netscc id - national institute for health research

NETSCC ID 16/165/01 workHORSE

1

Title: workHORSE: working Health Outcomes Research Simulation Environment

Protocol Date: 14/09/2017

Funder NIHR-HTA project no: NETSCC ID 16/165/01

Start Date: 01-11-2017

End Date: 31-10-2019

Principal Investigator: Professor Martin Enrique O'Flaherty, Professor of Clinical Epidemiology, Department of Public Health and Policy, University of Liverpool [email protected] Co-Investigators: Dr. Ffion Lloyd-Williams, Research Fellow, Department of Public Health and Policy, University of Liverpool [email protected] Professor Rumona Dickson, Director, Liverpool Reviews and Implementation Group, Department of Health Services Research, University of Liverpool, [email protected] Prof Simon Capewell, Professor of Clinical Epidemiology, Department of Public Health and Policy, University of Liverpool [email protected] Professor Iain Buchan, Clinical Professor in Public Health Informatics, Division of Informatics, Imaging and Data Sciences, The University of Manchester [email protected]

Dr Angela Boland, Associate Director, Liverpool Reviews and Implementation Group, Department of Health Services Research, University of Liverpool [email protected]

Dr Brendan Collins, Health Economist, Department of Public Health and Policy, University of Liverpool [email protected] Dr Christodoulos Kypridemos, Research Associate, Department of Public Health and Policy, University of Liverpool [email protected] Dr Philip Couch, Research Fellow, School of Health Sciences, The University of Manchester [email protected]

1. Summary of Research Background The NHS Health Check Programme (NHS HCP) is a multifactorial ‘risk reduction’ programme offered

to adults in England aged 40-74. The Programme objective is the early identification and

management of otherwise healthy people at high risk of cardiovascular disease (CVD) and diabetes.

It is the largest, nationwide cardiovascular disease screening and prevention programme in the

world. However, studies of the NHS HCP suggest that it might perhaps be further improved by

including additional conditions and by facilitating local commissioning. This then raises the key

mailto:[email protected]










2

question: What is the potential for health and equity gains and cost effectiveness of the NHS

Health Check Programme?

Aim

This project will provide a validated open source/open access, flexible model enabling local

commissioners to quantify the potential and cost effectiveness for population health gain of the NHS

Health Check Programme, by building on the solid foundation of our existing IMPACTncd model.

Objectives

Objective 1. Further develop our proven and tested computer model to allow for developments and

changes to the NHS Health Check Programme and the diseases it addresses.

Objective 2. Update the evidence base to support model and scenario development.

Objective 3. Assess the effectiveness, equity and cost-effectiveness of the alternative strategies for

implementation of the NHS Health Check Programme, as specified in the NIHR call.

Objective 4. Recruit a diverse group of stakeholders to powerfully strengthen the user perspective to

inform desirable features of the user-friendly model and identify additional locally relevant scenarios

to test.

Objective 5. Develop a sustainability and implementation plan to deploy our user-friendly web-

based decision-support model at local level.

Research Plan

Key project activities are model development, evidence base update, effectiveness, equity and cost-

effectiveness analyses, scenario co-production with local stakeholders, implementation plan and

dissemination. We will also engage with our four lay advisers to inform study design and delivery

and ensure emerging findings and project results are appropriately disseminated to the widest

audience.

Objective 1: Model Development. We will further develop our proven and tested computer model

to allow for developments and changes to the NHS HCP and the diseases it addresses. We will

extend the IMPACTncd model (BMJ 2016), to develop a validated, fully stochastic generic

microsimulation environment, including a prototype user interface to model the implementation of

the NHS HCP at local level. The diseases included are heart disease, stroke, diabetes, and kidney

disease, plus certain types of dementia, atrial fibrillation; alcohol misuse, cancers and Chronic

Obstructive Pulmonary disease. The model will provide outputs to assess population health,

economics and equity gains (e.g. life years gained/disability burden avoided), Quality adjusted life

years gains and cost-effectiveness and absolute and relative inequalities. In addition, the model will

be able to report cases detected (hypercholesterolemia, hypertension, and previously undiagnosed

CVD conditions). The model will also support sophisticated scenario specifications, compare

different implementation approaches and be adaptable to support local commissioning.

Objective 2: Updating the evidence base to support model and scenario development. We will

update and conduct a series of focused systematic reviews (SRs) to help inform model development

and delivery. Building on the considerable, existing evidence base informing IMPACTncd, we will

focus on identifying and using existing evidence based guidelines and systematic reviews. We will

conduct new SRs using rapid scoping methods, with a research plan refined after reviewing the

workHORSE final parameter specification. This will give priority to key parameters reflecting

programme implementation (eligibility, annual coverage, uptake, risk profile, prescription and

adherence rate to treatment, brief interventions, referrals to lifestyle services and recruiting


3

practices). As new evidence appears, workHORSE will facilitate frequent updates of the evidence

base through explicit input parameters.

Objective 3: Assess and compare the alternative strategies for implementation of the NHS Health

Check Programme, as specified in the NIHR call. We will formally examine several measures to

explore the effectiveness of the Health Check Programme: cases detected, cases prevented, deaths

prevented or postponed, number needed to attend to prevent a case; life expectancy and disease-

free life expectancy; QALYs. All results will be stratified by age, gender, ethnicity and socioeconomic

position to facilitate equity impact analysis.

Health economics analysis will be based on cost utility (cost per QALY) analysis presented as an

incremental cost-effectiveness ratio compared to the best current intervention or a do-nothing

scenario. Outcomes will also be presented as net monetary benefit. WorkHORSE will conduct

extensive probabilistic sensitivity analyses (2nd order Monte Carlo) and validation.

Objective 4: Stakeholder engagement to powerfully strengthen the user perspective to inform

desirable features and scenario design. We will recruit and engage a diverse group of stakeholders

in four workshops. This will powerfully strengthen the user perspective, refine desirable features of

the user-friendly model and identify additional scenarios to test. This process will be closely

integrated with model development and evidence synthesis activities.

Objective 5: Develop a sustainability and implementation plan to deploy workHORSE as a user-

friendly web-based decision model at local levels. We will develop implementation and

sustainability options to ensure open access, facilitate the integration of the model with data at local

level, refine scalability strategies and provide recommendations for local scenario design and

implementation. This plan will be informed by insights co-produced with stakeholder and then by

piloting a secure cloud-based platform to allow local commissioners to run their own scenarios.

Important outcomes

Our project will deliver an open source, scalable model for use with local or national data, with

flexible scenario design features to accommodate and assess changes in Programme delivery and

eligibility criteria. We will provide a detailed development and implementation plan with

recommendations for deployment, sustainability and open access.

Furthermore, our study will explore and report on the potential for increased effectiveness and cost-

effectiveness of the NHS HCP for population health gain.

Outcome Dissemination We will publicise our findings via stakeholder workshops, diverse organisations, Health Check users,

appropriate websites, media, conferences and papers in academic and general journals.

2. Background and Rationale Non-communicable diseases are imposing a heavy and increasing burden on our society. 1 However,

much of this chronic disease burden is eminently preventable; it thus demands urgent attention. 2,3

Yet, how best to maximise the potential for prevention is still debated. Prevention is broadly

recognized as the most effective and cost-effective way of reducing NCD burden. However, the

specific interventions and ways of delivering them may have different degrees of effectiveness.

The NHS Health Check Programme (NHS HCP) in England represents a key programme to achieve

this prevention goal. The Programme’s objective is the early identification and management of

otherwise healthy people at high risk of cardiovascular disease (CVD) and diabetes.2 It is the largest


4

nationwide CVD screening programme in the world. To date, the programme has invited over 10

million people to participate (95% of all eligible participants thus far). The original modelling

assumed an uptake of 75%. The current uptake of approximately 48% overall, reflects significant

variations at local authority level. However, the programme has improved its performance in terms

of invitation to participate, with higher participation amongst the socially disadvantaged. It has

detected approximately 7800 new cases of hypertension (38/1000 Checks), 1930 new cases of type 2

diabetes (9/1000 Checks) and 800 new cases of chronic kidney disease (4/1000 Checks). 4

Programme costs are also substantial. There is thus an urgent need to make the programme even

more effective, efficient and equitable.

