How do counts of deaths in ICE custody compare across U.S. administrations when adjusted for detention volume and reporting methods?
Executive summary
Counts of deaths in ICE custody spiked to the highest absolute level in more than two decades in 2025 — roughly 30–32 deaths reported by media trackers and advocates and acknowledged in ICE notices — coinciding with a rapid expansion of the detained population to roughly 65,000–68,000 people by late 2025 [1] [2] [3]. Comparing administrations requires normalizing by detention volume and reconciling inconsistent reporting practices (fiscal year vs. calendar year), plus accounting for documented undercounting practices such as releases shortly before death; available sources show a worsening picture in both raw counts and signals that per-capita mortality rose, but they do not provide a single, validated cross‑administration death rate [4] [5] [6].
1. What the headline numbers say: record deaths and record detentions
Multiple news organizations and trackers concluded that 2025 produced the most deaths in ICE custody since 2004, citing about 30–32 in-custody fatalities announced over the calendar year and agency notices — a total that made 2025 the deadliest year in two decades (The Guardian, Reuters, Newsweek) [1] [2] [7]. That rise occurred as ICE’s detained population ballooned: reports place the detained census near 65,000–68,440 in late 2025, a roughly 78% jump from mid‑December 2024 levels and record highs for the agency [3] [8] [9].
2. Why raw counts mislead: the need for per-capita rates and standardized denominators
Experts and scholars emphasize death rates per admissions or per person‑days as the correct comparators — not raw yearly totals — because ICE’s detained population and the mix of facilities have fluctuated sharply across administrations and across years [6]. A peer‑reviewed update covering FY2021–2023 calculated death rates per 100,000 admissions and found the death rate decreased over that narrow window, illustrating how year-to-year comparisons can move opposite to raw counts when the denominator or reporting window changes [6]. No single public source in the reporting set provides a validated, administration‑wide death rate standardized across multiple presidencies, which limits definitive arithmetic comparisons [6] [4].
3. Reporting irregularities and undercounting: systemic caveats
Several sources document how ICE reporting choices and practices complicate comparisons: ICE posts death reports on a detainee‑death page and uses fiscal‑year reporting that is sometimes out of date, the agency has been criticized for “releasing” people from custody shortly before their deaths which can avoid inclusion in official tallies, and independent trackers have compiled higher totals than the agency’s public list [4] [5] [7] [9]. Scholarly work specifically flagged released‑then‑deceased cases and concluded that official reports likely underestimate mortality associated with detention [6].
4. Contextual drivers: detention expansion, fewer inspections, and facility mix
The 2025 spike aligns with policy decisions that expanded detention capacity — reopening and adding facilities and dramatically increasing reliance on jails — and with a major drop in oversight inspections, which watchdogs warn elevates risk of preventable deaths [3]. Human rights groups and medical reviewers have argued many detention deaths are preventable with timely care; other observers note the rapid scale-up can overwhelm medical systems inside facilities [10] [9].
5. Competing narratives and limits of the available evidence
DHS and ICE publicly state they follow medical protocols and contest characterizations that the trend represents a systemic failure, and they cite internal standards and reporting procedures [4] [7]. Advocacy groups and independent reporters counter with compiled tallies, FOIA‑based studies, and calls for independent investigations [1] [11] [10]. Importantly, the assembled sources do not furnish a uniform, cross‑administration dataset of deaths per standardized detention exposure (person‑days or admissions) that would permit a definitive numeric ranking after adjustment; that absence prevents an unequivocal, fully adjusted comparison across multiple presidencies [6] [4].
Conclusion: what can be said with confidence
Absolute deaths in ICE custody rose sharply in 2025 alongside an unprecedented detained population, and multiple independent trackers record totals well above recent years, producing both a raw increase and plausible increases in mortality risk per detainee given reduced oversight and system strain [1] [3] [2]. However, because reporting windows, denominators, and acknowledged undercounting practices differ across sources and administrations, the public record does not yet support a precise, audited cross‑administration death rate adjusted for detention volume and reporting methods; the balance of evidence points to a real and troubling deterioration in 2025 but stops short of a fully standardized, number‑for‑number verdict [6] [4].