COVID-19 National Daily Health Survey

Some folks at Stanford Medicine are conducting a National Daily Health Survey to learn and predict which geographical areas will be most impacted by coronavirus based on how people are feeling. It is a research study that will be used to inform local and national responses, such as redirecting medical resources or improving policies and public guidance. Upon first-time completion, some of the questions are as follows:

   

It appears that there are multiple research efforts that are competing for the public’s attention, such as the COVID Symptom Tracker that I discussed previously. While there are some similarities in the questions asked, both research studies ask about COVID-19 symptoms, exposure, health professional status, etc. Both research studies also request that participants fill out questionnaires on a daily basis.

I have been completing the COVID Symptom Tracker questionnaire daily through an iOS app. It takes about 1 minute to complete. The COVID-19 National Daily Health Survey can provide daily email reminders if you choose to provide your email address, and after the initial survey, subsequent assessments are estimated to take about 1 minute to complete as well.

It might be interesting if researchers partnered on crowdsourcing efforts like this to come up with a master list of questions and answer responses that could be used by multiple research teams to achieve their individual goals. This would only work if they are studying the same topic with overlapping survey questions.

COVID-19 Hospital Impact Model for Epidemics

As if I have not written about enough COVID-19 hospital resource calculators, I just learned about another one called COVID-19 Hospital Impact Model for Epidemics (CHIME) which was produced by Predictive Healthcare at Penn Medicine and described in a research publication in the Annals of Internal Medicine. Similar to other COVID-19 predictive models, CHIME uses a variety of inputs (hospital parameters, transmission parameters, and severity parameters) to predict hospital resource utilization. The forecasted utilization outcomes include new admissions, inpatient census, and the number of susceptible, infected, and recovered individuals in the hospital.







And as if this blog hasn’t provided sufficient coverage about hospital resource calculators, the authors of the CHIME research study compared the predictions from CHIME with those from other COVID-19 forecasting tools including this one and this one.

In conclusion, I suspect that there are many other COVID-19 hospital resource calculators out there that I haven’t covered, but it is probably unnecessary for me to continue writing about them unless they have unique features that I have not previously written about. I hope that few hospitals will exceed capacity, and for the ones that do, I hope that these kinds of forecasters had been used long ago so that hospital administrators could plan ahead for the anticipated surge in demand.

Apple Watch Fall Detection

There are many stories about the fall detection feature of Apple Watch saving people’s lives, including this one and this one. Unless you are at least 65 years old, the fall detection feature is turned off by default. About 6 months ago, I decided to manually activate fall detection on my watch to see what would happen. Fall detection was activated twice during this time.

The first time was when I was playing catch with my daughter before softball practice. Since I’m right-handed, I wear both my watch and my fielder’s glove on my left side. I misjudged a low throw and caught the softball in the palm of my hand which activated fall detection. I suspected that it was the combination of the impact of the ball jarring my hand/wrist as well as the downward motion of my arm as the impact occurred.

The second time was when my daughter and I were roughhousing at home. She hit my left arm, and while I do not recall making any significant downward motion, it activated fall detection. Here’s the full message after stitching together 2 screenshots:

After pressing the “I’m OK” button, it asks the user to confirm whether or not a fall occurred. I assume that this is to help Apple improve its fall detection algorithm.

So far, I’ve had 2 false positive activations and zero true positives, so I think Apple’s decision to deactivate fall detection in people younger than 65 is justified. However, if you’d like to turn on the feature, see the Apple support page for details.

Yet Another COVID-19 Calculator

I’ve previously written about web-based COVID-19 forecasting tools like this one and this one that predict hospital resource utilization and other factors. Adding to the mix, the folks at Rush University Medical Center have created a hospital resource calculator as well.

Based on data from Johns Hopkins University Center for System Science and Engineering which I’ve also discussed, this visualization forecasts the number of cases for each state. Based on a set of variables, the calculator can also estimate hospital admissions, census, and demand for personal protective equipment.

Rather than explain all the input variables, I’ll direct you to a blog that was jointly written by the folks who created the calculator. It describes the variables as well as their units of measure which might not be immediately apparent by just looking at the site.

As with any forecast, the accuracy is only as good as the fitness between the predictive model and what is happening in the real world. The spread of COVID-19 is complex, and we still do not fully understand the reasons why some people with COVID-19 do well and others to poorly. Nevertheless, when we broaden our view from patients to populations, we can see patterns against which we can create mathematical models to estimate what might happen in the future

COVID-19 vs. Top 15 Causes of Death

Here’s another interesting COVID-19 data visualization that shows the growth of COVID-19 as a cause of death relative to the other top 15 causes:



Source: https://public.flourish.studio/visualisation/1845748/

A Smartphone-based Approach to Contact Tracing

Contact tracing is a laborious process that is used to identify and reach out to people how have had contact with someone with a confirmed case of an infectious disease. As part of an effort to contain an outbreak, those contacts are isolated or quarantined. If the contacts develop into confirmed cases, then contact tracing is performed on their contacts too.

