Competition closed, winner announced
The problem: Pandemics, epidemics and outbreaks of infectious diseases are an escalating threat globally. When large scale epidemics or pandemics occur, often our systems cannot deal effectively with them, resulting in preventable deaths and serious illness, as well as catastrophic disruption to society, as seen with Ebola in 2014, and more recently with Zika virus and birth defects. Epidemic control has many facets, but starts with surveillance and early detection. In public health, the word "surveillance" has a specific meaning, outlined for you below. However, intelligence gathering in public health is similar in methodology and principles to intelligence gathering in many other areas. Traditional public health surveillance may use data from doctors, laboratories or sentinel networks. This yields more accurate data, but may not be timely. Earlier detection of outbreaks makes outbreak control more successful, and may even prevent large scale, catastrophic epidemics or pandemics. If an outbreak can be identified very early before it becomes large scale, it can be controlled rapidly.
ZikaHACK ’16 is sponsored by The NHMRC Centre for Research Excellence, Integrated Systems for Epidemic Response (ISER), for multidisciplinary student teams to come together to design and develop a tool early identification of disease outbreaks. The tool should be a technology solution that makes use of public domain information, with the early detection of the 2015/16 Zika virus outbreak used as a framework. The challenge is to see who can detect an abnormal surveillance signal earliest prior to the official recognition of the Brazilian Zika Virus outbreak. What signal do you look for? How do you identify such a signal? How early could you have picked the outbreak? These questions should be brainstormed within competing teams.
Information for applicants
Essential background Zika virus PH SurveillanceEssential background PH Surveillance
Phase 1: Shortlisting - concept designPhase 2: Product Development
Essential resourcesBibliography
Essential background on Zika virus:
Zika virus was first identified in monkeys in 1947, and then in humans in 1952 in Uganda (1). Prior to 2007, the virus was not considered a public health problem; it was occasionally reported in equatorial regions of Africa and Asia (2). In 2015, the virus was found to have spread widely throughout Brazil & South America and was associated with an increase in birth defects such as microcephaly (a small head). This prompted the World Health Organization (WHO) to declare the current Zika virus outbreak a public health emergency of international concern on 1st February 2016 (3). WHO/PAHO release an epidemiological alert for possible Zika virus infection in Brazil on 7th May 2015 (4) It is estimated that anywhere between 10-80% of the Brazilian population has been exposed to the virus (5-7).
Most people with Zika infection appear healthy and may have no symptoms at all (8). If signs and symptoms do appear, they are mostly mild and resolve by themselves (8-10). Commonly described symptoms include: fever, rash, joint pain and eye pain but more severe disease is rare. This makes disease detection difficult as these types of symptoms can look a lot like other diseases such as a mild flu, or more severe infections such as dengue and yellow fever. Zika virus infections have also been associated with Guillain-Barre Syndrome (a neurological condition that causes paralysis in people of any age) (11).
People become infected with Zika after being bitten by a mosquito carrying the virus. However not all mosquito species can carry the virus – species which are able to carry the virus include the Aedes Aegypti, commonly known as the Yellow Fever Mosquito, and the Aedes albopictus, commonly known as the Asian Tiger Mosquito. (12, 13) There is also evidence that the common Culex species mosquitoes can carry the Zika virus, however transmission has yet to be confirmed.
There is also a small chance of Zika transmission following sexual contact with an infected person (14, 15). Close, non-sexual contact with another infected person has not been reported as a risk factor for infection. Mothers infected with Zika virus during pregnancy transmit the virus to their newborns during pregnancy (16).
Current evidence suggests that Zika virus may cause microcephaly and other severe brain defects in babies born to infected mothers. (17, 18) Many women may not even know they have become infected until their baby is born with a birth defect after nine months gestation. The risk of microcephaly due to Zika is estimated to be low, ranging from 0-5%, however it may be nearer to 30% (5-7).
Babies born with birth defects, such as microcephaly, eye lesions and arthrogryposis (twisted limbs), have been more likely to have a mother with Zika virus infection or symptoms. (17, 18) Many women may not even know they have become infected until their baby is born with a birth defect after nine months. The risk of microcephaly due to Zika is estimated to be low, ranging from 0-5%, but may be nearer to 30%. (5-7) However, the infection has been so widespread in Brazil that between 1st of January 2016 and 2nd of July 2016 approximately 165,907 suspected and confirmed cases of Zika virus have been reported (19).