Most recent evaluations have focused on process measures and some intermediate outcomes. 4–12

However, there is no evaluation of the NHS HCP impact on disease incidence and mortality.

Furthermore, observational data and randomized clinical trials for similar programmes have

produced conflicting results, ranging from minimal to substantial efficacy and cost-effectiveness.13–

15 However, none of these evaluations are strictly comparable with the current implementation of

the Health Check Programme. They thus have a limited role in helping to determine the programme

effectiveness, cost-effectiveness and equity, particularly at local level. Neither do they allow

exploration of different implementation options.

Conducting an empirical evaluation of the entire programme might be challenging. Furthermore, it

would not provide rapid insights within the urgent timescales needed for decision cycles on

investment and changing public health priorities. The NHS is under stress from population growth,

increasing need, and challenging financial environments.

A simulation modelling approach might therefore provide rapid and useful insights to help

commissioners and planners identify which specific aspects of the programme might increase its

effectiveness, cost effectiveness and equity. Such modelling might also help to assess NHS HCP

synergy with other preventative activities happening at local and national levels.

Furthermore, given the complexity of the NHS HCP and its focus on implementation at local level, it

is imperative that any decision-support tool provides a local perspective.

However, most previous modelling approaches to assess this type of programme have been ad hoc,

short lived, proprietary, not comprehensive and not validated. Almost none have provided equity

outcome analysis.15–18 Furthermore, existing tools lack important features to realistically model the

changing population risk profile over time, and the interaction between diseases that share common

determinants but which operate on different timescales, such as CVD and cancers. These crucial

factors might have substantial implications for the overall cost-effectiveness of the programme.

Our extensive preparatory work. Towards a comprehensive environment to simulate the

NHS Health Check Programme at national and local levels.

National Evaluation

We have developed and validated IMPACTncd, a discrete time, dynamic, stochastic microsimulation

model of non-communicable diseases. IMPACTncd simulates the life course of synthetic individuals

under different counterfactual scenarios.19 During the simulation, CVD incidence and CVD and non-

CVD mortality are recorded. Furthermore, we quantified the uncertainty of model outputs using

second-order Monte Carlo simulation, and reported 95% uncertainty intervals. The results were

stratified by year, five-year age group, sex, and fifths of index of multiple deprivation. Five scenarios

were considered: baseline scenario (assuming current trends in risk factors will continue into the

future); universal screening; screening concentrated only in the most deprived areas; structural


5

population-wide intervention; and the combination of population-wide intervention and

concentrated screening.

We used this IMPACTncd model to explore the effectiveness and equity impact of different options to

reduce CVD burden, including cardiovascular screening strategies based on the NHS HCP principles.

Compared with the baseline scenario, universal screening might prevent or postpone approximately

19 000 cases and 3000 deaths; concentrated screening 17 000 cases and 2000 deaths; and the

combination of the population-wide intervention and concentrated screening 82 000 cases and 9000

deaths. The most equitable strategy would be the combination of concentrated screening plus the

population-wide intervention.19

Local Evaluation

We recently successfully adapted our microsimulation IMPACTncd model to examine the cost-

effectiveness and equity of the NHS Health Checks cardiovascular disease prevention programme

using real-world data from the deprived northern city of Liverpool. We recently presented our

analyses at Public Health England’s NHS Health Checks scientific conference. 20

We first extended IMPACTncd to tackle lung and gastric cancer, to test the model ability to incorporate

diseases with different time scales. 21,22 We then modelled the current implementation of HCP using

local and national data on effectiveness, costs, and participant risk profiles. Disease costs and health

state utilities were drawn from standard sources and discounted at 3.5% annually using a healthcare

perspective.

We modelled three scenarios representing A) continuing the current implementation of the NHS HCP;

B) an optimal implementation of HCP assuming optimal coverage, uptake, treatment and lifestyle

change; and C) combining scenario A with structural policies increasing dietary consumption of salt,

sugar, fruit and vegetables.

The model suggested that over 15 years the total benefits ranged across the three scenarios between

approximately 300 and 1,750 CVD cases prevented or postponed.

Cumulative discounted net costs for the three scenarios likewise ranged from a cost of £2m to a cost-

saving of £17m, depending on specific scenario content. The current implementation or its optimized

implementation might become cost-effective by 2031, while a combination of the current

implementation with structural policies might become cost-saving by 2030. An optimized

implementation would be neutral in terms of equity, while the combined implementation would

almost certainly reduce inequalities.

Overall, this work represents a robust proof of concept of the core modelling engine prior to

embarking on the proposed workHORSE project.


6

This proposal: building workHORSE

We will now build on this extensive preparatory work. Using the solid foundation of our existing,

validated model, IMPACTncd, we will create workHORSE (Working Health Outcomes Research

Simulation Environment). This will provide a flexible, open-source tool enabling local commissioners

to quantify the potential for population health gain and cost-effectiveness of the NHS HCP.

This decision-support tool will develop and validate additional modules to quantify benefits and

costs, both overall and for the nine specified conditions. Model outputs will be stratified by risk

groups.

Our model development will follow three strategic principles:

Openness (to foster transparency and promote continuous development of the tool by

interested stakeholders),

A solid evidence base (explicitly linking model parameters to best effectiveness evidence) and

Up-to-date (exploiting the growing availability of local health surveillance data and new

research).

3. Evidence explaining why this research is needed now The original Department of Health modelling assumed an uptake of 75% in their cost-effectiveness

calculations (Green, 2005). However, this level of uptake has not yet been achieved in the real-world

implementation of the programme.8 However, the initial modelling assumptions around specific

levels of costs, sustainability and compliance have been questioned. Each of these factors might

have considerable cumulative weight when trying to best estimate cost-effectiveness of the current

NHS HCP implementation, or when exploring alternative ways of delivering the programme in future.

Furthermore, since the original implementation of the programme, new evidence on cardiovascular

screening has emerged, and multiple empirical observations of the NHS HCP delivery have become

available. The cost-effectiveness of the concept has also been studied, and a recent systematic

review identified 14 economic evaluations (5 RCT based, 7 observational and 2 modelling studies). 15

The randomized evidence highlighted the need for sustained, long-term risk factor changes to

achieve cost-effectiveness. Most observational and modelling studies suggested that CVD screening

programmes might be cost-effective. However, some relied on assumptions on costs, uptake,

compliance, and sustainability of the therapeutic interventions, which might not be entirely

consistent with empirical observations or with the actual implementation approach of the NHS HCP. 23,24 Happily, a more recent analysis with updated primary care data suggested that modifications to

the current strategy including opportunistic case finding, and strategies prioritising and targeting

patients by age or prior estimate of cardiovascular risk might be cost-effective.25

A more recent, comprehensive review by Usher-Smith et al, commissioned by PHE has also shown

that coverage varies substantially across different groups in England, with consistent higher uptake

in older groups, women and deprived communities. Uptake also varies significantly across region,

suggesting substantial practice variations which could inform redesign of key processes of the

programme (such as invitation or uptake). Promising interventions and best practices are starting to

become available, but the evidence base is small. Importantly, referrals to lifestyle advice are

relatively low and there is little evidence on the effect of the NHS HCP intervention on subsequent

diet, physical activity or alcohol intake. Usher-Smith et al also provide suggestions for further

empirical studies. Furthermore, this review also represents a major opportunity to explore the

potential of using modelling approaches to integrate evidence from diverse sources to identify the


7

potential impact of improving different aspects of the NHS HCP processes. Possible improvements

might include further characterisation of best practices, better measurement of process

performance and outcome, programme redesign or action research. 8

Furthermore, none of the studies have looked at other diseases that could benefit from the

Programme, such as chronic kidney disease (CKD) or common cancers which could potentially be

prevented by weight reduction or smoking cessation. Formally including these additional benefits

would likely improve the overall effectiveness of the programme and increase value for money.

