Restoring Trial Competency: Which Drug Satisfies Sell
Nobody in the room disputes that involuntary medication is legally available here. The disagreement is narrower and harder — which drug, at what dose, by which route, actually satisfies every clause of the test that makes it available.
R.M., a 42-year-old man, spent eleven years as a structural engineer before he stopped going into the office, convinced — first quietly, then in a stream of increasingly specific emails — that his former employer had begun routing surveillance through the building's HVAC system to document him for a lawsuit that didn't exist. The emails, sent over four months to six people at the company, escalated from aggrieved to explicitly threatening; no weapon was ever recovered, no victim was ever approached in person, and R.M. has no prior arrests of any kind. He was indicted on one federal count of transmitting threatening communications in interstate commerce. He has been in pretrial detention nine weeks.
Two competency evaluations, one by each side, agree on the diagnosis and on the finding: delusional disorder, persecutory type, encapsulated — his affect, employment history, family relationships, and reasoning outside the delusional system are otherwise intact, which is exactly what makes the delusion itself so hard to dislodge without medication. He cannot discuss a plea with his attorney because he is convinced the attorney has already been contacted by the same people surveilling him; every attempt to talk strategy is redirected into a fresh iteration of the conspiracy. He has never been aggressive in the facility, has no history of violence beyond the charged conduct itself, and the institution's own risk review found him not dangerous under its ordinary involuntary-medication policy — the administrative route this facility already has for dangerous patients simply does not reach him. That is precisely the gap the government's motion is built to close.
He has refused all psychotropic medication twice when it was offered voluntarily, each time citing the same belief that drugs are "how they finish the job." Nine weeks in, the case has not moved, and it cannot move until either his delusion resolves or the trial proceeds around it — which the law does not permit for a defendant this impaired.
The Sell hearing, drafting the government's proposed order
Sell v. United States sets four factors, and the first and third aren't seriously contested here — an important governmental interest is at stake in prosecuting a federal threats charge, and no less intrusive alternative exists once voluntary treatment has failed twice. The real work is factor two: substantially likely to restore him, and substantially unlikely to produce side effects that interfere with his defense. Herbel and Stelmach's 2007 review of 22 involuntarily medicated federal defendants with delusional disorder found 77 percent restored to competency — a diagnosis long assumed to respond poorly to medication, restored at a rate the court can rely on. A standard oral antipsychotic, dosed and titrated the ordinary way, is what I'd put in the order.
I'm not disputing the 77 percent — I'm pointing out it answers only half of factor two. The Herbel and Stelmach cohort measures whether delusional thinking resolved, not whether the specific agent used left the defendant able to sit at counsel table without visible sedation or restlessness. Oral haloperidol — the obvious high-potency first-generation choice here — carries real extrapyramidal liability at the doses this kind of restoration typically requires — a defendant who is factually competent but presents as stiff, akathisic, or flattened in front of a jury has had exactly the harm factor two exists to prevent, just delivered by the choice of drug rather than by the decision to medicate at all.
"Substantially unlikely to interfere with his ability to assist counsel" is doing real work in the text of factor two, not decoration around the restoration number — a second-generation agent with a materially lower EPS burden at comparable antipsychotic effect is the more defensible order, even if it means a slightly longer titration before we can represent to the court that he's ready.
Both of you are arguing about the right pill. I'd rather we argue about whether he'll actually take it. He believes the medication is the mechanism of the conspiracy against him — that's not garden-variety reluctance, it's the delusion itself pointed directly at the treatment. Sell requires the government to show restoration is likely within a reasonable period of time, and an oral regimen that depends on daily voluntary swallowing from a man who thinks swallowing it completes the plot against him is a fragile basis for that representation, whichever agent we pick.
I'm not arguing for the injectable purely on adherence grounds either — an LAI takes longer to reach steady state, and if the court wants restoration on a predictable timeline, that has to be weighed honestly against oral's faster path if he does take it. But "he'll take it because the order says he must" is not the same claim as "he'll take it," and the order should say which one we're actually counting on.
Agreed: the proposed order names risperidone, oral, titrated over four weeks with competency reassessment at eight, and states haloperidol decanoate as the named fallback if two consecutive weeks show non-adherence rather than waiting for a full titration failure to declare it.
Not agreed, and left as a genuine open disagreement in the filing rather than smoothed into consensus: whether the court should be told up front that the fallback exists, or only informed if it's actually triggered. The pharmacist argued that naming a fallback in advance is honest disclosure of a real contingency; the pharmacologist worried that a court handed two options at once will simply pick whichever sounds less restrictive without the clinical picture that would justify the choice either way when the moment actually comes. The psychiatrist didn't take a side — noting only that whichever way it's filed, the 77 percent figure describes the diagnosis' response to being medicated, not this specific order's own probability of working exactly as written.