§ 4243 Commitment After NGRI: Verbal Threats, Episodic Bipolar Disorder, and Release-Plan Instability Can Support “Substantial Risk”

1. Introduction

This decision addresses the federal civil-commitment framework governing defendants found not guilty by reason of insanity (NGRI). Joshua Richard Potenza was acquitted by reason of insanity on a charge of transmitting a threatening communication under 18 U.S.C. § 875(c), after leaving voicemails threatening to kill a former roommate/coworker (C.W.) and threatening C.W.’s family.

Following the NGRI verdict, the district court conducted the statutorily required dangerousness inquiry under 18 U.S.C. § 4243. Potenza argued he should be released because: (i) his past conduct was “only verbal threats,” (ii) he had not been violent during extended custody, (iii) risk-assessment instruments were an imperfect fit, and (iv) his symptoms were stable/remitted at the time of evaluation.

The key legal issues were:

• Dangerousness standard: whether Potenza proved that release “would not create a substantial risk of bodily injury … or serious damage to … property” due to a present mental disease or defect (18 U.S.C. § 4243(d)).
• Meaning of “present mental disease or defect”: whether an episodic condition (bipolar disorder with psychotic features) can satisfy the “present” requirement even when stable in a structured setting.
• Evidence and methodology: how much weight a court may place on expert testimony and structured risk tools (PCL-R; HCR-20-V3) in forecasting risk under § 4243.
• Appellate review: whether the district court’s factual findings were clearly erroneous.

2. Summary of the Opinion

The Eleventh Circuit affirmed the commitment order. Applying clear-error review, it held the district court’s findings were “plausible” given the full record, including:

• the history of threats and prior protective-order issues,
• evidence linking threatening behavior to mental illness,
• Potenza’s lack of insight and resistance to treatment, and
• expert testimony that immediate, unsupervised release would increase destabilizing stressors (housing, employment, access to care) that elevate risk.

The court rejected arguments that (i) threats alone can never support a substantial-risk finding, (ii) the expert’s limited experience rendered her opinion weightless, (iii) the HCR-20-V3’s broader conception of “violence” made its use improper, and (iv) “remission” or “between episodes” precluded a finding of a present mental disease. It emphasized that Congress requires courts to make difficult forward-looking dangerousness assessments under § 4243 and that episodic disorders can still be “present.”

3. Analysis

3.1 Precedents Cited

Shannon v. United States, 512 U.S. 573 (1994)

The panel cited Shannon v. United States for the proposition that § 4243 is part of a broader “statutory scheme” governing insanity acquittees. While Shannon is often associated with instructions to juries regarding the consequences of an NGRI verdict, the Eleventh Circuit invoked it here to situate Potenza’s commitment hearing within Congress’s comprehensive post-verdict framework: NGRI triggers mandatory custody pending a hearing, with release dependent on the acquittee meeting a statutory burden.

United States v. Wattleton, 296 F.3d 1184 (11th Cir. 2002)

United States v. Wattleton supplied two key doctrinal anchors:

• No rigid factor checklist: § 4243 “does not provide a checklist of factors” for assessing dangerousness. Courts may consider any factor that “relate[s] to the risk of danger caused by a present mental disease or defect.” This principle supported the district court’s broad evidentiary lens—incorporating threats, treatment noncompliance risk, insight deficits, and prospective instability (housing/employment/support).
• Release-support environment matters: Wattleton recognized courts may consider “lack of support” upon release and the dangers stemming from that lack (e.g., destabilization and treatment dropout). The Potenza panel used this to validate the district court’s focus on the practical conditions of immediate release, not merely current symptom severity.

