healthcare data

Get off Caché and Ensemble and onto IRIS before the maintenance releases stop.

We move Caché and Ensemble to InterSystems IRIS and keep IRIS and HealthShare upgraded after that. Our engineers build HL7 and FHIR interfaces in Health Connect, and the InterSystems developers we staff hold current certifications.

Talk to a InterSystems IRIS & HealthShare engineer What changed recently

platform dossierInterSystems IRIS & HealthShareHealthcare data

Vendor: InterSystems
Next deadline: Chrome and Edge expected to drop XSLT on 17 Nov 2026
Extended Maintenance: IRIS 2026.1, GA 25 Mar 2026
Continuous Delivery: IRIS 2026.2, GA 22 Jul 2026

release watch

Chrome and Edge drop XSLT, and Message Viewer breaks without the fix17 Nov 2026
XSLT fix builds (planned): 2026.1.6, 2025.1.6, HS 2026.2, 2024.1.7, 2018.1.13; 2026.3.0 expected 18 Nov28 Oct 2026

connects with

Epic OH

in short·15-second read

Cohash moves Caché and Ensemble to InterSystems IRIS before 31 March 2027, after which InterSystems plans no more maintenance releases for them.

We also upgrade IRIS, IRIS for Health and HealthShare, help you choose between Extended Maintenance and Continuous Delivery, build HL7 and FHIR interfaces in Health Connect, deploy Unified Care Record and staff certified InterSystems developers.

you probably need this if

Caché or Ensemble still runs production, and the last maintenance release is close. Nobody has started the inventory.
Upgrades keep getting skipped, because nobody ever chose a release track.
Health Connect productions run interfaces nobody has documented.

terms on this page FHIRExtended Maintenancevector

Ask an engineer facts checked 2026-11-03

how it runs: 01Code and production inventory
02Release track and plan
03Rehearsal
04Cut-over and handover
you keep: An inventory of ObjectScript code, productions and interfaces, each with a named owner
Release track decision with upgrade calendar
Caché or Ensemble to IRIS runbook, timed on the rehearsal so the cut-over window is real
+ 2 more documents
who does it: InterSystems ObjectScript developerHealth Connect interface engineer (HL7, FHIR)HealthShare Unified Care Record specialist InterSystems IRIS Development Professional InterSystems ObjectScript Specialist
platforms: Epic OH
proof: 1 case study on this
background: field note · 12 minChrome drops XSLT on 17 November. What breaks in InterSystems, and the fix for your version

intersystems · release timeline

25 Mar 2026 IRIS 2026.1 (Extended Maintenance) It's the Extended Maintenance release for IRIS, IRIS for Health and HealthShare Health Connect. If you're waiting on the XSLT fix, InterSystems plans to put it in update 2026.1.6.
22 Jul 2026 HS2026-01 alerts for HealthShare, IRIS for Health and Health Connect Six alerts in one bulletin, five of them rated high risk for security, privacy or operations. Two to look at first are logins through the FHIR server APIs that never get logged, and chained FHIR searches that can show data a user isn't allowed to see. Affected versions differ per alert. Details are in InterSystems' alert PDF, so check your build against it.
22 Jul 2026 IRIS 2026.2 (Continuous Delivery) Out for IRIS, IRIS for Health and Health Connect as a Continuous Delivery release, build 2026.2.0.217.0. It adds FHIR Explorer for building and testing FHIR queries. It also removes the Apache UIMA integration, deprecated since 2023.1, so code that calls UIMA has to come out before you upgrade.
11 Sep 2026 C-CDA replacement can wipe earlier patient data In Unified Care Record and Information Exchange before 2026.2, a CDA or C-CDA flagged as a replacement (RPLC) can clear patient data ingested earlier. Query-based Edge Gateways are one of the cases InterSystems names. Version 2026.2, or an ad hoc patch for older versions, adds an import setting called blockPatientReplaceActionCode. It's off by default, so someone has to switch it on.
8 Oct 2026 Kits move to downloads.intersystems.com InterSystems put its software kits on a new download site, with an API for scripted downloads. The old WRC distribution site stays up during the transition, and nobody has given it an end date yet. If build scripts still pull kits from the WRC, move them to the new API while both sites work.
28 Oct 2026 XSLT fix builds (planned): 2026.1.6, 2025.1.6, HS 2026.2, 2024.1.7, 2018.1.13; 2026.3.0 expected 18 Nov InterSystems plans to roll this out in stages. 2026.1 and 2025.1 are first in line, planned for 28 October, with HS 2026.2 a day later. The 2024.1 and Caché/Ensemble fixes are planned for 12 November, five days before the browsers are expected to change. The advisory warns some lines might only get fixes soon after 17 November.
today · Nov 09, 2026
17 Nov 2026ahead Chrome and Edge drop XSLT, and Message Viewer breaks without the fix Chrome and Edge are expected to stop running XSLT in the browser from this date. The Management Portal's Message Viewer and Visual Trace need it to show XML messages, and so do Zen Reports. InterSystems has a fix planned for the maintenance lines back to 2024.1 and for Caché and Ensemble. HealthShare's fix is planned for HS 2026.2. If you can't patch in time, the XML Source option still opens the raw message.
31 Mar 2027ahead Caché and Ensemble maintenance ends InterSystems plans no Caché or Ensemble maintenance releases after this date, and it hasn't set an end date for product support. Ad hoc corrections stay best-effort. A 2018.1.13 release with the XSLT fix is planned for 12 November 2026.

