How many immigration detainees have died in ICE custody since 2020?
This fact-check may be outdated. Consider refreshing it to get the most current information.
Executive Summary
ICE-reported data and independent analyses do not yield a single, definitive tally of detainee deaths since 2020, but combining available figures produces a known minimum of 53 detainee deaths in ICE custody from calendar/FY 2020 through 2025, with important gaps and reporting caveats that prevent a precise total [1] [2] [3]. The discrepancy arises from differing year definitions, undercounting of people who died after release, and missing year-by-year public reporting for 2021–2024 [4] [2].
1. What the major sources actually claim — a concise extraction that matters
The independent reporting and research present three key numeric anchors: ICE itself and a journalism account show 21 deaths for fiscal year 2020 and report at least 20 deaths in calendar year 2025 so far, with 2025 described as the deadliest year since 2004 [1]. A peer‑reviewed analysis published in 2024 documents 12 deaths between FY2021 and FY2023, and contrasts that with 38 deaths in FY2018–2020, implying lower mortality in the immediate post‑2020 period [2] [3]. A civil‑society report cites at least 70 deaths between Jan 1, 2017 and June 20, 2024, but does not break that total out by calendar year for 2021–2024, and it warns that ICE reporting omits deaths that occurred shortly after release [4]. These discrete figures are the factual anchors available.
2. Adding the figures — why you can compute a minimum but not a final total
Summing the available figures yields a defensible minimum: 21 deaths in FY2020, plus 12 deaths in FY2021–FY2023, plus 20 deaths recorded in 2025 to date, equals 53 known deaths attributable to ICE custody across the covered periods [1] [2]. This sum is explicitly a lower bound because the civil‑society dataset and ICE reporting practices leave 2024’s in‑custody toll unclear, and both ICE and some analyses exclude individuals who died after release or during processes not counted as “in custody,” creating systematic undercounts [4]. Consequently, any single, definitive cumulative number since 2020 cannot be asserted from the present record without additional year‑by‑year disclosures.
3. Why counting is complicated — definitions, timing and excluded cases change the math
Public data are affected by at least three technical counting issues. First, ICE often reports deaths by fiscal year while media reports sometimes use calendar years; that mismatch shifts which deaths fall into “2020” or “2021” depending on the convention [1]. Second, ICE’s official tallies historically exclude deaths that occur after release or deportation shortly before death, which advocacy groups and researchers say creates a blind spot in mortality assessments [4]. Third, independent research uses admission‑based denominators and distinguishes medical versus nonmedical causal attributions, producing different perspectives on trends even when raw counts are similar [2]. These definitional gaps materially affect any attempt at a single cumulative number.
4. What the trend data suggest — lower rates after 2020, but a concerning 2025 uptick
Peer‑reviewed work reports a decline in the aggregate death rate per 100,000 admissions in FY2021–FY2023 compared with FY2018–FY2020, with 12 deaths in FY2021–2023 versus 38 in FY2018–2020, suggesting some improvement in the immediate post‑2020 period and possible effects from policy and release practices [3] [2]. However, journalism and ICE statements indicate 2025 has been unusually deadly, with at least 20 in‑custody deaths reported, making 2025 the deadliest year since 2004 as of the reporting date [1]. The combination of lower death rates in 2021–2023 and a sharp 2025 rise complicates any simple narrative of sustained improvement.
5. Causes and context — COVID’s waning role, medical deaths, and reporting motives
Research finds COVID‑19 was contributory in a substantial share of deaths during the COVID era, accounting for 72.7% of medical deaths in FY2020 months studied, but a smaller share (25%) of medical deaths in FY2021–2023, indicating a shifting causal profile over time [5]. Advocacy and civil‑society reports emphasize that underreporting and exclusions bias public counts downward, a point that aligns with the methodological caveats in ICE’s disclosures [4]. Media outlets frame the 2025 spike as an urgent accountability issue, while ICE statements typically emphasize operational challenges; both perspectives rest on the same partial dataset but pursue different institutional agendas [1] [4]. Readers should treat causal claims with attention to these reporting limitations.
6. Bottom line for questioners — the precise answer and what’s still missing
The most defensible, evidence‑based reply is that at least 53 people died in ICE custody from FY2020 through the 2025 reporting cited, based on the combined published figures available [1] [2]. This is a minimum not a complete census: missing year‑by‑year public counts for 2024, ICE’s exclusions of post‑release deaths, and differing year definitions make a single verified cumulative total impossible from the provided sources (p1