Furthermore, no studies have looked at the impact of the Programme on health outcomes according

to socioeconomic circumstances. Substantial debate exists about the equity of this programme, with

suggestions that individual level interventions might increase inequalities.26 This is a complex issue,

because considerations of risk concentration and differential uptake might have a significant impact

on outcomes. 19,27

Empirical comprehensive evaluations of programmes like the NHS HCP focusing on a wider range of

diseases are needed. However, such studies will be large, of long duration, complex and costly. They

would also be unlikely to reflect a specific implementation at local levels. Thus, a comprehensive, up-

to-date, flexible model could help commissioners and stakeholders to make better informed

decisions, particularly if the model is easy to update as new evidence on key parameters becomes

available.

4. Aims and objectives

Aim: This project will provide a validated open source/open access, flexible model enabling

commissioners to quantify the potential and cost-effectiveness for population health gain of the NHS

Health Check Programme, by building on the solid foundation of our existing IMPACTncd model.

Objectives:

Our two-year project will address five objectives:

Objective 1. Further develop our proven and tested computer model to allow for developments and

changes to the NHS HCP and the diseases it addresses.

Objective 2. Update the evidence base to support model and scenario development.

Objective 3. Assess the effectiveness, equity and cost-effectiveness of the alternative strategies for

implementation of the NHS Health Check Programme, as specified in the NIHR call

Objective 4. Recruit a diverse group of stakeholders to powerfully strengthen the user perspective to

inform desirable features of the user-friendly model and identify additional locally relevant scenarios to

test.

Objective 5. Develop a sustainability and implementation plan to deploy our user-friendly web-

based decision-support model at local level.


8

5. Research Plan

Our open access/open source workHORSE modelling environment will integrate the epidemiology,

social epidemiology, economics and utilization data to support commissioners and other

stakeholders to explore the effectiveness, equity and economics of local implementation of the

NHS Health Check Programme.

The backbone of workHORSE is IMPACTncd. This versatile microsimulation model will be further

extended to evaluate comprehensive, preventative interventions for cardiovascular disease (CVD),

Atrial Fibrillation (AF), Chronic Obstructive Pulmonary Disease (COPD), dementia and common

cancers. Functionality will be added to model key processes in the Health Check Programme, to

enable users to easily explore different implementation scenarios.

We will test this decision-support tool with key stakeholders and co-produce additional, locally

relevant scenarios and obtain insights to inform a development and implementation plan for optimal

dissemination and use.

Below, we structure our Research Plan as tasks to deliver the five specific objectives:

Objective 1. Further develop our proven and tested computer model to allow for developments

and changes to the NHS Health Check Programme and the diseases it addresses.

The key activity in workHORSE is the development of a working, open source, flexible

environment to enable analysis of the effectiveness, cost-effectiveness and equity of the NHS HCP.

Key tasks will be model development, scenario design features, and ability to support local

implementation and model validation.

We will extend our current work with our validated IMPACTncd, a discrete time dynamic stochastic

microsimulation model of non-communicable diseases. IMPACTncd simulates the life course of

synthetic individuals under different counterfactual scenarios. Its ongoing extensions and pilot

local implementation will further extend its current capabilities to perform health economics

analysis.19 We will further develop a fully stochastic, generic microsimulation environment, with a

prototype user interface to model the implementation of the NHS HCP at local level. The diseases to

be included in workHORSE are heart disease, stroke, diabetes, kidney disease and certain types of

dementia, atrial fibrillation, alcohol misuse, cancers and COPD. The model will provide outputs to

assess population health, economics and equity gains (e.g. life years gained/disability burden

avoided), QALY gains and cost-effectiveness and absolute and relative inequalities. In addition, the

model will be able to report cases detected (hypercholesterolemia, hypertension, and previously

undiagnosed CVD conditions).

Model development We will further develop and extend our validated computer model to allow for developments and

changes to the NHS Health Check Programme and the diseases it addresses. The model will be able

to tackle effectiveness, cost-effectiveness and equity questions at local level.

We will build on our ongoing work with IMPACTncd (Kypridemos et al BMJ 2016). In summary

IMPACTncd is a discrete-time dynamic stochastic microsimulation that is actively developed in R and

C/C++ under the GNU v3 licence (open source). It samples synthetic individuals from the synthetic

population which mirrors the English population and simulates their life courses under different

counterfactual scenarios. It currently models the age, gender, socioeconomic status (social class,

household income, Index of Multiple Deprivation), behavioural risk factors (diet, physical activity,


9

smoking), and biological risk factors (body mass index, systolic blood pressure, and total cholesterol)

to model lifetime disease outcomes.

For every simulated year, these parameters are updated to model the ageing of the synthetic

individual and the projected calendar trends. Then, IMPACTncd estimates the risk of each synthetic

individual to develop any of the modelled diseases (currently CHD, stroke, lung cancer and gastric

cancer) based on past exposures and accounting for the time lags between exposure and disease.

Subsequently, the model calculates disease incidence from the estimated disease risk, and disease

prevalence. In any simulated year, synthetic individuals are at risk of dying from any modelled

disease they develop, or any other non-modelled disease (competing risk framework).

The IMPACTncd model will be extended for workHORSE. It will include the five current NHS HCP

‘conditions’ in addition to heart disease and stroke (diabetes, kidney disease and certain types of

dementia), plus additional ‘conditions of interest’: atrial fibrillation, alcohol misuse, COPD and, in

addition, the five commonest cancers related to preventative interventions as recommended in the

International Agency for Research on Cancer, Continuous update project on Cancer Prevention And

Survival, and Parkin reports.28,29 In addition, further key risk factors will be incorporated including

alcohol intake and Hb1Ac.

Relevant outputs will include cases detected, deaths prevented, QALYs, total life expectancy, healthy

life expectancy, disability burden avoided, costs, and cost-effectiveness; all stratified by age, gender

and socioeconomic circumstances. We will present the outputs in graphic and tabular form, and raw

outputs will be available for further analysis. A crucial feature is its ability to provide robust

socioeconomic stratification (enabled by the synthetic population methodology, described in section

11 “Sampling”, page 13). This permits equity analysis by relating the impact of different

interventions on measures of inequalities and measures of health gains, in the equity plane graph.19

The model will track risk factor trends and thus be able to model any type of preventative

intervention and other key features of the NHS Health Check pathways, as detailed below. Explicit,

user modifiable inputs to reflect local costs of implementation will inform health economic and

equity analyses from a health care and societal perspective.

The simulation can be run up to 50 years in the future, and outcomes can be reported for any

intermediate years. All outcomes will be presented with 95% uncertainty intervals (see Sensitivity

and Uncertainty Analysis below, page 11)

Scenario Specification features

workHORSE will allow sophisticated scenario specification to model the different NHS Health Check

approaches through six parameters:

1. The eligibility criteria for the intervention

2. The annual coverage (proportion of the eligible population that is invited every year)

3. The uptake (proportion of those invited who attend a Health Check)

4. The risk profile of participants (which can be easily summarised using their QRISK score)

5. The prescription rate of treatment after a Health Check and subsequently the adherence

and continuity of each medication

6. The rate of referrals for brief interventions and lifestyle services and their subsequent

effectiveness.


10

This will be implemented as a Scenario specification tool. Interacting with this tool will be facilitated

by building a prototype simple user interface using sliders to represent choices for the above six

parameters and delivering basic outputs in graphical and tabular format.

For the development and implementation proposal (Objective 5), we will propose a “roadmap” to

develop a more comprehensive user interface to facilitate interacting with all key parameters in the

model, including intervention effectiveness, costs and delivery of outputs.