the story · intersystems iris & healthshare estate

From the first call to a setup your own team can run.

Here's what usually prompts a call about your InterSystems IRIS & HealthShare estate, what we do about it and in which order, and where you end up.

01 · where it hurts02 · what we change03 · how it runs04 · where you end up

incident board · intersystems iris & healthshare…3 open

openCaché or Ensemble still runs production, and the last maintenance release is close. Nobody has started the inventory.

recurringUpgrades keep getting skipped, because nobody ever chose a release track.

blockedHealth Connect productions run interfaces nobody has documented.

exposure

Caché and Ensemble to IRIS…IRIS and HealthShare upgradesInterface development in Health…Unified Care Record deploymentVector search and AI on IRISCertified staffing

startrehearsehandover

Code and production inventory

Release track and plan

Rehearsal

Cut-over and handover

we are here

ready to hand over

Every interface now runs on IRIS, and each one is written down.
Each patient shows up once in the clinician viewer now.
An inventory of ObjectScript code, productions and interfaces, each with a named owner
Release track decision with upgrade calendar

chapter 01 · where it hurts
Why teams reach out
- Caché or Ensemble still runs production, and the last maintenance release is close. Nobody has started the inventory.
- Upgrades keep getting skipped, because nobody ever chose a release track.
- Health Connect productions run interfaces nobody has documented.
chapter 02 · what we change
How we solve it
- Caché and Ensemble to IRIS migration
- IRIS and HealthShare upgrades
- Interface development in Health Connect
- Unified Care Record deployment
- Vector search and AI on IRIS
- Certified staffing
chapter 03 · how it runs
Our approach
- Code and production inventory
- Release track and plan
- Rehearsal
- Cut-over and handover
chapter 04 · where you end up
What changes
- Every interface now runs on IRIS, and each one is written down.
- Each patient shows up once in the clinician viewer now.
- An inventory of ObjectScript code, productions and interfaces, each with a named owner
- Release track decision with upgrade calendar

what we do

Our InterSystems IRIS & HealthShare work

6 kinds of InterSystems IRIS & HealthShare work, each run by engineers who have done it on live tenants.

An inventory of ObjectScript classes, productions, globals and integrations, a conversion plan and rehearsal on a copy of production before the switch.
Upgrade plans that pick Extended Maintenance or Continuous Delivery releases on purpose, with regression testing of productions and interfaces.
HL7 v2, FHIR and X12 interfaces built as productions, with business rules, data transformations and alerting documented.
HealthShare Unified Care Record set up across sources, with patient identity matching, data feeds and consumer access planned per organization.
Embeddings stored next to the source records in IRIS, with access rules kept on the source tables and retrieval tested against real questions.
ObjectScript developers, interface engineers and HealthShare specialists, onsite in the US or from nearshore and offshore teams across time zones, within your data access rules.

capability 01 / 06in production

Caché and Ensemble to IRIS migration

An inventory of ObjectScript classes, productions, globals and integrations, a conversion plan and rehearsal on a copy of production before the switch.

talk to an engineer about this →

capability 02 / 06in production

IRIS and HealthShare upgrades

Upgrade plans that pick Extended Maintenance or Continuous Delivery releases on purpose, with regression testing of productions and interfaces.

talk to an engineer about this →

capability 03 / 06in production

Interface development in Health Connect

HL7 v2, FHIR and X12 interfaces built as productions, with business rules, data transformations and alerting documented.

