3 years ago, when everyone was still thinking about a beer with corona, Professor Aura Timen (RIVM, Radboudumc & VU University Amsterdam) and Associate Professor Teun Zuiderent-Jerak (Athena Institute, VU University Amsterdam) spoke about the future of developing guidelines. ‘The many discussions about vaccinations and thus the declining willingness to be vaccinated was a major concern,’ recalls Aura. ‘How can we integrate citizens’ questions into vaccination guidelines?’ they wondered. That was the driving force behind a study in which the question “How can citizens ‘and professionals’ experiential knowledge and values be integrated into vaccination guidelines?” was central. Following the launch, COVID-19 was consulted and the focus quickly shifted to the as yet undeveloped guideline for COVID-19 vaccinations. Now the researchers are looking back. What have they learned from this process?
Where do you find the voice of citizens?
Lea Lösch (Athena Institute, VU University Amsterdam), researcher and social scientist with knowledge of artificial intelligence, will also participate in the conversation. “She is indispensable in this research,” explains project manager Teun.
‘In the Netherlands, we are pretty good at giving patients a voice in general guidelines. Through a patient association, for example, you can find out what questions there are. But the situation is fundamentally different with vaccination guidelines. Then it is not about patients, but about citizens and there is no patient association that knows what is going on.
‘So where do we find that voice from the citizens, how do we reveal what is important to include in the directive? Artificial intelligence proved to be the answer. ‘
Simple answers to difficult questions
Aura: ‘Citizens may have questions and concerns that a scientist has no idea if only scientific sources are consulted. If the widespread concerns and relevant experiences of citizens are lacking in a vaccination guide, then you have a problem. ‘ ‘Sometimes a question does not seem relevant from a medical point of view,’ Teun continues, ‘but if many people are concerned about something, it is actually important to include that care in the guide. You want GGD doctors and general practitioners to explain as clearly as possible whether the corona vaccine alters your DNA; something many wondered when the first COVID-19 vaccines were given. With a new mRNA technique like this, we as a society would normally be reluctant to use it so widely. So it is understandable that it raised the question ‘.
How do you find relevant experiences?
‘There were discussions on different platforms’, Lea explains the working method. »On social media, especially Facebook, but also via RIVM Information Point for example. Our method, which we develop together with the VU’s Social AI group, distills experiential knowledge, which we call citizens’ questions and concerns, from online messages. You create a kind of automated filter that can quickly identify descriptions of experiences in large datasets based on characteristics and text patterns. With the answer “LOL” (laughs out loud) or a single emoticon, we can not do so much. But with an explanation of concerns. ‘
‘So we were able to identify the relevant experiences with artificial intelligence. Step 2 was to analyze what the discussions were about. What are common themes? What questions are there among a large number of people? ‘
‘The deployment of artificial intelligence is therefore not a matter of sitting back and seeing what comes out. This research also requires a qualitative analysis and a lot of self-reading. ‘
Healthcare professionals are also concerned citizens
“What surprised us,” says Teun, “was not so much the themes themselves. Several concerned citizens are very often health professionals. The idea that health professionals know how a vaccine works, or have no worries, is therefore not always They were the first to receive the vaccine and had many questions about it. The division is less sharp than we thought; as a doctor you can also just be a concerned citizen. Or – as you also saw with HPV vaccines, for example – a worried parent with questions for your child. ‘ Lea: ‘For the COVID-19 guide, we used different data sources and thus discovered that there were different issues. The risk of thrombosis, the rapid hospitalization procedure, the mRNA technique, post-vaccination contamination: We were able to make these themes transparent to the guideline developers. ‘
More widely applicable
Teun: ‘We have shown how to use artificial intelligence to extract experiential knowledge from large amounts of messages on social media and RIVM channels.’ ‘The big challenge we face now,’ Aura continues, ‘are the more complex issues that concern citizens and that cannot be included in an extra section of the guide.
‘Now that the project is halfway through, we have shown that it is therefore possible to integrate the citizen’s voice in the guideline development.’
Teun: ‘The application is even broader than just vaccination guidelines. Consider oncology guidelines. There is also some discussion about breast or prostate cancer. With our method, you discover what problems are alive at the moment. ‘ “Exactly,” says Aura, “the use of artificial intelligence is an exciting development that can not be stopped. We are really groundbreaking, because we have not found any other research that does this yet.”
Technology alone is not enough
“So it’s possible,” Lea says proudly. ‘Even when the time pressure is high! We are on the way to developing guidelines that are more in line with people’s concerns and experiences. ‘ ‘The only comment I have left,’ says Teun, ‘is that you still cannot find the voice of all groups. Vulnerable groups are not always digitally active. People with lower health skills also tend to be less digitally proficient. So their voice is not well represented. ‘ “It is true,” Aura concludes, “technology alone is not enough. More is needed to integrate experiential knowledge. But we have taken a huge step in the democratization of knowledge. So as researchers, we look forward to seeing what else can be done with this in guideline development. ‘
Innovation of guidelines
Guidelines describe (part of) a care process, where it is determined what is necessary to provide good care from the client’s perspective. The challenge is to keep these guidelines up to date and practicable in the light of the rapid development of international knowledge. In addition, practical data are increasingly available that can be used to update the guidelines. Therefore, we fund 9 projects that provide knowledge on improving the timeliness, accessibility and practical applicability of guidelines. This project is one of them.
> See the 9 projects
Editorial Marie Stegenga, final editing SunMw