Whether Long Covid, Covid 19, H5N1 or future pandemics, there is a need for the strategic development of holistic crisis communication strategies in which AI should also play a crucial role.
1. Conduct even more targeted practice-oriented research
Not only this website, but also many other private websites, contribute significantly to advancing the understanding and treatment of Long Covid, ME/CFS and other chronic illnesses through their great commitment and wide range of content.
All of this knowledge transfer clearly shows how important it is to develop customized treatment approaches that are tailored to the many individual needs of patients.
Case studies and success stories
On countless (private) websites and channels worldwide, for example, on the Medical Medium website, you can find many “case studies” and practical reports. An incredible wealth of knowledge is available to the scientific community on all these pages and channels and could probably fill entire medical volumes for miles. From personal experiences, detailed genetic analyzes and laboratory parameters, health data, holistic approaches, ecosystems, successes, setbacks and recommendations. The presentations of specific cases and success stories show how personalized therapeutic approaches (could) work in practice and what positive results have already been achieved.
Everything can be found on the World Wide Web and so many valuable case studies could be created from it. All of these valuable resources are not only very informative, but in many cases also practical.
Take advantage of this invaluable potential, always keeping in mind that there are already more than 400 million people affected by Long Covid and ME/CFS worldwide.
2. more strategic crisis and prevention communication
The comprehensible presentation of these complex topics and the provision of clear, comprehensible instructions and recommendations can have a direct and positive influence on patient care.
400 million Long Covid and ME/CFS sufferers should be a strong driver to access all the in-depth resources already available on these websites and channels. To help as many sufferers as possible and possibly make better decisions so that they receive the best possible care.
Interdisciplinary approaches: continuous research and adaptation
Continuous collection and analysis of all this information helps to constantly improve treatment strategies and adapt them to the latest scientific findings.
The scientific community and clinical medicine should draw on all these valuable and extensive holistic considerations in order to be able to master the great challenges of Long Covid. An interdisciplinary approach that integrates medical, psychological and social aspects is necessary to develop truly comprehensive solutions. Only in this way can innovative and sustainable strategies be produced.
3. Best Practices - The labor market cannot afford it!
In view of the increasing problem of the shortage of skilled workers and the rising numbers of Long Covid and ME/CFS sufferers, the economy cannot afford not to draw on best practice experiences from those affected. Successful examples of those affected who work from home/remotely could serve as a role model and show companies how they can implement similar models so that long Covid and ME/CFS sufferers have a chance to be available on the market again with all their existing know-how.
Workplace adjustments: developing flexible working models that allow affected professionals to work from home without endangering their health.
Health promotion in the workplace: Introduction of health promotion programs tailored to the special needs of employees with chronic illnesses.
Research and education: Funding of research into the links between environmental factors and chronic diseases, as well as educational programs to raise awareness of these issues.
Environmental factors and infection rates, i.e. pesticide applications and Covid-19, because the correlation between high pesticide applications and increased Covid-19 infection rates raises very important questions about environmental factors and their role in the spread of disease. This suggests that reducing exposure to environmental toxins could have more than just a positive impact on public health.
Signaling pathways and common disease mechanisms
Pesticides, xenobiotics, EBV, Parkinson's disease, dementia, Covid-19, and micronutrients: the fact that the same signaling pathways are affected shows the complexity and the links between different disease patterns. Further research (SLC and ABC transporters, signaling pathways, viruses, pesticides and microbiome) into the links between genetic mutations, environmental factors and chronic diseases is crucial to improving understanding and treatment options.
4. more interdisciplinary digital platforms are needed
The use of digital platforms to promote knowledge sharing and collaboration between all those affected, clinics and researchers should be a priority. A great many sufferers are waiting for medical help and support, have no medical supervision and a great deal of valuable knowledge remains unused and unheard.
For those who have been lucky enough to be heard, interactive tools should be available that allow users to track and collaboratively adjust their symptoms and treatment plans.
Accessibility: Digital platforms should always be set up in multiple languages so that language barriers can be overcome and non-native speakers can also access valuable information and resources. Please don't forget that for many people, the internet is their only connection to the outside world and many are still without medical support and assistance.
An intensive build-up of community initiatives for long Covid, ME/CFS, multiple chemical and chronic illnesses TOGETHER (not separate approaches, because they are all closely “related” to each other): Building initiatives and networks that support patients and their families and provide them with access to the latest knowledge and treatment options.
