What does it take to make integrated care work ?

Within the EU, there are 10 million people with chronic heart failure (CHF), 20 million people with chronic obstructive pulmonary diseases (COPD), and 60 million people with diabetes (DM). These three conditions alone cost the EU healthcare systems around €125 billion a year. Care coordination and telehealth (CC&TH) services are an effective way to treat these patients in their homes. However, the adoption of these services is relatively slow. A primary reason that is hindering the adoption is the integration of these services into the care delivery structures of exiting routine care practices. So far, CC&TH has been directed at improving health outcomes, administrative efficiency and cost effectiveness. The Advancing Care Coordination & Telehealth Deployment (ACT) Programme is the first to explore the organisational and structural processes needed to successfully implement care CC&TH services on a large scale. This EU-funded programme includes a consortium of clinical experts, universities, industry partners, and authorities from 5 leading healthcare EU regions. The objective of the ACT programme is to analyse what needs to be done to effectively integrate CC&TH services and identify best practices in Europe. Five leading EU healthcare regions with broad experience in CC&TH care delivery have shared their experiences of program deployment and provided data for the assessment. Each region has a population of at least 3000 patients (CHF, COPD and DM). The participating regions are Lombardy (IT), Basque Country (ES), Catalonia (ES), Northern Netherlands (NL), and Scotland (UK). The ACT evaluation framework considers final outcomes on efficiency and efficacy, as well as process outcomes to measure the integration of the CC&TH deployment into routine care with respect to stratification, care coordination and workflow optimization, patient adherence, and staff engagement. Data (qualitative and quantitative) were collected at baseline (month 6) and during several iterations (month 14, 18 and 24). Data collection, storage, analysis and visualization is supported an evaluation engine. There are yet no generally accepted indicators for measuring the successful deployment of CC&TH at European level. Hence, such data is not easily shared or compared. Such limitations need to be acknowledged when interpreting results. Despite marked differences between participating programmes (e.g. target population, maturity level, funding schemes), a common way to present and evaluate these programmes with the terminology, data models and its evaluation framework and engine was developed during the ACT programme. This allowed the presentation and comparison of programmes in a common framework, where (relative) differences were highlighted, and enabled generalization of findings and knowledge transfer. Over 2500 survey responses and more than 90 datasets were collected and analysed. The insights, conclusions, recommendations and best practices were then incorporated into a ‘cookbook’ for large-scale deployment of CC&TH, from which some high level examples are given below. Our findings suggest that for staff engagement it is necessary to provide interventions in training, early engagement of staff, communication feedback loops, and recognition of professional expertise. Tailored adherence strategies that address the patient’s needs are required to achieve adherence, not only for clinical needs, but also social support and socio-economic needs. Furthermore, health risk assessment should employ a population-based approach. Although some of these findings may be appear intuitive or self-evident, it is important to stress that there actually is evidence for these findings. They are based on data from various operational deployments across Europe. Further details and examples are available in the cookbook and online (www.act-programme.eu). The assessment process using the evaluation framework and engine in the regions has been useful for the regions. Some examples of their practical impact are: - demonstration of good program outcomes to policy makers, - evidence of shifts towards primary care in utilization, but not in resources, - knowledge transfer between regions after highlighting relative strengths and weaknesses, and - benchmarking of programs within a region. Due to their complexity, each CC&TH deployment is unique. Despite the differences, the methodology and tools to assess these deployments, and a common way to present their relative strengths and weaknesses was successfully developed within the ACT programme. This approach has proven to be effective in creating knowledge transfer within and across the regions. As a next step, the ACT consortium intends to apply the framework and tools to monitor progress of deployment performance during iterative process improvement cycles. In addition, the consortium seeks to promote the structured development of generally accepted quality indicators for successful deployment of care strategies across Europe.

Patient adherence: We discuss the promotion of self-managed care, identify its complications and benefits, and review how patients define the new role.

