Contracting and Funding for Crisis Diversion: SLAs, Payment Models, and Governance That Prevents Cost-Shifting

Why crisis diversion contracts often underperform

Crisis diversion is a multi-organization workflow, which means it fails at the seams: unclear eligibility, denial without accountability, delays that create ED boarding, and post-diversion gaps that drive repeat crisis. Many systems try to manage this with “goodwill” and informal relationships. That works until volume rises, staffing becomes fragile, or a high-profile safety event forces scrutiny. At that point, the contract becomes the operating manual—if it is specific enough.

In practice, the most common contracting error is paying for activity (a visit, a day, a bed) while assuming outcomes (reduced ED use, fewer arrests, fewer repeat crises) will follow. A high-functioning contract makes incentives explicit, defines required workflows, and creates a governance loop that fixes problems without waiting for annual renegotiation.

Oversight expectations you should design around

Expectation 1: Clear accountability for medical necessity and utilization management. Medicaid agencies and managed care plans typically expect defined criteria for level of care, documentation that supports it, and a process for resolving disputes. If diversion settings cannot demonstrate appropriate use and safe discharge/step-down, payers may tighten authorizations or claw back payment—disrupting the entire crisis continuum.

Expectation 2: Data governance and privacy compliance with measurable performance. Funders increasingly require timely encounter reporting, consistent identifiers, and audit-ready records. Contracts should specify who captures what data, how quickly it is submitted, and how corrections occur. When data duties are vague, performance disputes become political rather than factual, and systems revert to ED default because it is the only “visible” pathway.

Core contract components for diversion governance

At minimum, diversion contracts should define: referral sources and acceptance expectations; response times; minimum staffing and credentialing; required clinical workflows (risk assessment, safety planning, medication processes); handoff and follow-up requirements; escalation routes; incident reporting; and performance reporting cadence. If you cannot describe the day-to-day workflow in the contract, you cannot reliably buy it.

Operational Example 1: A diversion SLA that makes denial accountable (not just “no beds”)

What happens in day-to-day delivery

The contract requires that every denial or deflection is coded in a standardized set of reasons (medical exclusion, acuity mismatch, staffing limitation, capacity full, documentation incomplete, or safety concern), with a short free-text note and the name/role of the decision-maker. Denials are transmitted back to the referring entity (mobile crisis/ED) within a defined timeframe and automatically appear in a weekly denial dashboard. The governance meeting reviews denial patterns by provider, shift, and reason code, and triggers corrective actions (workflow change, staffing adjustments, or criteria clarification).

Why the practice exists (failure mode it addresses)

This practice prevents “soft denials,” where providers effectively refuse cases through delay, vague responses, or inconsistent criteria application. Soft denials push people into ED/jail without visibility, making diversion look ineffective and undermining trust between partners.

What goes wrong if it is absent

When denials are not tracked and categorized, partners argue from anecdotes. ED staff stop attempting diversion because it feels unpredictable. Providers, meanwhile, may tighten criteria informally to manage risk, which concentrates higher-acuity cases in EDs and increases involuntary interventions—exactly the outcomes diversion was meant to reduce.

What observable outcome it produces

Within a quarter, systems typically see fewer “unknown” denials, reduced denial-to-ED escalation, and faster resolution of criteria disputes. Evidence includes denial trend reports, improved acceptance timeliness, and measurable reductions in ED boarding for diversion-eligible referrals.

Operational Example 2: Payment terms that reward throughput and safe stabilization, not length of stay

What happens in day-to-day delivery

The contract uses a blended payment model: a base per-diem or case rate for crisis receiving/stabilization plus a small performance component tied to defined outcomes (timely acceptance, completion of safety plans, follow-up within target window, and low rates of avoidable re-presentation within a defined timeframe, risk-adjusted where possible). Providers submit monthly performance files alongside encounters, and the payer/county validates the figures against ED claims or crisis platform data. Disputes are handled through a defined reconciliation process with timelines.

Why the practice exists (failure mode it addresses)

This design addresses the failure mode where providers are financially indifferent to delays or to weak discharge planning. Pure per-diem payment can unintentionally reward longer stays or “parking” people, while pure case rates can incentivize premature discharge. The blend creates an incentive to stabilize quickly and safely, with documented handoffs.

What goes wrong if it is absent

Without aligned incentives, systems see either prolonged stays that choke capacity (leading to diversion denials and ED boarding) or rapid discharges without real stabilization (leading to repeat crisis, safety events, and increased law enforcement involvement). In both cases, funders perceive diversion as “expensive” without evidence of value, and contracts become vulnerable to cuts.

What observable outcome it produces

Observable outcomes include improved bed turnover without increased adverse events, fewer boarding hours, and reduced repeat crisis contacts for discharged cohorts. Evidence includes occupancy and throughput dashboards, incident trends, and claims-based analysis showing reduced ED utilization for diversion-eligible groups.

Operational Example 3: Contracted post-diversion follow-up as a closed-loop requirement

What happens in day-to-day delivery

The contract requires a documented warm handoff to an ongoing care team (community mental health, primary care integration, ACT, case management, or peer support), plus a first follow-up contact within a defined window (for example 24–72 hours depending on acuity). The provider must confirm medication access (pickup or delivery), verify appointment scheduling, and document barrier resolution attempts (transportation, housing, benefits, domestic safety). A shared referral tracker records referral creation, acceptance, appointment completion, and non-attendance follow-up actions.

Why the practice exists (failure mode it addresses)

This requirement prevents the “stabilize-and-release” failure mode where crisis services do good work in the moment but leave the person to navigate a fragmented system afterward. Without closed-loop follow-up, repeat crises become predictable and diversion loses credibility with ED partners and funders.

What goes wrong if it is absent

If follow-up is not contracted and measured, it becomes optional and inconsistent. People miss appointments, can’t obtain medications, or return to unsafe environments without support. Systems then experience rising repeat 988/911 calls and ED presentations, which prompts tighter utilization management and less willingness to fund diversion expansion.

What observable outcome it produces

Systems should see improved follow-up completion rates, fewer repeat crisis contacts within defined windows, and clearer accountability when referrals fail. Evidence includes closed-loop referral dashboards, medication access verification rates, and cohort analyses demonstrating reduced re-presentation and improved stability indicators.

Governance cadence written into the contract

Contracts should specify how governance happens: who attends, how often, and what decisions can be made (criteria updates, corrective action plans, staffing remediation, and data fixes). Include escalation pathways for urgent concerns (safety events, refusal patterns, sudden capacity drops) and define what constitutes a reportable incident versus a performance variance. This prevents “surprise” findings and helps partners solve problems collaboratively.

Finally, build in change control. Crisis systems evolve quickly—new capacity, new 988 routing, new state expectations. A contract that cannot adapt will be bypassed operationally. A contract that adapts through defined governance will keep diversion credible, safe, and fundable over time.