Restoration note, September 3, 2026: This is a technically restored version of a MIT News profile published on August 21, 2020. Adoption, donation, partnership, attendance, and comparative-performance figures are preserved as claims reported by Kinsa or founder Inder Singh at that time, not as independently audited results. Kinsa’s school program has since ended. Most importantly, excess fever is a nonspecific syndromic signal: it may indicate unusual illness activity, but it cannot identify SARS-CoV-2 or another pathogen without independent testing and surveillance.
Kinsa was founded in 2012 by MIT alumnus Inder Singh MBA ’06, SM ’07 with the mission of collecting more timely information about when and where infectious illness is spreading.
The system started with families. In August 2020, MIT News reported that more than 1.5 million Kinsa “smart” thermometers had been sold or given away across the United States, including hundreds of thousands distributed to families in lower-income school districts. Those were contemporaneous company figures. The thermometers connected to an app that recorded temperatures and symptoms and offered age- and symptom-based guidance, including when a user might consider seeking medical attention.
For users who opted into data sharing, readings could be deidentified and aggregated into community-level signals. Kinsa shared summaries with participating parents and school officials to help them see when feverish illness was increasing locally.
MIT News also reported that Kinsa was working with more than 2,000 schools and numerous businesses. Data from the thermometer network were used to develop models of influenza-like illness at finer geographic scales than many conventional flu forecasts.
Technical note: The original article characterized Kinsa’s work as predicting flu spread 12 to 20 weeks in advance at the city level. The linked June 2020 preprint actually evaluated 30-week forecast ensembles for nine cities. It reported that the ensembles contained skillful variants and that the strongest variants could potentially be identified early in the season, but its principal fits were selected in hindsight. The study was preliminary, had not been peer reviewed, and disclosed that several authors were Kinsa employees or shareholders.
When COVID-19 emerged in the United States, Kinsa compared observed fever levels with baselines derived from earlier cold and flu seasons. The residual—feverish illness above the expected range—was treated as a possible warning of unusual disease activity. In August 2020, Kinsa said it was working with health officials in five states and three cities.
“By the time the CDC [U.S. Centers for Disease Control and Prevention] gets the data, it has been processed, deidentified, and people have entered the health system to see a doctor,” Singh said. “There’s a huge delay from when someone contracts an illness and when they see a doctor. The current health care system only sees the latter; we see the former.”
That contrast captured Kinsa’s intended advantage but was too broad. By 2020, CDC and state and local agencies already operated near-real-time syndromic surveillance based on emergency-department and other health-care data. Kinsa’s more specific distinction was that household temperature readings could be collected before a person sought care or received a laboratory test.
Kinsa also made its HealthWeather maps available to the public, media, and researchers. The company’s April 2020 preprint described forecasting expected influenza-like illness at the county level and flagging observed values above the expected range as anomalies. Those anomalies were possible indicators of COVID-19 activity, not measurements or diagnoses of COVID-19 itself.
Singh argued that Kinsa’s signal could complement testing, contact tracing, masking, and other public-health measures.
Better data for better responses
Singh told MIT News that his first exposure to MIT came while he was a graduate student at Harvard University’s Kennedy School of Government.
“I remember I interacted with some MIT undergrads, and we brainstormed some social-impact ideas,” Singh recalled. “A week later I got an email from them saying they’d prototyped what we were talking about. I was like, ‘You prototyped what we talked about in a week?’ I was blown away, and it was an insight into how MIT is such a do-er campus. It was so entrepreneurial. I was like, ‘I want to do that.’”
Singh subsequently studied through the Harvard-MIT Program in Health Sciences and Technology and earned his MIT master’s and MBA degrees. After graduation, according to the original profile, he joined the Clinton Health Access Initiative and worked on agreements intended to lower the cost of medicines for HIV, malaria, and tuberculosis in lower-resource countries.
“The world tries to curb the spread of infectious illness with almost zero real-time information about when and where disease is spreading,” Singh said. “The question I posed to start Kinsa was, ‘How do you stop the next outbreak before it becomes an epidemic if you don’t know where and when it’s starting and how fast it’s spreading?’”
Kinsa’s founding insight was that a sensor embedded in an existing household behavior might collect earlier information while also providing direct value to the person using it.
