June 23rd, 2021Erika Tyagi and Liz DeWolf

The challenges of interpreting vaccination data reported by carceral agencies

In January, our team began collecting data on vaccines administered in prisons and jails, information which has proved essential to tracking the state of COVID-19 behind bars. The addition of vaccine data to the data on cases, deaths, and tests we have been collecting through the pandemic allows us to better interpret facility-level trends and to monitor progress toward broader immunity within facilities.

On our homepage, we’ve been posting vaccination data in a simplified format via a table that shows the numbers of people who have received at least one dose of a vaccine for each state and federal agency that reports this information. We also estimate rates based on recent population data. In our raw data, available on GitHub, we include more granular information at the facility level (when reported by agencies) on the number of individuals who are partially and fully vaccinated, as well as the total number of doses that have been administered.

Unfortunately, in collecting vaccination data, we have encountered the same lack of transparency and consistency in data reporting that has hindered our COVID-19 data collection efforts since the beginning of the pandemic.

A persistent lack of transparency and consistency

Through our Data Reporting & Quality Scorecards, we’ve noted the chronic lack of transparency from carceral agencies when it comes to reporting COVID-19 data. When agencies do share data, they rarely define variables in a clear or consistent way, making it difficult to meaningfully compare data across systems. This problem has only been exacerbated as agencies have refused to report the nuances in vaccine administration, nuances which are critical to interpret and make meaning from the data that are reported.

The Colorado Department of Corrections (DOC), for example, added a figure — currently, 10,541 — labelled simply “Inmate Vaccinations” to its COVID-19 tracking dashboard. Though at first glance it may appear straightforward, such a vague label raises a number of questions: Does this number refer to the number of doses administered, or does it refer to the number of people who have received doses? If the latter, does it only include people who are fully vaccinated, or does it also include individuals who are partially vaccinated? Are only individuals currently in the agency’s custody included in the data, or are doses given to individuals who have since been released also counted? If an individual was vaccinated before entering the prison, are they included in the data?

Another example comes from the Tennessee DOC. In April, the agency added a box to its daily COVID-19 update reporting the total number of doses administered across the state’s prisons. Given that the agency has been administering both one-dose and two-dose vaccines, reporting only the number of doses administered leaves unclear how many people have been fully vaccinated. Of the 15,214 doses administered by the agency so far, were the majority of these one-dose vaccines, meaning as many as 66 percent of the roughly 23,000 people held in Tennessee’s prisons are fully vaccinated? Or were they mostly first doses of two-dose vaccines, meaning as little as 33 percent of the incarcerated population has been fully vaccinated?

These are just two examples of the issues we have encountered interpreting vaccine data. Others are listed below:

  • Reporting transparency: Currently, 25 agencies do not report any vaccination information on a regular basis. A handful of these agencies occasionally report updates through media reports, press releases, or social media posts; however, we do not consider this a suitable alternative to easily accessible and regularly updated data directly from the agencies.
  • Variables reported: 10 agencies report only the number of people who have received an initial dose of a vaccine, but not the number of people who are fully vaccinated. Four agencies — Arizona, South Carolina, Maine, and the Federal Bureau of Prisons — report only the number of people who are fully vaccinated, but not the number at least partially vaccinated. The Tennessee DOC, as noted above, reports only the number of doses administered, not the number of individuals who have received these doses.
  • Populations reported: 36 agencies do not report vaccine data for prison staff. The agencies that do report this data all add the caveat that staff are not required to report whether they have been vaccinated at an off-site location. All but three agencies do not specify who is included in the incarcerated populations they are reporting vaccine data for, which certainly includes those who were vaccinated while in and remain in custody, but may or may not include those who have since been released and/or those who received a vaccine before entering the prison.
  • Data granularity: 16 agencies report vaccine data for incarcerated people at the facility level, but 12 agencies only report these data at the systemwide level. While systemwide data may be useful for advocates to hold agencies accountable, it is all but useless from an epidemiological perspective. Behind bars, the virus spreads primarily within individual facilities, and as such, it is critical to know how many people within each facility have been vaccinated, not just across the state’s many facilities.

Challenges in calculating and interpreting rates

The purpose of collecting vaccine data is twofold: for advocates and the broader public to hold carceral systems accountable, but also for public health experts to evaluate the level of immunity within a given facility and hence the likelihood of future outbreaks. For the latter, we need to know the percentage of people currently in a facility who have been partially and fully vaccinated, regardless of when, where, and from whom they received their vaccine.

Unfortunately, very few carceral agencies report the data necessary to make these calculations, including the vaccination status of those currently in custody and up-to-date facility-level populations. Instead, most agencies that report vaccination data report the cumulative number of individuals who have ever been vaccinated by the agency, which could either under-count (by excluding those who were vaccinated before entering the facility) or over-count (by including those vaccinated by the agency who have since been released) the true number of people currently inside a facility who are vaccinated — and we often have no way of knowing which and to what extent each possibility exists.

As a result, we have settled for calculating rough estimates of vaccination rates, dividing the cumulative number of individuals vaccinated as reported by each agency by the most recent facility population (which, for some agencies, can be months or even years old). While this may have been a reasonable proxy for true vaccination coverage inside prisons early on, the cumulative vaccination rate will become less and less reliable as individuals continue to move in and out of prisons over the coming months. This problem is particularly acute in jails, where population churn is much higher than in prisons and where, concerningly, populations have been rising in recent months.

