It seems like a simple question — is the Covid-19 outbreak getting better or worse in Oregon?
But the daily case counts that many news organizations obsess over don’t actually tell us much about the outbreak, at least not by themselves.
I talked to Kim Toevs from the Multnomah County Health Department about what metrics matter the most and stalked Nate Silver’s Twitter feed to see how data scientists are tracking the pandemic.
Here’s what I learned about how to make sense of the numbers:
1. The daily confirmed case count
What it tells us: How many people tested positive for Covid-19 on a given day.
What causes it to increase: President Trump has blamed the rising case counts on increased testing, but that’s only one of three factors that make a big difference:
- Infection rate. On average, how many additional people does each carrier of the virus infect? If that infection rate is greater than 1, the number of new cases will grow. If it’s less than 1, new cases will decline.
- Testing rate. What turns a new case into a new confirmed case is testing. Health officials must test people to know they have the virus, so the more testing that states do, the more likely they are to identify new cases.
- Contact tracing. Identifying new cases also depends on testing the right people, i.e. people who have actually been exposed to the virus. Contact tracing helps health officials find those high-risk populations and make sure they get tested.
The relative weight of each of these factors will vary from state to state, which is why it’s important to look at other metrics (like test positivity and hospitalizations) to get the full picture.
How to interpret the data: The confirmed case count is most helpful for comparing the severity of outbreaks in different parts of the state — but only if you adjust for population.
For example, Union County currently has 282 confirmed cases of Covid-19, and Multnomah County has 1,893 confirmed cases. So at first glance, you might think Multnomah is getting hit harder.
But after adjusting for population size, Union County has 105 confirmed cases per 10,000 residents, and Multnomah has 23 confirmed cases per 10,000 people.
So if someone tries to tell you that Covid-19 is just an urban problem, that’s the number to throw back at them.
Where to find the data: The Oregon Health Authority publishes daily case counts on its website, and this OHA data dashboards shows the per-capita case count for each county.
The New York Times also created a national map that shows where case counts are rising the fastest, but it’s not adjusted for population, so it’s only somewhat useful for comparing Oregon to other parts of the country.
2. The test positivity rate
What it tells us: The percentage of Covid-19 tests that come back positive.
(Reminder: A “positive” test means the virus was detected. A “negative” test means the virus was not detected.)
What causes it to increase: This metric is closely tied to the infection rate, i.e. if more people are contracting the disease, more people are going to test positive.
But again, it also matters who is getting tested.
For example, when Oregon hospitals started performing elective surgeries again in early May, many of them required patients to get a Covid-19 test before their surgery.
Those patients were very unlikely to test positive for Covid-19 (they didn’t have symptoms, they’d been pre-screened for exposure, etc.), so testing them likely caused the test positivity rate to go down.
On the other hand, Oregon counties also started hiring more contact tracers in May, and those contact tracers helped identify and test more people with known exposure to the virus, i.e. people who were more likely to test positive and cause the test positivity rate go up.
So long story short, the test positivity rate isn’t a foolproof metric.
But barring any dramatic changes in who’s getting tested statewide, it’s probably the best single barometer of whether Oregon’s outbreak is getting worse or better.
How to interpret the data: If the test positivity rate goes up in Oregon, it’s a strong sign that our outbreak is getting worse. And if it goes above 10 percent, it means we’re experiencing one of the worst outbreaks in the country.
Where to find the data: The Oregon Health Authority keeps track of the cumulative test positivity rate in each county, but that’s not as useful as knowing how test positivity is changing from week to week.
To get that data, you’ll have to crunch the numbers yourself — or become a Bridgeliner member and we’ll do it for you in our members-only newsletter.
3. The rate of untraced cases
What it tells us: The percentage of confirmed cases that cannot be traced back to another known case.
What causes it to increase: Ineffective contact tracing, which generally means that the state doesn’t have enough contact tracers to follow up with everyone who tests positive to investigate their source of exposure.
States can sometimes fix that problem by hiring more contact tracers, but once the number of new cases reaches a certain level, it becomes virtually impossible to hire enough people to keep up.
