General operational and educational information for corrections professionals. Not legal, medical, or compliance advice, and not a certification of compliance with any law or standard. Policies and standards vary by agency and jurisdiction; follow your facility's policy and your own legal, medical, and professional advisors.
Courts and juries understand that people die in custody, from pre-existing conditions, from suicide, sometimes from something nobody could reasonably predict. The exposure is almost never the death itself. It is what the agency knew, or reasonably should have known, and failed to act on. That is the legal question in nearly every case that reaches a courtroom, and it is built long before the death, out of routine documents nobody thinks are important at the time.
The fear sitting in the back of every jail administrator's mind is an in-custody death and the lawsuit that follows. The honest answer to where the exposure sits is: not in the death. It is in whether staff had information about a risk and either ignored it or failed to escalate it. That is the entire question a wrongful-death case turns on.
Liability is built out of intake screening forms, medical request slips, cell-check logs, and shift pass-down notes. In nearly every case worth studying there is a specific moment in that paper trail where someone had information, a reported chest pain, a family member's call, a cellmate's mention of erratic behavior, and it either did not get escalated or got escalated verbally and never documented as escalated.
You can do everything right operationally in the moment and still lose the case if the file does not show it. You can make an honest, defensible judgment call under real pressure and still lose if there is no record the call was ever made. Exposure is decided in the thirty, sixty, and ninety days before an incident, not during it.
The first is intake. Medical and mental-health screening at booking is often rushed, especially in high-volume jails, and frequently done by staff without the clinical background to catch subtle indicators. A screening that takes four minutes and asks yes-or-no questions will not catch a person in opioid withdrawal who is minimizing symptoms out of fear, or a person in a mental-health crisis who has learned how to mask it for a few minutes in front of a stranger with a clipboard.
The second is the gap between complaint and response. A person says they are not feeling well and it gets treated as a routine annoyance, because officers hear many complaints a week and most genuinely are not emergencies. The problem is you do not get to know in advance which one is the exception. Agencies that reduce exposure build in a hard rule: every medical complaint gets logged and gets a documented response time, no exceptions and no in-the-moment judgment about whether it sounds serious enough to act on.
Family and third-party reports are the most consistently overlooked signal. A family member calls worried about a loved one's medication or mental state, and that call either becomes a logged piece of information routed to medical and custody staff or an emotional phone call that gets quietly forgotten. Families often know a person's baseline and history in ways intake never captures, which makes those calls among the earliest clear warnings available.
Close the intake gap. Put medical professionals on the screening wherever the budget allows, and where it does not reach every booking, make the questions behavioral and observational rather than pure self-report, so staff are not entirely dependent on a frightened or intoxicated person reporting their own condition accurately.
Build a real audit trail for cell checks and medical requests: timestamped, specific, and reviewed by a supervisor on a regular schedule, rather than sitting untouched in a binder until litigation asks for it.
Log every medical complaint with a documented response time, including the ones that sound minor. The record has to show the response, not just the complaint.
Build a formal process for logging and routing third-party calls: not just that a call happened, but what was said and what was done with it, with the follow-up itself documented.
None of this is exotic or expensive relative to what a single wrongful-death judgment costs. It is discipline, applied consistently, and written down every time.
Monitoring technology closes the gap between believing checks happened and being able to show they happened. In litigation, a log that says a check occurred and a record that can confirm it occurred are treated very differently by a judge and a jury.
This is not about distrust of staff. It is about building a system that does not depend entirely on any one person's memory or judgment holding up perfectly, under pressure, every time, because no one's does. A verification layer protects a good officer's account as much as it catches a bad one.
Every event raised for attention is reviewed by a person before action, and no clinical judgment is made by software. No facial recognition is used.
This question usually comes from agency leadership after a review board or a grand jury looks at a death and the finding comes back preventable. That word is the one that keeps administrators up at night, because it means the death was not inevitable. Somewhere in the chain, something could have gone differently. After 15 years in custody and operations, including time in a facility with an attached medical unit, I have come to see in-custody medical deaths less as single dramatic failures and more as chain failures. Several small gaps, each survivable on its own, that happened to line up on the same person at the same time.
