Introduction video and May HL7 Connectathon information
What is the basic idea?
We need to know how to reduce the negative impact of COVID-19. The sooner we know what works the sooner we overcome. We learn what works by carefully evaluating all the relevant research and observations. We have a lot to learn and we have many people who know how to do this careful evaluation to learn it.
One way we accelerate our knowledge about COVID-19 is to share what we learn and coordinate among the many working to develop our knowledge. This way we reduce excess duplication and cover more needs. Part of the COVID-19 Knowledge Accelerator is for people creating the systematic reviews of our research to share and coordinate their efforts. Let's help each other learn how to do it faster while we are doing it.
Another way we accelerate our knowledge about COVID-19 is to use computers in advanced ways like what may be called artificial intelligence, machine learning, machine intelligence, natural language processing, supercomputers, or high-performance computing. Computers are not able to do this yet because we have not yet trained a machine how to evaluate scientific research to determine what works. The people doing the systematic reviews can train the machine on what they do, and the machine can be developed to help the people do it faster. Let's help each other learn how to do it faster while we are doing it.
For a 2-page summary of this idea see COVID-19 Knowledge Accelerator: Help Us Help Us Overcome COVID-19 Impact at the Speed of Thought
Where do we start?
Our suggested first project is a scoping review of the existing research about COVID-19 to systematically identify the research that includes people with COVID-19 and catalog within this research the Populations that were studied, the Exposures that were reported, and the Outcomes that were reported. This scoping review effort can continue for as long as new research is produced about COVID-19 so there is no anticipated ending date. Results can be reported continuously as the work is processed.
How can I help?
Click here to sign up for a team to work on this task. You can sign up as an individual or as an organization to register your offer to help. As the team is assembled, we will figure out how precisely to get started.
- Identify research about COVID-19 — We can start with the COVID-19 Open Research Dataset (CORD-19) but someone needs to (1) manage the effort to know which articles have been considered for our scoping review and which are in the queue, and (2) seek additional research sources to expand the effort beyond CORD-19.
- Select research that is about people with COVID-19 infection (Human interpretation) — People need to look at each article and determine whether it is or is not about people with COVID-19 infection. During the project we will develop computer support for this article selection effort so the system for doing it will change. But at the core effort we need people to read the article for this basic judgment whether or not to include the article in this scoping review.
- Develop software support for article selection — Software may include the user interface that people use to report their inclusion/exclusion judgment, the management system for coordinating inclusion/exclusion decisions across multiple people and multiple articles, and the use of machine intelligence to increase the efficiency of the article selection process.
- Describe the Populations, Exposures and Outcomes reported in the selected articles (Human interpretation) — People need to describe these concepts in precise and unambiguous terms from a standard set of terms so we can apply the same way of cataloguing the research across all the research. Although this is difficult at first it will get easier and faster as we apply software support and use what we learn from doing it.
- Develop software support for specification of Populations, Exposures and Outcomes — Software may include the user interface that people use to report their definitions of these evidence variables, the management system for coordinating the effort across multiple people and multiple articles, and the use of machine intelligence to increase the efficiency of the evidence variable specification process.
- Reporting the results — We will need to make the results of this scoping review available in multiple ways, both at the data level (so people can find specific research of interest as defined by Populations, Exposures and Outcomes) and the summary level (to report the scoping review overall progress)
- Project coordination — There is a lot to do.
- Project funding — As of March 27, 2020 there is no specific funding for this project and there is no specific plan for seeking funding for this project. There is no solicitation being made right now and no formal mechanism to process financial contributions to accelerate this effort. This could change as the problem is large and universal and there is clear value in accelerating our ability to overcome the problem.
- Project awareness — Modify the website pages describing this project so it is easier for everyone to understand, and spread the word to raise awareness.
Who is behind this?
Soon we should be able to answer this with a big list, derived by compiling the lists of everyone signing up for any of the project support roles noted above. As of March 27, 2020 this is an idea offered by Brian S. Alper, MD, MSPH, FAAFP, FAMIA. Ideas are derived by building on the contributions of others and this idea is developed by Dr. Alper's roles or interactions with the following groups (in alphabetical order):
Email Brian Alper: firstname.lastname@example.org
COVID-19 Knowledge Accelerator Google Drive
Meetings on Mondays at 4pm (EST): Calendar Invite