Comment on page
Manual vs. digital
Why not just do it with an app?
Exposure notification uses information shared via mobile device to notify people who might have had contact with someone infected. While manual contact tracing can take time due to the need for individual interviews, exposure notification makes it possible to identify hundreds or thousands of at-risk persons almost immediately. This is usually done by logging “close contact” events between Bluetooth-enabled mobile phones. If someone is diagnosed, their log of contact events could be used to alert anyone physically near them within the infectious period.
Because most applications so far only notify contacts--without collecting information about symptoms, providing isolation/quarantine information and support, and following up over the recommended two-week period--manual contact tracers will still be required. For that reason, Bluetooth-enabled apps are better described as “exposure notification” than "digital contact tracing".
- Immediate, once automated
- Based on objective distancing/Bluetooth data
- Removes notification burden from human workforce, empowering staff to focus on neediest cases
According to research, 60–80 percent adoption is required to make exposure notifications effective. This level of adoption would require massive, population-level awareness, trust, and engagement of the sort few apps or features on the market enjoy. For reference, Android’s latest version, “Pie”, took a year to achieve 20% adoption. However, iOS has 60% market share and a faster adoption cycle.
Here are the principles that are critical for apps or settings to achieve adoption:
- Data privacy. People are understandably wary of sharing private health and location data with corporations and the government. To address this, solutions should implement solutions like contact event numbers and decentralized local storage. This unified set of data and privacy guidelines for contact tracing and exposure notification was created to educate developers and inform end-users of the data rights they should expect any effective solution to respect.
- Interoperability. Apps should be capable and willing to share data formats to enable data to be aggregated for broader public health use. Apple and Google allow one app per jurisdiction (state, city, or county) to deploy exposure notifications APIs, meaning there could be dozens of apps available. No single app is likely to reach 60%+ adoption, but all apps working together have a much better chance. Developers should use openly published protocols and work collaboratively to ensure their solutions are verifiable and interoperable.
- Trustworthy. Brand matters. If an app is nefarious or perceived as such, it could create broad public distrust and lowering likelihood that any exposure notification technology gets adopted.
Digital contact tracing solutions are built by and for smartphone users. However, given that COVID-19 disproportionately affects the elderly and underprivileged, the impact these apps can have will be limited. For example, 19% of the US doesn’t have a smartphone. Rural areas have 71% smartphone adoption and 53% of those over the ages of 65 have a smartphone.
Even with 100% adoption, opt-in, and Bluetooth enablement, there are still challenges with efficacy.
A false positive occurs when a contact is notified who wasn't at meaningful risk of transmission. This can happen for a variety of reasons including:
- The app isn’t aware there’s a wall between you and someone infected.
- Not every contact is a transmission.
False positives can hurt adoption due to distrust. For example, if someone chooses to quarantine, but doesn’t get sick, there’s a strong incentive to stop using it. Tests can alleviate this problem by limiting quarantines to those that are sick, but still create disincentive given the time, discomfort, and potential cost needed for the test.
A false negative occurs when an app fails to register a transmission. These happen for a variety of reasons, including:
- An infected person transmitted the virus without being in close contact (e.g., by touching the same surface).
- Transmission occurred via someone who didn't have the app.
- Bluetooth must always be on, your phone must always be in your pocket, and powered on.
- For both exposure notifications and manual contact tracing ubiquitous, cheap, fast, and accurate testing is critical. Otherwise, false positives and quarantines will be significantly higher than needed. For exposure notifications, this means decreasing public trust and lower adoption.
In the long-term, digital contact tracing and exposure notification have the potential to streamline the contact tracing workflow, enabling the workforce to focus more on providing support to individuals in need. That said, given the challenges described above, there is no effective digital replacement for a manual contact tracing workforce. Governments should continue to aggressively ramp up manual contact tracing as an immediate solution.