The first cluster of microcephaly cases in Brazil were reported in August 2015 (20), suggesting a substantial epidemic at least 9 months before. WHO/PAHO issued an epidemiological alert in November 2015 and asked countries to report increases of congenital microcephaly and other central nervous system malformations.
Essential background on Public Health Surveillance:
Effective intervention during epidemics such as Zika virus relies on the rapid and early detection of outbreaks through public health surveillance. Public health surveillance is defined by the World Health Organization (WHO) as “the continuous, systematic collection, analysis and interpretation of health-related data needed for the planning, implementation, and evaluation of public health practice”. (21-23) Traditional public health disease surveillance, such as collated reports from laboratories and doctors, is often not timely enough for early intervention, because test results have to be validated and checked. Most forms of public health surveillance include elements of case detection, reporting, analysis, validation and dissemination to detect outbreaks and inform control measures.
The sheer increase in public domain data, driven by the internet, smart phones and social media use, mean that a lot of information about our day to day activities is readily accessible. In this deluge lies useful information relevant to various aspects of human life including public health surveillance and epidemiology. Research has implicated the potential use of public domain data in public health surveillance that can complement traditional surveillance. (24, 25) The term “Digital Epidemiology” has come to encompass the use of online data for the purposes of population health surveillance. For example, Twitter has been used to track the incidence of Influenza A H1N1 in the USA, anonymous web logs for screening of pancreatic cancer, Facebook activities for predicting mental illness, and much more.
See essential resources for more detail on these approaches.
One of the most successful health surveillance tools that utilize public domain data is Health Map. This system mines online media sources in 15 different languages for the purposes of monitoring disease outbreaks in real time, and provides a map of these outbreaks online. Other well-known surveillance efforts include Google Flu trends and Google Dengue Trends. Unfortunately, Google’s efforts are no longer maintained amid concerns about the validity and accuracy of their disease outbreak estimates.
Despite these limitations, public domain sources, such as online search engines, social media and blogs, can provide timely alerts for early detection of disease outbreaks (24, 26, 27). The WHO reports that more than 60% of their initial outbreak reports come from unofficial sources (28, 29). Because effective outbreak response relies on the rapid and early detection of disease outbreaks, the use of invalidated yet timely open source information can provide vital early warnings for authorities during epidemics.
Competition Details:
We challenge teams to develop a computer system/software tool to mine public domain information to detect the 2015 Zika outbreak faster than traditional surveillance methods using any technical platform of your choice (such as Python and R) The team that develops the best tool, will win $15,000 and will be able to work with the EpiWATCH team to further develop their concept.
The event is open to a team consisting of a minimum of 3 and a maximum of 6 members comprising of current students from any University around the world. The team should contain both undergraduate and postgraduate students. Teams must be multidisciplinary and must have students from at least 2 disciplines, one in a health related discipline (such as medicine, nursing, public health, allied health) and one in a STEM (science, technology, engineering, mathematics) discipline. Other disciplines can also be included in the team. We suggest you approach your university to assist with coordinating and connecting you with other students interested in forming a team or try connect using the ZikaHACK Facebook event page here: https://goo.gl/cYoHuY. Part of the competition is about forming teams across faculties, and across postgraduate and undergraduate programs.
Eligibility criteria
- Student team has 3 to 6 enrolled students (who must all be enrolled at the time of the Phase 1 Submission Date of 30 November 2016)
- There is a single nominated team leader
- Team includes undergraduate and postgraduate students
- Team includes students from the following two discipline areas: STEM (science, technology, engineering, mathematics) and health related (medicine, nursing, public health, allied health) disciplines
- Must be studying at a registered university and recognised within its country as a university.
- Letter of support including verification of the student’s status of enrolment using the template provided for each team member is attached to the application. [Student Verification Template].
- Application is in English
- No team member has a direct connection with any investigator or affiliate of ISER (such as a student-supervisor relationship)
- The work has been done entirely by the student team, with no other assistance.
