Integrating a novel biosensor into
cardiac care
Lactate is one of the most important indicators of cardiac recovery, yet clinicians had no way to monitor it continuously at the bedside. I designed the interface for the first real-time lactate integration on a clinical heart monitor.
A new sensor, and a new kind of data to design for.
The Impella is the world's smallest heart pump, used when a patient's heart can't sustain blood flow. The Abiomed Auxiliary Monitor (AxM) shows the care team real-time vitals, trends, and alerts from the pump.
Why Lactate Matters
Lactate is a blood metabolite produced when organs aren't receiving enough oxygen. It's one of the clearest indicators of whether a cardiac patient is recovering (currently measured via blood draws every 4–6 hours). A continuous real-time sensor changes that gap significantly.
Three Design Constraints
The Original Interface
The existing AxM pulled from a single data source: the heart pump. The team was also planning to introduce algorithmic decision tools and external biosensors, and the existing interface had no way to support them.
The Solution
I restructured the layout to accommodate new data types and future integrations, introducing a trend system that represents lactate meaningfully while supporting existing clinical workflows.
Modular Monitoring Layout
The AxM uses a tile-based design system. Integrating lactate meant introducing a data type that behaves differently from everything else on screen without disrupting the visual hierarchy clinicians rely on during high-pressure moments.
Lactate Tiles
Two tile variations were explored to support different monitoring needs: a simplified view focused on current value and directional change, and an expanded view that also included change from baseline for added clinical context.
Lactate Trend Page
Selecting a lactate tile opens a dedicated trend view where clinicians can inspect values across adjustable time ranges, tap to view exact measurements, and drag across the graph to compare changes over time.
Multi-Trend Comparison
The multi-trend view allows clinicians to compare lactate alongside other hemodynamic metrics, helping identify relationships between signals while maintaining the same interactive inspection behaviours across graphs.
Critical Monitoring Mode
A high-frequency monitoring mode was introduced for severely critical patients, updating lactate values every 20 seconds to 2 minutes. This design decision was informed by contextual inquiry findings showing that rapid spikes could be missed in standard monitoring intervals.
The project advanced to its next development stage.
Advanced Development 1 (AD1) is J&J's internal gate for demonstrating clinical and technical feasibility before committing further R&D resources. The prototypes passed, showing that this integration was something physicians would actually use, for a sensor that doesn't yet exist in hospitals.
How did I get here?
There were no existing UX patterns for continuous lactate on a clinical monitor, so I started by just trying to understand the space: who actually uses this data, when it matters, and what they do with it when it does.
Getting up to speed on the clinical side
Before sketching anything, I spent time on literature reviews and internal artifacts to get a handle on biomarkers, hemodynamics, and how lactate fits into the Impella workflow. I also spoke with two principal scientists from the academic research team and the Director of the Critical Care Platform, people who knew the clinical workflows deeply. Together this gave me enough context to ask the right questions before talking to clinicians directly.
Where does lactate actually come up?
Lactate means different things at different points in the patient journey. I mapped out both workflows to get a clearer picture of when it's being watched, by whom, and what they're deciding based on it.
Running a workshop with the broader team
We brought together the algorithms and engineering teams to generate concepts around tile structure, action triggers, and trend interpretation. After dot-voting, we landed on a lactate-focused flow that 5/5 clinical consultants preferred in early feedback — mainly because it was simple and didn't require much cognitive overhead to act on.
Sketching out what the interface could look like
With a baseline understanding of the clinical context, I started generating early wireframe concepts exploring different ways to display continuous lactate data on the existing AxM screen. This included tile layout variations and different approaches to showing trend over time.
Testing early concepts with clinicians
I ran semi-structured interviews with 5 clinical consultants to check whether the early wireframe concepts matched how clinicians actually think about lactate during patient monitoring. The goal was to figure out which interaction model better supports real decision-making moments — and where the concepts were off.
Critical Care Physician
Critical Care Nurse
Interventional Cardiologist
Clinicians prefered the metric-focused layout as the interaction of tapping a tile to drill into a dedicated trend page felt more intuitive and kept clinicians focused on one metric at a time.
Clinicians responded well to the large number display and directional arrow, but wanted change from baseline rather than percentage change.
Watching the prototype in a real clinical scenario
I had the opportunity to participate in a lactate-focused animal study with clinicians and stakeholders. During the session, we found that the data update frequency we'd been using, based on earlier clinical consultant feedback, caused people to miss significant spikes in lactate, particularly in more critical cases. Because of this, I added a high-frequency monitoring mode specifically for critical conditions. It was a straightforward fix, but one that wouldn't have come up in a regular prototype review since it only became visible with fast-moving live data.
I introduced a high-frequency mode with more granular timeframes and rapid real-time updates (every 20 seconds to 2 minutes) for patients in critical condition. Standard intervals weren't fast enough to catch sudden lactate spikes.