 programme is a complex three-part programme that comprises: i) an education module that 134 teaches school students the science behind health messages through a 2-week module with a 135 "hands-on" practical one-day visit to a teaching laboratory at University Hospital 136 Southampton or in school while COVID-19 restrictions apply; ii) training for teachers in 137 skills to support behaviour change; and iii) access to a specially-designed, interactive 138 smartphone app with game features. Although there is good epidemiological evidence of long-term tracking of health behaviour, 5 9 151 11 12 school-based trials rarely follow up for more than a year. 1 21 22 The persistence of 152 intervention effects is therefore uncertain. We take a conservative approach and focus on 153 potential impacts of improved diet and physical activity likely to manifest in the short to 154 medium term: up to a maximum time horizon of 20 years. We also explore alternative 155 assumptions about the persistence of effects on behaviour after trial follow-up. 156 physical activity compared with usual schooling for a cohort of adolescents. The model 161 focussed on four potential short to medium-term benefits of healthy eating and physical 162 activity in this age group: better mental health outcomes, higher earnings and reduced 163 incidences of adverse pregnancy outcomes and Type 2 diabetes. 164 165 The model assumed that improved diet quality and increased physical activity would impact 166 these four health outcomes via reduction in BMI. Discussion with key project stakeholders 167 reiterated these four benefits as the most relevant obesity-related effects in this population. 168 The model did not include later life cardiovascular disease or other chronic diseases as 169 outcomes since the likely impact on these of an intervention undertaken as an adolescent was 170 uncertain. The model also investigated independent effects of physical activity on diabetes 171 and depression (i.e. direct impacts not mediated by BMI). Information relating to 172 epidemiology, mortality, effectiveness, health-related quality of life and costs was obtained 173 from a variety of sources and used to inform the model parameters and assumptions. See Death. Outcomes associated with mental health, loss of earnings and adverse pregnancy 185 outcome were incorporated as model events. Mental health encompasses a wide spectrum of 186 conditions. Therefore, a pragmatic approach was adopted to include the two most relevant 187 mental health events for adolescents: clinical depression (henceforth, referred as depression) 188 and General Anxiety Disorder (GAD). These events were categorised as: chronic (history of 189 persistent mental illness), intermittent (experiencing intermittent episodes), and new onset (a 190 one-time episode). Adverse pregnancy outcome was defined by pre-term delivery categorised 191 as: extremely pre-term (delivery <28 weeks); very pre-term (delivery <33 weeks); and 192 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. England, 2015. 24 The incidence of Type 2 diabetes in UK was based on an analysis of 229 longitudinal electronic health records in the Health Improvement Network (THIN) primary 230 care database. 25 The prevalence of depressive episodes and GAD was taken from the Adult 231 Psychiatric Morbidity Survey 2014 26 and from Mental Health of Children and Young people 232 in England 2017. 27 The proportion in each category was assumed as follows: 17% had a 233 chronic and 40% had a fluctuating (intermittent) course, while 43% remitted (new one-time 234 episode). 28 Those individuals with depression or anxiety are at higher risk of suicide than the 235 general population. 29 The excess death rate for those with depression and anxiety was 236 calculated by using the suicide rate in the UK from Office of National Statistics (ONS) 2017 237 and applying a relative risk of 10.9 for depression and anxiety. 29 The proportions of pre-term 238 deliveries, obtained from ONS 2017 data, were assumed as follows: 0.5% of total births as 239 extremely pre-term (<28 weeks), 1.2% very pre-term birth (28 to <33 weeks) and 6.3% 240 moderately pre-term birth (33-36 weeks) respectively. 30 241 242 The economic model assumed a positive correlation between increased BMI and the risks of -244 Type 2 diabetes, depression and GAD, pre-term delivery and loss of earnings. We fitted 245 equations to the BMI relative risks. Hazard ratios, obtained from the Medicare Current 246 Beneficiary Survey 1991-2010, were used to estimate the increased risk of individuals with 247 higher BMI developing Type 2 diabetes. 31 The odds of depression and GAD in obese and 248 overweight adolescents compared with normal-weight adolescents were obtained from 249 Sutaria et al. 32 This systematic review included 22 observational studies published between 250 2000 and 2017, representing 143,603 children. The relative risk of pre-term birth for mothers 251 with overweight and obesity was obtained from Mcdonald et al. 33 The risk was assumed to be 252 the same for all three type of pre-term births. 253 254 The direct effect of physical activity on developing depression is modelled independent of the 256 effect of physical activity via BMI. The odds ratio of developing depression was assumed to 257 be 0.83 (95% CI 0.79-0.88) in those with high levels of physical activity compared to those 258 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 8, 2021. ; https://doi.org/10.1101/2021.05.05.21256672 doi: medRxiv preprint with lower levels. 34 Furthermore, an increase from being inactive to achieving the 259 recommended physical activity level (150 minutes of moderate-intensity aerobic activity per 260 week) was assumed to lower the risk of Type 2 diabetes incidence by 26%, after adjustment 261 for body weight. 35 The pooled odds ratio between Type 2 diabetes and risk of depression was 262 1.33 (95% CI, 1.18-1.51). 