Blog

Language Matters: Reporting COVID-19 in Prison Systems

CPP collects and analyzes data on five primary variables reported by 53 sources: each state prison system, ICE, the Federal Bureau of Prisons, and Puerto Rico. In an analysis of definitions available on each system’s website, discrepancies in language used to report COVID-related data were identified. Specifically, definitions of the number of people who are incarcerated who are tested for COVID-19 (“Inmates Tested”)  and positive cases in staff (“Staff Positive”) vary. These differences in terminology are important: reporting the number of tests given does not capture how many of them are re-tests of the same individual, due to either re-exposure to the virus or sentinel surveillance testing. 

Historically, CPP has defined “Inmates Tested” as the total number of incarcerated individuals in prisons who have received a COVID test. This was mostly the case early in the course of the pandemic wherein testing was slow to ramp up. However, more recently, as reductions in population have occurred and more robust testing efforts have been deployed, systems have begun defining their testing data disparately. Our team recently did a content analysis of reporting across all of the systems we are tracking. What we found is detailed here in Table 1. 

Definition of “Inmates Tested”Percentage  
Number of COVID tests given 23%
Number of people tested26%
Undefined 30%
Variable not reported by DOC 26%
Table 1. System Definitions of “Inmates Tested.” Note: Colorado, Vermont, and Washington report both the number of people tested and the number of tests given. 

Very few of the systems reporting data are providing information relevant to staff testing. Out of 53 systems, only 7 are reporting the number of staff that have been tested (defined by CPP as “Staff Tested”). Of these 7 systems, only 1 defines the variable as the number of DOC-administered tests to staff. For the remaining systems, 3 leave “staff tested” as undefined, and 3 specify that testing is self-reported by staff members. 

Systems should aim to be clear in how they define variables related to COVID-19 testing and cases, particularly when it comes to re-exposures and retesting incarcerated people and staff members. In light of these findings, CPP will begin to report two categories of data for relevant systems: both the number of people tested and the number of tests given. For more insight into how systems and CPP define these COVID-related variables, check out our “Data Dictionary” here. We continue to re-evaluate how these definitions differ between systems and what it means for the standardization of data on CPP’s platform.  

COVID Case Watch October 28th, 2020

This graph shows confirmed positive cases of COVID-19 per 1,000 individuals in the prison and general populations in each state as of October 28, 2020. The left side (orange) refers to the prison population and the right side (blue) refers to the general population. 

The rate of COVID in the general population is 26.61 per 1,000 individuals. In the prison population (in state prisons), the rate is 113.35 per 1,000 individuals. This means that, on average, the rate of COVID infections in prison populations is 4.26 times the rate of COVID infections in the general population. Arkansas, Kansas, and Vermont continue to have a prison infection rate that is over 10 times that of the general population. 

Despite a surge of COVID cases in the general population this month, in 45 of the 50 US states, the rate of COVID infections among those imprisoned exceeds the rate of COVID infections among the general population. As cases have risen in the Midwest, the general population of Montana, North Dakota, South Dakota, and Wisconsin have seen an increase in case rates of 4.9-7.9 per 1,000 in the past week. However, the case rates in Midwestern prisons have increased much more. Montana, North Dakota, South Dakota, and Wisconsin prisons have seen an increase in case rates in the past week of 54 per 1,000, 16 per 1,000, 302 per 1,000, and 38 per 1,000, respectively. 

Importantly, states have varying testing strategies within prisons and for their general population, indicating that these rates likely reflect a falsely low disease incidence with some states’ rates being more accurate than others.

COVID Case Watch October 21, 2020

This graph shows confirmed positive cases of COVID-19 per 1,000 individuals in the prison and general populations in each state as of October 21, 2020. The left side (orange) refers to the prison population and the right side (blue) refers to the general population. 

The rate of COVID in the general population is 25.06 per 1,000 individuals. In the prison population, the rate is 106.84 per 1,000 individuals. This means that, on average, the rate of COVID infections in prison populations is 4.26 times the rate of COVID infections in the general population. Arkansas, Kansas, and Vermont continue to have a prison infection rate that is over 10 times that of the general population. In 42 of the 50 US states, the rate of COVID infections among those imprisoned exceeds the rate of COVID infections among the general population.

Importantly, states have varying testing strategies within prisons and for their general population, indicating that these rates likely reflect a falsely low disease incidence with some states’ rates being more accurate than others.

COVID Case Watch October 14-15, 2020

This graph shows confirmed positive cases of COVID-19 per 1,000 individuals in the prison and general populations in each state as of October 14 and 15, 2020.* The left side (orange) refers to the prison population and the right side (blue) refers to the general population. 

