diff --git a/Dockerfile b/Dockerfile
index 5c84d1c..6aa19c6 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -15,4 +15,4 @@ ENV FLASK_APP=isacc_messaging.app:create_app() \
 
 EXPOSE "${PORT}"
 
-CMD gunicorn --bind "0.0.0.0:${PORT:-8000}" ${FLASK_APP}
+CMD flask upgrade && gunicorn --bind "0.0.0.0:${PORT:-8000}" ${FLASK_APP}
diff --git a/isacc_messaging/api/isacc_record_creator.py b/isacc_messaging/api/isacc_record_creator.py
index e72f5ea..162d11f 100644
--- a/isacc_messaging/api/isacc_record_creator.py
+++ b/isacc_messaging/api/isacc_record_creator.py
@@ -342,22 +342,18 @@ def execute_requests(self) -> Tuple[List[dict], List[dict]]:
         sent = 0
         for cr_json in next_in_bundle(result):
             cr = CommunicationRequest(cr_json)
-            # as that message was likely the next-outgoing for the patient,
-            # update the extension used to track next-outgoing-message time
             patient = resolve_reference(cr.recipient[0].reference)
-            patient.mark_next_outgoing()
 
             # Do not interact with CR if the patient is inactive or sending date is past the cutoff
             if not patient.active or cr.occurrenceDateTime.date < cutoff:
                 cr.status = "revoked"
                 cr.persist()
-                revoked_reason = ""
+                revoked_reason = "Past the cutoff"
                 if not patient.active:
                     revoked_reason = "Recipient is not active"
-                else:
-                    revoked_reason = "Past the cutoff"
                 cr.report_cr_status(status_reason=revoked_reason)
                 errors.append({'id': cr.id, 'error': revoked_reason})
+                patient.mark_next_outgoing()  # update given state change
                 continue
 
             # Otherwise, create a communication
@@ -384,6 +380,7 @@ def execute_requests(self) -> Tuple[List[dict], List[dict]]:
                     extra={"Updated Communication": stopped_comm},
                     level='debug'
                 )
+                patient.mark_next_outgoing()  # update given state change
                 continue
 
             # Otherwise, update according to the feedback from the dispatch
@@ -419,6 +416,7 @@ def execute_requests(self) -> Tuple[List[dict], List[dict]]:
                     extra={"resource": f"Communication/{comm.id}", "statusReason": e},
                     level='exception'
                 )
+            patient.mark_next_outgoing()  # update given state change
 
             # Flooding system on occasions such as a holiday message to all,
             # leads to an overwhelmed system.  Restrict the flood by processing
diff --git a/migrations/versions/recalculate-next-outgoing-times.py b/migrations/versions/recalculate-next-outgoing-times.py
new file mode 100644
index 0000000..3a016c2
--- /dev/null
+++ b/migrations/versions/recalculate-next-outgoing-times.py
@@ -0,0 +1,24 @@
+# Migration script generated for recalculate-next-outgoing-times
+revision = '17f3b60d-0777-4d0e-bb32-82e670b573a7'
+down_revision = 'None'
+
+import logging
+from isacc_messaging.models.fhir import next_in_bundle
+from isacc_messaging.models.isacc_patient import IsaccPatient as Patient
+
+logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
+
+
+def upgrade():
+    # iterate over active patients; confirm/set next-outgoing extension
+    active_patients = Patient.active_patients()
+    for json_patient in next_in_bundle(active_patients):
+        patient = Patient(json_patient)
+        patient.mark_next_outgoing(persist_on_change=True)
+    logging.info('corrected next-outgoing-times')
+
+
+def downgrade():
+    # Nothing to undo
+    logging.info('downgraded')
+