Points to note and issues to watch out for in AI-supported crisis communication
Bias detection and avoidance: Use of techniques to detect and avoid bias in data and algorithms.
Bias detection tools: Tools such as IBM's AI Fairness 360, which were developed to detect and mitigate bias in AI systems, would come to mind “spontaneously” on this topic.
Secure Multiparty Computation: a cryptographic method that allows multiple parties to jointly analyze data without exchanging it among themselves. Or explainable AI (XAI) systems, which make their decision-making transparent and can explain how they arrive at their results. (here you would have to research specific systems – depending on the use case).
What else is there?
Google's “What-If Tool”: This is an example of a tool that helps developers understand AI models and analyze bias. Firstly, the “What-If Tool” provides an interactive visual interface (visualization) that allows to visualize and explain how models work (1) This is particularly important in health crisis communication as it helps decision makers and the public to understand the basis for certain actions and decisions. The tool can be used to run through different scenarios to simulate the impact of actions and identify the best strategies and make informed decisions. The tool also helps us to recognize and avoid bias in models. This is particularly important in health crisis communication to ensure that communication is fair and unbiased. The “What-If Tool” can be integrated with existing models deployed on the Google Cloud Platform or other platforms (1). This enables seamless integration into existing systems and improves the efficiency of crisis communication. This tool also allows us to analyze and adapt in real time. This is of course of paramount importance, especially in the rapidly evolving health crisis (1). It helps us to react quickly to new information and changes and then adapt our communication accordingly.
Let's imagine that a health department uses a model to predict infection rates and plan vaccination campaigns. With the “What-If Tool”, they could then run through various scenarios to see how infection rates would develop under different conditions. They could also analyze the impact of planned vaccination campaigns on different population groups and then adapt their communication accordingly. Google's “What-If Tool” could well be a comprehensive and flexible solution for analysis and communication in health crises/pandemic situations, promoting transparency, explainability and real-time adaptation.
Of course, all this would only be possible if strict security measures were implemented to protect the integrity of the data and prevent unauthorized access. Continuous monitoring and adaptation of the measures is crucial to keep pace with ongoing developments and threats on a daily basis.
Please always keep algorithmic bias in mind, as it can have a significant negative impact on our crisis communication.
Data bias, stereotype reinforcement and misinterpretation of data/data sets: Algorithms learn from historical data, which is often biased. If this data contains systematic inequalities or prejudices, the model will adopt these biases.
In our crisis communication, this can then lead to certain groups (which would also correspond to a risk group in the health sense) being systematically disadvantaged, e.g. through less access to important information or unfair treatment in the distribution of resources, e.g. vaccines, off-label medication and much more. In terms of “stereotypes”, this means that this can lead to discrimination and inequality in the respective crisis situation, e.g. if certain population groups are classified as less trustworthy or less needy. AI algorithms could also interpret data incorrectly, especially if they have not been optimized for the specific context. This in turn can then lead to wrong decisions and miscommunication of the crisis, which could potentially exacerbate our (health) crisis instead of mitigating it.
Have all local contexts really been considered?
We have to ask ourselves very specifically whether our AI model has only been developed and used globally or whether it already takes local, cultural and social contexts into account. If these important references are missing, it can have massive and harmful effects on our local communication strategy, which is then not tailored to the specific needs and circumstances of the affected community.
One dramatic consequence of this could be in the provision of medical care: Populations that are considered less at risk based on historical data may have less access to vaccines, medicines and medical care. Vaccination campaigns could prioritize the distribution of vaccines to populations that are considered at high risk, while completely neglecting others who are also at high risk.
Or even in the area of protective measures, i.e. protective materials such as masks, air filter systems and disinfectants could be distributed unevenly, making certain groups more susceptible to infection.
Or let's imagine what happens in the area of prevention when only insufficient information is available. If algorithms, for example, make false assumptions about the vulnerability of certain groups, this could then lead to these groups being less informed about preventive measures or to incorrect modeling. This can then lead to the spread in certain (local) regions or population groups being completely underestimated, making an already existing crisis even worse.
Finally, the EU's Digital Services Act is a corresponding legal framework that aims to create more transparency and accountability in the handling of digital data and AI.
(1) https://cloud.google.com/blog/products/ai-machinelearning/introducing-the-what-if-tool-for-cloud-ai-platform-models