How to interpret the data: The Oregon Health Authority’s goal is to keep the rate of untraced cases below 30 percent statewide, but lower is even better.
Where to find the data: The chart in the bottom-left corner of this dashboard shows the rate of untraced cases statewide.
4. The number of hospitalizations
What it tells us: How many people are currently hospitalized with confirmed or suspected cases of Covid-19.
Why it might go up: The vast majority of people who test positive for Covid-19 never get hospitalized, so what really matters here is the infection rate among vulnerable groups, especially the elderly and people with pre-existing health conditions.
If more outbreaks start spreading at nursing homes, prisons, and other high-risk communities, the number of hospitalizations will likely go up.
But if we’re able to protect those groups from infection, hospitalizations could stay flat or decline, even if overall cases are rising.
How to understand it: We know that about 20 percent of Covid-19 patients who get hospitalized in Oregon end up dying, so tracking hospitalizations is one way to predict the death toll.
But it’s also about making sure our hospitals don’t get so overwhelmed that they have to start turning people away or providing substandard care.
Oregon has fewer hospital beds per capita than most other states, so if hospitalizations start rising steadily, the Oregon Health Authority may need to build temporary medical centers, acquire more ventilators, or take other steps to allow more patients to get care.
Where to find the data: The Oregon Health Authority addresses hospital capacity on Page 2 of its daily reports.
5. The death rate
What it tells us: The percentage of confirmed or suspected Covid-19 cases that ultimately result in deaths.
Why it might go up: More people in high-risk groups contract Covid-19, or, in the worst-case scenario, the virus mutates in a way that makes it more deadly.
On the flip side, the death rate could go down if more effective treatments emerge, if the virus mutates in a way that makes it less deadly, or if we get better at protecting people in high-risk groups from infection.
How to interpret the data: We probably won’t know the true death rate in Oregon for a long time, because so many early cases went undetected.
But over time, comparing the death rate of Covid-19 to the death rate of other viruses, like the seasonal flu, could help us make decisions about how much and how fast to ease restrictions on businesses and events.
Where to find the data: You can calculate the death rate with OHA’s numbers by dividing the number of reported cases by the total number of reported deaths.
But remember that many cases (and at least some deaths) have gone unreported in the U.S., so this number should be taken with a grain of salt.
6. The demographics
What it tells us: What types of people are most at risk of contracting Covid-19, and of dying from it.
How to understand it: We’ve learned that class and race are strong predictors of whether someone will get sick with Covid-19, with low-income Americans and people of color being infected at disproportionate rates.
We also know that people with preexisting health conditions are by far the most likely to die from Covid-19, and we know that about 60 percent of the deaths in Oregon have happened at nursing homes and other long-term care facilities.
One takeaway from that data is that preventing more deaths in Oregon starts with taking more safety precautions at long-term care facilities and providing stronger oversight.
But it’s also a reminder that the issues being highlighted by the Black Lives Matter movement right now belong front and center in the coronavirus conversation.
Because it’s not like Covid-19 randomly discriminates against people with dark skin.
The real reason we see racial disparities in the U.S. is that we’ve built a racial class system though generations of homesteading, redlining, “urban renewal,” environmental racism, overpolicing, mass incarceration, etc.
That class system is why Black and Latino Americans are more likely to work and live in the places with the highest risk of infection, and it’s why we’ve seen such disproportionate case counts in communities of color.
And one final caveat: Devin and I mentioned in Monday’s newsletter that Portlanders who identify as Black or Hispanic aren’t statistically any more likely to die from the virus, despite being much more likely to get sick with Covid-19.
We’d hoped that might be a success story, but Kim Toevs from Multnomah County Public Health says it’s too early to tell, because the death toll in Oregon is still quite low, and it’s also quite concentrated at a handful of long-term care facilities that aren’t necessarily representative of the state’s demographics.
Where to find the data: The Oregon Health Authority publishes demographic data for the entire state, or you can use this dashboard to see numbers just for Portland.
Want to follow our weekly updates?
Become a Bridgeliner member to get our weekly rundown of key metrics delivered to your inbox, and use the discount code DATADIVE to get your first month for free.