This is truly yeoman’s work by Dave Dixon. I believe it is quite accurate, despite the difficulty of working with DOH data. As before, there is a lengthy explanation of the steps taken to get these rates, very lengthy, but we want people to replicate our work, and if we made mistakes, let us know. Minnesota DOH could fix the data issues in a second if it wanted to, and its failure to do so can only be attributed to a desire to hide from people the true age bracket case rates. The general message is that getting vaccinated is protective for older age groups. But older people even when vaccinated still get sick and hospitalized and die from CV-19. These case rates are consistent with the proportion of events chart, also published today, which show rising percents of events among the vaxed.
- The table and charts below display the Covid hospital admissions case rates, and Covid death case rates, for Minnesota sorted by age group, comparing breakthrough and non-breakthrough cases. The case rates are simply the percentage of those who test positive for Covid go on to be admitted to a hospital or to die. These are cumulative rates over the course of the pandemic, over the time period for specimen collection dates of cases from 1/19/2021 to 11/15/2021 for hospital admissions, and from 1/30/2021 to 10/30/2021 for deaths. These date ranges were selected in order to cover the time period during which there were breakthrough events reported, and during which there is a reasonable likelihood that admissions or deaths would be reported, as described more fully below. Choosing different time periods for the analysis would generate different results of the analysis.
- These tables and charts are an update of the data table published originally on 11/15/2021.
- On 10/28/2021 we had requested that the Minnesota Department of Health (MDH) provide us, under the Minnesota Data Practices Act (https://www.health.state.mn.us/data/datapractices/index.html), information regarding Minnesota’s breakthrough Covid cases (breakthrough cases being defined as someone who tests positive for Covid 14 or more days after completing their vaccination series). Specifically, for each breakthrough case, we asked that MDH provide us the date of specimen collection, the date of hospital admission, the date of death, and the age of each individual. This Data Practices Act request (ID No. 21-211) was separate from a much more comprehensive request submitted some weeks ago, and we hoped that perhaps a simple, more focused request might be acted upon. On Wednesday, Nov. 10, 2021 MDH provided an Excel spreadsheet containing information about 57,023 breakthroughs cases. The data provided was the date of specimen collection, the age of the individual, and Yes/No flags for whether the case was hospitalized or had died. We were not given the date of hospitalization or death, and we followed up with a request for these dates as well.
- On 11/24/2021 MDH provided an updated version of the same data file originally provided on 11/10/2021, in response to a request for updated information. This new data file contains the specimen collection date, age, and yes/no flags for hospitalization and death for 83,933 breakthrough Covid cases, identical to the data provided in the original data file.As of 11/22/2021, MDH’s Vaccine Breakthrough Report web page https://www.health.state.mn.us/diseases/coronavirus/stats/vbt.html also reports a total of 83,933 cumulative breakthrough cases, exactly matching the total cases in the data they provided (Note that it is necessary to use the Internet Archive https://archive.org/web/web.php to look at past data on the Vaccine Breakthrough web page).
- In the new breakthrough data provided by MDH, the earliest breakthrough case is 1/20/2021, and the most recent is 11/14/2021. The first version of the data file provided by MDH had the earliest breakthrough case on 3/08/2021. For this reason the time period over which we calculative the case rates has been extended to match the range of breakthrough cases in the data file.
- Compared to the first version of this table published on 11/15/2021, there are several major changes. 1). The basic analysis method is similar. We are simply comparing the breakthrough cases, hospital admissions, and deaths to the overall cases, hospital admissions, and deaths, sorted by age group. 2). All of the underlying breakthrough data is revised using the new data file received on 11/24/2021. 3). The time period analyzed is much longer. The new data file as an earliest breakthrough case of 1/20/2021, while the older data file had an earliest case of 3/08/2021. The new file also has 2 weeks of newer data on the recent end of the file.
- Because MDH did not provide the dates of hospital admission and death, only yes/no flags, we do not know the exact timing when these events occurred. We have assumed that hospitalization occurs 3 days on average after the specimen collection date of a positive case, and that death occurs on average 14 days after specimen collection date of a positive case. Using assumed time lags rather than actual dates should have minimal impact on the accuracy of the analysis, since we are calculating cumulative totals over the entire time period being analyzed.
