COVID-19 reporting on campus
I evaluated 50 college and university COVID-19 dashboards based on what metrics they were reporting. Unsurprisingly, I found significant discrepancies from school to school.
|Betsy Ladyzhets||Aug 31|
Welcome back to the COVID-19 Data Dispatch.
I’m in your inbox a little late today, because I’m trying something new this week: compiling a new data resource. I evaluated 50 college and university COVID-19 dashboards based on what metrics they were reporting, and how often they were reporting them. Plus: more on K-12 COVID-19 data and a new reporting mandate for hospitals and nursing homes.
This is a baby newsletter, so I would appreciate anything you can do to help get the word out. If you were forwarded this email, you can subscribe here:
How are colleges and universities reporting COVID-19?
Screenshot of UNC-Chapel Hill’s COVID-19 dashboard. The school has made headlines in the past two weeks for facing outbreaks almost immediately upon students’ return to campus.
Across the country—and despite the warnings of numerous faculty members and public health leaders—colleges are reopening. Freshmen are returning to campus eager to meet their classmates, and upperclassmen are returning eager to see friends after months in their hometowns.
Most college students will certainly do their best to wear masks, socially distance, and meet outdoors wherever possible. But it is inevitable that COVID-19 will hit campuses, and when it does, school leaders need to be prepared: not only with testing and tracing systems and room for students to self-isolate, but also with a robust data system that can let their communities know exactly what’s going on.
For this issue, I surveyed the COVID-19 dashboards of 50 higher education institutions across the country. This analysis expands upon the work of Andy Thomason, an education journalist at the Chronicle of Higher Education who has been compiling dashboards on Twitter for the past two weeks:
Andy Thomason @arthomasonI’m collecting colleges’ public testing dashboards but have only come up with a handful. Posting the ones I know about in this thread but @ me if you know of more.
The 50 dashboards I examined include public and private schools in 26 states. The most heavily represented states are North Carolina (nine schools) and Massachusetts (4 schools). I plan on expanding this analysis to include all 82 schools currently on Andy Thomason’s list, as well as any other dashboards that I can find, but 50 seemed like a solid number to discuss for today’s issue.
Rather than attempting to count the number of COVID-19 cases occurring at colleges and universities, I chose to focus on how schools are reporting: what COVID-19 metrics are they making public, and how often are these metrics updated? (If you’d like to see sheer case numbers, I recommend this dashboard by the New York Times, which includes counts from both public sources and additional NYT reporting.)
Here’s the spreadsheet with my work. I encourage you to check it out for yourself, especially if you attend or have friends or family attending one of the schools on the list. But I’ll include some summary statistics here because, yeah, you’re probably skimming through this on Monday morning and have about 900 other things to do.
41 schools report some form of cumulative COVID-19 cases that have occurred on campus, usually since the beginning of August. Most of these (29) separate out student and faculty/staff cases.
24 schools report COVID-19 cases per day. 20 of these separate out student and faculty/staff daily cases.
18 schools report active COVID-19 cases. Definitions for “active” vary, from cases identified in the past 14 days to all cases minus those which reportedly no longer need to isolate thanks to a negative test.
No schools report COVID-19 deaths. Thankfully, no school yet needs to.
29 schools report cumulative COVID-19 tests. The majority of these (21) do not specify whether their test counts are reported in specimens, people, or some other unit.
No schools report antibody or antigen tests.
23 schools report some form of test positivity. 4 schools report daily figures, 11 report weekly figures, and 8 others report some form of cumulative figure. (Test positivity over a range greater than 14 days is not particularly useful, as this metric is used to track testing capacity over time.)
19 schools report counts of students and/or staff currently undergoing quarantine or isolation.
19 schools update their dashboards daily, 6 update Monday through Friday, 15 update on a twice- or three times-weekly schedule, and 7 do not clearly state how often updates are made.
And here are a couple of examples of notable dashboards:
Roger Williams University: This dashboard focuses on testing. It’s the only dashboard in my initial analysis to report tests in both “tests administered” (specimens) and “total participants” (people). It also comprehensively reports test wait times, which few schools are tracking.
University of Nevada at Reno: This school has less of a “dashboard” and more of a series of COVID-19 reports. Click into a given month to find one-line statements about new cases, sorted by day. No overall statistics are reported.
University of Wisconsin at Madison: This dashboard includes charts and a table that show total tests, positive tests, and percent positive by day. I’m a fan of the dashboard’s Data Notes section, which clearly defines all the dashboard’s terms and encourages users to reach out with questions.
University of Connecticut: UConn has five campus across the state, and each one gets its own section in this dashboard. The organization is very clean; it’s easy to navigate and compare campuses.
University of North Carolina at Chapel Hill: UNC-Chapel Hill gets a shout-out for being the only dashboard I analyzed to report COVID-19 cases by residence hall. The university’s overall COVID-19 response may have been a clusterfuck, but at least their dashboard gives students information to help them manage it.