The critical need for improved data reporting standards

The transparency crisis around data from carceral agencies did not begin with this pandemic. Corrections agencies have long failed to report reliable data on a host of metrics directly implicating the physical and mental health of the incarcerated, including the use of solitary confinement, a practice employed by DOCs nationwide although condemned by the United Nations Special Rapporteur on Torture; the frequency of uses of force by staff against those in custody; the number and causes of deaths among incarcerated people; access to medical care; and other basic information necessary to monitoring the health and safety of incarcerated people.

The COVID-19 health data that have been reported by carceral agencies during the pandemic are largely piecemeal, unreliable, and woefully incomplete. That they have nonetheless provided the public with an unprecedented glimpse into the health of incarcerated people is an indication of just how secretive and opaque carceral institutions ordinarily are. Among the pandemic’s many lessons is the vital need for transparency as to what goes on inside prisons and jails — not only concerning COVID-19, but as to all aspects of carceral life that bear on the health and safety of the people who live and work inside.

As policy and health researchers, we are hopeful that even after the pandemic, agencies confining human beings will develop and implement standardized and clearly defined metrics and tools for contemporaneous reporting of data about the well-being of those in their custody.

Overview of data reported by agencies

The table below summarizes the data on vaccine administration reported by the 50 state correctional agencies, the District of Columbia’s Department of Corrections, the Federal Bureau of Prisons (BOP), and U.S. Immigration and Customs Enforcement (ICE) as of June 23, 2021.

Carceral AgencyIncarcerated people
(At least one dose)
Incarcerated people
(Fully vaccinated)
Staff
(At least one dose)
Staff
(Fully vaccinated)
BOPNo dataBy facilityNo dataBy facility
ICENo dataNo dataNo dataNo data
AlabamaBy facilityNo dataBy facilityNo data
AlaskaBy facilityBy facilityNo dataNo data
ArizonaNo dataSystemwide onlyNo dataNo data
ArkansasNo dataNo dataNo dataNo data
CaliforniaBy facilityBy facilityBy facilityBy facility
ColoradoSystemwide onlyNo dataSystemwide onlyNo data
ConnecticutSystemwide onlyNo dataSystemwide onlyNo data
DelawareSystemwide onlySystemwide onlySystemwide onlySystemwide only
District of ColumbiaNo dataNo dataNo dataNo data
FloridaNo dataNo dataNo dataNo data
GeorgiaBy facilityNo dataBy facilityNo data
HawaiiNo dataNo dataNo dataNo data
IdahoBy facilityBy facilityNo dataNo data
IllinoisNo dataNo dataNo dataNo data
IndianaNo dataNo dataNo dataNo data
IowaNo dataNo dataNo dataNo data
KansasBy facilityNo dataBy facilityNo data
KentuckyNo dataNo dataNo dataNo data
LouisianaNo dataNo dataNo dataNo data
MaineNo dataSystemwide onlyNo dataNo data
MarylandBy facilityBy facilitySystemwide onlySystemwide only
MassachusettsSystemwide onlySystemwide onlySystemwide onlySystemwide only
MichiganBy facilityBy facilityNo dataNo data
MinnesotaBy facilityBy facilityBy facilityBy facility
MississippiNo dataNo dataNo dataNo data
MissouriSystemwide onlyNo dataNo dataNo data
MontanaNo dataNo dataNo dataNo data
NebraskaNo dataNo dataNo dataNo data
NevadaNo dataNo dataNo dataNo data
New HampshireBy facilityBy facilityNo dataNo data
New JerseySystemwide onlyNo dataSystemwide onlyNo data
New MexicoNo dataNo dataNo dataNo data
New YorkNo dataNo dataNo dataNo data
North CarolinaSystemwide onlySystemwide onlySystemwide onlySystemwide only
North DakotaBy facilityBy facilityNo dataNo data
OhioSystemwide onlyNo dataSystemwide onlyNo data
OklahomaNo dataNo dataNo dataNo data
OregonNo dataNo dataNo dataNo data
PennsylvaniaBy facilityBy facilityBy facilityBy facility
Rhode IslandNo dataNo dataNo dataNo data
South CarolinaNo dataBy facilityNo dataNo data
South DakotaNo dataNo dataNo dataNo data
TennesseeNo dataNo dataNo dataNo data
TexasNo dataNo dataNo dataNo data
UtahNo dataNo dataNo dataNo data
VermontNo dataNo dataNo dataNo data
VirginiaSystemwide onlyNo dataSystemwide onlyNo data
WashingtonSystemwide onlyNo dataSystemwide onlyNo data
West VirginiaBy facilityBy facilitySystemwide onlySystemwide only
WisconsinBy facilityBy facilityNo dataNo data
WyomingNo dataNo dataNo dataNo data

next post

June 25th, 2021Erika Tyagi and Poornima Rajeshwar

Counting the Shots: The State of COVID-19 Vaccinations Behind Bars

As we continue to survey the vaccine rollout in prison systems across the country, we have identified several key takeaways that raise cause for concern.