☰

Baxter International Inc.

Stabilise operations, then connect the new assets

Industry
Medical Devices, Pharmaceuticals and Connected Care
Headquarters
Round Lake, Illinois, United States
Public information as of
January 2026

A4BEE prepared this analysis from publicly available sources. It reflects our own reading of Baxter International Inc.'s published strategy and is not endorsed by, or produced in cooperation with, Baxter International Inc.. Company website

Strategic priorities

Baxter International is rebuilding after two structural shocks: the 2024 Hurricane Helene flooding of its North Cove, North Carolina plant, which supplied roughly 60 percent of US IV fluids and peritoneal dialysis solutions, and the 2025 divestiture of its Kidney Care segment to form Vantive. In 2025 the CEO described the route forward as tightly tied to the Baxter GPS Growth and Performance System, a continuous-improvement programme that starts from standardised metrics and simpler processes across the medical products, connected care and pharmaceuticals businesses.

Capital is being concentrated on resilience and rationalisation. The company became the first manufacturer to earn the Healthcare Industry Resilience Collaborative's Gold Resiliency Badge for IV solutions, nutrition solutions and premix drugs, a signal that redundancy planning is now a purchasing conversation with hospital customers. The CIO has made the consolidation of thousands of legacy applications left over from the Hillrom acquisition and the Vantive split an explicit mandate, and the firm has guided to a 2030 carbon-neutral target for direct operations.

The third on-going programme is the integration of the 2021 $10.5 billion Hillrom acquisition. The HST connected-care segment combines Hillrom's smart beds, patient monitoring, and care communications with Baxter's own infusion and fluid management platforms, and the work to make Centrella, Voalte and Welch Allyn share data with hospital electronic medical record (EMR) systems remains incomplete. The remaining priorities for digital work — operations, consolidation, and integrated connected care — all sit on the same underlying capability: manufacturing and clinical data that can be read outside the equipment and software that produced it.

Challenges we see

  • Operations Manufacturing

    Restoring IV fluid supply after the North Cove loss

    Hurricane Helene destroyed access bridges and breached a levee protecting the North Cove site in Marion, North Carolina in late 2024. The facility had supplied roughly 60 percent of US IV fluids and peritoneal dialysis solutions, and the disruption was followed by a national IV-fluid shortage and a multi-year reconstruction plan that includes distributed manufacturing nodes in Spain, Mexico, Thailand and Singapore.

    Where a single site carries a majority of the supply of a clinically essential product, the option set after disruption is reactive allocation rather than planned substitution. Reconstructing the line with redundancy designed in shifts the next event from a supply crisis to a planned re-route.

  • Compliance Regulatory

    Restoring data integrity at the Ahmedabad pharmaceutical facility

    Following an FDA warning letter and Form 483 observations, the Ahmedabad pharmaceutical facility has data-integrity items to address, relating to electronic data capture on visual inspection machines and to the reconciliation of operator logs and badge swipe records. India-sourced product volumes and US import status are tied to the resolution of those observations.

    Where instruments already produce the data they need to record, capturing it at the instrument and binding it to an audit trail makes the difference between an inspection that finds evidence and one that finds a gap.

  • Digital Integration

    Connecting legacy OT assets to enterprise systems

    Manufacturing floors still contain legacy PLCs, chillers and mixing tanks that are air-gapped from enterprise systems, while the CIO has committed to rationalising thousands of legacy applications inherited from the Hillrom acquisition and the Vantive carve-out. The first step is reading current-state data; the second is collapsing redundant applications on top of a connected data path.

    Decisions about predictive maintenance and overall equipment effectiveness (OEE) currently depend on manual reconciliation across systems. Once legacy equipment is read through a documented protocol, the same data can serve maintenance, quality and engineering rather than each building its own extract.

  • Digital Operations

    Stabilising software behaviour across connected infusion devices

    The Novum IQ infusion pump platform has been subject to recalls relating to software behaviour linked to race conditions during user-interaction sequences. Each recall event triggers an FDA software submission cycle and a commercial discussion with hospital customers.