ACT Evaluation Framework
ACT uses a holistic framework for evaluating the scaled deployment of CC&TH.Its success relies on the analysis of drivers and outcomes, which address the qualitative and quantitative aspects of the deployments.
• The Key Performance Indicators are the quantitative outcome indicators that capture CC&TH performance.• The Key Drivers are the qualitative indicators, which describe anything that affects CC&TH performance.This holistic framework is what drives the ACT evaluation engine.

ACT Evaluation Engine
The ACT evaluation engine is a useful tool for data collection, storage, and analysis of CC&TH data.Developed by Philips Research and the University of Thessaloniki, it utilises a number of open source tools, and custom-developed modules.It enables centralised data collection, an on-line survey tool, and interactive dashboard for data analysis and visualisation of key indicators.The platform is secure, flexible, and user-friendly.

www.act-programme.eu
The work leading to these results has received funding from the European Community's Health Programme under grant agreement n° 20121209.
The ACT programme is fully aligned with the European Innovation Partnership in Active and Healthy Ageing objectives to deploy integrated care for chronically ill patients.

Insights and Conclusions
A number of insights and conclusions were identified by the ACT programme.These will prove useful and valuable for informing the large-scale deployment of CC&TH.Targeted at populations of chronic patients and elderly people, these insights and conclusions are a useful benchmark for implementing and exchanging best practices across the EU.
• Perceptions between managers, frontline staff and patients do not always match.• Organisational structure does influence the views and experiences of patients.• Successful patient adherence happens when staff are engaged.• There is a willingness by patients to participate in healthcare programmes.• Patients overestimate their level of knowledge and adherence behaviour.• The ability to track the use of resources is a useful feature of a stratification strategy, however, current regional case finding tools are difficult to benchmark and evaluate.• Data availability and homogeneity are the biggest challenges when evaluating the performance of the programmes.

Patient Adherence
Healthcare systems, programmes, personalities, and -importantly -patient profiles vary substantially.
For this reason, we need to design adherence strategies that are adapted around the needs of the patient.Not only clinical, but also social support and socio-economic needs.

Stratification
Health risk assessment should employ a populationbased approach.Both for risk prediction modelling and indicators.This is a priority for the scale up of integrated care at an EU level.

Assessment
European regions should agree on a minimum dataset of outcome indicators to be collected by all CC&TH programmes, addressing data collection challenges (comparability, ambiguity), with a small set of strictly defined indicators.The EC should facilitate unified data collection through interfaces, conforming to data privacy and protection legislation, and promote further initiatives in the implementation of evaluation of integrated care programmes.

Regional Good Practice Examples
Basque Country: Population Intervention Programmes Regional risk stratification strategies indicate how to match populations, and individual requirements, to appropriate service levels.

Lombardy: CREG Programme
Personalized care plans improve patient satisfaction and perception of safety, as well as their capacity to manage their condition.

Catalan Agency for Quality in Healthcare
There is a need for proper KPI systems (data + IT system) and a mandate to review the progress of interventions, guide decision-making, and identify areas that need management attention and resources.

Northern Netherlands: Embrace
A novel, population-based, integrated care model for the elderly community, aimed at helping them live at home as long as possible.With comfort and independence.This programme demonstrates an ongoing approach to encouraging and acting upon frontline staff feedback to promote staff engagement.

Scotland: REACT programme
A community service for the frail and elderly, which supports their desire to remain at home.Avoiding the need for unplanned emergency admissions.

ACT Consortium
The

Good practices
Regional Programmes Good Practice Factsheets 1 2 3 " Patients overestimate their level of knowledge and adherence behaviour.They are often unaware of the impact of their own behaviour on their health."

Primary challenge
Quality of life for patients who are chronically ill is a primary challenge for healthcare systems in EU member states.They are faced with a population that is rapidly ageing.Plus, the related burden of chronic illness growing to pandemic proportions.
In the EU, some 10 million people suffer from heart failure, 20 million have chronic obstructive pulmonary disease (COPD), and 60 million live with diabetes.These three conditions cost EU healthcare systems around EUR 125 billion each year.

