Preventable ED revisits often look like “non-adherence” or “poor engagement,” but the root cause is frequently simpler: the system did not see the ED event in time to respond. When event notifications arrive late (or not at all), primary care and HCBS teams end up working from outdated information—medication changes are missed, follow-up appointments are not scheduled, and new risks are not governed. In avoidable utilization governance, ADT alerts and real-time event feeds are not a technology project; they are a reliability mechanism that turns “we should follow up” into a defined, owned workflow. Done well, they also strengthen primary care and care coordination by making transitions visible and actionable across settings.
What “good” looks like: alerts that trigger action, not anxiety
An ADT alert is only valuable if it produces a predictable next step. The operating question is not “Did we receive an alert?” but “Did the alert reliably produce the right follow-up for the right person, within the right timeframe, with proof?” That requires three design choices.
First, define scope and thresholds. Not every notification warrants the same response. A best-practice approach stratifies by risk (e.g., complex conditions, polypharmacy, cognitive impairment, recent discharge, prior ED use) and by event type (ED registration, discharge, observation stay, inpatient admission). Each tier has a time-bound response standard.
Second, assign ownership. Alerts must land in a monitored queue with named roles (not a shared inbox with implied accountability). Ownership includes who performs outreach, who reconciles meds, who schedules follow-up, and who closes the loop with primary care.
Third, require auditable closure. “Attempted contact” is not closure. Closure means the event was reviewed, risk was assessed, tasks were assigned, and outcomes (including failures) were recorded in a way that stands up to payer and regulator scrutiny.
Oversight expectations you should design for
Expectation 1: payers and state oversight will increasingly expect transition reliability, not narrative assurance. Whether driven by Medicaid managed care contracting, value-based arrangements, or state quality strategy priorities, the common direction of travel is measurable follow-up and reduced preventable utilization. Organizations should be prepared to show response timeliness, closure rates, and exception handling—especially for high-risk members.
Expectation 2: governance must protect safety, rights, and privacy while enabling coordination. Event feeds can expose sensitive information; governance must define minimum necessary access, role-based visibility, and documentation standards. The control is not “don’t share,” but “share safely and consistently, with clear purpose and accountability.”
Operational example 1: ED registration alert triggers same-day outreach and risk triage
What happens in day-to-day delivery. An ED registration alert arrives in a monitored work queue that is checked at set intervals (e.g., every hour during business hours, and via an on-call protocol after hours for the highest-risk tier). The care coordinator reviews the member’s current risk tier, recent contacts, and care plan, then initiates outreach using a scripted-but-clinically-informed call flow. If the person cannot be reached, the workflow routes to alternate contacts, HCBS staff on shift, or a home visit request depending on program design. The coordinator documents the reason for the ED visit (as known), confirms immediate safety, and opens follow-up tasks (medication check, appointment scheduling, transportation, durable medical equipment, caregiver briefing).
Why the practice exists (failure mode it addresses). Without early visibility, systems learn about ED use days later—after the discharge plan is already failing. The registration alert is designed to prevent “silent ED episodes” where nobody in the community-based team knows the person is in crisis, and therefore nobody prepares the next steps that reduce repeat visits.
What goes wrong if it is absent. The first post-ED contact becomes reactive and fragmented: HCBS staff arrive to normal visits without knowing the person was in the ED; primary care remains unaware of medication changes; and warning signs (e.g., dehydration, infection, falls, behavioral escalation) are not addressed until they trigger another urgent event. In practice, this shows up as repeated 911 calls, missed follow-up appointments, and escalating caregiver distress.
What observable outcome it produces. Teams can evidence shorter time-to-first-contact after an ED event, higher rates of completed follow-up tasks within defined timeframes, and fewer “unknown reason” ED episodes. Audits show a clear trail: alert received, outreach attempted/completed, risk decision recorded, and next-step tasks closed or escalated.