talk to an engineer about this →

capability 04 / 06in production

Unified Care Record deployment

HealthShare Unified Care Record set up across sources, with patient identity matching, data feeds and consumer access planned per organization.

talk to an engineer about this →

capability 05 / 06in production

Vector search and AI on IRIS

Embeddings stored next to the source records in IRIS, with access rules kept on the source tables and retrieval tested against real questions.

talk to an engineer about this →

capability 06 / 06in production

Certified staffing

ObjectScript developers, interface engineers and HealthShare specialists, onsite in the US or from nearshore and offshore teams across time zones, within your data access rules.

talk to an engineer about this →

typical situations

InterSystems IRIS & HealthShare situations we walk into

We pieced these together from InterSystems IRIS & HealthShare situations that come up a lot. They're patterns. You won't find a specific client in them.

case file 01

Ensemble still runs a hospital's interfaces, and March 2027 is getting close

composite · not a client story

Beforethe situation

Ensemble runs most of a hospital's HL7 interfaces. Productions were built by staff who have since left, and no maintenance releases are planned after 31 March 2027.
What we dothe work

We inventory every production, business rule and transformation and write down what each interface actually does. Then we rehearse the move to IRIS for Health on a copy of production and regression-test the interfaces against recorded messages.
Afterwhere you end up

Every interface now runs on IRIS, and each one is written down. The team is on a supported release track, so the next upgrade is a planned job instead of a scramble.

case file 02

Unified Care Record stuck on patient matching

composite · not a client story

Beforethe situation

A regional network is rolling out HealthShare Unified Care Record across several organizations, and patients keep showing up twice. Each source formats its identifiers its own way, so the matcher sees two people where there's one.
What we dothe work

We profile identifiers per source, agree normalization rules with each organization and tune matching in a test environment with real duplicate cases. Data feeds are corrected at source where possible.
Afterwhere you end up

Each patient shows up once in the clinician viewer now. The next organization to join doesn't start from scratch, because the matching pattern is written down.

case studies

InterSystems IRIS & HealthShare case studies

All case studies

implementation Industry platforms

Epic integrations, FHIR APIs and TEFCA connectivity

Building and governing the HL7 and FHIR integrations, SMART on FHIR apps and TEFCA exchange around Epic, with an interface catalog and scope reviews.

read the case study

approach

How a InterSystems IRIS & HealthShare engagement runs

We don't start a step until the last one has produced something you can check.

step 01
Code and production inventory

First we pull a list of every ObjectScript class, production, business rule, transformation and integration. Then comes the slow part, finding who owns each one and whether anybody still uses it.
step 02
Release track and plan

We sit down with you and choose between Extended Maintenance and Continuous Delivery. Then we write the migration or upgrade plan around that choice, upgrade dates included.
step 03
Rehearsal

The conversion or upgrade runs on a copy of production first. Interfaces get regression-tested there, and the timings set the cut-over window.
step 04
Cut-over and handover

We're there for cut-over and keep watching the productions for a while after. By the time we hand over, every interface is documented and your team owns the runbooks.

ai on intersystems iris & healthshare

Where AI fits on InterSystems IRIS & HealthShare, and where it needs guardrails.

InterSystems puts AI into the data platform itself. Vector search is part of the core engine. IRIS AI Hub, embedded in IRIS and IRIS for Health, is meant for building and deploying agents and governing them. IntelliCare, on the EHR side, has generative AI built in. For us the practical part is that retrieval data can stay inside IRIS instead of being copied to a separate vector store.

AI agentasks for data or an action

Scoped identityonly the access it needs

MCP serverthe one door into the platform

Logged, approveda person signs off on writes

InterSystems IRIS & HealthSharedoes the work

InterSystems doesn't currently offer a supported MCP server. A community IRIS MCP Server Suite is on Open Exchange. It's pre-release, and InterSystems doesn't support it. In production we'd have agents use IRIS REST or FHIR endpoints, with a dedicated user and minimal roles, audit logging, and approval for writes.

Vector search in IRIS

IRIS works as a vector database, so embeddings sit right next to the clinical or operational data they describe. For us the design work is which records get embedded. Access rules stay on the source tables, and we test retrieval against real questions.

IRIS AI Hub for agents

InterSystems describes AI Hub as a set of capabilities for building, deploying and governing AI agents and agentic workflows, embedded in IRIS and IRIS for Health. For us, what your version and licence include gets confirmed before we design around it.

