Is there a way to configure bot logging to a file or stdout? Even when

auditEventDestination

includes "log" and vmcontext is the runtime version, I can't seem to understand where the bot is actually logging to when running the server locally (in this scenario, with Docker`). It seems they are not to be found in the logs of the Medplum server.

Are there any specific considerations or migration scripts for migrating from v4?

Hi everyone Just read the new docs regarding [scheduling](https://github.com/medplum/medplum/blob/main/packages/docs/docs/scheduling/index.md) and most specifically on [defining availability](https://github.com/medplum/medplum/blob/main/packages/docs/docs/scheduling/defining-availability.md) which is something our team has been working on. This modeling aligns pretty well with our initial thoughts about it, but they are some use cases that could be considered. For instance, some more complex scenarios would require and extended version of

scheduling-parameters

that can include more complex scheduling constraints (e.g. by body site, etc). What is the recommended approach to expand this extension to include more scheduling parameters? Also, what is the status for the

$find

,

$hold

,

$book

and

$cancel

Appointment booking Operations? We've been developing our own quite analogous operations, and it would be interesting to know what is the scope of these and release plans, if already defined.

I can obviously go to http://localhost:3000 to get to the medplum app, but curious if it's possible to see the provider view locally?

Can I use Firebase ID Token as an external access token for token exchange? In the dashboard setup, I need to provide authorizeUrl and tokenUrl, but I'm encountering 400 errors. Digging deeper, I'm seeing Firebase isn't a standard OAuth 2.0 IDP "Because you are using Firebase signInWithEmailPassword and getting an ID token, your flow is not a typical OAuth2 'authorize → token' endpoint sequence. In particular" whats required to resolve this ? Or alternative since my app uses the Firebase signInWithEmailPassword to verify users

Hey! I saw this forum uses Linen to sync Discord threads to the web. I’m building [Paladin](https://trypaladin.com/), a smoother, cleaner alternative that makes Discord forums easier to browse on the web. It’s free to use and can run alongside Linen, so you don’t have to switch anything to test it. Would you be interested in trying it out? 😀 (example preview)

A customer is experiencing issues where users are getting invalid/expired token errors, particularly when they have multiple browser tabs open. The problem seems to occur when tokens get refreshed in one tab, causing issues in other tabs. The customer can't reproduce it consistently but has received multiple reports from clinicians about this issue.

Hi, Shawn from Superpower here. We currently are on a v3.1 community plan and would love to upgrade RBAC (AccessPolicy, ProjectMembership, etc.) related function to v4 to leverage all the great improvements. Is there best practices and/or any caveat we should pay attention to if we don't want to upgrade other parts to V4? @rahul16019 Do we need to go through the full pre v3.3 -> v3.3 -> migration -> v4 path?

A developer hit the 10,000 offset limit while running a script and wanted to know if this limit could be removed or increased. They were concerned about performance impact but needed to process larger datasets. Initially they tried cursor-based pagination but ran into the limitation that it only works with

_lastUpdated

sorting in ascending order, which didn't fit their use case since they needed a variety of sorting options.

A developer is evaluating the best FHIR resource approach for team-based task assignments in their healthcare system. In their current (non-Medplum) system, they have several "team" email addresses that function as logical entities managing 10-100 tasks per team. Their current workflow is: 1. Task gets created and assigned to a team 2. Team member picks up the task and reassigns it to themselves 3. Member completes the task and marks it as done They're torn between three approaches: - **Option A**: CareTeam resources (coordinated units with specific responsibilities) - **Option B**: Group resources (collections of individuals) - **Option C**: Individual Practitioners (their current approach - simpler but doesn't represent team structure accurately)

A developer is implementing a feature where one Practitioner can complete a Task that's owned by a different Practitioner. They want to give users two options: 1. Complete the task while keeping the original owner attribution unchanged 2. Complete the task and update the attribution to the Practitioner who "took over" They're looking for the preferred design pattern and which fields to update on the Task resource to properly track this workflow. The use case is to enable smoother workflows where non-owners can complete Tasks under urgent conditions, with potential financial compensation implications later.

A developer was trying to send photos through the Medplum API and had two main questions: Do photos need to be Base64 encoded when using the API? They were hitting size limits when bundling photos with encounters and documents in transactions - should photos be sent individually instead? Their current setup stores images on S3, and they were initially trying to bundle everything together in a single transaction.