Local Adaptation

We will initially develop the model functionality using locally available data from Liverpool. We will

then implement the model in an additional local authority (identified early in the stakeholder

engagement process). We will then coordinate with representatives of PHE and local authorities to

further explore potential data issues at local levels. That advice will feed into the document

summarizing our recommendations for deploying the model and use across local authorities

throughout England. (See section 6 “Research Plan”, Objective 5, page 13)

Validation and modelling standards

We will follow the International Society for Pharmacology and Outcomes Research (ISPOR)

recommendations.30 We will explore face validity through our stakeholder engagement process,

conduct checks for internal validity of the mathematical equations and coding, and compare outputs

with similar models to assess and explain similarities and differences. To check external validity, we

will conduct comparisons of model outputs to empirical independent data on process and outcomes

when feasible. An example of the proposed validation programme can be obtained from

Kypridemos et al BMJ 2016. In summary, it will compare model key outputs (mortality, prevalence),

with independent sources when possible.

Furthermore, we fully agree with Vemer et al, in response to the ISPOR guidelines, that what we also

seek in our validation programme is to check if the model is fit for purpose to explore decisions using

the analysis of effectiveness, cost-effectiveness and equity enabled by the model and as the model

evolves with new features and data.31

We will also develop an initial proposal of specific modelling standards for NCD prevention. We will

involve colleagues and build on ISPOR and Brighton Declaration principles to facilitate and

standardize model comparisons and provide suggested guidance for future exploitation of the open

source/open access model.30,32

More details on the methodology to represent the target population in the model are described

below in section 13 “Sampling”, page 15.

Objective 2. Update the evidence base to support model and scenario development

Building on the existing considerable evidence base informing IMPACTncd. We will update

and conduct a series of focused systematic reviews (SRs) to help inform model development and

delivery. This will provide estimates of baseline risks (overall, and in specific groups), adherence to

interventions, the effectiveness of diverse interventions (pharmacological and behavioural e.g.

diet, smoking, alcohol and physical activity exercise) for the nine ‘conditions of interest’ specified

above, and quantify likely costs and cost-effectiveness.

The workHORSE Systematic Review Programme will be conducted by experienced researchers

supervised by Prof Dickson and Dr Boland, and will operate in close collaboration with the model

development team.

Specifically, the overall focus of the programme is to provide evidence to inform model parameters.

The current epidemiological simulation IMPACTncd model is already benefitting from a substantial

set of SRs informing key epidemiological parameters, as well as many of the specific NHS Health


11

Check Programme interventions (hypertension control, smoking cessation, weight reduction

strategies and statin use). The new parameters reflecting the NHS Health Check Programme process

checkpoints (invitation, uptake and compliance) will be a priority for the systematic review team,

including a focus on existing evidence that highlights best local practices (mostly in the grey

literature).

Given the breadth of areas to review, we will focus on first identifying and using existing evidence

based guidelines and SRs. We will conduct new SRs using rapid/scoping methods, with a research

plan refined after reviewing the workHORSE final parameter specification. This will give priority to

key parameters reflecting programme implementation, adherence and invitation methodologies.

Our special focus will be to inform the six key parameters to simulate the implementation of NHS

HCP approaches (eligibility, annual coverage, uptake, risk profile, prescription and adherence rate of

treatment) plus the rate of brief interventions and referrals to lifestyle services and their subsequent

effectiveness, and best practices for recruiting patients (including grey literature).

All the parameters in the model will be modelled explicitly and in a format that allows easy

updateability to ensure long-term relevance of the evidence base for the model. This will also be

documented in a transparent way and clearly linked to the specific model inputs.

Objective 3. Assess the effectiveness, equity and cost-effectiveness of the alternative

strategies for implementation of the NHS Health Check Programme, as specified in the

NIHR call

workHORSE will provide a rich set of co-produced scenarios that will help to explore

effectiveness, equity and cost-effectiveness of different ways of implementing the NHS

HCP and generate outputs that can support real world decision making.

Effectiveness and equity analyses

We will look at several measures to explore the effectiveness of the health checks: cases detected,

cased prevented, number-needed-to-screen, and deaths prevented or postponed.

These analyses will include incidence, prevalence and mortality of the modelled diseases; cases and

deaths of all the conditions of interest prevented or postponed; deaths prevented or postponed;

cases detected; number needed to attend to prevent a case; life expectancy and disease-free life

expectancy; and Quality Adjusted Life Years. Each will be presented as absolute numbers and as

rates, as appropriate. All results will be stratified by age, gender and socioeconomic position.

Our IMPACTncd model incorporates key concepts and ideas from Social Epidemiology. It particularly

takes into account the multiple levels at which health determinants operate. It implements the

concept of social production of disease developed by Diderichsen, Whitehead et al, understanding

that health inequalities are the result of differential social exposure, differential vulnerability and

differential outcomes in the production of health inequalities. 33 This framework is particularly useful

to explore equity issues in complex care pathways, such as the NHS HCP.

For workHORSE, we will continue the development of our social epidemiology approach to analyse

the equity impact of the NHS Health Check Programme based on the equity plane presented in

Kypridemos et al.19 This will be done by relating measures of absolute and relative inequalities to

incremental cost-effectiveness ratios, and will provide features to support formal distributive equity

analyses following the principles laid out by Cookson et al.34


12

Health economics analyses

The economic evaluation will follow the guidelines stipulated in the Consolidated Health Economic

Evaluation Reporting Standards (CHEERS) statement.35

The main method will be cost utility (cost per QALY) analysis presented as ICERs (incremental cost

effectiveness ratios) compared to the best current intervention, represented as a baseline scenario

or a do-nothing scenario. Outcomes will also be presented as net monetary benefit (NMB) based on

combining net costs with a valuation of QALYs gained. This will allow us to better understand the

‘Value for Money’ of the NHS Health Check Programme. Cost-effectiveness will be stratified by age,

gender, ethnic group and deprivation group to enable understanding of whether enhanced targeting

of certain groups (for example over 50s or South East Asian populations) would increase cost-

effectiveness for the overall programme. We will also consider the opportunity cost in terms of

potential competing interventions, recognising that NHS HCP is most often commissioned by local

authority public health budgets, and are most often provided by GPs who would otherwise be

working in the NHS.

Intervention costs of the NHS Health Check Programme, additional costs of treatment associated

with NHS HCP (such as continued monitoring, blood pressure and statin medication), and costs of

CVD events and treatment will be collected in line with HTA and ISPOR guidelines. Social care costs

will also be modelled, as will indirect costs of lost productivity through illness and early deaths, and

informal care costs. Effectiveness of NHS HCP will be drawn from the systematic reviews.

QALYs will be estimated using standard sources using EQ-5D population preference based indices.

We will also present disaggregated outcomes (costs and consequences) for the conditions of interest

in terms of modelled incidence, prevalence, mortality, life years gained, disability burden avoided,

QALYs, disease costs, and indirect costs, in an impact inventory, (as recommended by the second US

panel on cost-effectiveness). These will be disaggregated by the five current conditions considered in

the remit of the NHS HCP (CHD, diabetes, kidney disease, stroke and dementia) and the four

additional conditions of interest: atrial fibrillation, alcohol misuse, cancers and COPD.

Scenario analysis will be conducted using different willingness to pay for QALY thresholds of

£13,000, £20,000, £30,000 and £60,000. These respectively equate to estimates of the opportunity

cost of spend in the NHS, the NICE willingness to pay thresholds for public health programmes and

for new technologies, and the UK Treasury’s valuation of health policy interventions. Scenario

analysis will also use discount rates of 0% and 6% as well as the base case scenario rate of 3.5%.

Furthermore, we will seek advice from our stakeholders on best ways to present the effectiveness,

cost-effectiveness and equity outputs to ensure better support of local, real world decision making

and facilitate translation to non-specialists, including return of investment calculations and provide

information on opportunity costs when relevant.

Sensitivity and Uncertainty Analyses

A crucial activity in any analyses based on a simulation model is to fully understand the

implications of the inevitable uncertainty in the parameters used to inform the model. These

features will also allow us to pursue value of information analysis to provide the research and

public health communities with insights to influence research agendas to provide data more

relevant to decision support.