- Since the data file provided contains breakthrough cases, hospital admissions, and deaths by age group over a time period, we then need to compare this data to the overall cases, hospitalizations, and deaths by age group over the same time period. The Minnesota Weekly Report (https://www.health.state.mn.us/diseases/coronavirus/stats/index.html) provides weekly totals for hospital admissions and deaths by age group. However, this data is based on the date of report, and not when the hospital admission or death actually occurred. For this reason, in order to match actual date of events, we assume that there is a 1 week lag between actual date of hospitalization and appearance in the Minnesota Weekly Report. In addition, we assume there is a 2 week lag between date of death and appearance in the Minnesota Weekly Report. These lags are consistent to the lags that found for data being published on the Minnesota Situation Update web page (https://www.health.state.mn.us/diseases/coronavirus/situation.html). Because the actual lag between actual hospital admission and death, and publication in the Minnesota Weekly Report is unknown, the total hospital admissions and deaths, and therefore the unvaccinated hospital admissions and deaths, may be different than the totals appearing in the table below. Since the only age group data we have for hospital admissions and deaths is based on date of report we have no way around this source of error currently.
- The Not Fully Vaccinated Hospital Admission Case Rate is calculated as the number of non-breakthrough hospital admissions divided by the number of non-breakthrough cases, resulting in the percentage of non-breakthrough cases that are hospitalized. Non-breakthrough hospital admissions are found by taking the total hospital admissions (computed by taking differences between successive cumulative totals from the Minnesota Weekly Report https://www.health.state.mn.us/diseases/coronavirus/stats/index.html ) minus the breakthrough hospital admissions computed for the time period. We have determined in other analysis that there is an average lag of 1 week between hospitalization and reporting in the MDH data files and reports. Non-breakthrough cases are found by computing the total cases by age group from the data file wagscd.csv published on https://www.health.state.mn.us/diseases/coronavirus/stats/index.html#wagscd1 and subtracting the breakthrough cases described above.
- The death case rates are found by a similar process to that used to find the hospital case rates, but using a 2 week assumed lag from date of death to reporting in the Minnesota Weekly Report.
- Potential sources of error:
It is likely that there are people who may have contracted a mild enough Covid infection that they did not seek medical attention or get tested, causing an undercount of the true number of breakthrough infections.
We know that there are likely to be breakthrough cases that have not been identified by MDH, because the vaccination occurred out of state, by the VA, or by a Native American medical facility. In response to a query to MDH on 9/23/2021 we were informed that MDH has access to vaccination records of Minnesotans vaccinated in Iowa, Wisconsin, and North Dakota if they gave the provided a Minnesota address. MDH does not have access to vaccination records in any other state. MDH does not have access to Veterans Administration vaccination records, and only partial or limited access to Tribal health organizations vaccination records.
We are unsure how complete the data matching process is for MDH even for the vaccination records in their possession. We suspect, for example, that on 11/08/2021 that MDH reported additional breakthrough cases as early as 5/02/2021, by examination of the breakthrough case, hospital admission, and death rates associated with the charts and tables on the Vaccine Breakthrough Report web page https://www.health.state.mn.us/diseases/coronavirus/situation.html#vbt2 . It is reasonable to assume that in the future additional older breakthrough cases will be identified and published.
The time period used for the analysis may not have complete breakthrough case reporting.
MDH publishes no data on the ages of those given booster shots over time. We are only given the cumulative number of booster shots administered each week day on the Vaccine Data web page (https://mn.gov/covid19/vaccine/data/index.jsp). We therefore cannot take the effects of booster shots into account in this analysis
- By comparing the Vaccinated and Not Fully Vaccinated Case Rates for each age group we can judge the relative risk between vaccinated and unvaccinated people in each age group, and the relative risks between age groups.
- We can see, not surprisingly, that the older age groups are at greater risk of hospitalization and death, whether vaccinated or not, and the non-vaccinated are at higher risk of hospitalization and death than the vaccinated, once testing positive for Covid.
- The Not Fully Vaccinated death case rate for the 100+ age group is 116.67%. There were 14 deaths and only 12 cases among the Not Full Vaccinated population in the 100+ age group. This illustrates the potential errors that are occurring by being forced to assume the lag from date of specimen collection to death, and then from death to publication in the Minnesota Weekly Report.
- The overall hospital admission case rate for the vaccinated is 4.33% and for the not fully vaccinated is 5.38%. This means that in general the vaccinated are less likely to be admitted to a hospital than the unvaccinated. The ratio of the rates, 4.33%/5.38% = .805, means that the vaccinated are 80.5% as likely to be admitted to a hospital, when averaging over the entire population over the entire time period analyzed.
- The death case rates are the percent of positive cases who go on to die. We can see that the unvaccinated have much higher case rates than the vaccinated for all of the age groups that are 20 years old and older. However, because we included the under 20 age groups in the overall average, of overall death case rate for the unvaccinated is lower than for the vaccinated, 0.71% vs. 0.88%.