GE HealthCare · Enterprise Analytics
An analytics platform spanning ICU and Labor & Delivery clinical workflows. The core design challenge: clinical dashboards serve different cognitive tasks than operational dashboards — and conflating them is a patient safety issue.
Healthcare analytics systems are often built as if all data serves the same purpose. They are not. A clinician at a bedside reading a patient's fetal heart rate trace is making a different kind of decision than a charge nurse looking at unit occupancy across a shift. These are not variations in the same task. They are fundamentally different cognitive acts.
The Clinical & Operational Insights platform was built on a single organising principle: separate these two paradigms, design for each one's actual cognitive requirements, and resist the pressure to merge them into a single unified view that serves neither well.
Clinical dashboards support time-critical decision-making in the presence of high uncertainty. The design requirements follow from this: information needs to be available before it is consciously sought, status needs to be readable at a glance from distance, and the visual design must not produce alert fatigue by treating everything as equally urgent.
The ICU and Labor & Delivery workflows had different density requirements, different glanceability needs, and different time horizons for the decisions being made. The platform accommodated these differences rather than flattening them.
Operational dashboards support a different kind of reasoning — trend analysis, resource planning, shift handover, compliance review. The temporal horizon is different (shift, day, week rather than minute, hour), the decision-maker is different (charge nurses, unit managers, administrators), and the cost of being wrong is different (inefficiency rather than immediate patient harm).
Designing operational dashboards for clinical contexts requires understanding that the same people sometimes use both kinds of dashboards — and that switching between cognitive modes is itself a design problem. Context-switching cues, visual differentiation, and navigation architecture all need to support the transition.
The information architecture challenge in this project was not just "where does each piece of data live?" It was "how do the two paradigms relate to each other — and when should they intersect?"
There are moments when clinical and operational data are genuinely connected: a sudden increase in acuity in the ICU has operational consequences. The platform needed to make these connections available without collapsing the distinction between paradigms. This required an architecture that could surface cross-paradigm signals without requiring users to manually navigate between modes.
NDA note
Screens and specific design artefacts from this project are under NDA and cannot be shared publicly. The content above describes the design architecture and key decisions without reference to specific implementation details or proprietary clinical data structures.
Platform scope
The core insight
Clinical and operational analytics are different cognitive paradigms. Merging them doesn't serve both — it fails both.
Related work
CareLogic — Clinical Decision Support Platform →