Data Update

Massachusetts Data Update November 3, 2020

Happy Election Day. I’ve had some time to digest the new dashboard from Massachusetts and the file download that accompanies it. I’m disappointed in the changes to the dashboard, but it isn’t as bad as it looked at first glance. However, it seems to me that the state still is trying to highlight statistics that paint the pandemic in Massachusetts in a more positive light, as I noted yesterday.

First, new individuals tested and their test positivity rate are not shown on the dashboard (this is what I reacted to yesterday). The newly-tested positivity rate is significantly higher than the positivity rate for repeat testers and about 75% of all tests currently are for repeat testers. However the testing information for new testers is still in the data download, if not on the dashboard itself.

The second significant change is to the hospital reporting. The state now only reports confirmed covid hospitalized patients, ICU patients, and intubated patients. This led to a fairly sizeable drop (almost 30%) in the overall patient count based on October 31st data.  Presumably, the suspected case patients are still in the hospital, regardless of whether they are counted by the state in the official tally.  However, the state did provide a backhistory of these confirmed hospitalization statistics, so it is possible to compare these data over time.  In addition, the state dropped the ‘New Daily Admissions Report’ altogether.  I had used this in my data updates, and have a less than satisfactory replacement.

In general, the state’s reporting of hospitalizations has been quite poor. The state’s cumulative hospitalization figures never matched that of hospitals, and once the state went to weekly reporting of hospitalizations by age group, the fourteen day totals in the weekly report never came close to matching the totals from the race / ethnicity report. I had used the race / ethnicity report to calibrate my estimates for hospitalizations by age group.  Furthermore, the state dropped the race / ethnicity report from the dashboard in this latest update.  As a result, it’s unclear that I’ll be able to provide meaningful estimates of hospitalizations by age group going forward. 

Table 1: Massachusetts Testing Statistics
7 Day  Trailing Average
November 3, 2020
Testing Statistic   Current 7 Days Ago 4 Weeks Ago
Test Positivity Rate (Individuals)   6.3% 6.3% 3.6%
Test Positivity Rate (Include Suspected)   7.0% 6.7% 3.8%
Test Positivity Rate (All Tests)   1.8% 1.9% 1.1%
Test Positivity Rate (Newly Tested)   6.3% 6.3% 3.6%
Test Positivity Rate (Repeat Testers)   0.4% 0.4% 0.2%
Percentage Repeat Testers   76.1% 74.8% 74.0%
Test Positivity Rate (Higher Ed)   0.1% 0.1% 0.1%
Test Positivity Rate (Non Higher Ed)   3.1% 3.3% 2.1%
Percentage Higher Ed Testers   45.2% 44.6% 48.9%
Newly Tested (Lagged 1 Week)   17,005 16,206 15,831
Higher Ed Tests (Lagged 1 Week)   30,062 32,177 29,654
All Tests (Lagged 1 Week)   67,410 67,827 58,373

Test positivity rates have held steady for the past several days – the first somewhat positive news in the last several weeks.  With the dashboard revision, the state appears to be providing daily, rather than weekly, updates on higher educational testing (not all the changes made by the state were misguided).  I’ve included them in this report.  As I’ve indicated in prior posts, higher education positivity rates are extremely low, and almost half the testing in the state since September 1st has been for higher education ( This higher ed testing has kept “all test” positivity lower than it would be otherwise, as Table 1 shows. 

Table 2: Massachusetts Hospitalization Statistics
7 Day Trailing Average
November 3, 2020
Hospitalization Statistic   Current 7 Days Ago 4 Weeks Ago
Confirmed  Patients Hospitalized   432 369 264
Confirmed  Patients in ICU   85 78 55
Confirmed  Patients Intubed   46 36 22
Net New Confirmed Patients   12 8 5
Net New ICU Patients   2 2 (2)
Net New Intubated Patients   1 1 0
 Percent ICU / Hospitalized   20% 21% 21%
 Percent Intubated / ICU   54% 46% 39%

I’ve revamped the hospitalization report shown in Table 2.  As noted previously, I’m now using the state-provided data for confirmed cases only.  In addition, I’m now showing the average daily net change in hospitalizations, ICU patients, and intubations over the last week.  For example, there has been an increase of 12 net hospitalizations over the past week, meaning that there are now approximately 84 (12 times 7) more hospitalized confirmed patients than one week ago.  The increase in the number of hospitalizations is greater now than it was four weeks or even one week ago. 

Table 3: Massachusetts Reported Case and Death Statistics
7 Day Trailing Average
November 3, 2020
Statistic   Current 7 Days Ago 4 Weeks Ago
Total Deaths Including Suspected   21 19 16
Total  Deaths Confirmed Only   21 18 16
Deaths in Long-Term Facilities (All Cases)   12 8 14
Percent from Long-Term Care   58% 45% 87%
Total Cases Including Suspected   1222 1085 625
Total Confirmed Cases   1135 1009 588

The structure of Table 3 is unchanged.  The 7 day average of reported confirmed cases remains above 1,000, even though the state reported only less than 1,000 confirmed cases for the past two days.  The death rate is holding steady, but has gone up a bit over the past several weeks.  Now that the state has converted all the graphics on its ‘Dashboard of Public Health Indicators’ graphics to 7 day trailing averages, you will notice some difference between my figures and those provided by the state for the number of cases and number of deaths. 

This is because I am calculating the 7 day averages using the cases and deaths reported each day, and the state is calculating its averages using the cases and deaths on the day they occurred (the “as-of” date), rather than the day they are reported.  I favor the state’s approach, but only with a lag to account for the fact that there is a lag in the reporting of deaths and cases so that the most recent days undercount the eventual totals for those days.  Consequently, I expect that the numbers reported on the ‘Dashboard of Public Health Indicators’ will tend to be lower than those I report.

Data Update Methodology

Massachusetts Reporting Change November 2, 2020

Well, the state did it again. That Charlie Baker certainly is sly. The state revised the daily dashboard to eliminate useful information and generally provide a more upbeat assessment of the state of coronavirus in Massachusetts. This is a quick summary of the changes as I’ve identified them so far. I will perhaps have more later as I work through the changes, and I will also have to modify the data updates to accommodate these changes.

There are three significant changes as far as I can tell. First, the state dropped information about the number of people tested for the first time each day, and the total number of people tested. This means that the newly tested positivity rate (which was significantly higher than the all test positivity rate) can no longer be determined, nor can the breakdown between newly tested individuals and repeat testers. While the state is still reporting new confirmed and cases,, they are no longer reporting how many new people have been tested to determine a positivity rate. [Update.  Likely my mistake. Although the state did drop this from the dashboard, I might have missed the first time tester information in the files to be downloaded.  I still might be able to calculate this].

Second, they eliminated the race/ethnicity report. This was a way to double check and properly scale the weekly breakdown of cases, hospitalizations, and deaths by age group. They are providing certain information by age, but I don’t think at first glance that it will be sufficient to continue to do this calculation.

Finally, they eliminated suspected hospitalizations from the hospital count. Thus, you will see a significant drop in hospitalizations today. That is fake news. In fact, 33 confirmed patients were added to the total today.

Disappointing on the whole. In my opinion, the state should be adding information to the report, not eliminating it.