User:Ian Furst/Knowledge sharing

This year, the theme of Wikimedia movements annual summit Wikimania will be "Stronger Together: Wikimedia, Free Knowledge and the Sustainable Development Goals".

Wikipedia is an online encyclopedia, free of commercial interest, created and maintained by volunteers, and guided by the principals of neutrality and free-use. It's reach, relevance, and accuracy have made it the most frequently used source for health care knowledge on the internet, and an invaluable resource for health care providers.

The rash of Dengue

Wikipedia's dengue fever article is an excellent example of the resource provided. Dengue is a life-threatening neglected tropical illness, endemic to Southeast Asia that affects between 50 million and 528,000 million people a year. It is grossly under reported, often unrecognized, and hits hardest the developing world.

Looking at a multi-year graph of article visits, interest is clearly greatest during rainy seasons, when Dengue is most active. In 2018 alone the English language version of the article had over 1.5 million visitors (over 4,200 per day), was monitored by 333 volunteer editors, and had 52 updates. An article on the topic is available in 78 different languages, partly through a workforce of volunteer translators.

This single article acts as a vital resource to share knowledge both with the public and health care workers around the world. Looking just a little further than Dengue, a short list of Wikipedia medical articles (pneumonia, tuberculosis, cancer, typhoid fever, dengue fever, measles, hypertension, major depressive disorder, influenza, hepatitis C), collectively average 1.2 million hits per day, rank as some of the greatest health care challenges facing the developing world, and are just 10 of the more than 30,000 English articles monitored by Project Medicine volunteers. Health content on Wikipedia exists in various degrees in another nearly 300 languages.

Challenges

Internet access

Global internet access.

Despite the richness and reach of Wikipedia, it still has some significant limitations. The greatest is that internet access is not universal, especially in developing nations, conflict zones, and rural areas where knowledge sharing could have the most meaningful impact. As of April 2019, only 56.1% of the worlds population had internet access, and 41% of peoples in developing countries. This equates to more than 3 billion people unable to share knowledge quickly and efficiently. Wikipedia has identified over 800,000 schools without access to the internet, which would benefit from Wikipedia offline.

In health care training the problem is more acute. Much of the scientific research published is in English. Without access to current research, or online translation tools, it is nearly impossible for non-English speaking health care providers in developing nations, to be aware of, let alone implement current best practices.

Illiteracy

Compounding the problem of internet access is illiteracy. With more than 15% of the world's population unable to read, a text-based knowledge resource such as Wikipedia is of little use. This leaves more than 750 million people (mostly in developing nations, and women) unable to access information even when it is made available. India for instance has a 26% illiteracy rate. With a population of 1.4 billion, that equates to 350 million people, in a single country who cannot access Wikipedia because it's text-based.

Actionable data

Last, while Wikipedia is a passive tool for public good, it is not particularly reflective. Data analytics are limited and, for the most part, not shared. Currently, aggregate visitor data is visible on a global basis for each article, or for all of Wikipedia by country. There is no easy means to drill down to data by language, region and article, nor any means to alert editors when articles are getting high use such as during an infectious disease outbreak.

Solutions

IIAB deployment in Peru with Global EPMA

Wikiproject medicine's mission is to share up-to-date biomedical knowledge with all people. That mission furthers the Sustainable Development Goals (SDGs) 3.4, 3.5 and 3.7. Beyond Wikipedia's medical content, it would also further SDGs 4.2, 4.3, 4.A, 5.6, 5.B, 6.B as well as others where knowledge transfer is required. But as discussed above, the platform excludes roughly 60% of the worlds population, with a heavy skew that negatively affects rural, developing, and female populations. Three projects are discussed;

Internet-in-a-box

Internet-in-a-Box (IIAB) is a low-cost digital library, consisting of a wireless access point with storage. Nearby users can connect to IIAB with a computer or mobile device, then surf a locally stored version of Wikipedia (in any language), as well as other material such as Open Street Maps and the CDC website. It is primarily for distribution in places where access to the Internet is limited or controlled. The device includes software which creates a Wi-Fi hotspot that offers closed platform access.