13

IMPACTncd and its adaptation for workHORSE already has a state of the art probabilistic sensitivity

analysis based on 2nd order Monte Carlo simulations. This propagates population heterogeneity,

inputs parameter uncertainty and partly structural uncertainty of the model to model outputs,

enabling value of information analysis to highlight future research areas which can reduce

uncertainty and lead to better decision making. We will also consider costs uncertainty in our

sensitivity analyses (see immediately below).

Economic uncertainty analysis will be conducted where costs and utility are fitted to probability

distributions to account for first and second order uncertainty in the estimates. A series of iterations

will be run with Monte Carlo sampling from the probability distributions of all input parameters to

capture uncertainties in the estimates, generating 95% CIs around all outcomes, (as recommended

in ISPOR modelling guidelines).36 This will be used to present the uncertainty around cost-

effectiveness, and the probability that each strategy is dominant at different willingness to pay

thresholds. This will also be presented visually in a cost-effectiveness plane and cost-effectiveness

acceptability curves (CEACs).

Objective 4. Recruit a diverse group of stakeholders to powerfully strengthen the user

perspective to inform desirable features of the user-friendly model and additional locally

relevant scenarios to test

Tapping into our extensive networks, we will recruit and engage local stakeholders, users and lay

persons to co-produce scenarios that reflect current and potential implementations of the NHS

Health Check Programme, including alternative options and strategies to illustrate the range of

features that workHORSE will offer. Their insights will powerfully influence model development,

evidence update, scenario analysis and implementation plans.

Establishing the Stakeholder Group We will build on our existing networks and expertise of working with stakeholder groups and

increasing public participation in research.37–39 We will identify relevant stakeholders at local level

(including LA and NHS commissioners, local government association) particularly those with expertise

in the commissioning (local authority commissioners) and provision of the NHS HCP (including GPs,

practice nurses and community pharmacists), third sector organisations (e.g. British Heart Foundation,

Diabetes UK, and Alzheimer’s Research UK) and national organisations (including PHE, DH, NICE).

Stakeholder Recruitment Stakeholders with an interest or experience in the areas of CVD, diabetes, COPD, and related cancers

and health checks will be particularly targeted. Stakeholders will thus be identified and recruited via

direct approaches - personal contacts, snowballing via existing stakeholders, letters, emails, phone

calls, and indirectly via postings on web sites, blogs, and newsletters.

Stakeholders will be offered brief training covering their activities and tasks and other

responsibilities. The stakeholder group will be established using established ground rules of

acceptance, mutual respect and equality, and will be facilitated by experienced researchers with

training in supporting these principles (led by Lloyd-Williams). Stakeholders will receive

reimbursement to cover travel and subsistence expenses.

Stakeholder Engagement and Mobilisation, eliciting perspectives and priorities.

The Stakeholder Group will meet at four workshops throughout the project (in months 2,8,14 and

23). Our methodology will be based upon the “Exploratory scenario-building approach” six steps of

scenario setting40 and will reflect our co-production model for scenario specification. 41At


14

successive stages, the Group will be invited to develop a shared understanding of the current and

potential implementations of the NHS Health Check Programme, including alternative options and

scenarios; produce scenarios based upon the research evidence to explore, develop and prioritise

specific, evidence-based scenarios; to analyse with workHORSE, focussing on local relevance,

acceptability, feasibility and affordability and facilitate dissemination of emerging findings, final

outputs, and the decision support tool itself.

Workshop Data Collection and Analysis

An experienced qualitative and participatory research team led by Lloyd-Williams will facilitate the

workshops, moderate the discussions and take notes. The research assistant will be responsible for

administration, audio-recording and detailed note taking.

All data will be transcribed and analysed using NVivo software for qualitative data analysis. Output will

help to identify and prioritise the scenarios of greatest relevance to local NHS Health Check

Programmes, and will inform the ongoing development of the model. Building on our previous

successful stakeholder studies, we will use the Framework Method which follows five pre-defined

stages: (1) familiarisation, (2) identification of a thematic framework, (3) indexing, (4) charting, and (5)

mapping and interpretation.42,43

Objective 5. Develop a sustainability and implementation plan to deploy our user-friendly

web-based decision-support model at local level.

With the insights gained from building workHORSE and the inputs from our stakeholder

engagement process, we will propose principles to ensure the sustainability and implementation

needs of the decision-support tool at local level.

The plan will explore options to facilitate the integration of the model with data at local level,

propose implementation and scalability strategies, provide guidelines for scenario design based on

insights gained during the stakeholder engagement process, and detail a development plan to

facilitate interactions with the decision-support tool by users. Furthermore, we will propose specific

modelling standards for non-communicable disease prevention simulation modelling , building on

International Society for Pharmacology and Outcomes Research and the Brighton Declaration.32,44

To pilot key aspects of the plan, we will ensure that the Model architecture will be future-proofed to

maximize its potential integration with NHS HCP surveillance data. Our recent work at local

authority level in Liverpool suggests that this is feasible (See Section 3 “Background and Rationale”,

page 3). However, further work on defining standards and pre-processing of inputs to run in the

model will be essential. In the dissemination plan, we will therefore provide our insights gained

through the model development process and aligned with the current efforts on data procurement

led by Public Health England. To future proof workHORSE, we will ensure its ability to provide

additional outputs and analysis through future modular extensions (e.g. additional diseases,

disability, multi-morbidity, and enabling distributive equity analysis).

To inform implementation options, scalability and interactivity with the model, we will pilot a secure

cloud-based platform to allow local commissioners to run their own scenarios by providing the

essential inputs in an electronic form. We will use the private-cloud infrastructure of @HeRC and

eventually link our implementation with the e-lab platform, through our collaboration with Buchan

and Couch. The local commissioners will have the options to run pre-specified scenarios (i.e. based

on best practice) or modify them and hence obtain estimates of effectiveness, cost-effectiveness

and equity for the local populations. We will continue working with Liverpool data for development

purposes, and we will seek through the stakeholder engagement process to partner with one

additional local authority to further develop the local prototype and explore implementation and


15

scalability issues, and as a proof of concept to ensure the sustainability of open access. This process

will be supported by the development of a prototype simple user interface providing intuitive

interaction with the scenario design features proposed in Objective 1 (See Section 6 ”Research Plan”,

page 7). This pilot will help us to propose a workable solution to ensure sustained open access to

the model, and a realistic estimate of resources needed for broad implementation at local level.

6. Health technologies being assessed CVD Screening programmes and preventative interventions for NCD. As our baseline case, we will

employ the current PHE recommendations for implementing the NHS Health Check Programme.

Specifically, we will explore scenarios based on the current implementation of the NHS Health Check

Programme, options to increase uptake and adherence (conservative, realistic, and optimal, using

real life exemplar practices) and options to maximize equity impact (targeted screening, or

proportionate universalism. Furthermore, workHORSE will have the flexibility to explore alternative

scenarios for NHS HCP processes through its scenario design features, to accommodate scenarios co-

produced with stakeholders.

7. Design and theoretical/conceptual framework Evaluating the effectiveness, equity and cost-effectiveness to support decision making is unlikely to

be practically solved through traditional design methods alone. Thus, simulation modelling based on

best, up-to-date evidence is a useful approach to explore alternative implementations of the NHS

HCP as an “in silico” experiment.

Most modelling approaches have not been transparent, easily adaptable to other uses or realistically

reflecting the implementation of a complex programme like the NHS Health Checks.

A modelling approach also has the added benefit of identifying areas of uncertainty that could guide

more targeted research efforts. For instance, designing empirical studies to test specific

implementation proposals that promise the best returns on investment as suggested by modelling

analyses using workHORSE.

8. Target population The model adaptation for workHORSE is designed to be primarily used at local level, following the

transfer of responsibility for commissioning and implementation to local authorities in 2013.

