Smartphone Apps Test and Track Infectious Diseases

Work / Technology & tools ADAPTED FROM GETTY FROM ADAPTED SMARTPHONE APPS TEST AND TRACK INFECTIOUS DISEASES The prevalence, power and portability of smartphones make them valuable tools for pathogen monitoring and citizen science. By Sandeep Ravindran

ebojyoti Chakraborty took just a quantify bands using machine learning, and to deploy, and they get real-time insights few months to develop a COVID-19 export the results to the cloud. Called TOPSE1, directly from people on the ground,” says diagnostic test that worked in his their app laid the foundation for a test that has John Brownstein, a computational epide- lab; the challenge was to optimize it now been approved by the Drugs Controller miologist at Boston Children’s Hospital in for the field. General of India. “You can actually do this Massachusetts. “That kind of data can out- DBased on the gene-editing technology test in local pathology labs in places that are pace what traditional surveillance provides.” CRISPR, the test produces a band on a paper resource-limited,” Chakraborty says. “Per- strip if viral RNA is present. But Chakraborty, haps one day it can be done even at home.” Portable epidemiology who heads an RNA biology group at the Insti- The billions of smartphones in use world- Smartphone science didn’t start with tute of Genomics and Integrative Biology in wide offer unprecedented opportunities COVID-19. But the pandemic has spurred New Delhi, says he and his colleagues couldn’t for disease tracking, diagnostics and citizen researchers to fast-track citizen-science always agree on whether a faint band counted science, as Chakraborty learnt. Exam- efforts that use smartphones to gather as a positive signal. What they needed was an ples include apps that enable phone users information about the disease. Volunteers objective — and, ideally, portable — quanti- to monitor COVID-19 symptoms, count can regularly log details about their symp- fication tool. Then Chakraborty realized he disease-carrying mosquitoes and detect toms, testing status and location through already had one: his smartphone. microscopic pathogens. They could even help apps or websites. For instance, data from The researchers developed an app to the world to prepare for the next pandemic. 5 million users of Brownstein’s crowdsourced photograph the paper strip, identify and “These tools can be incredibly cheap tracker for influenza and COVID-19 — called