    Releasing software that connects to delivery hardware at the patient bedside raises the bar for catching interaction-level defects before market. Validation that exercises the device as a system rather than feature-by-feature narrows the gap between the test corpus and the real-world interaction space.

  • Digital Integration

    Closing the Hillrom–Baxter connected-care integration

    Three years after the $10.5 billion Hillrom acquisition, the HST segment still runs proprietary protocols for Centrella, Voalte and Welch Allyn that do not publish to hospital EMR systems in a standard form. The 2025 mandate from the CIO to rationalise thousands of applications is the commercial framework for finishing this work.

    Where device protocols and hospital EMR systems don't share a transport, every integration is a project. Translating proprietary outputs to HL7/FHIR (Health Level Seven / Fast Healthcare Interoperability Resources, the healthcare data exchange standards) once turns the next hospital deployment into a configuration rather than a development project.

Opportunities, by urgency and business impact

Each bubble is one opportunity, numbered to match the list below. Further right means it bites sooner; higher means a bigger effect on the business. A bigger bubble means a bigger implementation effort.

Source: A4BEE analysis of public sources

  1. Digitising the North Cove reconstruction in parallel with the rebuild

    The North Cove, North Carolina facility is being rebuilt physically while the US IV-fluid supply continues to run on the remaining sites and on international manufacturing nodes. Standing-up a new line with conventional commissioning creates a long ramp-up period during which the supply gap is filled with emergency allocation.

    A facility digital twin built in parallel with the physical rebuild allows the new automation logic, control sequences and operator procedures to be exercised against a virtual plant before the physical machines are installed, and the same model becomes the disaster-recovery exercise surface for the next disruption.

    • Baxter North Cove Fact Sheet, May 2025
    • Hurricane Helene Updates, Baxter newsroom

  2. Capturing instrument data directly into the lab system of record

    The Ahmedabad facility's visual inspection machines already produce electronic readings, but those readings are not captured into the system of record. Inspection results are instead transcribed manually, which the FDA warning letter relates to the data-integrity observations.

    Connecting the visual inspection machines and laboratory instruments through a documented protocol, and binding the result to the operator who performed the inspection and the batch under test, generates the audit trail the regulatory framework expects and removes the gap that the inspection identified.

    • FDA Warning Letter, Baxter Ahmedabad facility
    • Baxter 2024 Corporate Responsibility Report

  3. Reading legacy equipment through a documented connection

    Legacy PLCs, chillers and mixing tanks across the manufacturing network collect data locally and do not send it to enterprise systems. The first software and AI work for condition monitoring depends on the data arriving in a documented form in the first place.

    An edge connectivity layer that wraps existing equipment with a documented protocol such as OPC UA (Open Platform Communications Unified Architecture) or MQTT (a lightweight publish-subscribe messaging protocol commonly used in industrial settings) extracts the health data without modifying the underlying machine, and creates the data path that predictive maintenance and OEE reporting can later be built on.

    • Baxter Reports Third-Quarter 2025 Results
    • Baxter 2024 Form 10-K

  4. Exercising infusion-pump software against a virtual device

    Novum IQ recalls relating to software behaviour indicate that interaction-level defects can reach the field. Each released defect is expensive in regulatory rework and in customer trust, and the time between findings is dominated by the test cycle.

    A virtual representation of the pump's mechanical and user-interface behaviour, paired with automated regression coverage, lets thousands of user-interaction sequences be exercised overnight so that defects surface in the development cycle rather than in the field.

    • Baxter Important Product Updates
    • FDA recall documentation for Novum IQ infusion pump

  5. Translating connected-care device data into the EMR

    Centrella, Voalte and Welch Allyn devices communicate using proprietary protocols, and integrating them with hospital EMR systems has historically been a project per deployment. The 2025 mandate to rationalise thousands of legacy applications is the natural moment to standardise the data path.

    An HL7/FHIR-based edge gateway translates device output into the healthcare standard once, so that subsequent hospital deployments become configuration rather than reintegration, and so that the same data can feed clinical reporting, device management and the IT application rationalisation programme.