We will initially develop the model functionality using locally available data from Liverpool. We will

then implement the model in an additional local authority (identified early in the stakeholder

engagement process). We will then coordinate with representatives of PHE and local authorities to

further explore potential data issues at local levels. We will then feed this into the document

summarizing our recommendations for deploying the model and use across local authorities

throughout England. As a default, we will use data from regional or national level as appropriate.

9. Inclusion/Exclusion Criteria We will define our inclusion and exclusion criteria for eligibility for the NHS HCP based on the

current specification from PHE. The programme includes adults aged 40-74, and excludes those with

pre-existing vascular disease (hypertension, ischaemic heart disease, stroke or transient ischaemic

attacks, atrial fibrillation, heart failure, peripheral arterial disease, chronic kidney disease, familial

hypercholesterolemia, diabetes or already being treated with statins).


16

10. Setting/context workHORSE is primarily designed to model delivery of the NHS HCP through the primary care

system. However, it is possible to also model other types of delivery settings (e.g. care assistants,

pharmacies, etc.), if suggested during the stakeholder engagement process.

The model will particularly focus on the needs of Local Authorities’ commissioning and public health

teams in terms of available outputs and the health economic analysis perspective.

11. Search strategy (in the case of projects involving evidence synthesis): Given the interactive nature of model and scenario development, search strategies targeted to

updates of the evidence base will be developed by the Systematic Review team as a result of the

interaction with stakeholders and modellers building specific scenarios.

12. Sampling In order to create a representation of the eligible population for the NHS Health Check, we use the

statistical framework originally developed for the European Union Statistics on Income and Living

Conditions by Alfons et al.45,46 The method allows us to expand the sample of a nationally

representative population survey to a ‘close-to-reality’ synthetic population, while preserving the

statistical properties and respecting the correlation structure of the original survey. The two main

advantages of this framework are: 1) it considers the hierarchical structure of the sample design of

the original survey (i.e. individuals within households, within larger geographical areas); and 2) it

can generate trait combinations which were not present in the original survey (due to limited sample

size) but are likely to occur in the real population. The latter is particularly important, because it

avoids bias from excessive repetition of a limited set of trait combinations present in the original

survey sample. (For example, the original survey may only have two 80 year old male participants,

both ex- smokers). Unlike other methodologies, the approach proposed by Alfons et al. can generate

80 year old male synthetic individuals, who are never or current smokers even though these

combinations of traits were not present in the original, limited survey. It also ensures that extreme

outliers present in the original survey are not overrepresented in the synthetic population.

We have evolved the original statistical framework to better reflect epidemiological principles. Our

approach considers the widely-accepted framework of the ‘wider determinants of health’ and allows

socioeconomic conditions to influence lifestyle choices (i.e. diet, smoking, etc.), which in turn

influence biological risk factors (i.e. blood pressure, cholesterol, etc.).33 Hence, crucially, the

synthetic population replicates the clustering of risk factors in the most deprived socioeconomic

groups that has been observed in England. This method will be further adapted to create synthetic

populations at local level, using deprecated Strategic Health Authorities to adapt Health Survey for

England data, supplemented by local authority data when available.

This flexibility will enable a productive interaction during the stakeholder engagement process to co-

produce additional, locally relevant scenarios for modelling, including alternate definitions for

eligibility criteria.

13. Data collection The current epidemiological and intervention effectiveness parameters including effect sizes are

already sourced from existing, published systematic reviews and meta-analyses (see Kypridemos et

al 2016, reference 19).

The additional data needed for the project are publicly available and include the Health Survey for

England datasets for 2001-2014, ONS mortality and demographics using published data. We will also

ask for updates on our previous ad hoc data requests. All data will be stratified by age, sex and IMD.


17

We will liaise with PHE to obtain data for the baseline for the six parameters for the scenario builder.

For the Local Adaptation of the model, we will use deprecated Strategic Health Authorities to

calibrate Health Survey for England data for local use, and local level mortality from ONS to calibrate

the model. We will work with colleagues from PHE and the prototype local authorities to obtain data

for the baseline for the six parameters for the scenario builder.

Given the comprehensiveness of the model needed to respond to this HTA call, it is expected that

local data may not always be available to inform specific model parameters. In this situation, we will

rely on regional or national data, or failing that, robust and transparent assumptions (tested in

sensitivity analyses).

14. Data analysis All the model outputs will be stratified by year, age, gender, sex, QIMD, and, when appropriate,

disease for both national and local models. These will include: incidence, prevalence and mortality

of the modelled diseases; cases and deaths of all the conditions of interest prevented or postponed;

deaths prevented or postponed; cases detected; number needed to attend to prevent a case; life

expectancy and disease-free life expectancy; QALYs; Health Check costs; direct and indirect disease

costs; ICER and NMB and distributional equity outcomes.

We have summarized the different analysis strategies above for health economics (See “health

economics and cost-effectiveness methods”, page 11), equity (“Equity Analysis”, page 10) and

stakeholder engagement (See “Workshop Data Collection and Analysis “page 13). All the outputs will

be provided with a central value and a 95% uncertainty analysis (see section 7 Research Plan,

“uncertainty analysis”, page 11)

15. Dissemination and projected outputs

We will primarily disseminate our findings via stakeholder workshops, academic papers, a final

report, mass media, project website, Facebook, and Twitter feeds. We will also invite our lay

advisors to comment on the production of press releases and other publicity materials.

The final workshop will provide an excellent opportunity to organize a media event, supported by

the University of Liverpool Media team.

Our scenario analyses will help users to identify the strategies most likely to be feasible, cost-

effective and equitable. We will disseminate our results widely to local and national stakeholders,

decision makers, policy makers, public health communities, the third sector and the wider public. For

the academic community, we will produce a methods paper and results papers describing scenarios

and findings, all targeted at high impact, open access journals.

To maximally engage with key target audiences and ensure high impact, we will also link with key

individuals at leading organisations including the British Heart Foundation, Alzheimer’s Research UK,

UK Health Forum, Faculty of Public Health, Public Health England and the Richmond Group.

Dissemination via these networks will offer diverse “pathways to impact” for decision makers and

commissioners at local, regional and national levels.

16. Plan of investigation and timetable Milestones:

Milestone 1 (Month 9): Working model prototype with documentation

Milestone 2 (Month 10): Simple user interface and documentation


18

Milestone 3 (Month 11): Updated evidence base

Milestone 4 (Month 12): Stakeholder / User designed scenarios

Milestone 5 (Month 13): Analysis of stakeholder designed scenarios and emerging results

Milestone 6 (Month 22): Finalised results

Milestone 7 (Month 23): Model report, deployment and sustainability planning proposal, draft

journal manuscripts

Month 1-3 4-6 7-9 10-12 13-15 16-18 19-21 22-24

TASK Model development X X X X X X Evidence Review X X X X X X X Stakeholder Workshops X X X X Scenario Design X X X X X X Analysis X X X X X X Dissemination X X X X X X X X

17. Project management We will create a workHORSE Project Advisory Group (PAG) to guide and advise on the

Implementation of the overall project. The aim will be to ensure continuous quality improvement in

the delivery of the research. The group activities will include: advising on recruitment of participants,

methodology, reviewing detailed financial expenditure against timetable, delivery of project and

research activities, ensuring adherence to schedules and deadlines, disseminating key findings, and

helping to ensure it is delivered on schedule and within budget.

The PAG will include relevant topic experts, policy makers and stakeholders (academics, health

economists, local/regional government and NHS decision makers and planners, NGO experts, and

the lay advisers). The PAG will appoint an independent chair and include the Principal Investigator

(Martin O’Flaherty), Co-investigators, and workHORSE researchers.

The PAG will meet formally face-to-face or by Video Link on a quarterly basis. The PAG will also

communicate informally where independent advice or comment is required.

