What is possible · Further vignettes
Recurring challenges.
Made visible. Made manageable.
These vignettes illustrate how CauseACTION translates the hidden complexity of daily work into visible, actionable insight. Each reflects a real pattern within cardiac services — not isolated incidents, but systemic challenges that repeat because the system has no way to see them clearly enough to address them.
Operations · Theatre flow
Coordinating theatres and critical care — replacing phone calls with live context
One of the most persistent operational challenges within cardiac services is managing critical care bed availability throughout the day. For Band 7 coordinators, this requires balancing multiple moving parts across several theatres simultaneously — conducted through a constant stream of phone calls, entirely reactive, introducing risk and contributing to theatre downtime and late finishes.
CauseACTION addresses this by applying live data from the EPR and theatre systems to automatically prioritise cases and surface real-time bed availability to all teams simultaneously. ITU update bed readiness directly. Fewer phone calls, fewer late finishes, senior nurses freed for clinical decision-making.
Learning · NLP intelligence
Revealing hidden clinical patterns — what free text knows that databases do not
Certain intraoperative challenges remain invisible to traditional reporting systems because they are described only in free text notes — never coded, never counted, never fed back into learning. A powerful example involves cases where myocardial ischaemia management deviates unexpectedly: an arrest not sustained, multiple re-arrest attempts, conversion to full bypass without conventional arrest, later found to be due to anomalous coronary drainage. The finding was previously documented — it was there in the operative notes. But written in unstructured free text, it didn't automatically surface and didn't reach the people or processes that could have learned from it.
CauseACTION applies NLP to operative notes, perfusion records, and post-case summaries — identifying and classifying similar descriptions across cases, flagging clusters of events that might indicate system weaknesses previously unrecognised. Tacit knowledge becomes explicit organisational learning.
Safety · Coordination failures
The case that nearly disappeared — process failures and communication gaps
A high-risk patient was scheduled for a complex aortic procedure. The referral arrived, an initial request for further details was sent, and nothing came back. Nine days later, contact was re-established the evening before surgery — leaving hours rather than days for essential preparation. This is a structural pattern: complex referrals that fall between systems, between people, between working weeks. The information needed to prevent the problem exists. It has no mechanism to surface itself.
CauseACTION prevents this through live referral tracking, automated escalation when responses are not received within defined windows, and shared accountability. Key risk factors and preparation tasks are visible to all stakeholders before the day of surgery — not because someone remembered to chase, but because the system made forgetting impossible.
Workforce · Complexity capture
Capturing operational complexity — the adaptations that never get recorded
Within a busy cardiac unit, countless small adaptations occur daily — staff running late, sickness cover arranged informally, theatre allocations changed at short notice. None are routinely recorded. Yet each shifts the operational picture: workload peaks, staffing ratios change, the risk profile of the list quietly increases while remaining invisible to anyone not physically present. Operational complexity can be precisely defined: the volume of change against a planned baseline. Currently that cost is absorbed silently.
CauseACTION makes complexity measurable by overlaying real-world operational data onto planned schedules — surfacing workforce pressures as they unfold, enabling proactive rebalancing, and generating a retrospective dataset that makes invisible patterns visible for the first time.
Evidence · Health Technology Assessment
Why technologies appear not to work — and how to tell the difference
Cerebral near-infrared spectroscopy (NIRS) monitors oxygen delivery to the brain during cardiac surgery. Two centres both record it as "in use." At one centre, when the NIRS signal drops, the team responds — adjusting flow, pressure, or haemoglobin. At the other, the signal is observed but rarely acted upon systematically. In both cases, the HTA dataset records the same thing: NIRS used. The analysis concludes: NIRS has no effect. But the conclusion is wrong. The technology did not fail. The implementation did.
This is one of the hardest problems in real-world evidence — what might be called implementation fidelity bias. Health Technology Assessment typically captures whether a technology was present and what the outcome was. It does not capture whether the technology was used as intended, whether the signals it generated were interpreted correctly, or whether the appropriate responses followed. Variation in implementation is systematically misread as variation in effectiveness. Technologies are dismissed. Guidance is wrong. Patients are affected.
The same failure mode runs across devices, drugs, and pathways. In every case, outcome = intervention × implementation × context. Current HTA can measure the first and the last. It cannot see the middle.
CauseACTION introduces fidelity-aware real-world evidence. For the NIRS example it would capture four layers for every case: the signal (NIRS dropped below threshold), the context (patient haemoglobin, bypass flow, pressure at that moment), the response (was action taken — and what action), and the outcome (AKI incidence, stroke, length of stay). This creates three analytically distinct groups that current HTA cannot distinguish: NIRS present and acted upon; NIRS present and not acted upon; no NIRS. The question shifts from "does NIRS work?" to "does NIRS work when used as intended?" — which is the question that actually matters, and which produces guidance that is conditional, context-aware, and clinically actionable rather than a blunt yes or no. The same principle applies to every monitoring technology, every pathway protocol, and every drug where dose, timing, and response are as important as prescription. CauseACTION does not just measure care. It can explain why interventions appear to work — or not — in the real world.