The rate of COVID in the general population is 23.81 per 1,000 individuals. In the prison population, the rate is 103.49 per 1,000 individuals. This means that, on average, the rate of COVID infections in prison populations is 4.35 times the rate of COVID infections in the general population. Arkansas, Kansas, and Vermont have a prison infection rate that is over 10 times that of the general population. In 41 of the 50 US states, the rate of COVID infections among those imprisoned exceeds the rate of COVID infections among the general population.

Importantly, states have varying testing strategies within prisons and for their general population, indicating that these rates likely reflect a falsely low disease incidence with some states’ rates being more accurate than others.

*Based on when data is updated, general population data is from October 14th and state prison data is from both October 14th and October 15th

COVID Case Watch September 30, 2020

This graph shows confirmed positive cases of COVID-19 per 1,000 individuals in the prison and general populations in each state as of September 30, 2020. The left side (orange) refers to the prison population and the right side (blue) refers to the general population. 

The rate of COVID in the general population is 21.78 per 1,000 individuals. In the prison population, the rate is 96.11 per 1,000 individuals. This means that, on average, the rate of COVID infections in prison populations is 4.41 times the rate of COVID infections in the general population. Arkansas, Kansas, Michigan, and Vermont all have a prison infection rate that is over 10 times that of the general population. In 40 of the 50 US states, the rate of COVID infections among those imprisoned exceeds the rate of COVID infections among the general population.

Since these rates were last calculated by our team on August 19th, the general population has experienced an increase in cases of 5.23 per 1,000 population whereas the prison population has experienced an increase in cases of 26.11 cases per 1,000 population. 

Importantly, states have varying testing strategies within prisons and for their general population, indicating that these rates likely reflect a falsely low disease incidence with some states’ rates being more accurate than others.

Covid Prison Hotspots

This week 22 out of 53 prisons systems are classified as Covid Prison Hotspots. This is our fourth posting.

We were inspired by the Kaiser Family Foundation‘s characterization of states as “covid hotspots” based on recent changes in cases and test positivity for the general population. So we applied their analysis to prison systems! For each prison system (50 states, FBOP, ICE, and Puerto Rico), we calculated the 14-day percent change in covid cases, the 14-day percent change in test positivity (i.e., the percent of tests that return positive), and the 7-day “rolling” average for test positivity, all for incarcerated people (not staff). If a prison system reports a 5% or greater increase in covid cases over the two week period AND reports a 1% or greater increase in test positivity over the two week period OR reports an average test positivity of 10% or higher, then they are classified as a Covid Prison Hotspot.

Covid Prison Hotspots

This week 23 out of 53 prisons systems are classified as Covid Prison Hotspots. This is our fourth posting.

We were inspired by the Kaiser Family Foundation‘s characterization of states as “covid hotspots” based on recent changes in cases and test positivity for the general population. So we applied their analysis to prison systems! For each prison system (50 states, FBOP, ICE, and Puerto Rico), we calculated the 14-day percent change in covid cases, the 14-day percent change in test positivity (i.e., the percent of tests that return positive), and the 7-day “rolling” average for test positivity, all for incarcerated people (not staff). If a prison system reports a 5% or greater increase in covid cases over the two week period AND reports a 1% or greater increase in test positivity over the two week period OR reports an average test positivity of 10% or higher, then they are classified as a Covid Prison Hotspot.

Covid Prison Hotspots

This week 22 out of 53 prisons systems are classified as Covid Prison Hotspots. This is our fourth posting.

We were inspired by the Kaiser Family Foundation‘s characterization of states as “covid hotspots” based on recent changes in cases and test positivity for the general population. So we applied their analysis to prison systems! For each prison system (50 states, FBOP, ICE, and Puerto Rico), we calculated the 14-day percent change in covid cases, the 14-day percent change in test positivity (i.e., the percent of tests that return positive), and the 7-day “rolling” average for test positivity, all for incarcerated people (not staff). If a prison system reports a 5% or greater increase in covid cases over the two week period AND reports a 1% or greater increase in test positivity over the two week period OR reports an average test positivity of 10% or higher, then they are classified as a Covid Prison Hotspot.

COVID-19 Tests per 1,000

In the last week, Wyoming has begun reporting testing information within prisons. Of the 41 states reporting testing information, 21, including Wyoming have administered more than 500 COVID tests per 1,000 inmates. Eight states have administered more than one test per inmate with Minnesota administering over three tests per inmate. Multiple prison systems are still testing fewer inmates per 1,000 than their state’s general population.  

COVID-19 Tests per 1,000

In the past two weeks, no additional states have reported testing information within prisons. Of the 40 states reporting testing information, 20 have administered more than 500 COVID tests per 1,000 inmates with eight administering more than one test per inmate. Multiple prison systems are still testing fewer inmates per 1,000 than their state’s general population.