36 263 The intervention effect was based on three systematic reviews and meta-analyses that 266 estimated the overall effects of school-based obesity prevention interventions. 37 38 39 The 267 results of the meta-analyses were found to be significantly different between groups based on 268 General population mortality, adjusted for age and gender, was based on ONS 2020. 41 43 Adults with diabetes was assumed to have a disutility of 283 -0.161 where the pooled mean EQ-5D score for individuals with Type 2 diabetes was 0.67 at 284 a mean age of 60 years. 44 We estimated the disutility by comparing the mean general 285 population EQ-5D score with that for diabetes. A disutility of -0.087 is used for adolescents 286 with Type 2 diabetes based on a Swedish cohort of adolescents aged 13-18 years. 45 A utility 287 decrement of 0.188 was assumed for those with intermittent episodes of mental health 288 conditions; 46 a decrement of 0.488 for those with persistent/chronic depression; 47 a decrement 289 of 0.094 (half of the decrement for intermittent depression) was assumed for those with a new 290 episode of depression. For pre-term delivery, a mean utility decrement of 0.066 was applied 291 throughout the model time horizon. This was based on a systematic review and meta-analysis 292 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Utilities Index (HUI) Mark 2 (HUI2) or Mark 3 (HUI3) measures as their primary health 294 utility assessment method. 48 We found no evidence for quality of life loss in parents of pre-295 term babies. Therefore, we assumed that the quality of life decrement would be similar to 296 intermittent mental health condition and lasts for the first two years. 297 Costs 299 The intervention cost for our illustrative analysis was based on LifeLab Plus. Further Table 1 . 316 The current and future costs for Type 2 diabetes were sourced from Hex et al. 50 Direct health 318 costs and indirect societal and productivity costs were estimated using a top-down approach. 319 We assumed that individuals with diabetes would not incur costs for complications as these 320 are likely to affect individuals who have diabetes for a longer period. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Total societal costs for children born at 32-33 weeks and 34-36 weeks gestation from birth to 326 24 months were taken from a study that compared these costs to those for children born at 327 full-term. Costs for children born <28 weeks and 28-33 weeks were estimated, assuming that 328 they varied in the same way. 53 The cost used in the model was the mean societal cost 329 difference between these two groups. Carlo simulations of 1000 iterations were run for the PSA to assess the combined effects of 359 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Effectiveness parameters were assigned beta and lognormal distributions, utilities were 366 assumed to follow a beta distribution, and costs were assigned a gamma distribution. The horizon and intervention costs also influenced the base case results, but to a lesser extent. The 392 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. results of the PSA presented as the cost-effectiveness scatter plot showed that the simulations 393 lie in the North-East quadrant of the cost-effectiveness plane. See Figure 3 . This implies that 394 the intervention-LifeLab Plus is likely to produce health benefits at an additional cost. At a 395 willingness-to-pay threshold of £20,000 per QALY gained, the probability of LifeLab Plus 396 being cost-effective was 69% compared with usual schooling at 31% respectively. analyses showed that if the duration of the treatment effect was not sustained to this extent, 406 the intervention would be less cost-effective. See Table 2 . 407 408 Previous cost-effectiveness studies evaluating dietary and physical activity interventions for 410 adolescents have largely consisted of within trial analyses that have not considered the 411 benefits beyond the trial period. [56] [57] [58] [59] [60] The cost-effectiveness of these intervention vary between 412 cost saving of NZ$835 per child for the low intensity program 56 to £120,630 per QALY for 413 the HELP intervention. 59 However, comparison between studies is difficult because of the 414 differences in the study designs, the interventions considered and the outcomes reported. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 8, 2021. ; https://doi.org/10.1101/2021.05.05.21256672 doi: medRxiv preprint a conservative assumption that treatment benefits for adolescents from such multicomponent 426 intervention as LifeLab Plus do not persist beyond 20 years. 427 The study addresses an important public health question by examining if interventions 430 targeting healthy eating and doing more physical activities provide value for money from a 431 societal perspective. The study incorporates existing evidence on the effect of improvement 432 in adolescent health behaviours on four high prevalence short-to-medium term benefits 433 relevant to young people: improved mental health, higher earnings, improved pregnancy 434 outcomes and prevention of Type 2 diabetes. Sources for data used within our model were 435 identified from a targeted literature review. Where data were not available for adolescents, 436 we have used data from the adult population. Model structure and assumptions were informed 437 by this review and discussions with public health experts. UK-specific incidence rates were 438 approach to estimating income lost due to obesity. We have adopted a simple approach; 454 however, this is a complex interacting bi-directional system. We have not fully explored 455 whether income loss is due to obesity or whether obesity is caused by income loss through, 456 for example, unemployment. Finally, where data from adolescent age groups were 457 unavailable, those from adult populations were used to inform model parameters. 