- All students have agreed to be named as part of the team
- The application coversheet, eligibility criteria checklist and declaration is completed and submitted with the application. [Application Coversheet]
The competition will consist of two phases: (i) shortlisting-concept design phase; and (ii) product development phase. Shortlisted proposals will proceed to phase two and have the opportunity to interact with the ISER team to further develop their ideas prior to product showcase and final demonstration in April 2017. The winning team will have the opportunity to collaborate with ISER on the development of a commercial product based on their Showcased Tool.
Key Dates:
- Date for proposal submission (Design Phase): 5pm (AEDT), 30th November 2016.
- Shortlist notification to participants: 15th December 2016.
- Product Showcase Submission & Demonstration: April 2017.
- Evaluation of Showcased Products by EpiWATCH team: May 2017.
Phase 1: Shortlisting - concept design Phase:
In this phase we invite participants to submit a detailed proposal on their method and approach to solving the problem. Each team will also need to address the specific problem of capturing early signals for the Zika epidemic. Proposals will be judged and shortlisted on the basis of the quality of the submission and originality of the solution. The proposal must be written in English and not exceed 3,000 words. Proposals must be submitted in .doc, .docx, or .pdf file format. Teams are also required to submit a (i) CV not exceeding one page (ii) a letter of support from their University confirming they are an enrolled student at the time of submission; and (iii) a cover sheet. All proposals should be original work and will be checked for plagiarism by automated software. All files should be submitted electronically, via email to EpiWATCH-ISER@unsw.edu.au
Submissions are due 30th November 2016.
Phase 2: Product Development Phase:
Phase 2 is open to shortlisted applicants ONLY. Successful applicants will be contacted by 15th December 2016 with more details.
During Phase 2, teams will have the opportunity to interact & collaborate with the EpiWATCH team at ISER to develop their solution. In this phase, each team will (i) gather data gathering relevant to the disease outbreak (ii) prepare and clean the data and (iii) design and utilize the best algorithms to detect outbreaks from the data preparation step (ii).
Criteria for evaluation by the EpiWATCH team:
- The ability of the tool to identify disease outbreaks early (as measured against retrospective data from the Zika virus epidemic).
- The quality of the test data submitted to the EpiWATCH team.
- The ease of use of the tool.
- Clarity and transparency of the tool.
- System documentation.
- Demonstrated multi-disciplinary approach to developing the tool.
Final solutions are due to be presented in April 2017.
The winning team will be announced in May 2017.
Essential resources:
A systematic review of using social media to track outbreaks:
Al-garadi MA, Khan MS, Varathan KD, Mujtaba G, Al-Kabsi AM: ‘Using online social networks to track a pandemic: A systematic review. Journal of Biomedical Informatics. 2016; 62:1-11.
A review of the traditional origins, methods, and evaluation of public health surveillance:
S. Declich and A.O. Carter: “Public health surveillance: historical origins, methods and evaluation” available at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2486528/pdf/bullwho00413-0101.pdf
An example of using Twitter to track Influenza outbreaks in the USA:
Signorini A, Segre A M and Polgreen P M, "The Use of Twitter to Track Levels of Disease Activity and Public Concern in the U.S. during the Influenza A H1N1 Pandemic" available at http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0019467
An example of using weblogs to screen for pancreatic cancer:
Paparrizos J, White R W and Horvitz E, "Screening for Pancreatic Adenocarcinoma Using Signals From Web Search Logs: Feasibility Study and Results" available at http://research.microsoft.com/en-us/um/people/horvitz/JOP_June_2016.pdf
An example of using Facebook to monitor mental health:
Park S, Lee S W, Kwak J, Cha M, and Jeong B, "Activities on Facebook Reveal the Depressive State of Users" available at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3806432/
A link to Health Map:
http://www.healthmap.org/site/about
A link to Google Flu Trends:
https://www.google.org/flutrends/about/
FAQs:
Q: I’m interested in participating but my friends aren’t, is there a way I can connect with other interested students to form a team?
A: Yes. We want to ensure that all students who are interested in participating can. We have set up a Facebook page (https://goo.gl/cYoHuY) where updates to the competition will be regularly posted. Individual students are welcome to use this page to network and form teams.
Q: Will prizes be given for second or third place?
A: No. Only the winning team will receive a cash prize $15,000. Teams who are shortlisted in Phase two will receive a certificate of merit from ISER.