Validation of the BEAMER Model – factors influencing adherence behaviour 

The BEAMER project is an Innovation Medicine Initiative (IMI) funded public-private partnership that focuses on understanding factors that impact a patient’s adherence across various disease areas. Our primary objective is to develop a model that identifies patient needs and preferences, which in turn enables patient-specific support for improving treatment adherence. The BEAMER project’s proposed model has reached the maturity where real-world validation will start.

Nine healthcare organisations and universities are participating in the BEAMER project as pilot/validation sites. Activities are just to begin, with exciting and challenging times ahead. In this blog post, we share an overview of what we wish to achieve when we validate the BEAMER Model. The ongoing planning process will help the BEAMER team get more information about validation strategies and choose methods that are feasible for all the pilot sites.  

The reasoning behind validating a model, such as the BEAMER Model, is to collect evidence (data) to assess the proposed model’s performance, meaning that we want to understand if the model consistently delivers what it is designed to deliver under different disease conditions or treatment settings (hospitals, in-clinic or in-home/online settings).

The validation activities will help us understand the accuracy of the output of the BEAMER Model. In other words, they help evaluate how well the BEAMER Model facilitates the identification of patient needs and preferences for improving treatment adherence. These activities also investigate the effects on adherence of providing patients with personalised and targeted support based on the outcome of the BEAMER model. By validating the BEAMER model, we provide empirical proof that it can successfully improve patient adherence and therefore be implemented in practice. 

The BEAMER model aims to be disease-agnostic, meaning that it is useful or applicable to any condition. For the validation of the BEAMER Model, we will use data from six selected therapeutic areas; cardiovascular, endocrinology, immunology, neurology, oncology and rare diseases. The data are collected from nine pilot sites in six countries: Norway, Portugal, the Netherlands, Spain, Italy and Germany. In addition, we will use real-world data collected from other partners to validate the BEAMER Model. This will help us to confirm that the model can cope with the wide range of heterogeneity of the real world and successfully explain adherence behaviour regardless of the conditions or settings. 

We are carefully considering how to choose wisely among different research methods and designs to balance advantages, concerns and risks. For example, some research designs are costly and time-consuming, and the BEAMER project has a pre-decided budget and a period of approximately a year of validation activities of the BEAMER Model. The goal of the planning process is to review every step of the validation, to fully appreciate the opportunities and be aware of all possible risks. This allows to implementation of mitigation measures to maximise success and decrease the negative impact of any identified risks.  

During the planning process, we actively engage with relevant healthcare organisations, healthcare providers, and other key stakeholders. This collaborative approach helps us gain a comprehensive understanding of the pilot sites, their eligible patients, and specific validation activities at the different pilot sites. By doing so, we can assess the specific capabilities and select the most suitable research methods and designs to validate the BEAMER Model. 

Ultimately, to validate the proposed BEAMER model, we will implement practical, effective and appropriate research methods and designs at the pilot sites. This approach aims to minimize risks and maximise opportunities at the test activities with the nine healthcare organisations involved in the BEAMER project! 

Authors: Silje H. Henni, Anne Moen, Line Løw (UiO), and Jim Ingebretsen Carlson (PredictBy)