Care coordination and telehealth services
Chronically ill people can now be treated in their own homes.One of the most effective ways to manage these patients is through the deployment of integrated care.Care coordination and telehealth services (CC&TH) provide patients with independence, freedom, and control over their health and lifestyle.Using remote management systems, and an integrated network of caregivers, patients can help themselves.Resulting in measurably improved health.

Promising results
Clinical studies around CC&TH are promising.These integrated systems can reduce the economic burden of chronic care, and maximise delivery of clinical support.Positive benefits of CC&TH are linked to how well organisational change is implemented.CC&TH has the potential to reduce hospital admissions, days in hospital and mortality rates.Despite the shortage of skilled professionals within European healthcare systems.

Barriers to deployment
CC&TH has been limited to pilot programmes and test installations.This is due to the complexity of translating evidence into practice.Deployment on a wider scale can only be achieved with new behaviours, routines and ways of working.Which require significant organisational change.So far, this has been directed at improving health outcomes, administrative efficiency and cost effectiveness.With a more proactive experience for patients and health professionals.
The Advancing Care Coordination and TeleHealth Deployment Programme (ACT) is the first of its kind.It is specifically designed to help overcome the barriers surrounding the large scale deployment of care coordination and telehealth services (CC&TH).

Foundation to overcome barriers
The ACT programme brings together a powerful Pan-European consortium.Which includes healthcare authorities, clinical experts, universities and industry partners.
In full support of the ACT programme, their efforts are directed at greater health outcomes for the chronically ill.

Beyond the trial setting
With the objective of identifying good practices, the ACT programme goes beyond the trial setting to assess the structural and organisational drivers, and their barriers, in five EU healthcare regions: Lombardy (IT), Basque Country (ES), Catalonia (ES), Northern Netherlands (NL), and Scotland (UK).Each participating region manages, via integrated care programmes, a group of at least 3,000 heart failure, COPD, diabetes, and comorbid patients.

Large-scale deployment of CC&TH
This 'cookbook' of good practices identifies best-in-class processes, structures and ways of working from these five partner European regions.Their insights, conclusions and recommendations, presented here, can be used to build a blueprint for largescale CC&TH deployment.

Evaluation of CC&TH deployments
In order to transfer the advantages of CC&TH services into routine clinical practice, it is essential to understand how to measure CC&TH service performance over time.It is also important to understand how the performance is driven, so that large-scale deployment can be guided by the configuration of these outcomes.
The ACT programme gathers regional data by using an Evaluation Framework and Evaluation Engine to help investigate three core questions: 1. How do CC&TH services work around Europe? 2. What needs to be done to make them work better? 3. How to deploy high-quality CC&TH services in new European regions?

Evaluation Framework
The Evaluation Framework refers to a holistic approach for assessing qualitative and quantitative (effectiveness and efficiency) performance of CC&TH services deployed in a local healthcare system.The framework of measurement indicators is structured in domains and subdomains over several areas, as presented in Figure 1.

Scope
Introduction and purpose of the ACT Project

Evaluation Engine
The ACT evaluation engine is responsible for capturing, monitoring and evaluating CC&TH deployment, using collected data.This web platform is able to publish, manage, and analyse on-line surveys, collect quantitative indicators via on-line and off-line templates, visualise stored information in multiple views and export data for further statistical analysis.The data is organised as key drivers (generic aspects of healthcare programmes determining their quality with respect to CC&TH) and key indicators (measurable healthcare outcomes used for assessing the drivers).These drivers and outcomes are important considerations when deploying CC&TH at scale.
Figure 3 presents the elements and context of the evaluation engine.The engine has been built on open source technology to offer a content management system with a database.It offers surveys, statistical analysis and interactive visualisations.The analysis performed by the domain experts has been added in the evaluation engine, and runs on the collected data.

The engine supports:
• A central point of authorised data entry and navigation in the system.
• An online survey tool, with questionnaires for different users, regions, and groups of indicators.• Import modules for survey data from the survey tool, and external sources.
• An import module for quantitative data templates.• A data model, which supports integration of data originating from different sources into a common database.• Raw data and descriptive visualisations of programme details, stakeholder perceptions, and comparison to a reference.• Full interactive implementation of the evaluation framework, which can be used for hypothesis generation of good practices.Figure 4 depicts an example of an interactive visualisation provided by the engine.
Visit the ACT portal at http://www.act-program.eu.Here you will find in-depth details regarding the methodology, evaluation framework, tools, additional background information, and news related to the ACT programme.
The ACT Programme Risk prediction modelling is a priority for the implementation of integrated care at a European level.