I’d like to clarify some terms to put things into perspective. An outbreak is when an infectious disease occurs in unexpectedly high numbers but is limited to one geographic area. When the infection spreads to multiple areas, it is called an epidemic. When the infection spans multiple countries or continents, it is called a pandemic.

Contact tracing is a containment strategy for a disease outbreak. When an infectious disease becomes an epidemic or pandemic, contact tracing is not feasible, and we switch to mitigation strategies such as school closures, limiting restaurants to take-out or delivery, and other social distancing measures. The United States is predominantly addressing COVID-19 through mitigation strategies. However, if we get the disease sufficiently under control and can begin to relax social distancing measures, we will need to ramp up contact tracing to focus once again on containment.

The Margolis Center for Health Policy at Duke University released guidance on a COVID-19 containment strategy that addresses testing, surveillance, contact tracing, isolation, and quarantine. As described in this NPR article, contact tracing is a resource-heavy and time-consuming process.

It was announced today that Apple and Google are working together on a contact tracing infrastructure using smartphones. Here are identical press releases from Apple and Google. More information is available via both technical documentation and a layperson illustration.

The first phase of the collaboration will include APIs that will enable data sharing across Apple’s iOS and Google’s Android devices. This would allow app developers to implement the APIs and share data between the apps and health authorities. The second phase will include a Bluetooth-based contact tracing platform that, according to The Verge, would detect signals from nearby phones at 5-minute intervals and store those connections but not their physical locations. If one person develops COVID-19, it would be possible to determine, based on those connections, who else might have been recently exposed.

An important aspect of this approach that I alluded to is that unlike GPS that tracks actual locations, this Bluetooth-based approach only tracks other smartphone users in their proximity but does not know the physical location of the phone. This might be a critically important aspect of the design of the solution that would allay the fears of people who feel strongly about protecting their privacy. I look forward to seeing the implementation of both phases of this collaboration, and the timing of this collaboration is good because I am optimistic that in the coming weeks/months we can shift from mitigation back to containment.

COVID-19 Shenanigans

As with any crisis situation, a small number of “opportunists” (or shall we say “sociopaths”?) prey on public anxiety and fear for their personal gain. Amidst the COVID-19 pandemic, I’ve seen various kinds of shenanigans and will summarize some of them here.

Malware

Although there is a perennial battle against malware and cyberattacks, some hackers have impersonated the CDC and WHO as part of phishing scams. Others have launched COVID-19 email campaigns to bait recipients into clicking malicious links or have created fake coronavirus maps that look similar to authentic ones to attract unsuspecting clicks.

Hospitals are especially busy now, and new ransomware attacks are targeting remote employees to get through VPN connections and exploit other security vulnerabilities. Interestingly, some ransomware groups have pledged not to target hospitals during the COVID-19 pandemic, but I would guess that this represents a minority of hackers. Nevertheless, it is interesting to see that some criminals have enough of a conscience to hold back—unfortunately it is hard for me to feel positive about this news. Meh.

Credit Card Fraud

In a similar vein, coronavirus-related credit card fraud is also on the rise, so be on the lookout for scams and make sure you manage your credit cards wisely.

Additionally, I’ve seen a variety of restaurants offer special deals if you place orders through their mobile apps. I would recommend that if you take advantage of these kinds of offers, check to see if the app requires you to store your credit card or if you can enter your credit card for payment and choose not to store your card number. The more companies that have your credit card information, the more likely your personal information will be compromised in a security breach at some point. Balance the convenience of saving your credit card information against the risk of your credit card information getting into the wrong hands.

Zoombombing

Due to social distancing efforts, many workplaces have transitioned partly or completely to remote work, and most schools have switched to online learning. This has resulted in a sudden and dramatic increase in demand for videoconferencing. While there are a large number of video conferencing solutions, Zoom has seen explosive growth due in large part to its ease of use. In its most basic meeting configuration, participants can join a Zoom video chat by simply entering a meeting ID. Although password protection is an option, not everyone has been aware of it, so meeting crashers have engaged in a variety of activities ranging from disruptive to obnoxious to salacious, resulting in coinage of the term Zoombombing.

As a result, Zoom has created a website to provide recommendations on how to prevent Zoombombing. While most video conferencing solutions share a core set of common features, check with your video conferencing solution about both corrective and preventive actions you can take to video conference safely.