Operational example 2: ED discharge alert triggers medication reconciliation and care plan update
What happens in day-to-day delivery. When an ED discharge alert is received, the workflow assigns a medication reconciliation task to a designated role (e.g., nurse care manager or trained med rec specialist) within a defined window (commonly 24–48 hours, sooner for high-risk). The specialist requests the discharge summary if not automatically available, compares pre-ED and post-ED medication lists, and contacts the member/caregiver to validate what is actually being taken. Discrepancies are routed to the prescribing clinician or pharmacist support. The care plan is updated with new red flags, monitoring requirements, and follow-up appointment details, and HCBS direct care staff receive a brief “what changed” summary to support safe day-to-day delivery.
Why the practice exists (failure mode it addresses). Medication-driven utilization often comes from reconciliation failures: duplicate therapies, stopped meds that continue, new meds that are never started, or high-risk medications introduced without monitoring. The discharge alert is the trigger that prevents ED medication changes from staying trapped in the hospital record.
What goes wrong if it is absent. People return home with confusing instructions and multiple bottles, caregivers guess, and frontline HCBS staff are left to interpret changes without authority. The predictable results are adverse drug events, falls, uncontrolled symptoms, avoidable urgent care visits, and frustrated primary care teams receiving partial information after harm has occurred.
What observable outcome it produces. Organizations can track reconciliation completion rates, discrepancy rates resolved within target timeframes, and downstream reduction in medication-related ED revisits. Reviewers can see consistent documentation: list comparison, member confirmation, clinician escalation where needed, and updated monitoring tasks.
Operational example 3: Exception management for alerts that do not convert into follow-up
What happens in day-to-day delivery. A weekly (or twice-weekly) exception report identifies alerts without documented closure—no outreach, no follow-up appointment scheduled, no reconciliation completed, or unresolved “unable to reach” cases beyond a defined threshold. The report is reviewed in a short operational huddle with decision authority: barriers are categorized (wrong contact details, language access needs, transportation failures, capacity constraints, member refusal, provider availability), and an escalation path is applied. Some exceptions trigger rapid management actions (updating contact data, assigning alternate staff, initiating welfare checks), while others trigger system fixes (workflow redesign, staffing adjustments, training, or payer/provider coordination).
Why the practice exists (failure mode it addresses). Closed-loop systems fail when they assume completion. Exception management exists to prevent the most dangerous failure mode: a task that “should” happen but never does, leaving high-risk members in the community without coordinated follow-up.
What goes wrong if it is absent. The organization becomes dependent on individual vigilance. A missed alert might not be noticed until a serious adverse event occurs, and leaders cannot distinguish between “rare misses” and structural breakdown. Over time, staff lose confidence in the system and revert to ad hoc workarounds that are not scalable or auditable.
What observable outcome it produces. Leaders can evidence closure rates, reduced backlog of unaddressed alerts, and trend improvement in specific barrier categories (e.g., fewer wrong-contact failures after data quality interventions). Audit trails show that exceptions were identified, assigned, escalated, and resolved with documented decision-making.
Implementation notes that prevent common failure patterns
Design for reality, not ideal data. Expect incomplete event details and build workflows that can start with minimal information, then enrich as records arrive.
Make the queue visible. A daily “alerts received vs. alerts closed” view is a governance tool. If leaders cannot see backlog, they cannot manage risk.
Protect staff with decision rights. If frontline staff are asked to “handle” ED alerts without authority to schedule, escalate, or request records, the system will stall. Decision rights must be explicit.
Prove it works under pressure. Test the process during capacity strain (weekends, staffing gaps, high-volume periods). Reliability in easy weeks is not the standard.
Bottom line
ADT alerts and real-time event notifications reduce preventable ED revisits only when they are treated as governed pathways: defined thresholds, named ownership, time-bound tasks, and exception controls that make failure visible. The payoff is not just fewer ED returns—it is defensible coordination that primary care, HCBS operations, and payers can all see and rely on.