In practice, this means that up to 32 users who are within a couple of hundred meters of the hotspot can connect to the device and access or download the content that exists on the device: Wikipedia slices, medical knowledge, videos and books. IIAB is not intended to further connect users to the Internet or to content, beyond what is presented in the Box itself. It is ideal for remote schools and health care facilities who lack connectivity to the internet.

At present, the financial model for this technology is unclear. Approximately 300 have been produced and distributed to various sites around the world. The devices cost about $40 USD per unit to make, with that funding coming partly from donations to Wiki Project Med Foundation, and partly from those who are requesting a device. They have proven useful in school settings and the health care environment. One of the goals is to partner with NGO's that work in settings that lack proper internet access, and have them support funding and distribution.

VideoWiki

Dengue in simplified English

Dengue in Hindi

Wikipedia has long struggled with the move to video. Even though it is well recognized that video is a preferred method to digest information, and it increases accessibility to those who are illiterate there are some major hurdles. The greatest, is that information on a video cannot be easily edited, which results in many videos on the same subject with conflicting perspectives. This is the antithesis of the Wikipedia pillars to allow anyone to edit content, and build consensus.

Recently, a Wikipedian from Mumbai came up with a solution called VideoWiki, with a mission to visualize the sum of all of human knowledge. His solution, rather than creating video for import into Wikipedia articles, created script pages with embedded content that could be programmatically transformed into video. His engine, imports the media for each slide (or section of video), and using text-to-speech cloud assets to read the script. This has the added benefit of allowing the translation task force to work on video, the same way they do for articles. It also makes video created quick and simple. A 3 minute video can be created in 1-2 hours, and translated in minutes.

To date, the top 10 medical articles have summary videos, and the dengue fever video has been translated into 9 different languages. VideoWiki will be presented in Stockholm at Wikimania2019 in an effort to build consensus with other Wikipedians about its use and deployment. Funding from the Wikimedia foundation will be required for further platform development.

Knowledge sharing beyond content

Wikipedia, while an open platform, does not make public most of the usage data. Article statistics are view-able based on the API's of a few technical editors. While this is necessary from a privacy perspective, a shift to share limited, and partially aggregated data would allow Wikipedia to contribute to SDG 3.D;

"Strengthen the capacity of all countries, in particular developing countries, for early warning, risk reduction and management of national and global health risks."

Dengue pageviews with spikes during epidemics

Many countries don't have the infrastructure to create meaningful early warning systems. Wikipedia, on the other hand, regularly sees spikes in article usage that ties tightly to outbreaks. Limited regionalized aggregate data (which would preserve privacy) from Wikipedia could be leveraged to better identify early outbreaks. Alternatively, a partnership could be considered with a responsible third party (ICRC, Gates Foundation, WHO) to parse and distribute partially aggregated data.

Previously mentioned was the annual pattern seen with dengue fever. Each year epidemics of dengue coincide with spikes in article use. What if this data was available on a regional basis so that humanitarian organizations could recognize local outbreaks, and deploy countermeasures to the disease. What if automated notifications where attached to rare disease articles, so that when local communities visit beyond their typical volumes, organizations can take a closure look?

Take for instance, an emergency physician in rural India. That emergency physician could have a patient die, and list sepsis (widespread infection) as a cause of death. As a single death the information is unactionable. But if 2000 people in the same area, simultaneous trip across the same article for tularemia there is a trend to identify. This could weigh heavily on SDG 3.3 ("By 2030, end the epidemics of AIDS, tuberculosis, malaria and neglected tropical diseases and combat hepatitis, water-borne diseases and other communicable diseases") by not just sharing information, but by alerting humanitarian organizations, and pushing out information early in an epidemic.

Beyond strengthening the capacity for epidemic mitigation, there is also a role to play with micro-targeted knowledge sharing. For instance, infant mortality in many areas can be decreased with simple measures such as kangaroo care and having a birth attendant. Humanitarian organizations could target an area for knowledge sharing, and monitor the uptake based on site views.

This idea has not been brought forward to the community yet, but if well received will require funding to build monitoring tools, and relationship development with partner organizations who can act on information.