A part-time project administrator will support the Project Advisory Group (agenda, papers, minutes,

venues, and travel). The administrator will ensure the PAG meetings are timely and all

communication resulting from the meetings is accurate and circulated to all group members

highlighting action points to ensure continued efficient project delivery. They will liaise with all PAG

members to ensure continued engagement with the project aims, objectives and outcomes. The

PAG will have oversight of the entire research activities. Together with the Principal Investigator,

they will oversee the budget, ensuring it is being spent appropriately and on time. They will support

the principal investigator by requesting and reviewing quarterly update reports from each of the four

researchers. They will help support selection and appointment of the project researchers (advertising,

short-listing for interviews). They will also assist in inviting additional stakeholders for Workshop 4,

dissemination of workHORSE results, and supporting media activity.

18. Approval by ethics committees We will adhere to the University of Liverpool (UoL) Research Ethics Policy,47 and will follow

appropriate ethical procedures throughout. All participants will be recruited voluntarily.

Stakeholders will be provided with a subgroup specific (i.e. professionals and members of the public)

"Study Information Form". All gathered information will be anonymised and the principles of

confidentiality explained in terms of data coding, disposal, sharing and archiving. If willing to

participate, participants will then be asked to sign the study consent form. Stakeholders will receive

initial training to participate in the workshops, then debriefing later.


19

This project does not pose any direct risks or impacts to participants, nor do we envisage any issues

concerning disclosure. However, if any arose they would be dealt with immediately. Participants will

not be named except by their express permission. Ethical approval will be sought from the UoL

Research Ethics Sub-Committee for Non-Invasive Procedures which meets bi-monthly. All project

researchers working with the stakeholders will undergo background checks via the "Disclosure and

Barring Service".

Data issues. Only the research team will have access to the data. All named data will be stored

securely in the Department of Public Health and Policy, University of Liverpool. Handwritten notes

and audio tape recordings will be stored in locked cabinets until transcribed and then destroyed. All

transcribed data will be anonymised and kept on password protected project specific computers.

Twelve months after project completion, project outputs will be transferred to the UoL "Liverpool

Elements", an online tool allowing researchers to capture, collate and showcase their research. All

individual level data will be destroyed.

We do not perceive any conflicts of interest for this research.

19. Patient and Public Involvement (PPI) We have already recruited our Lay Advisers via Healthwatch and the People in Research website. To

enable our advisors to get to know the research team and vice versa, we held an informal meeting

on February 27th 2017 with our four lay advisers – (a retired female, volunteer Healthwatch Wirral, a

retired female, service user/volunteer Macmillan cancer information centre, a retired male, previous

PPI primary care research, and a male patient with diabetes, service user and patient involvement in

mental health). One key comment was that the proposal was unique as similar projects do not

usually involve stakeholders and this should be highlighted in the proposal. The Lay Advisers

discussed how NHS Health Check Programme service provision and uptake might be improved and

how this project could achieve this by demonstrating cost-effectiveness. They also provided

feedback on the content of the Summary in Plain English. They discussed their role in the project

tapping on their own experiences and skills, in helping interpret the results, and acting as a voice to

disseminate the research to a broad audience. They also offered suggestions on additional

stakeholders we should invite. All feedback has been incorporated. The full notes from the meeting

(approved by the Lay Advisers) are available.

To maximise the relevance of our proposed research, the four Lay Advisers will be invited to be

active members of the Project Advisory Group. They perceive their role as a “sounding board”,

advising on strategies to increase stakeholder involvement and successful workshop delivery, and

ensuring emerging findings and project results reach a wide audience. Their involvement will be

crucial in planning the study design and research delivery, especially reviewing documents for the

development, implementation and interpretation of the project. In disseminating the research

findings, we will encourage and support the Lay Advisers to contribute to the preparation of press

releases, abstracts and papers for publication. Their level of involvement will be entirely voluntary

and very flexible. Their input will be sought via email as well as face to face meetings. Training and

support will be provided throughout, including attending courses provided by the North-West People

in Research Forum and ongoing support and guidance from members of the research team. We will

also invite our Lay Advisers to shadow specific researchers if they feel they might gain useful

experience or have input to offer. Lay Advisers will be appropriately reimbursed for their involvement.

References

1. Murray CJ, Richards MA, Newton JN, et al. UK health performance: findings of the Global Burden of Disease Study 2010. Lancet. 2013;381(9871):997-1020. doi:10.1016/S0140-6736(13)60355-4.


20

2. Expert Scientific and Clinical Advisory Panel. Emerging evidence on the NHS Health Check: findings and recommendations - NHS Health Check. http://www.healthcheck.nhs.uk/latest_news/emerging_evidence_on_the_nhs_health_check_findings_and_recommendations/. Accessed March 22, 2017.

3. Kontis V, Mathers CD, Rehm J, et al. Contribution of six risk factors to achieving the 25×25 non-communicable disease mortality reduction target: A modelling study. Lancet. 2014;384(9941):427-437. doi:10.1016/S0140-6736(14)60616-4.

4. Robson J, Dostal I, Sheikh A, et al. The NHS Health Check in England: an evaluation of the first 4 years. BMJ Open. 2016;6(1):e008840. doi:10.1136/bmjopen-2015-008840.

5. Artac M, Dalton ARH, Babu H, Bates S, Millett C, Majeed A. Primary care and population factors associated with NHS Health Check coverage: a national cross-sectional study. J Public Health (Bangkok). 2013;35(3):431-439. doi:10.1093/pubmed/fdt069.

6. McNaughton RJ, Shucksmith J. Reasons for (non)compliance with intervention following identification of “high-risk” status in the NHS Health Check programme. J Public Health (Bangkok). 2015;37(2):218-225. doi:10.1093/pubmed/fdu066.

7. Baker C, Loughren EA, Crone D, Kallfa N. A process evaluation of the NHS Health Check care pathway in a primary care setting. J Public Health (Bangkok). 2015;37(2):202-209. doi:10.1093/pubmed/fdv053.

8. Usher-Smith J, Martin A, Harte E, et al. NHS Health Check Programme Rapid Evidence Synthesis. Cambridge; 2017. www.healthcheck.nhs.uk/document.php?o=1251.

9. Dalton AR, Soljak M, Samarasundera E, Millett C, Majeed A. Prevalence of cardiovascular disease risk amongst the population eligible for the NHS Health Check Programme. Eur J Prev Cardiol. 2013;20(1):142-150. doi:10.1177/1741826711428797.

10. Dalton ARH, Bottle A, Okoro C, Majeed A, Millett C. Uptake of the NHS Health Checks programme in a deprived, culturally diverse setting: cross-sectional study. J Public Health (Bangkok). 2011;33(3):422-429. doi:10.1093/pubmed/fdr034.

11. Robson J, Dostal I, Madurasinghe V, et al. The NHS Health Check programme: implementation in east London 2009-2011. BMJ Open. 2015;5(4):e007578. doi:10.1136/bmjopen-2015-007578.

12. Forouzanfar MH, Afshin A, Alexander LT, et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388(10053):1659-1724. doi:10.1016/S0140-6736(16)31679-8.

13. Ebrahim S, Taylor F, Ward K, Beswick A, Burke M, Davey Smith G. Multiple risk factor interventions for primary prevention of coronary heart disease. In: Ebrahim S, ed. Cochrane Database of Systematic Reviews. Chichester, UK: John Wiley & Sons, Ltd; 2011. doi:10.1002/14651858.CD001561.pub3.

14. Jørgensen T, Jacobsen RK, Toft U, Aadahl M, Glümer C, Pisinger C. Effect of screening and lifestyle counselling on incidence of ischaemic heart disease in general population: Inter99 randomised trial. BMJ. 2014;348.

15. Lee JT, Lawson KD, Wan Y, et al. Are cardiovascular disease risk assessment and management programmes cost effective? A systematic review of the evidence. Prev Med (Baltim). 2017;99:49-57. doi:10.1016/j.ypmed.2017.01.005.

16. Department of Health Vascular Policy Team. Economic Modelling for Vascular Checks. http://www.nhshealthcheck.nhs.uk/i/assets/Economic Modelling.pdf. Published 2008.

