Co-Responder and Alternate Dispatch Models

What It Is

Co-responder and alternate dispatch programs route certain 911 calls – typically those involving mental health crisis, substance use, or behavioral health – to non-law-enforcement responders, either instead of or alongside police. The two primary models are:

Co-responder: A mental health clinician or crisis worker is paired with a law enforcement officer and responds jointly. Both arrive at the same time. The clinician leads or supports the mental health assessment; the officer provides safety coverage and handles any law enforcement component.

Alternate dispatch (diversion): Calls meeting specific criteria are routed entirely to a non-law-enforcement team – typically a mobile crisis unit staffed by clinicians, peer support specialists, or paramedics. Law enforcement is not dispatched unless requested or conditions change.

Some programs use a hybrid: initial response by a civilian crisis team, with law enforcement available for immediate back-up on request.

Program Models

CAHOOTS (Eugene, Oregon)

The Crisis Assistance Helping Out On The Streets program has operated since 1989. CAHOOTS dispatches two-person teams – a medic and a crisis worker – to calls involving behavioral health, welfare checks, and non-violent crises. In 2019, CAHOOTS handled approximately 24% of all Eugene Police Department calls for service, with law enforcement backup requested in roughly 1% of CAHOOTS responses.

CAHOOTS is widely cited as a national model but is frequently mischaracterized. It operates in a mid-size city with a centralized dispatch system that was specifically redesigned to support call diversion. Its call criteria exclude calls with weapons present, active violence, or requests for law enforcement. It is not a template that transfers without adaptation.

STAR (Denver, Colorado)

Denver’s Support Team Assisted Response program launched in 2020, pairing a mental health clinician with a paramedic for calls involving behavioral health crises. In its first six months, STAR handled over 700 calls with no police backup requested and no arrests. The program has since expanded in call volume and hours of operation.

Co-Responder Programs

Co-responder programs exist in hundreds of agencies. The most studied model pairs a mental health clinician with a patrol officer – the clinician rides along or meets officers on scene. Evidence on co-responder outcomes is mixed but generally positive for subject satisfaction, reduced arrests, and increased connection to services. Officer satisfaction with co-responder partners tends to be high.

A consistent limitation is capacity: most co-responder programs operate limited hours (often daytime, weekdays) and cover only a fraction of calls that might benefit from clinician involvement.

What the Evidence Shows

Diversion volume. Programs that have been implemented with sufficient capacity and call criteria consistently show that a meaningful fraction of 911 calls – often 10–25% in urban environments – can be handled without law enforcement. Most of these are welfare checks, behavioral health contacts, and low-acuity calls.

Safety. There is no documented instance of a CAHOOTS or STAR responder being killed or seriously injured during a response. Critics argue that serious-incident calls are excluded from these programs by design – which is true, and is exactly the point of structured call criteria.

Recidivism and service connection. Evidence on whether alternate dispatch programs reduce repeat calls or improve long-term outcomes is limited and methodologically uneven. Programs that include active referral and follow-up outperform those that treat each call as a standalone event.

Cost. Alternate dispatch programs cost less per call than law enforcement response. Their fiscal impact on the overall law enforcement budget depends on whether diverted call volume leads to officer reallocation or reduction – a policy question, not an operational one.

Limitations and Failure Modes

Call Criteria Gaps

Every alternate dispatch program defines which calls are eligible. Calls involving reported weapons are almost universally excluded from civilian-only response. This creates a structural limitation for calls involving persons with edged weapons: if a caller mentions a knife – even a knife that poses no active threat – the call becomes ineligible for diversion under most programs, regardless of whether law enforcement is operationally necessary.

Agencies that want to extend diversion to a broader category of edged-weapon calls need call criteria that distinguish weapon present from weapon threatening, and that allow dispatcher judgment to override a default exclusion when warranted.

Intake Classification Dependency

Alternate dispatch programs only receive calls that are correctly classified as eligible at intake. Agencies that report low utilization of their mobile crisis unit frequently find the bottleneck is at the 911 call-taker stage, not in the team’s availability. A call classified as a “disturbance” will not route to a crisis team, even if the actual situation is entirely a mental health matter.

Capacity Constraints

Most alternate dispatch programs operate during limited hours and cover limited geography. A program that handles 8am–10pm calls on weekdays covers only a fraction of when behavioral health crises actually occur. After-hours calls revert to law enforcement dispatch by default.

Escalation Protocols

Civilian crisis teams need clear, tested protocols for when to request law enforcement backup and how quickly that backup will arrive. Programs that have not pre-positioned law enforcement resources – or that dispatch law enforcement from a distant location – may create gaps when escalation is needed.

Relationship to Edged-Weapon Incidents

The majority of fatal edged-weapon encounters in the United States begin as welfare checks, mental health calls, or disturbance calls – not reported weapon assaults. A dispatch system that routes behavioral health calls to trained crisis responders before law enforcement involvement reduces the number of situations in which an armed officer is the first contact with a person in psychiatric crisis who may be armed.

This does not eliminate edged-weapon risk – crisis responders do encounter subjects with weapons. It does change the context of that encounter: a clinician arriving as a primary responder, with law enforcement staged nearby, approaches a situation differently than an officer arriving to a call coded as an armed disturbance.

Operational Considerations

For dispatchers and call supervisors: - Maintain current eligibility criteria in a searchable, accessible format. - Flag calls that are borderline-eligible for supervisor review rather than defaulting to law enforcement dispatch. - Track calls where law enforcement backup was requested by a crisis team – these are data on where call criteria may need adjustment.

For law enforcement officers working alongside co-responder programs: - Understand the program’s call criteria and staging protocols before you need them. - When a crisis team is already on scene, follow their lead on approach and communication. - Backup requests from crisis teams are not escalations to a criminal response – they are requests for a safety presence.

Policy Implications

Agencies evaluating or expanding alternate dispatch programs should:

1. Map call volume against eligibility criteria. Identify the proportion of current law enforcement calls that would qualify for diversion under proposed criteria.
2. Build capacity to match demand. Undercapitalized programs serve as proof-of-concept but do not change outcomes at scale.
3. Integrate eligibility into CAD at the classification stage. Call-takers should see diversion eligibility flags in real time.
4. Define and publish escalation protocols. Clear, public protocols for when law enforcement backup is requested reduce ambiguity for crisis responders and officers alike.
5. Collect and publish outcome data. Calls handled, backup requested, arrests, injuries, and service referrals should be reported annually at minimum.

How to cite

Edged-Weapon Response Lab. (2026). Co-Responder and Alternate Dispatch Models. In EWRL Bulletins (Version 1.0). https://droptheknife.org/bulletins/co_responder_alternate_dispatch