Outbreaks Near Me — provided early evidence student on Williams’s team who works on the carrying, which could help to predict the risk of the benefits of masking2. “You get these project. of disease transmission and severity. Says impressive insights that would be harder to Williams has since added the option for Ott: “The phone is doing everything: it’s the collect from traditional health care quickly,” volunteers in Australia to eschew traps and camera that actually acquires the results, the he says. instead use a third-party app, iNaturalist, to computer that analyses the results, and it can Such projects can be rapidly deployed, an photograph and identify the insects. This led to send the result to the cloud for storage.” advantage in a fast-moving public-health crisis. a nationwide trial in February, dubbed Mozzie Cardiologist Gregory Marcus at the University Month. “We’re hoping to have the first national Ensuring trust of California, San Francisco, was able to take citizen-science mosquito-surveillance pro- Smartphones are a powerful way to do science, his team’s COVID-19 Citizen Science app from gramme,” he says. “The smartphone is at the but gaining users’ trust is crucial if an app is concept to 50,000 participants in under a year. centre of that.” going to be widely used and provide good data. “That would be impossible in a conventional Developers and other experts can help with research study,” he says. And because users Diagnosis by smartphone this, as well as with the technological hurdles. enrol by downloading an app, such studies are Smartphone science comes down to three App development is often tricky, but apps easy to scale up and tweak: researchers can features: computing power, connectivity and that collect health data require extra levels add questions about new vaccines or virus cameras. of privacy, security, support and regulatory variants, for instance. “Cameras on phones have become more compliance. Likewise, citizen-science projects Claire Steves, who studies ageing at King’s and more sensitive over the past ten years, must take care to anonymize and encrypt College London, analysed data from the and what I needed a microscope for before I users’ data. “If we don’t get those things right, United Kingdom’s COVID Symptom Study can do with my phone now,” says Rashid Bashir, then it’s going to be impossible to build trust app, which has been downloaded by some a bioengineer at the University of Illinois in and get people to use these tools, so it’s a 4.5 million people. She and her team used Urbana-Champaign. In 2017, Bashir and his priority for us,” says Brownstein. data from the app to develop a predictive team harnessed smartphone cameras in a Such tools also generate a lot of data. The model that found that a loss of the sense of point-of-care test that uses blood samples COVID Symptom Study app, for instance, smell (known as anosmia) was a predictor of to check for viral infections. The cameras was developed with the London-based testing positive for COVID-19 (ref. 3). The team data-science firm Zoe, which helped with also used data from the app to identify COVID- “We duplicated the both app design and data analysis. “You’re 19 hotspots in the United Kingdom4 and to having responses from four-and-a-half establish that long COVID, in which people bench-top fluorescence million people potentially on a daily basis, experience persistent symptoms, was more microscope for less so it’s very big data and it can be quite com- likely in app users who had reported experi- than $40.” plicated,” says Steves. encing more than five COVID-19 symptoms in Marcus helped to create a research platform the first week of illness5. called Eureka, funded by the US National Insti- Young children and older adults, can spot fluorescent signals generated from tutes of Health, which includes infrastructure who often don’t have smartphones, are amplified viral RNA, enabling detection of that enables scientists to build and deploy under-represented in the COVID Symptom Zika, dengue and another mosquito-borne mobile apps for health research. Eureka Study, Steves and her colleagues found. But virus called chikungunya7. He has now updated hosts the COVID Citizen Science Study, as thoughtful outreach, statistical analyses and the test to detect COVID-19 (ref. 8). well as more than 30 others. “The point is to cross-validation against other research can In some cases, smartphone-based help researchers utilize mobile technologies blunt those biases, she says. By comparing diagnostics can match expensive and bulky without having to have technology expertise, their results to the findings of UK COVID-19 lab equipment for speed and accuracy, while so they can focus on their clinical research and tracking studies that used conventional remaining portable enough for field settings. medical expertise,” he says. designs, the team was able to validate its Jeong-Yeol Yoon, a biomedical engineer at the That could help users and public-health models4. “We’ve been able to show that our University of Arizona in Tucson, coupled a efforts, too. Yoon envisions a future in which data very much reflect the data from [those] smartphone with some off-the-shelf micro- consumers pair over-the-counter kits with big population studies.” scope components and the MATLAB Mobile apps on their phones for convenient and reg- Smartphones can track other diseases, app to detect norovirus, which causes about ular testing, which could help to control future too, such as malaria, Zika and dengue, which 20 million cases of gastrointestinal illness pandemics. “We need to identify the people are spread through mosquitoes. Instead each year in the United States9. “We dupli- who get infected, and if you can actually do of sending technicians to trap and collect cated the bench-top fluorescence micro- it at home, I believe that can greatly decrease mosquitoes, Craig Williams, a public- and scope, which can easily be US$50,000, for the spread of the disease,” he says. “We need environmental-health scientist at the less than $40,” says Yoon, who is currently to be ready for the next pandemic.” University of South Australia in Adelaide, setting up clinical trials for both norovirus posted traps to 126 volunteers in southern and COVID-19 versions of the test. Sandeep Ravindran is a science writer in Australia and asked them to e-mail him smart- Melanie Ott, a virologist at the Gladstone Washington DC. phone photos of the trapped insects. Institute of Virology and Immunology in San 1. Azhar, M. et al. Preprint at medRxiv https://doi. The project, called Mozzie Monitors, pro- Francisco, California, co-developed a smart- org/10.1101/2020.09.13.20193581 (2020). 10 vided large-scale mosquito surveillance at phone-based COVID-19 diagnostic test with 2. Rader, B. et al. Lancet Digital Health 3, 148–157 (2021). 20% of the cost of a comparable professional Daniel Fletcher, a bioengineer at the University 3. Menni, C. et al. Nature Med. 26, 1037–1040 (2020). surveillance programme6. “It has been sur- of California, Berkeley. According to Fletcher, 4. Varsavsky, T. et al. Lancet Public Health 6, 21–29 (2021). 5. Sudre, C. H. et al. Nature Med. 27, 626–631 (2021). prising that citizen scientists were able to the test is more sensitive than standard lab- 6. Sousa, L. B. et al. Sci. Total Environ. 704, 135349 (2020). collect a higher abundance of mosquitoes oratory plate readers, and faster than the 7. Ganguli, A. et al. Biomed. Microdevices 19, 73 (2017). than professional programmes in the same gold-standard reverse-transcription PCR test 8. Ganguli, A. et al. Proc. Natl Acad. Sci. USA 117, 22727–22735 (2020). period, through a low-cost project from their for SARS-CoV-2. The smartphone-based test 9. Chung, S. et al. Nature Protoc. 16, 1452–1475 (2021). backyards,” says Larissa Braz Sousa, a graduate can also quantify how much virus a person is 10. Fozouni, P. et al. Cell 184, 323–333 (2021).