Every idle minute in an operating theatre is expensive. A scrubbed team stands ready, a sterile room sits empty, and somewhere down the corridor a patient — and a waiting list — pays for the delay. So when a study sets out to shrink those waits by combining two of the most-discussed ideas in healthcare operations — Lean and artificial intelligence — it earns a careful read.
A new open-access paper by Karunakaran and colleagues (2026) in the International Journal of Health Care Quality Assurance does exactly that. It asks how “Lean 4.0” — Lean management fused with Industry 4.0 AI tools — can cut patient waiting times in operating theatres. Below: what it found, the one idea worth borrowing, and — just as usefully — the questions it leaves wide open.
What the study did
The researchers took a qualitative route. Nine hospital operations and management professionals, recruited largely through LinkedIn, sat for semi-structured interviews; a scoping review (190 records narrowed to ten studies) set the backdrop; and the interviews were analysed thematically.
The portrait of why theatres run late will be familiar to anyone who has worked near one: patients arriving late, incomplete pre-operative preparation, thin staffing, and — above all — departments that don’t talk to each other. Against this, participants described AI’s promise: predictive analytics that forecast case durations, overruns and cancellations, letting a list be reshaped before the day unravels. Lean tools — 5S, value stream mapping, Kanban, Kaizen — stabilise the process; AI adds foresight. One participant called the pairing “digital Kanban.”
The one idea worth borrowing: a map of where you stand
The paper’s most useful contribution isn’t a finding — it’s a map. The authors plot any operating-theatre department on two axes: how far it has gone with technology (AI and real-time data), and how mature its processes and culture are (Lean discipline). That yields four recognisable types — and, more helpfully, a sense of which move comes next. The figure accompanying this piece plots all four.
The reactive firefighter (low technology, low process) is the theatre where lists slip because someone arrived late and no two departments share the same information. There is no standard work and little data to lean on. The instinct here is to buy software; the study’s advice is the opposite — stabilise first. Get the basics in place before any algorithm can help.
The lean traditionalist (low technology, high process) is a genuinely well-run theatre. Instruments live in fixed, labelled places (5S); staff “save minutes” by pre-staging trolleys and IV lines (Kaizen). What’s missing is foresight — scheduling still runs on manual lists and whiteboards. For this department, the participant-reported “85% accuracy” of predictive scheduling is the missing link that turns a reactive culture into a proactive one.
The digital island (high technology, low process) is the hospital that bought the predictive tool and then watched clinicians ignore its alerts. The technology is real; the adoption isn’t. The fix isn’t more technology but more people — train staff, involve them in planning, and make sure the data actually feeds decisions.
The lean 4.0 exemplar (high technology, high process) is the destination. Lean’s Kanban and value stream mapping are “supercharged” by AI-driven, real-time resource allocation. The paper points to academic medical centres — Mayo and Cleveland Clinic, as recalled by participants — where pairing value stream mapping with predictive scheduling reportedly trimmed waits by around a quarter.
The value of the matrix is the arrow it implies. Firefighters stabilise their way up to traditionalists; traditionalists add intelligence to become exemplars; digital islands have to build the culture they skipped. It is a diagnostic an operating-theatre manager could run on a Monday morning.
Where the evidence is still thin
Here is the part worth being honest about. The study’s title promises reduced waiting times, yet nothing is actually reduced in it. No AI model is built or validated; no Lean 4.0 intervention is run; no waiting time is measured. The headline numbers — “85% accuracy,” Mayo’s “25% reduction” — are single participants recalling things second-hand, not documented data. The sample is nine people, fairly homogeneous, recruited on LinkedIn, with a couple of voices doing much of the talking. And in a study about patient waiting times, the patient never speaks — nor do the surgeons, anaesthetists and nurses at the sharp end.
None of this makes the paper less valuable. It makes it a starting line rather than a finish line. So let’s read those gaps as a research agenda.
Five questions the study sets up
1. From perception to proof. Move from what managers believe to what actually changed. Interrupted time series, stepped-wedge cluster trials, or even disciplined before-and-after designs can measure real outcomes — turnover time, on-time first-case starts, cancellation and utilisation rates. And rather than quoting claimed accuracies, build and externally validate the predictive models (calibration, AUC, decision-curve analysis). Discrete-event simulation and queueing models can even test interventions before they touch a real theatre.
2. Whose voice, in which system? Multi-site, multi-country studies that separate public from private systems — and that finally include frontline clinicians and patients. Because theatre delay is deeply context-dependent, the health system has to be named, not assumed.
3. Turn the typology into a measurable model. The two-by-two is elegant but untested. Develop validated scales for its two axes, then test the typology empirically — structural equation modelling, with latent-class or mixture analysis to see whether those four types really exist in the data, and whether sitting in a given quadrant predicts performance.
4. Take culture, cost and equity seriously. The paper leans on culture and leadership without an implementation-science frame; CFIR, Normalization Process Theory or RE-AIM would sharpen it, and longitudinal designs could track change over time. It also never costs anything — yet draws its exemplars from elite US centres while claiming relevance to resource-constrained settings. Cost-effectiveness, budget-impact and digital-divide studies would close that gap.
5. Govern the algorithm. Bias, data quality, privacy, explainability and regulatory compliance get a nod, not a chapter. Fairness audits of theatre-scheduling tools, and studies of how clinicians accept — or quietly reject — explainable AI, are a research stream of their own.
The bottom line
Karunakaran and colleagues haven’t settled whether Lean 4.0 cuts operating-theatre waits — they have drawn the map for finding out. The typology, the tensions they surface, and the voices they gather are fertile ground for exactly the rigorous work the field now needs: measurement development, structural modelling, implementation science, health economics and simulation. The next move — and it is overdue — is the step from perception to proof.
Discussed article: Karunakaran, A.P., Antony, J., McDermott, O., Sony, M., Sutoova, A., Kaul, A., Roy Ghatak, R., Demir, S., Islam, D. and Ciliberto, C. (2026), “Reducing patient waiting times in operating theatres via Lean 4.0: a qualitative study”, International Journal of Health Care Quality Assurance, Emerald, DOI: 10.1108/IJHCQA-01-2026-0005 (CC BY 4.0).