InterSystems IntelliCare

IntelliCare is InterSystems' EHR, with generative AI at its core. If you're evaluating it, for us the work is integration with your existing HealthShare and IRIS interfaces, and how clinicians review AI output.

AI around HealthShare interfaces

AI can document old production classes and data transformations, explain HL7 and FHIR mappings and draft tests for interface changes. For us, engineers check every change in a test namespace first.

architecture

IRIS and HealthShare sit between the EHRs and everything that reads from them

Health Connect productions pass HL7 and FHIR messages, plus X12, between the EHRs and departmental systems. Unified Care Record builds one shared patient record from them. IRIS holds the data and the vector embeddings that analytics and AI teams pull from.

use cases

Typical InterSystems IRIS & HealthShare projects

The work teams usually bring us, and the platforms InterSystems IRIS & HealthShare has to talk to while we do it.

INT-01 You still run Caché or Ensemble, and no maintenance releases are planned after March 2027 common request
INT-02 Nobody on the team has picked between Extended Maintenance and Continuous Delivery releases yet common request
INT-03 Health Connect interfaces were built years ago and nobody documented them common request
INT-04 You're rolling out HealthShare Unified Care Record across several organizations and the matching isn't settled common request
INT-05 Privacy won't let clinical data go to a separate vector store, so AI retrieval has to stay inside IRIS common request

InterSystems IRIS & HealthShare · connected platforms

Epic Oracle Health EHR Snowflake Databricks Microsoft Entra ID MuleSoft

you keep

What a InterSystems IRIS & HealthShare engagement leaves behind

These are working documents, written against your configuration, so your team can run them without us.

inventory01

An inventory of ObjectScript code, productions and interfaces, each with a named owner
yours to keep
plan02

Release track decision with upgrade calendar
yours to keep
runbook03

Caché or Ensemble to IRIS runbook, timed on the rehearsal so the cut-over window is real
yours to keep
tests04

Interface regression pack for Health Connect productions, so you can rerun it on every upgrade
yours to keep
plan05

Unified Care Record matching plan, so the next source to join doesn't create duplicates
yours to keep

talent and delivery · intersystems iris & healthshare

Certified engineers, wherever you need them.

Some clients want a project team, some want a couple of engineers on theirs. Either works. Onsite, nearshore or offshore, with handovers across time zones that keep things moving.

Onsite

Engineers at your sites in the US

Nearshore

Overlapping hours with US teams

Offshore

Overnight progress on your backlog

All time zones

Coverage handed over across time zones

24-hour coverage · US Central time

OnsiteUS business hours

Nearshoreoverlapping US hours

Offshoreovernight progress

Combinedhandover, not hand-off

certifications our engineers hold

InterSystems IRIS Development ProfessionalInterSystems
InterSystems ObjectScript SpecialistInterSystems
HealthShare Unified Care Record Deployment SpecialistInterSystems
InterSystems EHR Integration SpecialistInterSystems

roles we staff

InterSystems ObjectScript developer
Health Connect interface engineer (HL7, FHIR)
HealthShare Unified Care Record specialist
IRIS system administrator
Caché to IRIS migration lead

due diligence · intersystems iris & healthshare

Questions to ask any partner, including us.

Put these to anyone bidding on your InterSystems IRIS & HealthShare work. The ones who've done it answer differently from the ones who've read about it.

Q1
How many Caché or Ensemble instances do we run, and what runs on each?
No maintenance releases are planned after 31 March 2027. You can't plan the move until someone has counted the instances and listed what each one runs.
Q2
Should we be on Extended Maintenance or Continuous Delivery, and who on our side makes that call?
It decides how often you upgrade and retest. When nobody makes the choice on purpose, upgrades tend to get skipped.
Q3
How will interfaces be regression-tested during the move to IRIS?
Replay recorded production messages against the new instance. A broken transformation shows up there before cut-over, instead of in a downstream system after it.
Q4
Who on the team is certified by InterSystems, for which exam, and is it still current?
InterSystems retires exams now and then. A certificate under an old exam name might not match the product as it is today, so ask what the engineer has worked on since.
Q5
We've got duplicates in Unified Care Record. How do you tune patient matching?
Once clinicians see the same patient twice, they stop trusting the shared record. Tune the matching rules on real duplicate cases pulled from your own sources.

01We're still on Caché. When do Caché and Ensemble maintenance releases end?

InterSystems plans no Caché or Ensemble maintenance releases after 31 March 2027, and it hasn't given an end date for product support. Without maintenance releases you don't get fixes, though. Most organizations plan to be on IRIS before then.

