Since a new variant wave is starting to hit the US, I spent a lot of time trying to find data that could help paint a larger picture of what is happening in a particular state given that much of the CDC data has gone extinct. I’ve stayed off social media to get this done. It was far more work than expected. In the past, it was possible to download 3-4 national data sets and then pull from them locally. Now, the way the data is configured online, I have to download individual data sets for each state each week. It’s frustrating.
Here’s a breakdown of the visualizations on the US page. Given the changes in data for many of these metrics, the most important thing is to look at the change over the last couple of months as opposed to comparing one particular point to one a few years prior. This example is California.
- Early Indicators Tab
- The background is the percentage of each variant colored using the same color scheme as can be found on covariants.org.
- The black line is the PCR positivity rate which is on the right axis.
- The brown line is wastewater and is really more of a relative scale, especially how the baseline that the CDC uses isn’t constant.
- The hollow black line is COVID emergency department visits.
- Google Searches – This is the newest set of data I’ve added to the site. It’s been available for a long time, but given that there was plenty of hospitalization data until recently, it seemed a bit superfluous. The search terms are listed in parentheses. This data comes from the Delphi Group at Carnegie Mellon University.
- Control (type 2 diabetes, urinary tract infection, hair loss, candidiasis, and weight gain) – These are medical search terms that have nothing to do with COVID, although one could argue that searches for diabetes will be increasing over time as a result of COVID infections.
- Nasal (nasal congestion, postnasal drip, rhinorrhea, sinusitis, rhinitis, and common cold)
- Oral (cough, phlegm, sputum, and upper respiratory tract infection)
- Fever (fever, hyperthermia, chills, shivering, and low grade fever)
- Throat (laryngitis, sore throat, and throat irritation)
- Lung (shortness of breath, wheeze, croup, pneumonia, asthma, crackles, acute bronchitis, and bronchitis)
- Senses (anosmia, dysgeusia, and ageusia – these are the proper terms for the loss of smell, foul taste, and loss of taste, respectively). The Delphi Group made a particular point that “The symptoms in this set showed positive correlation with cases, especially after Omicron was declared a variant of concern by the WHO.”
- Medical Care – these are the percentages of diagnosed COVID patients in each of these settings.
- Outpatient
- Emergency Department
- Hospital admissions – this is a particularly lagging indicator because it is coming from claims data.
Thanks Mike. This is clearly been a great deal of work and is very helpful. It is so unbelievable to me how much data the government has stripped away and has left everyone flying blind. In my county, Anne Arundel Maryland, they use to report 7 wastewater sites. Now all I can find is 2 for the state of Maryland. It looks like they may be bringing more back online but from what I can find the only way they are reporting this out is by test site #, not wastewater plant name. And I can’t find the key to the test site numbers. They are doing everything possible to obscure the date for ordinary folks. And now they are going after the masks.
Thank you for all the work you put into this. What I find most interesting is how the brown wastewater line consistently is predictive of the upcoming rise in PCR positivity. With lack of reporting of cases in the community, all we have left is wastewater monitoring (if your community is lucky). Thus communities should start to ‘buckle up’ when they see the ww levels rising. All communities should have this barometer to gauge how they proceed with caution in public spaces (increase masking and ventilation, consider remote options for meetings, test when sick and stay home if positive, etc). We use weather predictions to alter what we wear for the changes in weather, thus the same should apply with covid monitoring. We should include ww data with weather and pollen counts in weather apps and programming.