    • Baxter Reports Second-Quarter 2025 Results
    • WakeMed Health & Hospitals case study, Hillrom

What we'd propose

  • Digital CDMO

    Facility digital twin for the North Cove reconstruction

    A digital twin of the North Cove facility built in parallel with the physical rebuild, used to exercise automation logic, control sequences and operator procedures against a virtual plant before machines are installed, and kept as the disaster-recovery exercise surface for the next site disruption.

    • Virtual commissioning of the new line

      Exercising automation before the physical build

      Run the new line's PLC (programmable logic controller) logic, sensor responses and control sequences against a virtual plant, so commissioning debug cycles happen on the model rather than on the physical line where each iteration is a day of production.

    • Disaster-recovery scenario library

      The next disruption rehearsed in advance

      Model flood, utility and logistics scenarios on the same twin, including production re-routing paths to the Spain, Mexico, Thailand and Singapore nodes, so a future event is handled as a planned re-route rather than a national shortage.

    • Continuous OEE feed for Baxter GPS

      Performance data flowing into the operating system

      Stream availability, performance and quality metrics from the rebuilt line into the Baxter GPS visual management system so the continuous-improvement programme measures the same numbers the line reports.

    • The new line is commissioned against the model before the physical machines are installed, shortening ramp-up.
    • The same digital twin is reused as the disaster-recovery exercise surface, so the next shock is a planned re-route.
    • Performance data from the line feeds Baxter GPS without an additional manual collection step.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    Closed connectivity for the Ahmedabad inspection and lab instruments

    A direct connection from the visual inspection machines and laboratory instruments at the Ahmedabad facility into a secure laboratory information management system (LIMS), with each result bound to the operator and the batch, so the lab record is generated by the instrument rather than completed by hand.

    • Instrument-to-LIMS data capture

      Results captured at source

      Connect visual inspection machines, balances, pH meters and other laboratory instruments through a documented protocol, and write the result into the LIMS with instrument identity, method version and timestamp attached rather than reading the value off a screen and re-keying it.

    • Operator binding at the instrument

      Who performed the inspection

      Replace paper log-ins with badge or biometric authentication at the machine human-machine interface (HMI), so the inspector's identity is bound to the result in the system record.

    • GXP-aligned audit trail

      An inspection-ready record

      Implement the audit trail, electronic signature and versioning handling expected by GXP (good practice quality) and 21 CFR Part 11 (the US rule on electronic records and signatures), so the data-integrity gap the FDA identified is closed by the system itself.

    • The audit trail is generated by the instrument, not by transcription, which is what the FDA observation asked for.
    • Inspection prep becomes a query into the LIMS rather than a reconstruction across paper logs.
    • The same connectivity is reusable at the other pharmaceutical sites on the network.

    We've done this before

    See how we do Digital Lab
  • Enterprise AI

    Edge connectivity for legacy OT equipment

    A documented edge connectivity layer that wraps the existing PLCs, chillers and mixing tanks with OPC UA or MQTT publishing, so the equipment's health and process data land in an enterprise data hub without modifying the underlying machine.

    • Protocol translation at the edge

      Legacy equipment speaks a modern protocol

      Wrap existing controllers with a documented edge layer that speaks OPC UA and MQTT, presenting the equipment's process and health data in a vendor-neutral form to the enterprise data hub.

    • Secure OT-to-IT segmentation

      IEC 62443 zones from day one

      Define zones, conduits and remote-access rules to IEC 62443 (the international standard for industrial automation and control system cybersecurity), so the connection is engineered against the OT security standard rather than added as a side door.

    • Equipment data hub

      One stream of plant data

      Land the streaming data from the wrapped equipment in an enterprise data hub, where maintenance, quality and engineering can each read the same source rather than each building their own extract.

    • Predictive maintenance and OEE work start from a documented data stream rather than from re-keyed observations.
    • The same connection path is reusable across the manufacturing network, including the new North Cove lines.
    • Maintenance, quality and engineering read the same data, which simplifies the application rationalisation programme.