“The behavior in the home when someone gets sick is to grab the thermometer,” Singh said. “We piggy-backed off of that to create a communication channel to the sick, to help them get better faster.”
The company sold thermometers and created a sponsorship program through which corporate donors funded distributions to Title I school communities. Title I directs federal support toward districts and schools serving substantial numbers or percentages of children from low-income families. Singh said that 40 percent of families receiving a thermometer through Kinsa’s program had not previously had one at home; the original article did not cite an independent survey supporting that figure.
Kinsa also said its school program improved attendance while supplying years of fever data for comparison with official surveillance. The original article did not provide methods, uncertainty estimates, or independent evidence for the attendance claim.
“We had been forecasting flu incidence accurately several weeks out for years, and right around early 2020, we had a massive breakthrough,” Singh recalled. “We showed we could predict flu 12 to 20 weeks out—then March hit. We said, let’s try to remove the fever levels associated with cold and flu from our observed real-time signal. What’s left over is unusual fevers, and we saw hot spots across the country. We observed six years of data and there’d been hot spots, but nothing like we were seeing in early March.”
Kinsa quickly made its maps public. Singh recalled a March 14, 2020, call with former health officials and a researcher involved in Taiwan’s COVID-19 response. According to his account, the other participants initially questioned fever hot spots in places where reported COVID-19 cases remained low.
“I said, ‘There’s hot spots everywhere,’” Singh recalled. “They’re in New York, around the Northeast, Texas, Michigan. They said, ‘This is interesting, but it doesn’t look credible because we’re not seeing case reports of COVID-19.’ Lo and behold, days and weeks later, we saw the COVID cases start building up.”
The recollection illustrates the hypothesis behind the system, but by itself it does not establish forecast accuracy. Evaluating an early-warning method requires prespecified predictions, independent outcome data, geographic coverage measures, uncertainty estimates, and accounting for false alarms.
A tool against COVID-19
Singh described Kinsa’s data as offering an unusually early view of illness spreading through a community. He also made a comparative-performance claim that the original article did not independently benchmark:
“We can predict the entire incidence curve [of flu season] on a city-by-city basis,” Singh said. “The next best model is [about] three weeks out, at a multistate level. It’s not because we’re smarter than others; it’s because we have better data. We found a way to communicate with someone consistently when they’ve just fallen ill.”
In August 2020, Kinsa said it was helping health departments and research groups interpret its data and was working with businesses preparing to bring employees back to offices. The company was also considering international expansion.
“I started Kinsa to create a global, real-time outbreak monitoring and detection system, and now we have predictive power beyond that,” Singh said. “When you know where and when symptoms are starting and how fast they’re spreading, you can empower local individuals, families, communities, and governments.”
What subsequent evidence clarified
The mechanism behind connected-thermometer surveillance received empirical support, although not every claim in the 2020 profile was validated. A 2018 peer-reviewed study found that readings from the Kinsa network were strongly correlated with CDC-reported influenza-like illness nationally and reflected regional and age-specific patterns. A 2020 California study found that smart-thermometer data improved predictions of influenza, but not predictions of influenza-like illness. Both studies included Kinsa-affiliated authors.
A 2021 Science Advances study evaluated Kinsa fever anomalies alongside several other digital traces and concluded that such signals could help identify changes in COVID-19 activity before confirmed cases and deaths reflected them. The authors also emphasized confounding, false alarms, reporting biases, and the advantages of combining multiple signals rather than relying on one proxy.
A 2023 cohort study of voluntary Kinsa users found that febrile episodes forecast new reported COVID-19 cases and demonstrated uses for studying household transmission. Its data still came from a self-selected population of device and app users, and the study inferred viral transmission from fever episodes rather than confirming the responsible pathogen in each household.
These findings support the original article’s central idea: household sensors can complement clinical and laboratory surveillance by providing a timely symptom signal. They also reinforce its essential limitation. As the CDC notes, COVID-19 and influenza symptoms overlap and cannot be distinguished by symptoms alone. Fever anomalies are therefore most useful as an early warning to investigate—not as proof of where a particular disease is spreading.
Source: Adapted and technically restored from Zach Winn, “Real-time data for a better response to disease outbreaks,” MIT News Office, August 21, 2020.
Responses