2.2
One of the key challenges generated by large-scale deployment of integrated care services is the need for dynamic health risk assessment, both at population level and in the clinical scenario.This helps to feed adaptive case management strategies.
Particularly those aimed at covering the evolving requirements of chronic patients over time.

Summary of results
All five ACT regions agree on the relevant role of population-based health risk assessment for regional deployment of integrated care.Please clarify the concept: There is also consensus on the use of population health stratification and not clinical or individual stratification.That is, health risk assessment tools generated from modelling the entire population of a given region (or geographical area) with a holistic approach.
The evolution of risk prediction modelling tools allows proper quantification of sensitivity/specificity of the estimations.

Regional risk prediction modelling tools
Our observations show the use of diverse regional risk prediction modelling tools.Together with the criteria for health risk strata classifications, these preclude comparability of risk pyramid distributions among ACT regions (Figures 5 and 6).Likewise, different problems associated with data management preclude appropriate comparisons of the recommended indicators.
We also identified issues with license binding constraints, and insufficient transparency of some computational algorithms.These may limit transferability of population-based health-risk assessment among regions.These two factors might also preclude the adaptation of current risk prediction tools.
Especially for the following evolving requirements: • Integration between healthcare and social services.
• Implementation of synergies between population-based and clinical health risk prediction modelling.
There are innovative strategies for individual health risk prediction models, which can be applied in different clinical scenarios.The ACT work has reported on the conceptual steps required to develop these strategies.Further studies are needed to evaluate their feasibility, added value and clinical applicability.Health risk assessment should employ a population-health approach.Both for risk prediction modelling and indicators.This is a priority for the scale up of integrated care at a European level.

Risk Stratification
Risk stratification is a statistical process to determine and quantify characteristics associated with an increased or decreased chance of experiencing unwanted outcomes.By identifying factors prior to the occurrence of an event, it is possible to develop targeted interventions to mitigate their impact.
Risk stratification provides a rationale for allocation of resources to those patients at highest risk.To improve outcomes, and change their cost structure, healthcare organisations must design population health management interventions that target high-risk, high-cost patients who need to be managed carefully and proactively.A critical aspect of this is understanding multimorbid conditions.
Risk prediction modelling is also useful for screening purposes.That is, looking for discovery of cases with non-manifest illnesses that may benefit from early diagnosis and cost-effective preventive interventions.
In the clinical scenario, risk prediction of well-defined medical problems can support health professionals in the decision making process.Moreover, clinical risk prediction may contribute to patient classification in the optimal healthcare tier.Helping to define shared care arrangements between primary care and specialists.The ACT Programme

FigureFigure 3
Figure 2 Data collection overview

Figure 4
Figure 4 Screenshots of the evaluation engine.a) Summary of the overall survey responses per question.b) Detailed overview of responses per question.c) The overall survey score for two selected regions.d) Overview of survey questions mapped to domains.e) Detailed scores per domain for two selected programmes.

Figure 5
Figure 5 Risk prediction modelling tools in the ACT regions* * The Northern Netherlands was not included in the figure because the integrated care programmes are not using population-based health risk predictive modelling tools.** The total population of Scotland is 5,295,000 inhabitants.

Figure 8
Figure 8 Stratification recommendations for best practice.
Advancing Care Coordination & Telehealth Deployment (ACT) Programme brings together a Pan-European consortium of leading companies, universities, hospitals and healthcare authorities.Initiated in February 2013, the 2.5-year programme defined good practices in care coordination and telehealth.The objective of this EU co-funded project is to overcome the structural and al barriers surrounding the deployment of coordinated care.
Figure 6 Risk prediction strategies and characteristics of data reporting for the study on top indicators in the five ACT regions