Inspiration
We were approached by frontline medical staff and health officials to help them improve epistemic resilience in the medical community. hey are asking Edgeryders for help. Marco – a doctor – and Simona – a public health manager – both in Italy, are on the frontlines, so I am going to do the writing here.
It turns out that the health care system, in many countries (but not in all), got overoptimized over the years. It was honed to be efficient in the context of a predictable environment, with a known mix of pathologies. Over time, it discarded redundancies and second lines. It is now badly equipped to deal with unknown unknowns and black swan events.
Overoptimization also happened in agent space. Not only hospital wards not running at full capacity were closed: the people who got promoted to leadership positions in public health tended to be narrow specialists, who build their career on, say, hepatitis B. Their epistemic horizon is also made for stable environments. They do not know how to move in an epistemic space where evidence is absent, or ambiguous; they are not used to questioning evidence, and dislike doing so.
The result was a situation where evidence from China was not sifted and interrogated, but mistrusted. As a result, many countries in Europe underestimated SARS-CoV-2 (“mortality is like a flu’s”, which is of course true if you have the resources to treat every affected person to high standards). At this point, in Marco’s reconstruction, the whole system went into emergency mode. The emergency, plus the epistemic rigidity, mean that the protocols that are being communicated to doctors on the frontline are shaky at best, and “magical thinking” at worst. Ward directors and even government guidelines (Italy) will tell doctors to “keep working on patients even if you test positive, as long as you asymptomatic” (!!!), or to wear makeshift masks if you do not have any proper ones “because they are better than nothing”. Intensive care wards are being erected in a few days, by reconverting space from, say, psychiatric ones. Additionally – and this is very, very serious – there are strong suspicion that actual data are being hidden or misrepresented. Hidden: I have some personal evidence, as the open data community in Italy (@napo, @piersoft) is desperately asking regional authorities to release the data with open licenses and proper documentation. Misrepresented: it turns out that, in places like Bergamo, only 10-25% of the extra mortality with respect to seasonally adjusted average is explained by the official deaths by COVID-19. It’s hard to really figure out what is going on, also because shifting criteria of access to testing (Italian) mean that there is no way to even observe the statistical trends.
Epistemic resilience means, instead, being critical of evidence, and disobeying if that saves lives or unnecessary suffering. In the case of masks, we do have evidence: a makeshift mask is to a mask what a blanket it to a parachute. It is not “better than nothing”, it is exactly the same as nothing. A makeshift hospital ward might have intensive care machines, but it risks infecting patients in nearby wards unless air circulation is addressed seriously, with proper ventilation and filters, which is a feature distinguishing makeshift wards from proper ones. Plus, these things might create a false sense of security – again, we have a literature on this, more from finance and from war medicine – and distract us from taking measures that, while far from optimal, are indeed “better than nothing”. Example: isolate people at home, and send them cheap DIY ventilators or some such.
Marco and Simona think that epistemic resilience for the medical community means three things.
Build a trusted knowledge base on practices to deploy in the face of the evidence we have, and that we do not. Right now, doctors are getting advice off YouTube channels, and randos on WhatsApp. Some solid stuff is there, but also people who say the government has created SARS-CoV-2 to kill all the NoVax. I think of this knowledge base as the kind of document that @matthias and @lucasg are good at maintaining. This is necessary, because in the face of bad instructions doctors and nurses need some evidence-based good knowledge to push back. This need is why they are all on WhatsApp. Marco thinks this would consist both of articles and of stories from the frontlines. Important: this needs to be multilingual, because “the doctor on the field in Heidelberg is going to look for evidence in German”. A lobbying effort, informed by this knowledge, to try to get the health care system to get unstuck. A way for these doctors, nurses, etc. to stage a symbolic protest. “We cannot strike during an emergency, but maybe we could carry a symbol, or something.” So, they asked me if Edgeryders could help, mostly with 1. A community is assembling around them; translation efforts could be asked to Translators Without Frontiers or something. But we could host the community, and put in some editorial/community management work, maybe?
What it does
It is an app that acts as a bridge between messaging applications that are used by frontline workers for informal fast communication to coordinate and share information( social media, telegram and whatsapp) and an online platform. People chat with each other on these channels where a chatbot then relays and aggregates the contents onto a platform. On the platform where human editors with medical skills, e.g medical studnts; can then put together a summary of what has been said and send it back to the people who participated. Over time we can use a machine learning solution to do this but the need is immediate so we use the fastest solution right now.
How we built it
In close dialogue with members of our online community platform who themselves are frontline workers. We built the a chatbot application written in Vue.js and an API based on AIML adapted to JSON. The goal of the coding work was it: Create a reactive lightweight Vue front end for chatbot interactions Establish a syntax inspired by AIML and using JSON to easily write chatbot instances. Provide an interopable API that works on the web and with messenger applications Interface with the Edgeryders Discourse platform to create and retrieve topics or users.
Challenges we ran into
We did not have enough time to work on the user experience, only on the basic code for the app. If we have a UX designer we can get this done in a week.
One of the challenges that remains is figuring out what business model and strategy would make sense for us in order to generate revenue to maintain and keep improving the software, train people etc. Do we do a software as service solution? Do we focus on just installing the software and teaching people to use it? Something completely different? We have been winging it all along. Another challenge we have is design and communication. Miraculously people have been using our platform and tools in spite of our never having had a design team. And while we have a social media channel, we never managed to find the resources to do PR or marketing.
Accomplishments that we're proud of
People are using our software and community space to do a wide range of things from activist initiatives, to coordinate building hardware businesses to co-create and deploy public policies. We have developed our technology in collaboration with leading universities and are consistently highly evaluated by scientific committees. Our work has been cited in national and international media in spite of our being terrible at promoting anything we do. The business is generating revenue in and we don't have any debts. The community of people using our platform has grown to 5000 in 80+ countries.
What we learned
It's important to build things that you really need when the need arises. If you solve a problem for yourself it is likely you will build something that is useful to others. The other thing we have learned is that people tend to underestimate the human and social aspects of technology use and focus too much on the code. You really need to have community and excellent community management skills for your work to soar. We are lucky to have had this as part of the core team from the beginning.
What's next for Edgeryders - Refine/further develop our existing stack of tools for distributed collaboration to serve everyone struggling right now because of disrupted workflows, but especially frontline health workers and volunteers
Edgeryders has been remote by default since we started. We initially developed this stack of tools, operational practices and culture for ourselves - starting with off the shelf open source software which we then did small hacks on to adapt them to our needs. Then when the crisis hit we have been hacking the tools we already have and know work to support our community's medical heros in their work.
Due to the current crisis, we started getting many requests to help people set up or improve their remote collaboration practices. We predict that this pandemic will have a lasting impact on how we all work. Including what it means to engage in entrepreneurship - with fundamental shifts in how business ventures are born, operated, financed and governed. Thriving in the new landscape means looking beyond the next fancy tool. It requires a conscious effort to get socialised in the new digital collaboration contexts available to us. There is an explosion of how to guides, manuals etc. But what they do not understand is that online distributed practices are not a poorer substitute for old-school ways of working. They enable entirely new work lifeforms to evolve. So we need to train people on what this means and help people to use the tools that are available, and adapt them to their needs.
Log in or sign up for Devpost to join the conversation.