458 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 8, 2021. ; https://doi.org/10.1101/2021.05.05.21256672 doi: medRxiv preprint Cost-effectiveness of physical activity interventions in adolescents: model development and illustration using two exemplar interventions Life course epidemiology: recognising the importance of adolescence A Systematic Review of Digital Interventions for Improving the Diet and Physical Activity Behaviors of Adolescents National Diet and Nutrition Survey: Results from Years 1, 2, 3 and 4 (combined) of the Rolling Programme Engaging teenagers in improving their health behaviours and increasing their interest in science (Evaluation of LifeLab Southampton): study protocol for a cluster randomized controlled trial Health effects of adopting low greenhouse gas emission diets in the UK Longitudinal tracking of adolescent smoking, physical activity, and food choice behaviors Tracking of obesity-related behaviours from childhood to adulthood: A systematic review Developmental origins of noncommunicable disease: population and public health implications Advancing the adolescent health agenda Global Accelerated Action for the Health of Adolescents (AAHA!): guidance to support country implementation Controlling childhood obesity: A systematic review on strategies and challenges Delivering in-school interventions to improve dietary behaviours amongst 11-to 16-year-olds: A systematic review A cluster-randomised controlled trial of the LifeLab education intervention to improve health literacy in adolescents Media use and attitudes report Engaging adolescents in changing behaviour (EACH-B): a study protocol for a cluster randomised controlled trial to improve dietary quality and physical activity Methodological considerations and impact of school-based interventions on objectively measured physical activity in adolescents: a systematic review and meta-analysis Modelling in Ecomomic Evaluation: An Unavoidable Fact of Life Health Survey for England Trends in incidence, prevalence and prescribing in type 2 diabetes mellitus between 2000 and 2013 in primary care: a retrospective cohort study Adult Psychiatric Morbidity Survey The natural course and outcome of major depressive disorder in primary care: the PREDICT-NL study Risk of self-harm and suicide in people with specific psychiatric and physical disorders: comparisons between disorders using English national record linkage Relation between BMI and diabetes mellitus and its complications among US older adults Is obesity associated with depression in children? Systematic review and meta-analysis Overweight and obesity in mothers and risk of preterm birth and low birth weight infants: systematic review and meta-analyses Physical Activity and Incident Depression: A Meta-Analysis of Prospective Cohort Studies Quantifying the Association Between Physical Activity and Cardiovascular Disease and Diabetes: A Systematic Review and Meta-Analysis Diabetes increases the risk of depression: A systematic review, meta-analysis and estimates of population attributable fractions based on prospective studies A systematic review and meta-analysis of the overall effects of school-based obesity prevention interventions and effect differences by intervention components Systematic review and meta-analysis of school-based interventions to reduce body mass index A systematic review and meta-analysis of school-based interventions with health education to reduce body mass index in adolescents aged 10 to 19 years Effectiveness of after-school interventions at increasing moderate-tovigorous physical activity levels in 5-to 18-year olds: a systematic review and metaanalysis National life tables: England Mortality in people with Type 2 diabetes in the UK Estimating EQ-5D utility values for major health behavioural risk factors in England The use of the EQ-5D preference-based health status measure in adults with Type 2 diabetes mellitus Population health status based on the EQ-5D-Y-3L among adolescents in Sweden: Results by sociodemographic factors and self-reported comorbidity Measuring Health-related Quality of Life in Teens With and Without Depression Efficacy and cost-effectiveness of a specialist depression service versus usual specialist mental health care to manage persistent depression: a randomised controlled trial Preference-Based Health-Related Quality of Life Outcomes Associated with Preterm Birth: A Systematic Review and Meta-analysis Unit costs of Health and Social Care Estimating the current and future costs of Type 1 and Type 2 diabetes in the UK, including direct health costs and indirect societal and productivity costs Effectiveness and cost-effectiveness of mindfulness-based cognitive therapy compared with maintenance antidepressant treatment in the prevention of depressive relapse or recurrence (PREVENT): a randomised controlled trial Economic costs associated with moderate and late preterm birth: a prospective population-based study The cost of preterm birth throughout childhood in England and Wales The lifetime costs of overweight and obesity in childhood and adolescence: a systematic review Guide to the methods of technology appraisal Economic evaluation of a multi-disciplinary community-based intervention programme for New Zealand children and adolescents with obesity Cost and Cost-Effectiveness of Students for Nutrition and eXercise (SNaX) Modeling The Economic And Health Impact Of Increasing Children's Physical Activity In The United States Cost-effectiveness of a community-delivered multicomponent intervention compared with enhanced standard care of obese adolescents: cost-utility analysis alongside a randomised controlled trial (the HELP trial) Cost-effectiveness of a multi-component school-based physical activity intervention targeting adolescents: the 'Physical Activity 4 Everyone' cluster randomized trial