02Why can't we view messages in the IRIS Message Viewer after Chrome's XSLT change?

Chrome and Edge are expected to drop XSLT on 17 November 2026, and the Message Viewer and Visual Trace won't render without it. InterSystems plans fix builds for the Extended Maintenance lines back to 2024.1, plus one for Caché and Ensemble, and HealthShare's is planned for HS 2026.2. On 2026.2, the fix is expected in 2026.3.0 on 18 November, after the change. Until you've patched, open the message with the XML Source option instead.

03We run query-based Edge Gateways. Can a replaced C-CDA wipe a patient's earlier data?

It can. Before 2026.2, Unified Care Record and Information Exchange can clear patient data ingested earlier when a CDA or C-CDA arrives flagged as a replacement. 2026.2, or an ad hoc patch, adds the blockPatientReplaceActionCode import setting. It's off by default, so turn it on.

04Which release track should we follow? We can't decide between Extended Maintenance and Continuous Delivery.

Pick by how often you can test and upgrade. Teams that upgrade once in a while are better off on Extended Maintenance releases like 2026.1. Continuous Delivery, like 2026.2, gets you features sooner, but you'll be upgrading more often and retesting each time.

05How long does a Caché to IRIS migration take?

It depends on code volume, how many productions and interfaces you run, and how much downtime you can take. We size it after an inventory and a rehearsal on a copy of production.

06Which InterSystems certifications do your engineers hold, and do you check they're still current?

InterSystems publishes its exams. They include InterSystems IRIS Development Professional and InterSystems ObjectScript Specialist, plus HealthShare Unified Care Record Deployment Specialist and InterSystems EHR Integration Specialist. We match each engineer's certification to the work.

07Do you also work on Epic or Oracle Health?

Yes. Health Connect often sits between Epic or Oracle Health and the departmental systems, so the work tends to touch both sides of the feed. Epic and Oracle Health each have their own page here.

services

Related services

Projects on InterSystems IRIS & HealthShare usually end up touching one of these, so they're worth a look.

you are hereInterSystems IRIS & HealthShare

fact-checked Checked against vendor and official documentation last checked 2026-11-03 content fingerprint 442ac24a

01Book a 30-minute working sessionsenior engineer
02Talk to an engineer about your setupreply within 24 hours
03See how similar projects wentcase studies

Reviewed by the Cohash team InterSystems IRIS & HealthShare is a trademark of its owner, used here to identify the platform. how we handle facts and data →

next step

Tell us what you're running. We'll tell you what we'd do first.

A 30-minute working session with a senior engineer. No slides.

Book a working session See the case studies

useful to bring

The platform and the version you're on
Where you need to get to, and by when
What's custom: rules, plugins, integrations
Any audit finding or deadline driving it

Get off Caché and Ensemble and onto IRIS before the maintenance releases stop.

From the first call to a setup your own team can run.

Why teams reach out

How we solve it

Our approach

What changes

Our InterSystems IRIS & HealthShare work

Caché and Ensemble to IRIS migration

IRIS and HealthShare upgrades

Interface development in Health Connect

Unified Care Record deployment

Vector search and AI on IRIS

Certified staffing

InterSystems IRIS & HealthShare situations we walk into

InterSystems IRIS & HealthShare case studies

Epic integrations, FHIR APIs and TEFCA connectivity

How a InterSystems IRIS & HealthShare engagement runs

Code and production inventory

Release track and plan

Rehearsal

Cut-over and handover

Where AI fits on InterSystems IRIS & HealthShare, and where it needs guardrails.

Vector search in IRIS

IRIS AI Hub for agents

InterSystems IntelliCare

AI around HealthShare interfaces

IRIS and HealthShare sit between the EHRs and everything that reads from them

Typical InterSystems IRIS & HealthShare projects

What a InterSystems IRIS & HealthShare engagement leaves behind

Certified engineers, wherever you need them.

Onsite

Nearshore

Offshore

All time zones

Questions to ask any partner, including us.

How many Caché or Ensemble instances do we run, and what runs on each?

Should we be on Extended Maintenance or Continuous Delivery, and who on our side makes that call?

How will interfaces be regression-tested during the move to IRIS?

Who on the team is certified by InterSystems, for which exam, and is it still current?

We've got duplicates in Unified Care Record. How do you tune patient matching?

Related services

Tell us what you're running. We'll tell you what we'd do first.