Amsterdam Public Health

(1)Solving disparities using digital technology

Hosted by Societal Participation & Health, Digital Health and Global Health

The central question during this session was ‘How can we reduce health disparities by using digital technology in healthcare, instead of currently increasing it?’ After several presentations from the perspective of the three APH research programs and a panel discussion with a diverse group of stakeholders, the messages that stuck out the most were:

1. data should be as open as possible and as closed as necessary,
2. central data storage and data donation is the future,
3. 1 out of 5 person in the Netherlands currently has insufficient access to a computer and thereby cannot keep up with society,
4. e-health can help decentralizing care, which is important for rural communities (e.g. in Eswatini),
5. data sharing is especially important for rare diseases,
6. the global impact of digital health should be considered, especially taking into account the different health systems and digital inequalities worldwide,
7. future studies should focus on involving vulnerable groups (e.g., migrants and people with low digital health literacy).

In conclusion, digital health has the potential to improve health globally, but there are still some challenges ahead for researchers, policy makers and practitioners.

(2) Intersectionality in digitalization

Hosted by Personalized Medicine

In this session the central question addressed was: How can we develop ‘fair’ studies, interventions, and policies in a digitalizing world?

The first part of the session included three short presentations to give an introduction into intersectionality and inequality in digitalization. Petra Verdonk started off with a brief outline of intersectionality and social (in)justice in digitalization. She was followed by Carlien Straat, who elaborated on the process of designing the eHealth application “IkHerstel” and the experiences with creating individualized feedback and notifications. Finally, Nicole Goedhart explained several factors influencing digital inequality at the individual level, as well as the impact of structural digital inequality on a more global scale.

Participants were then invited to participate in an “obstacle run”: groups were assigned to list obstacles for addressing intersectionality in different phases of research (project development, data collection, data analysis and dissemination of results) and to define solutions.

Participants concluded that one of the most important solutions to intersectionality is to strive for diversity in the academic workforce. But already we can do more to pay attention to diverse groups and cater to their needs. Be inclusive, creative and involve your target population!

(3) Evaluation of digital innovations in the clinic?

Hosted by Health Behaviors & Chronic Diseases and Quality of Care

Susan Oudbier – key messages

1. Digital tools can be used in all steps of the patient journey
2. All stakeholders should be involved in the development of digital tools
3. It is important to use terminology that is adapted to the user (patients) while evaluating a digital tool
4. Make sure you use previous knowledge on common barriers and facilitators when implementing a digital tool

Judith Jelsma – key messages

1. An European project on the evaluation of the use of tools for teams in the centralized hospital care and chronic disease management has recently been granted

2. Within the Lofit study, the platform “C-board” is used to communicate between health care professionals and the patient

Femke van Nassau – 6 steps of process evaluation development

1. Aim of the evaluation
2. Make a logic model with stakeholders
3. Choose your process indicators
4. Choose your methods (when & who)
5. Planning
6. Analysis and reporting

Recap after group assignment:

• It’s a lot of work to prepare your process evaluation
• Tip - Make a big matrix at the start, and then start selecting the most important elements (you cannot measure everything)
• There will be a more extensive workshop on process evaluations early 2023 – details will be communicated through the APH communication channels.

(4) Wearables

Hosted by Aging & Later Life and Mental Health

During the ‘Wearables’ session, Willem van den Brink from TNO research introduced the topic with a presentation on his research project on wearables. This presentation was followed by short presentations by researchers of the A&LL and MH research programs who shared their knowledge and experiences with using wearables in research and healthcare. Finally a panel discussion with the audience was held to discuss the do’s and don’ts. The main aim of the session was to share insights and experiences that can hopefully help other researchers when starting research using wearables.

Using wearables can facilitate the transition from ‘sick care’ to ‘proactive healthcare’, since they can measure outcomes continuously and in the real world. Before the use of a wearable, the measures where only episodic and in a clinical setting.

Before choosing a wearable for research, it is important to identify what you want to measure: You need to know the values of the patients. These are often qualitative outcomes and have to be translated into measurable outcomes; measurable with a wearable.

Additionally to the aspects above, choosing a wearable that is usable for the patient is also very important. The wearable should be easy to wear, easy to use (e.g. charging, synchronizing) and there should be possibilities to personalize the wearable.

(5) Federated Learning

Hosted by Methodology

Wessel van Wieringen (Associate Professor, Epidemiology and Data Science) introduced the topic Federated Learning, and explained the different types of learning that can be used for situations with limited access to data but where summaries of data are available (federated, sequential, and transfer learning).

Judith ter Schure (Research Associate, Clinical Epidemiology, Biostatistics and Bioinformatics) started her talk with an interactive game with dice and gambling. She presented how she and her research team had applied federated learning in ongoing clinical trials during the pandemic. Very interesting!

Hans Wenzel (Junior researcher at Integraal Kankercentrum Nederland) also showed a successful application of federated learning using cancer registry data from three countries. He discussed the practical challenges of each component of federated learning: infrastructure (e.g. each device has to be turned on at the same time), data (e.g. datasets must be aligned), and algorithms (e.g. can be time-consuming).

Afterwards, there was an interactive discussion between the audience and speakers, concluding that federated learning can be very appropriate to overcome challenges related to data sharing.