    We've done this before

    See how we do Enterprise AI
  • Enterprise AI

    Software validation harness for the Novum IQ platform

    An automated software validation harness for the Novum IQ infusion pump, combining a virtual model of the device's mechanical and user-interface behaviour with a regression test rig that exercises interaction sequences overnight.

    • Virtual pump behaviour

      A software model of the device

      Build a virtual representation of the pump's motor behaviour, fluid mechanic response and user-interface state machine, so interaction sequences can be exercised without the physical hardware.

    • Automated regression corpus

      Interaction sequences run overnight

      Drive the virtual pump through thousands of user-interaction sequences — including race-condition cases identified in field reports — so defects surface in the development cycle rather than in the field.

    • Validation evidence for the next submission

      Coverage data feeds the regulatory file

      Generate coverage and pass/fail reporting from the harness into the regulatory submission package, so the next FDA software filing evidences the depth of the test corpus that was run.

    • Interaction-level defects surface in the development cycle rather than in the field.
    • Regression coverage data feeds the regulatory file, shortening the next submission cycle.
    • The same harness pattern is reusable for the next connected device added to the portfolio.

    We've done this before

    See how we do Enterprise AI
  • Enterprise AI

    HL7/FHIR bridge for the Hillrom connected-care portfolio

    An HL7/FHIR-compatible edge gateway that translates data from Centrella, Voalte and Welch Allyn devices into the healthcare exchange standard, so each hospital deployment becomes a configuration rather than a custom integration.

    • Device-to-FHIR translation

      Proprietary output, standard format

      Translate proprietary device output from Centrella, Voalte and Welch Allyn into FHIR resources, so the EMR receives data in the standard it already consumes.

    • Edge gateway orchestration

      Coordinating many devices at one hospital

      Coordinate the data flow from many devices at a single hospital site through an edge gateway, so the bridge handles translation and buffering locally rather than raising the load on the hospital network.

    • Application consolidation view

      Footprint visible to the rationalisation programme

      Surface the functional footprint of each connected-care application through the gateway, so the CIO's rationalisation programme has a current map of which device-level functions are duplicated across the application estate.

    • Hospital deployments become configuration rather than custom integration.
    • The same gateway feeds the application rationalisation programme with a current functional map.
    • Future device additions translate once and inherit the same EMR contract.

    We've done this before

    See how we do Enterprise AI

Digital maturity: today and target

Scored out of 100 across six dimensions. The target is what Baxter International Inc.'s own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
Data integrity and compliance 38 → 88
The FDA observations on the Ahmedabad facility describe an inspection environment where the instrument capability exists but is not captured into the system of record. Resolution means the system of record is generated by the instrument, with the audit trail built in.
IT and OT convergence 42 → 82
Legacy PLCs, chillers and mixing tanks remain air-gapped across the manufacturing network, while the CIO has explicitly committed to rationalise thousands of legacy applications. The first step is a documented connection; the second is consolidating applications on top of it.
Predictive maintenance 32 → 76
Maintenance is currently largely run-to-failure outside the few sites that have been modernised. A documented data path from the equipment is the precondition for any predictive work to be commissioned.
Supply chain digitalisation 45 → 88
The North Cove event exposed the absence of a digital model of the supply chain that could be exercised for disaster scenarios. The reconstruction is the natural moment to build the digital twin that future planning work will live on.
Software quality automation 50 → 82
Novum IQ recalls relating to interaction-level defects indicate that race conditions are being caught after release rather than before. A virtual device + automated regression harness closes the gap between the test corpus and the real-world interaction space.
Connected care interoperability 40 → 78
Hillrom device data is still published in proprietary formats, and hospital EMR integration is still a project per deployment. A standard translation layer changes the unit of work from project to configuration.

Check this yourself

Our Service Portal has free self-assessments and market comparisons. These are the ones that line up with what we've read above — no sales call required.

Think we've read this right?

Talk to us

Related reading

This is an independent analysis prepared by A4BEE from publicly available information as of January 2026. It reflects A4BEE's own interpretation and opinion, is not affiliated with, endorsed by, or verified with Baxter International Inc., and may be incomplete or inaccurate. All company names and trademarks are the property of their respective owners. To request a correction or removal, contact [email protected].