United States v. McIntosh, 900 F.3d 1301 (11th Cir. 2018)

United States v. McIntosh influenced the decision in three ways:

• Standard of review framing: clear error is “deferential”; appellate courts do not reweigh evidence. This helped the panel dispose of Potenza’s challenges that largely contested the weight and interpretation of expert evidence.
• “Mental disease or defect” is a legal term: McIntosh emphasized the phrase is a legal classification applied by the court to facts, rather than a purely clinical label dictated by medical professionals. This undercut any argument that symptom remission or diagnostic nuances necessarily foreclose a “present mental disease or defect” finding.
• Permissible reliance on expert/risk reports: McIntosh recognized district courts may rely on risk assessment panel reports and psychologist testimony in § 4243 decisions. Potenza’s attack on the expert’s experience and the instruments’ fit thus became, in the panel’s view, an insufficient basis for clear-error reversal.

3.2 Legal Reasoning

(a) The statutory burden allocation did the work

The opinion repeatedly returns to the decisive structural feature of § 4243: after an NGRI verdict, release is not presumed. Instead, the acquittee must prove by a preponderance of the evidence that release would not create the requisite “substantial risk” due to a present mental disease or defect. The panel treated the district court’s conclusion as a conventional burden-of-proof determination: Potenza did not carry it.

(b) “Substantial risk” can be inferred from threats plus illness-linked escalation dynamics

Potenza argued his history involved “only verbal threats,” so the district court lacked a basis to predict substantial risk of bodily injury or property damage. The panel rejected a categorical separation between “threat-only histories” and “substantial risk” findings. It endorsed the district court’s inference that repeated violent threats, combined with evidence of mental illness decompensation and treatment resistance, can reasonably forecast escalation to harmful conduct if released.

Notably, the panel did not require proof of prior completed violence. It accepted that:

• the threatening conduct drew law enforcement involvement,
• there were prior protective-order problems, and
• expert testimony identified a “clear link” between mental illness and threatening/violent behavior.

(c) Prospective instability and lack of supervision were legitimate risk inputs

Relying on United States v. Wattleton, the panel approved the district court’s attention to “lack of stability” at release: limited capacity to arrange psychiatric care, difficulties with housing/employment, and the risk these stressors pose for medication noncompliance and relapse. This reasoning treats dangerousness under § 4243 as contextual and forward-looking: risk depends not only on diagnosis but on the real-world scaffolding available to maintain stability.

(d) The court treated expert “weight” disputes as non-reversible factfinding

Potenza attacked the expert’s credibility and qualifications (e.g., only four years as a psychologist; PCL-R administered five times). The panel held that such criticisms went to the weight of evidence—squarely within the district court’s province—rather than demonstrating clear error. Absent an evidentiary rule violation or an implausible factual leap, appellate courts typically do not second-guess credibility determinations under clear-error review, a principle reinforced here via United States v. McIntosh.

(e) Imperfect risk tools were still admissible/usable evidence

Potenza argued the HCR-20-V3’s broader “violence” definition (including threats) was a poor fit because § 4243 focuses on “substantial risk” of bodily injury or serious property damage. The panel’s response is pragmatic: while the tool is “not a perfect fit,” the district court did not clearly err by considering it. The court also noted the statutory scheme forces courts to attempt inherently difficult predictions; structured tools can be part of that evidence mix even if not precisely congruent.

(f) Episodic disorders can still be “present mental disease or defect”

Potenza contended there was “no evidence” of a present mental disease because his symptoms were “in remission” and he was “between mood episodes.” The panel held the record supported a finding of bipolar disorder and narcissistic personality disorder notwithstanding stability in a “highly structured environment.” Critically, it emphasized Potenza cited no authority requiring courts to treat episodic illnesses as non-“present” during stable phases. The decision thus implicitly recognizes that “present” can mean an existing underlying disorder with potential for recurrence, particularly where relapse risk is tied to noncompliance, stressors, and lack of insight.

