Skip to main content

Using Wastewater Data to Track Case Levels

By May 13, 2022Commentary

A bizarre, but interesting twist on CV-19 data collection is that residues of viral infections can be found in wastewater, in other words, we are all pooping out viral fragments when infected.  Tracking those levels can be a somewhat objective method of identifying possible case levels.  That assumes some continuity in average viral loads over time.  Assuming that wastewater identification methods are consistent (more on that in Dave’s notes) comparing the wastewater measures with actual cases reported may give you some information on changes in the ratio of cases detected to undetected.  With more home testing and more mild infections, you might expect that the wastewater would have a higher ratio now than earlier in the epidemic.  That would be reflected in an increasing divergence between the wastewater measure and the number of reported cases.  In fact, in various regions of the state, the relationship is all over the place.  I suspect something is screwed up in the wastewater measuring.  Hard to imagine the process is very controlled.

What we see is generally a consistent trend on cases reported across regions, some rising a little faster than others, but a less consistent trend on wastewater, including some dramatic reversals.  Since wastewater is likely to be a somewhat lagging indicator (although the U claims it is a leading one but I don’t see that on the charts), but here looks like a leading indicator in general, I just think something is out of whack in the measuring process.  But it is interesting to look at and we will keep an eye on it.  I am, however, disappointed that the data does not seem particularly useful for identifying changes in the ratio of detected cases.

Dave’s notes:

  1. The University of Minnesota publishes summary data of Covid sampling of wastewater in Minnesota here: https://experience.arcgis.com/experience/a8d269bd670a421e9fd45f967f23f13c?data_id=dataSource_1-17ed5c83cca-layer-5%3A2 The U of M is monitoring wastewater plants for the presence of Covid indicators. They state that the amount of Covid detected in wastewater is predictive of clinical cases of Covid 1 to 2 weeks later. They also state that they are studying whether their metrics also correlate to hospitalizations.
  2. In the following charts we are plotting the U of M metric “Mean Weighted Average (N and O genes)” of virus copies per liter (blue curves) along with the newly reported covid cases per 10,000 residents in each region (red curves). We have yet to find any documentation describing exactly what the mean weighted average represents. For the purposes of this post it is assumed that the weighting process accounts for the differing populations in each region. For this reason we have chosen to plot Covid cases as the cases per 10k residents, rather than simply number of Covid cases per region.
  3. It is important to note that the U of M states that they changed their methodology on 3/15/2022. This means that the older data they publish, from 1/1/2021 to 3/15/2022, is not directly equivalent with the post-3/15/2022 data. We are only publishing the newer data in these charts. Also, the mean weighted metric we are displaying only starts on 4/3/2022. We will likely plot and publish the pre-3/15/2022 data separately in the future.
  4. The cases per 10,000 plotted in the charts are not the cases published by the U of M. Rather, we are taking cases per week per county from the Minnesota Department of Health Weekly Report archive, https://www.health.state.mn.us/diseases/coronavirus/stats/index.html, which includes cumulative cases per county each week. From this data we calculate cases per week, and then cases per week per region per 10,000 population.
  5. The regions are defined by the graphic on the U of M wastewater surveillance web page, which we have reproduced in Fig. 8.
  6. It is interesting that there appears to fairly good correlation between the wastewater data and newly reported covid cases per capita. If the wastewater data were published in a timely manner then it could be predictive of the future rise of fall in Covid cases. Most recently the wastewater data for 5/8/2022 and 5/11/2022 were published at the same time on 5/11/2022. If this reporting schedule holds then every Wednesday we would have up-to-date wastewater data, which is more timely than the 5 to 7 day average reporting lag for Covid cases currently.
  7. The plots show the latest wastewater data released on 5/11/2022, but the case data is from the prior Thursday Minnesota Weekly Report from 5/05/2022. We can see that the wastewater virus data for the Metro, Southeast, Southwest, and Northwest regions all show decreasing amounts of virus, and the other regions are increasing. It will be interesting to see if the next county case data released on 5/12/2022 matches this prediction from the wastewater data.
  8. Another possible use for this data is the determination of Covid cases independent of the standard testing regime, which is subject to variations such as differing testing rates and the use of at-home tests which aren’t reported to the authorities. Unfortunately, the change in methodology likely prevents us from directly comparing the number of Covid cases to wastewater Covid levels over the entire pandemic. However, it is possible that wastewater virus data may be useful in indicating overall trends in the population.

Leave a comment