The first link is intake. Chronic conditions, diabetes, cardiac history, severe asthma, seizure disorders, have to be identified and flagged at booking, not discovered three days later when someone collapses on a tier. That requires a screening process thorough enough to catch both self-reported and observable conditions, and a system that actually routes that information to the people managing housing and medication, not a folder sitting in a records office nobody opens that shift.
The second link is medication continuity, and it is one of the single highest-risk points in the entire system. Someone arrives on insulin, blood pressure medication, seizure medication, or a mental health prescription, and there is a gap, sometimes hours and sometimes days, between arrival and that medication being verified and resumed. That gap has killed people repeatedly, across facilities nationwide. It is one of the best documented and most preventable failure points in correctional medical care, and it is also one of the most fixable, because at bottom it is a process and workflow problem rather than a resource problem waiting on new funding.
The third link is the handoff, between shifts and between custody and medical staff. A concern raised at 6 AM has to survive the shift change at 7, medical rounds at 9, and a housing move at 11 without getting lost or softened along the way. Every one of those handoffs is a place where information can quietly die even if the person does not, yet.
Fixing intake means treating medical screening as seriously as security classification, because a missed chronic condition is just as dangerous as a missed security risk, it simply kills more quietly and over days instead of minutes. Fixing medication continuity means a hard, tracked rule: verification and resumption of critical medications inside a defined short window after intake, audited regularly, not left to whenever medical staff get to it during a busy week. Fixing handoffs means building real redundancy, not a single verbal pass-down between two tired people at shift change, but a documented flag that follows the person through every shift and every housing move until it is formally closed out by medical staff rather than quietly handed off again to the next shift.
I want to name the cost plainly, because it clarifies the stakes better than anything else. A preventable in-custody medical death typically leads to a wrongful death suit, an investigation, and often a consent decree that reshapes how the entire agency operates for years afterward under outside monitoring, at far greater cost and far greater loss of local control than fixing those three links would ever have required. Agencies that treat medication continuity and intake screening as line items to trim during a budget crunch are trading a small visible saving today for a compounding cost later, measured in dollars, in oversight, and in a life that did not have to end that way.
If I were auditing an agency's prevention program in an afternoon, I would ask one question and follow it wherever it led. Pick any three people currently on critical medication and show me the documented chain of custody for that medication from intake to right now, today. Where it breaks down, a gap with no explanation, a day with no record either way, is where the next preventable death is most likely to originate, because that gap has already proven it can happen under the current process. Agencies that run that exercise on themselves honestly, without preparing the answer in advance, usually find the gap before it costs anyone their life rather than after.
One caution, and I will be direct because I build monitoring technology for a living and I do not want this read as a pitch. Technology does not fix a broken intake process or a culture that does not take medication continuity seriously. No camera or software system does that work for you. What it does is give you visibility into the hours between scheduled rounds that you did not have before, and give supervisors an independent way to verify that the human processes described above are actually happening rather than trusting that they are. A person still reviews what is flagged and a person still decides what to do about it. Used that way, alongside real process fixes, it is a genuine addition. Used as a substitute for fixing intake screening or medication continuity, it is a false sense of security that will be exposed the same way every other paper-only measure is.
Rarely. It turns on deliberate indifference: whether staff had information about a risk and ignored it or failed to escalate it. The death alone is seldom the legal question.
In the thirty, sixty, and ninety days before an incident, in whether intake caught what it should and whether concerns were documented when raised, not during the incident.
It is often rushed and done by staff without clinical background. A short yes-or-no screening misses withdrawal or a masked crisis, so behavioral and observational questions matter more than self-report.
Log it and record a response time, every time. You do not get to know in advance which complaint is the exception, so the rule cannot depend on an in-the-moment judgment.
They are frequently the earliest clear warning. Log what was said, route it to medical and custody staff, and document the follow-up, not just that a call came in.
It closes the distance between believing a check happened and being able to show it did. A confirmable record and a bare log are treated very differently in litigation.
Almost always several. A missed condition at intake, a delayed medication, a concern that did not survive a shift change, each survivable on its own, lining up on one person at the same time. Breaking the chain at any single link prevents most of them.
The gap between arrival and verified resumption of critical medication. It is one of the best documented preventable failure points, and it is a workflow problem rather than a funding problem, which is what makes it fixable.
