Two-way patient messaging escalation workflows differ less by whether messages can be sent than by how responsibility moves when a message needs attention. The retrieved documentation describes four distinct mechanisms: HealthShare Personal Community converts patient messages into Workbench tasks assigned to workflow groups; Avon uses triage teams, pages, and participant status changes; Medplum models ownership through FHIR Task resources with reassignment and escalation rules; and Salem Health specifies a 15-minute response obligation with a defined escalation chain for Epic Secure Chat. Each approach makes different assumptions about who owns a thread, what counts as escalation, and what evidence proves a response occurred. The practical question for software teams is therefore not which product supports escalation, but which ownership and routing model matches the organization's existing clinical and operational structure.

Task-Based Escalation in HealthShare Personal Community

HealthShare Personal Community routes patient-initiated messages through a mechanism called a message flow. When a patient sends a new message, the system determines whether it goes to the Workbench or to an external messaging system through a custom delivery interface [1]. For Workbench delivery, Personal Community creates a task for a given workflow group to manage. Only Workbench users associated with that workflow group can claim the task and respond to the patient [1]. This means escalation is expressed as task assignment within a bounded group, not as an open broadcast to all available clinicians.

The setup guide clarifies that each clinical group supporting patient-initiated messaging must have a message delivery flow defined [5]. Sites configure which providers and clinical groups are available to receive messages, the message delivery workflow for each clinical group, and the types of messages patients may send [5]. The documentation does not describe an automatic timeout-based escalation from one workflow group to another. Instead, the operational burden falls on workflow group membership and task claiming. A message that remains unclaimed stays within the group's task queue unless an administrator or agent reassigns it through the Workflows, Agents, and Tasks function, which the messaging guide references but does not detail [1].

Triage Teams and Page-Based Escalation in Avon

Avon's messaging module introduces a different escalation primitive: the page. Pages notify organization members that they should check a message thread as soon as possible [4]. The recommended flow begins with a triage team, defined as the set of people who serve as the patient's first point of contact. When a patient messages the care team thread, the triage team is notified. The triage team then pages in appropriate organization members to resolve the patient's questions, and those members are added to the message thread [4]. Escalation here is an explicit act of paging a specific person into an existing conversation.

Avon also distinguishes participants from observers. Participants have read and write access to the thread and receive notifications when new messages arrive. Observers see the thread in their chat list but are not notified unless an exception applies [4]. A page can be resolved by clicking a resolve button, and it can be unresolved later [4]. This creates a visible state change for escalation: a page is either open or resolved, and the resolution is attributed to a named member with a timestamp. The documentation does not specify automatic escalation rules, such as paging a second person if the first does not respond within a set interval. The escalation trigger appears to be a human decision by the triage team.

A Suggested Validation Scenario for Escalation Logic

A useful hypothetical validation scenario would test whether escalation state survives a change in ownership. Create a patient message that enters a queue with no owner. Claim it as a first responder, then reassign it to a second responder using the product's native mechanism. Record whether the second responder inherits the full thread history, whether the first responder retains visibility, and whether the audit trail shows the original owner, the reassignment reason, and the timestamp. Then simulate a no-response condition by leaving the message unclaimed for a defined interval and observe whether any automatic reminder or escalation occurs. This scenario does not test clinical outcomes or compliance; it tests whether the software's ownership and routing model matches the organization's documented escalation policy.

FHIR Task Ownership and Reassignment in Medplum

Medplum's decision guide frames escalation as a question of response ownership. It asks whether messages are assigned to specific providers, to a pool or queue by role, or both, and whether threads can be reassigned or escalated after initial assignment [8]. The guide maps these questions to FHIR Task mechanics. For individual assignment, Task.owner is set to a specific provider. For pool or queue routing, Task.performerType is set by role type while Task.owner is cleared, and providers claim from the pool [8]. Reassigning the same response obligation means updating the existing Task's owner with an If-Match version check, while adding a new conversation participant through the membership workflow and preserving existing participants [8].

The guide also addresses visibility after reassignment. Whether the previous owner retains access is decided through AccessPolicy and the membership workflow, with a review of existing child messages and files as well as the header [8]. This is a more granular ownership model than a simple queue. Escalation can be represented as a change in Task.owner, a change in performerType, or a change in thread membership, and each choice has different audit implications. The guide notes that if a structured audit trail of who owned a thread and why it was rerouted is needed, the Task data should support that requirement [8]. Medplum does not prescribe a single escalation policy; it supplies the resource model for implementing one.

A Clinical Escalation Standard with Time Thresholds

Salem Health's interim Interdisciplinary Communication Standard provides a concrete clinical escalation workflow for Epic Secure Chat. The standard classifies communication into categories: urgent matters requiring a response within 15 minutes use a page or direct phone call, while routine matters within one to two hours may use Secure Chat or a phone call [2]. Active staff members are expected to respond to phone, beeper, or answering service contacts within 15 minutes of initial contact. If there is no response after 15 minutes, a second attempt is made, and the switchboard directly pages or calls the provider. After another 15 minutes without response, the House Supervisor is contacted [2].

The standard also assigns responsibility to the sender. It states that a sent page or delivered Secure Chat message is not the same as a message received, and that the sender must confirm urgent or critical communication has been received and acted upon [2]. For nursing escalation, the first step is to page or call the provider and verify the page was directed to the correct provider with patient context included [2]. The document further requires that any order discussed in Secure Chat be entered and signed in Epic, because a Secure Chat message is not an order [2]. This standard is explicitly interim and pending a system-wide policy, so it represents one organization's current baseline rather than a universal requirement [2].

Comparing Escalation Boundaries and Unresolved Questions

The four models differ in where escalation authority lives. HealthShare Personal Community places it in workflow group membership and task claiming, with no documented automatic timeout escalation in the messaging guide [1]. Avon places it in a triage team's decision to page specific members, with page resolution as a visible state [4]. Medplum places it in Task ownership and membership changes that can be automated through bots, including cron-scheduled bots that flag threads with no response after a set number of days [8]. Salem Health places it in a policy-driven chain with fixed 15-minute intervals and named escalation targets [2].

Several questions remain unresolved by the supplied documentation. The HealthShare guides do not specify whether an unclaimed Workbench task can be automatically escalated to a second workflow group after a timeout. Avon's documentation does not describe automatic page escalation or response deadlines. Medplum's guide describes the mechanics for reassignment but leaves the choice of escalation policy to the implementer. Salem Health's standard is interim and specific to one health system. A software team evaluating these approaches should therefore ask whether escalation must be automatic or human-initiated, whether the audit trail must record ownership changes, and whether the organization already has a clinical response-time policy that the software must encode.