17. Schuetz CA, Alperin P, Guda S, et al. A standardized vascular disease health check in europe: a


21

cost-effectiveness analysis. Federici M, ed. PLoS One. 2013;8(7):e66454. doi:10.1371/journal.pone.0066454.

18. Cooney M-T, Dudina A, Whincup P, et al. Re-evaluating the Rose approach: comparative benefits of the population and high-risk preventive strategies. Eur J Cardiovasc Prev Rehabil. 2009;16(5):541-549. doi:10.1097/HJR.0b013e32832b38a1.

19. Kypridemos C, Allen K, Hickey GL, et al. Cardiovascular screening to reduce the burden from cardiovascular disease: microsimulation study to quantify policy options. BMJ. 2016;353:i2793. http://www.ncbi.nlm.nih.gov/pubmed/27279346. Accessed March 22, 2017.

20. Kypridemos C, Collins B, Capewell S, Parvulescu P, O’Flaherty M. The Impact of NHS Health Check Programme on Cardiovascular Disease Burden in Liverpool. A Health Economics Microsimulation to Quantify the Policy Options. Presented at the NHS Health Checks Annual Scientific Meeting "Getting Serious About Prevention 2017: Improving Cardiovascular Health Together" Organizes by PHE. Manchester 9/2/17

21. Kypridemos C, Guzman-Castillo M, Hyseni L, et al. Estimated reductions in cardiovascular and gastric cancer disease burden through salt policies in England: an IMPACTNCD microsimulation study. BMJ Open. 2017;7(1):e013791. doi:10.1136/bmjopen-2016-013791.

22. Allen K, Kypridemos C, Hyseni L, et al. OP11 The effects of maximising the UK’s tobacco control score on inequalities in smoking prevalence and premature coronary heart disease mortality: a modelling study. J Epidemiol Community Health. 2015;69(Suppl 1):A13.2-A13. doi:10.1136/jech-2015-206256.11.

23. Chang KC-M, Soljak M, Lee JT, et al. Coverage of a national cardiovascular risk assessment and management programme (NHS Health Check): Retrospective database study. Prev Med (Baltim). 2015;78:1-8. doi:10.1016/j.ypmed.2015.05.022.

24. Chang KC-M, Lee JT, Vamos EP, et al. Impact of the National Health Service Health Check on cardiovascular disease risk: a difference-in-differences matching analysis. CMAJ. 2016;188(10):E228-38. doi:10.1503/cmaj.151201.

25. Crossan C, Lord J, Ryan R, Nherera L, Marshall T. Cost effectiveness of case-finding strategies for primary prevention of cardiovascular disease: a modelling study. Br J Gen Pract. 2017;67(654):e67-e77. doi:10.3399/bjgp16X687973.

26. Adams J, Mytton O, White M, Monsivais P. Why Are Some Population Interventions for Diet and Obesity More Equitable and Effective Than Others? The Role of Individual Agency. PLOS Med. 2016;13(4):e1001990. doi:10.1371/journal.pmed.1001990.

27. O’Flaherty M, Capewell S. New perspectives on cardiovascular risk in individuals and in populations. J Epidemiol Community Health. 2012;66(10):855-856. doi:10.1136/jech-2012-201409.

28. World Cancer Research Fund International. Continuous Update Project. Cnacer Prevention & Survival.; 2016. http://www.wcrf.org/sites/default/files/CUP-Summary-Report.pdf. Accessed March 28, 2017.

29. Parkin DM. 1. The fraction of cancer attributable to lifestyle and environmental factors in the UK in 2010. Br J Cancer. 2011;105 Suppl:S2-5. doi:10.1038/bjc.2011.474.

30. Caro JJ, Briggs AH, Siebert U, Kuntz KM. Modeling good research practices--overview: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force--1. Value Health. 15(6):796-803. doi:10.1016/j.jval.2012.06.012.

31. Vemer P, van Voom GAK, Ramos IC, et al. Improving model validation in health technology assessment: comments on guidelines of the ISPOR-SMDM modeling good research practices task force. Value Health. 2013;16(6):1106-1107. doi:10.1016/j.jval.2013.06.015.


22

32. Webber L, Mytton OT, Briggs ADM, et al. The Brighton declaration: the value of non-communicable disease modelling in population health sciences. Eur J Epidemiol. 2014;29(12):867-870. doi:10.1007/s10654-014-9978-0.

33. Diderichsen F, Evans T, Whitehead M. The Social Basis of Disparities in Health. In: Challenging Inequities in Health: From Ethics to Action. Oxford University Press; 2009.

34. Cookson R, Mirelman AJ, Griffin S, et al. Using Cost-Effectiveness Analysis to Address Health Equity Concerns. Value Heal. 2017;20(2):206-212. doi:10.1016/j.jval.2016.11.027.

35. Husereau D, Drummond M, Petrou S, et al. Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement. BMJ. 2013;346. http://www.bmj.com/content/346/bmj.f1049. Accessed March 28, 2017.

36. Briggs AH, Weinstein MC, Fenwick EAL, Karnon J, Sculpher MJ, Paltiel AD. Model parameter estimation and uncertainty: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force--6. Value Health. 2012;15(6):835-842. doi:10.1016/j.jval.2012.04.014.

37. Taylor-Robinson DC, Milton B, Lloyd-Williams F, O’Flaherty M, Capewell S. Planning ahead in public health? A qualitative study of the time horizons used in public health decision-making. BMC Public Health. 2008;8. doi:10.1186/1471-2458-8-415.

38. Orton LC, Lloyd-Williams F, Taylor-Robinson DC, Moonan M, O’Flaherty M, Capewell S. Prioritising public health: a qualitative study of decision making to reduce health inequalities. BMC Public Health. 2011;11:821. doi:10.1186/1471-2458-11-821.

39. Lloyd-Williams F, Bromley H, Orton L, et al. Smorgasbord or symphony? Assessing public health nutrition policies across 30 European countries using a novel framework. TL - 14. BMC Public Health. 2014;14 VN-r:1195. doi:10.1186/1471-2458-14-1195.

40. Bock A-K, Maragkoudakis P, Wollgast J, et al. Tomorrow’s Healthy Society : Research Priorities for Foods and Diets : Final Report. Publications Office; 2014. https://ec.europa.eu/jrc/en/publication/eur-scientific-and-technical-research-reports/tomorrows-healthy-society-research-priorities-foods-and-diets. Accessed March 22, 2017.

41. Morgan W, Carey F, Ramsden S. A Co-production Model: five values and seven steps to make this happen in reality | Coalition for Collaborative Care. http://coalitionforcollaborativecare.org.uk/catherine-wilton/a-co-production-model-five-values-and-seven-steps-to-make-this-happen-in-reality/. Accessed March 22, 2017.

42. Pope C, Ziebland S, Mays N. Qualitative research in health care. Analysing qualitative data. BMJ. 2000;320(7227):114-116. http://www.ncbi.nlm.nih.gov/pubmed/10625273. Accessed March 22, 2017.

43. Ritchie J, Spencer L. Qualitative Data Analysis for Applied Policy Research. In: The Qualitative Researcher’s Companion. 2455 Teller Road, Thousand Oaks California 91320 United States of America: SAGE Publications, Inc.; :305-329. doi:10.4135/9781412986274.n12.

44. Caro JJ, Briggs AH, Siebert U, Kuntz KM. Modeling good research practices--overview: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force--1. Value Health. 2012;15(6):796-803. doi:10.1016/j.jval.2012.06.012.

45. Alfons A, Kraft S, Templ M, Filzmoser P. Simulation of close-to-reality population data for household surveys with application to EU-SILC. Stat Methods Appt. 2011;20(3):383-407.

46. Afons A, Kraft S. simPopulation: simulation of synthetic populations for surveys based on sample data [Internet]. 2013. http://cran.r-project.org/package=simPopulation.

47. University of Liverpool. Policy on ethical approval for research involving human participants, tissues or personal data. https://www.liverpool.ac.uk/intranet/research-support-


23

office/research-ethics/. Accessed March 24, 2017.

netscc id - national institute for health research

Documents