Hey everyone! Quick question about task management in Medplum - is it possible to create tasks and assign them to a particular AccessPolicy? The use case here is creating tasks that aren't assigned to a specific person, but rather to a role like "Care Coordinator" so that any care coordinator could pick up and work on those tasks. Basically looking for role-based task assignment rather than individual assignment.

How to handle when a doctor needs to correct a mistake in the original questionnaire submission? How to efficiently find and update the related Observation when a QuestionnaireResponse gets modified Whether to store the Observation ID back in the QuestionnaireResponse for easier lookups Best practices for handling corrections in clinical data Whether to process the form data client-side vs using Bots

A developer wants to set up webhook notifications when a specific resource's value changes in Medplum. They already have the webhook endpoint ready on their app side and are looking for the best practice approach to trigger these notifications from Medplum when a doctor changes an Appointment status.

Is it possible for a user to be both a Patient and a Practitioner for example in a given project? Or is this not supported.

Hey Medplum community! 👋 A developer recently needed help restoring a large batch of resources that were accidentally soft-deleted. They had the exact resource IDs and needed to restore: - 1204 Task resources - 111 Encounter resources - 111 Appointment resources

Hey Medplum community! 👋 A developer was working with compartments in Medplum and ran into some issues after the removal of user-defined compartments in v4.0.4. They previously used Organization compartments on Task resources as filters in GraphQL queries. Since Tasks don't have inclusion criteria for the patient compartment, they were looking for the recommended approach to achieve similar functionality. They also mentioned trying _tag as a filter previously, but experienced performance issues.

Hey Medplum community! I'm working on some changes to our insurance records and wanted to get some input on a few scenarios: 1. Flagging Coverage found via insurance lookup - When a patient hasn't uploaded pharmacy insurance or their uploaded insurance isn't accepted at our pharmacies, we use our pharmacy system to automatically look for alternative accepted insurance. What's the best way to flag coverages found through this process? 2. Using CoverageEligibilityRequest/Response - Is it appropriate to create these resources for the insurance lookup process described above? The docs make it sound like these are for tracking requests to insurers specifically. 3. "Archiving" Unaccepted Insurance - If we find a patient's insurance isn't accepted at our pharmacies, we want to "archive" it so our systems don't use it. Would setting Coverage.status to

cancelled

be appropriate, or should we use another approach?

Hey Medplum community! I have an open-ended question about referrals, particularly around tracking "prerequisites" and state for long-lived referrals. Background: When we're referring patients we need to share several data points with them by syncing to their EHR: Basic demographics Vitals Intake forms Lab result PDFs Once we've shared everything, patients can: Pay for the service Book an appointment The Challenge: We need to track what we've already shared and what's still pending (e.g., "Already shared X labs, but still need Y lab" or "All labs shared but intake pending"). We also want to record their invoice number and appointment details for future reference. What I've considered so far: Creating a ServiceRequest for the referral Adding lab results as .supportingInfo references to DiagnosticReport But I'm unclear on: How to annotate which prerequisite each item is meeting How to record external system IDs for uploaded data How to reference intake forms not yet in Medplum How to handle invoice numbers Any suggestions?

We're working on automatically revoking Medplum access when a user gets disabled in our SSO provider. We've tried deleting their ProjectMembership which seems to work - the Project disappears from their login screen and they can't log in anymore. However, we noticed this warning in the docs (https://www.medplum.com/docs/auth/user-management-guide#removing-memberships): "Do not edit or change the ProjectMembership resources directly..." Are we going down the right path by just deleting the ProjectMembership resource, or is there a better way to remove access programmatically?

Hey devs! 👋 A customer was asking about implementing different user roles in their EMR system. They have various roles like Admin, Physician, Registered Nurse, Care Coordinator, and Customer Support. Their main use cases were: Feature permissions (e.g., only physicians can do certain tasks) Task assignment logic (e.g., routing specific messages to nurses) Group filters/CareTeams (e.g., showing tasks assigned to specific roles)

A developer is asking if it's possible to enforce uniqueness on a resource on the Medplum side. They want to ensure only one record is ever created on the Medplum side while still allowing updates to happen.

Just wanted to share a solution for handling supply items in FHIR when integrating with an e-prescribing system. The Problem I needed to add non-medication supply items (like medical supplies) to our system and sync them with an e-prescribing platform. The supplies data looked something like:

SupplyId: 2
SupplyName: Poise Thin-shape Pad
IsRx: FALSE
IsObsolete: FALSE
NDC_UPC_HRI: 36000043127
NDC_UPC_HRI_TypeId: 2