INTRODUCTION

The practice of solitary confinement (SC) is so detrimental to mental health that the United Nations has determined that SC lasting greater than 15 days constitutes torture1. Solitary confinement has been shown to increase anxiety, hallucinations, and self-harm among those with and without prior history of mental illness. Per federal policy, SC should principally be implemented as a disciplinary measure for high-level offenses (e.g., assault); if used for administrative reasons (e.g., protective custody, pre-emptive isolation, suicide watch), SC should only be employed if restrictive housing is unavailable and for the briefest period possible2. However, the use of SC in US Immigration and Customs Enforcement (ICE) detention remains common, with 44,556 placements recorded between fiscal years (FY) 2015 and 2019.2 There are concerns that solitary confinement may be used during the SARS-CoV-2 pandemic for quarantine purposes or to control detainee unrest 3. This study investigates if the use of SC in ICE detention facilities changed during the pandemic.

METHODS

This analysis employs an interrupted time series model to determine whether the SARS-CoV-2 pandemic is associated with an increase in the use of SC. We digitally transformed and analyzed ICE Facility Significant Incident Reports from 2019 to 20224, produced as annual rolling retrospective reports detailing the monthly count of medical, mental health, disciplinary, and non-medical, non-mental health administrative SC placements from 128 active US immigration detention facilities. As this is public, ecological data, Institutional Review Board approval was not indicated.

The measure of interest was the total number of SC placements per 100 person-months, which trended between March 2019 and March 2022. To address autocorrelation, trends were analyzed as part of an interrupted time series with an autoregressive integrated moving average (ARIMA), under a linear model assumption, comparing trends before and after March 2020 (the beginning of the SARS-CoV-2 pandemic in the US). Analysis was done in R (version 4.1.3) with the astsa package.

RESULTS

There were 60,796 SC placements over the period studied. SC placements per 100 persons increased by a simple pre-post average of 215% between the pre-pandemic (mean 3.6 placements per 100 persons, standard deviation 0.4) and intra-pandemic (mean 11.3 placements per 100 persons, standard deviation 3.5) periods. Interrupted time series analysis shows a pre-post increase of 6.1 monthly placements per 100 persons (95% CI, 2.2, 10.1; monthly slope 0.08, 95% CI − 0.1, 0.3) (Fig. 1A). This increase is driven by a per-100 persons increase of 158.1% (pre-mean, 1.7, post-mean 4.4) in administrative placements, 6.8% (pre-mean 1.3, post-mean 1.4) in disciplinary placements, 129.4% (pre-mean 0.1, post-mean 0.3) in mental health placements and 905.6% (pre-mean 0.5, post-mean 5.4) in medical placements (Fig. 1B).

Figure 1
figure 1

Trends in Solitary Confinement in the US before and after March 2020 (the beginning of the US SARS-COV-2 pandemic). A Overall trends. B Trends by category (the reason for solitary confinement). Bands represent 95% confidence intervals with standard errors generated by local regression.

Full size image

DISCUSSION

This analysis has identified trends of increasing SC placements since March 2020, particularly in the medical and administrative categories. A report by the Governmental Accountability Office (GAO) analyzing a subset of SC placements found a FY 2019–2020 decline in all non-medical, non-mental health administrative SC placements besides “Other,” which grew 302% 5. Both facility-level and departmental officials are tasked with complying with federal policy to minimize administrative segregation. However, according to governmental investigations, 72% of ICE administrative placements did not review less restrictive options and several cases have lasted longer than 300 days2. Mental health patients have also been marginalized. According to the same GAO report, between the fiscal years of 2017 and 2021, the average SC placement among them (across all categories) was 31 days, with an average of 40 days for those with “serious” (i.e., incapacitating) mental health conditions, surpassing the United Nations 15-day limit defining torture5. Even medical segregation, while potentially necessary from an infection-control standpoint, might be more deleterious to mental health in a prison setting than in the community.6

This data has some limitations, principally that it depends on facility-level disclosures, leading to possible underreporting or miscategorizations for SC placements. However, our findings and the results of governmental investigations indicate significant central and facility-level failings to regulate SC. Beyond minimizing (if not eliminating) the use of SC altogether, ICE must better redirect individuals who have serious mental health conditions into intensive mental health care rather than SC. ICE must also consider placing a hard limit on the length of SC to minimize harm to detainees.