What Do We Mean When We Say Severity? Rating Homicide Risk in Clinical Practice

In my clinical work, the question threat assessment teams ask me most is not whether a concern is real. By the time a case reaches formal evaluation, the referral has already established that. What they are asking — in a dozen different forms — is how serious this is, and what that severity tells us about what to do next.

That is a question about severity. And severity matters for a reason that is easy to overlook: it determines the shape of the response. A lower-severity concern calls for monitoring, relationship-building, and ongoing care. A higher-severity concern calls for protective action now. Getting the severity wrong in either direction has real consequences — under-response leaves risk unaddressed; over-response can sever the trust relationships that produce the next referral.

What a Severity Rating Measures

A homicide severity rating scale, as I use the term, names a clinical construct, not a published instrument. There is no standardized “HSRS” that clinicians administer, and severity rating in this sense is not the VSRA. What I mean by it is a structured method for placing a person's current ideation, planning, and intent on a clinical continuum, from generalized, infrequent thoughts at the low end to specific, persistent, fully planned intent at the high end. It is not a prediction of what will happen. No instrument does that. It is a structured snapshot of current clinical state, designed to be updated at each contact as circumstances change.

The dimensions I weight most heavily are: the frequency and persistence of ideation (how often and for how long the person is engaged with lethal thoughts), the specificity of planning (a general scenario versus a named target with a rehearsed timeline), access to means (does the person currently have, or can they realistically acquire, the capability to act), and behavioral indicators of movement toward action — rehearsal, reconnaissance, or acquisition behavior. Each dimension can be assessed through structured interviewing and collateral information; none is sufficient in isolation.

Why Severity Is Not a Simple Count

One of the most common errors I encounter in applied settings is treating severity as additive — count the risk factors, assume a higher count means higher severity. The research consistently argues against that approach (Borum et al., 1999; Meloy & Hoffmann, 2021). Severity is configural: a single dimension can dominate the clinical picture in ways that a cumulative score misses.

A person with relatively low ideation frequency but a specific, identified target, confirmed means of access, and no protective relationships presents a different, and in my clinical judgment more urgent, profile than a person with high ideation frequency and no plan, no means, and three or more adults who maintain protective relationships with them (parents, teachers, coaches, supervisors, and other consistently involved adults). Both may score similarly on a simple count. A configured severity rating separates them.

The Safe School Initiative findings reinforce this in the school context: the attackers studied had communicated their intent in some form in the vast majority of cases, but those communications ranged from explicit statements to indirect behavioral signals recognized only in retrospect (Vossekuil et al., 2004). A severity framework trained on pre-attack behavioral indicators — not just verbal communication — captures a different and more clinically useful picture.

Protective Factors Matter as Much as Risk Indicators

The clinical picture is most meaningful when it weighs protective factors alongside risk indicators. Connected adults who are actively engaged with the person, a stable sense of meaning and belonging, access to mental health support, and the absence of a fixed grievance against a specific target can substantially moderate a severity rating even when ideation is present (Hart & Logan, 2011). An assessment that tallies only warning signs tends toward overreaction — and toward a surveillance-rather-than-support frame that rarely produces the outcomes we are working toward.

In my case series, some of the most concerning presentations at intake — high ideation, access to means, a history of prior threats — have resolved to manageable, lower-severity status within weeks when the right care relationships were put in place. The rating at intake was accurate. The rating three weeks later was different. Both were right.

What a Severity Rating Cannot Tell You

No severity rating tells you what a person will do. I want to say that plainly, because I have seen instruments used in ways that imply otherwise. A person can move from low to high severity quickly under acute stress, a triggering loss, a relationship ending, or the sudden acquisition of means. Equally, effective intervention can move a person's profile in the other direction. The case series is not small enough to ignore these dynamics — they appear regularly.

This means severity ratings require follow-up. A rating completed in October is not a clinical forecast for December. If circumstances shift materially — a job loss, a relationship rupture, a newly acquired firearm — the clinical picture has changed and the rating needs to be revisited. Point-in-time structured assessments are useful when used as intended: as explicit, documented snapshots supporting an ongoing management plan, not as permanent categorical labels.

Three Things I Would Emphasize for Practitioners

First: anchor your rating in behavioral evidence, not in the person's stated intent. People minimize, and stated denial is not clinically equivalent to the absence of ideation or planning. What is the person doing? What have they acquired? What do collateral contacts observe? The behavioral picture is more reliable than the verbal one.

Second: document what you are rating and why. A defensible severity rating can be explained item by item to a supervisor, a board, or a courtroom. If you cannot reconstruct the reasoning, the rating has limited clinical or legal value. Structured documentation is not bureaucratic overhead — it is the discipline that makes the rating defensible.

Third: share your reasoning with the team. A severity rating used only by the evaluator, without team transparency, loses the primary function it is meant to serve. Threat assessment is a team discipline. The rating is an input to a shared decision, not a conclusion handed down by the clinician.

Conclusion

The question how serious is this? is one of the most consequential in threat assessment practice. A structured severity rating gives practitioners a rigorous way to answer it — not with false certainty about what will happen, but with explicit, revisable clinical reasoning about where a person stands now and what the current picture calls for. That is what good threat assessment looks like, and it is the standard I try to hold in my own work.

References

Borum, R., Fein, R., Vossekuil, B., & Berglund, J. (1999). Threat assessment: Defining an approach for evaluating risk of targeted violence. Behavioral Sciences & the Law, 17(3), 323–337.

Hart, S. D., & Logan, C. (2011). Formulation of violence risk using evidence-based assessments: The structured professional judgment approach. In P. Sturmey & M. McMurran (Eds.), Forensic case formulation. Wiley-Blackwell.

Meloy, J. R., & Hoffmann, J. (Eds.). (2021). International handbook of threat assessment (2nd ed.). Oxford University Press.

Vossekuil, B., Fein, R., Reddy, M., Borum, R., & Modzeleski, W. (2004). The final report and findings of the Safe School Initiative. U.S. Secret Service and U.S. Department of Education.