3.3 Impact

(1) Threat-based histories may still sustain § 4243 commitments

The most practical takeaway is that § 4243 does not demand a history of completed violence. Repeated credible threats—especially coupled with illness-linked decompensation, protective-order violations, and treatment refusal—can suffice to uphold a “substantial risk” finding. This is especially consequential for defendants whose index offense is communicative (e.g., threats, stalking-type behavior) rather than overt battery.

(2) “Structured custody stability” is not dispositive

The opinion signals that good behavior and symptom stability in custody, while relevant, may be discounted if the court finds the stability is contingent on a structured setting that will vanish upon release. Future litigants should expect courts to ask: what concrete supports, supervision, and treatment access exist outside custody?

(3) Risk tools: not dispositive, but usable even when imperfect

The panel’s acceptance of HCR-20-V3 and PCL-R evidence—despite methodological critiques—encourages continued reliance on structured professional judgment instruments in § 4243 litigation. Challenges will likely be framed as “weight” disputes unless a defendant can show the tool’s use produced a legally irrelevant inquiry or was methodologically unsound to the point of unreliability.

(4) Episodic illness and the meaning of “present”

By refusing to equate “between episodes” with “no present mental disease,” the decision supports broader judicial discretion to treat chronic/episodic psychiatric disorders as “present” for § 4243 purposes, particularly when relapse risk is tied to poor insight and treatment refusal.

(5) Litigation strategy implications

• For acquittees: the strongest path to release often lies in proposing a concrete, verifiable conditional-release plan: identified treating providers, medication adherence mechanisms, stable housing, supportive supervision, and crisis plans—rather than relying on “no incidents in custody.”
• For the government/facility: documenting the link between illness symptoms and threatening conduct, plus barriers to outpatient continuity of care, can be persuasive even when the acquittee is currently stable.

4. Complex Concepts Simplified

NGRI (Not Guilty by Reason of Insanity)

An NGRI verdict means the defendant is acquitted of criminal responsibility because, at the time of the offense, a qualifying mental condition prevented appreciation of the nature/wrongfulness of the conduct. It does not mean automatic freedom; it triggers a separate civil-commitment process.

18 U.S.C. § 4243 commitment

After NGRI, the defendant is committed pending a hearing. At that hearing, the question is not guilt, but safety: whether release would create a substantial risk of bodily injury or serious property damage due to a present mental disease or defect.

“Preponderance of the evidence”

The defendant must show it is more likely than not that release would not create the prohibited substantial risk.

“Substantial risk”

The statute does not require certainty or proof beyond reasonable doubt. It requires a meaningful, non-trivial risk of serious harm, causally connected to a current mental condition.

“Present mental disease or defect”

“Present” does not necessarily mean acute symptoms on the hearing date. It can include an existing disorder that is episodic or managed in custody, especially if relapse risk is predictable under real-world conditions (stress, lack of treatment access, noncompliance, poor insight).

Clear-error review

On appeal, the question is not whether appellate judges would decide differently, but whether the district court’s findings were plainly mistaken. If the finding is plausible on the record, it will usually be affirmed.

Conditional release

Even when unconditional release is too risky, § 4243 allows conditional release under a prescribed regimen of care (treatment, monitoring, supervision). Noncompliance can lead to arrest and remand.

5. Conclusion

United States v. Joshua Richard Potenza reinforces the practical operation of § 4243: after an NGRI verdict, the acquittee bears the burden to prove safety for release, and appellate review is highly deferential. The decision underscores that (i) a history of repeated threats can support a substantial-risk finding when linked to mental illness and treatment resistance, (ii) stability in custody does not negate dangerousness if it depends on a structured environment unlikely to exist upon release, (iii) courts may consider imperfect but accepted risk instruments as part of a broader evidentiary picture, and (iv) episodic psychiatric conditions can still qualify as a “present mental disease or defect.”

Although unpublished, the opinion is a clear signal within the Eleventh Circuit that § 4243 dangerousness determinations may legitimately incorporate real-world release conditions (support, supervision, treatment continuity) and do not require prior completed violence to justify commitment.