As I mentioned earlier, I’d like to keep updating this spreadsheet as more COVID-19 school dashboards come online. If you’d like to add your school to the list or report an update or correction, you can do so using this Google form.
COVID-19 K-12 data continue to be patchwork
Since I reviewed COVID-19 data in public school systems last week, a few things have changed:
Alisha Morris’ crowd-sourced spreadsheet of COVID-19 cases, closures, and other reopening plan changes in K-12 schools has leveled up, thanks to collaboration with the National Education Association. Concerned students, parents, and teachers can now explore the data and report cases on a new NEA-run website. This NPR article explains more of the project’s background and findings.
Brown University and national education organizations are working together to track school pandemic responses. Their findings, planned to be released in a dashboard this September, will include: enrollment, poverty level, internet access, COVID-19 cases, and district opening status.
Duval County Public Schools, a district in Florida, was barred from reporting school-related COVID-19 cases by the county’s public health department. According to the Florida Times-Union, the county stated that the school district could not publish “school specific data related to COVID-19” without the state health department's permission.
Meanwhile, in Kentucky: this state’s COVID-19 dashboard now includes both a K-12 school public health report and a college and university public health report. Both reports, published daily, include active cases, deaths, and other statistics.
New reporting requirements for nursing homes, hospitals, labs
This week’s Department of Health and Human Services (HHS) update is a little indirect, in that it’s actually a Centers for Medicare & Medicaid Services (CMS) update.
Last Tuesday, CMS announced that nursing homes are required to test their staff and offer to test their residents when a COVID-19 outbreak occurs. Facilities that don’t test adequately can be fined. These new requirements follow the recent large-scale distribution of antigen tests to nursing homes, a move I’ve discussed in previous issues. It is still unclear, however, how exactly the results of all this testing will be reported. While the CMS COVID-19 dataset includes some new fields on testing this week, these are primarily yes/no fields, such as “Tested Asymptomatic Staff and/or Personnel Facility-Wide After a New Case.” Actual test counts are not yet reported.
The new CMS directive is also notable in that it includes new reporting directives for hospitals and COVID-19 labs. Hospitals were already required to report their COVID-19 patients, PPE needs, and other metrics to HHS, but now they are extra required to report them:
In March, Vice President Mike Pence sent a letter to all hospitals requesting that they provide the results of COVID-19 tests performed in their in-house laboratories to help better understand and track disease patterns. CMS’ new rules require such reporting of test results in order to ensure a more complete picture in the nationwide surveillance of COVID-19, as well as a more efficient allocation of PPE and other vital supplies. Hospitals will face possible termination of Medicare and Medicaid payment if unable to correct reporting deficiencies.
Outside laboratories conducting COVID-19 tests can similarly face monetary penalties if they fail to report to HHS: CMS will impose a fine of $1,000 a day for the first day a lab fails to report and $500 for each subsequent day. I and other folks at the COVID Tracking Project will be watching carefully to see how this directive impacts the completeness and accuracy of HHS data.
Featured data source
Charity Nagivator: I just have one featured source this week, but it’s crucial. Charity Navigator is a database that compiles and evaluates charities based on their financial health and accountability. The database has detailed rankings and search capability, and when a crisis hits—such as the recent California wildfires or Hurricane Laura—they put together quick lists of organizations where any interested samaritans can help the cause.
COVID source callout
In the COVID Tracking Project Slack server, we have emojis to express different challenges we face during data entry shifts. There’s :texas-4pm:, for days when Texas’ dashboard has not yet published new data by the state’s promised update time (4 pm Central). There’s :ltgovjoshgreen:, for the Hawaii Lieutenant Governor whose Instagram posts are, somehow, our most reliable source of hospitalization data for the state. And then, there’s my personal favorite: :fuckmichigan:.
I had to use :fuckmichigan: yesterday because Michigan’s dashboard hasn’t updated even though… by all intents and purposes… it sure did look like their dashboard had updated. The thing is, I’m pretty sure Michigan’s dashboard automatically changes its date to match the viewer’s time zone. The date has no bearing on the actual data displayed on that dashboard.
So if the state puts out its data update late, or updates its downloadable Excel files but not its dashboard, or updates its “Total Testing” page but not any other part of its dashboard, there’s no way to easily tell. You’ll just have to click through the dashboard ten times, download all the spreadsheets, compare every data point to the previous day’s data point, and start a meaningless Slack debate about timestamps just to finish out your goddamn shift.
Also, it’s been months, and Michigan’s hospitalization and PPE page is still labeled “Data About Places.” Hey Michigan, what the fuck does this mean.
More recommended reading
My recent Stacker bylines
States with the biggest nursing home outbreaks of COVID-19 (updated 8/26)
News from the COVID Tracking Project
That’s all for today! I’ll be back next week with more data news.
If you’d like to share this young newsletter further, you can do so here:
And if you have any feedback for me—or if you want to ask me your COVID-19 data questions—you can send me an email (firstname.lastname@example.org) or comment on this post directly:
Finally: this newsletter is a labor of love for me each week. If you appreciate the news and resources, I’d appreciate a tip: