Abstract
Background:
This quality improvement study was undertaken in the ophthalmology theatres at the Western Eye Hospital, Imperial College Healthcare NHS Trust, to address reduced capped theatre utilisation during the COVID-19 recovery period. Baseline utilisation was 75.8%, below the national Getting It Right First Time target of 85% for theatre touchtime utilisation by 2024/25. The aim was to identify the causes of underperformance and implement targeted interventions to improve theatre efficiency.
Methods:
A Lean Six Sigma DMAIC (Define, Measure, Analyse, Improve, Control) framework was used as the study design. Targeted interventions included increasing pre-assessment capacity to reduce under-booking, improving scheduling processes, reducing early finishes, and carrying out structured daily reviews of implant and equipment availability to minimise on-the-day cancellations. Multiple regression, factorial Design of Experiments, process capability analysis, Plan-Do-Study-Act cycles, and Statistical Process Control were used to analyse variation, test changes, and monitor performance.
Results:
The Analyse phase showed that early finishes, under-booking, and limited pre-assessment capacity were the main drivers of low utilisation, rather than commonly assumed factors such as late starts. Following the introduction of targeted interventions, capped theatre utilisation improved from 75.8% to 81%. Continuous monitoring through Plan-Do-Study-Act cycles and Statistical Process Control supported stability and informed further refinements.
Conclusions:
This project shows that combining Lean Six Sigma methodology with advanced statistical analysis can deliver measurable improvements in theatre utilisation. The findings challenged long-held assumptions about the causes of inefficiency and highlight the importance of optimising scheduling processes and pre-assessment capacity. The model offers a transferable, scalable approach for other units aiming to enhance theatre efficiency and achieve sustainable operational improvements.
Introduction
Imperial College Healthcare NHS Trust delivers acute and specialist services across five hospitals. The Western Eye Hospital (WEH) is a specialist ophthalmic centre providing elective surgery and emergency care.
In September 2021, WEH theatres had a capped utilisation rate of 75.8%, below the national target of 85% as per GIRFT (2021), which was similar to the average capped utilisation of other units of comparable size, such as the Eye Treatment Centre theatres at Whipps Cross University Hospital, Barts Health NHS Trust, which recorded 74%–78% capped utilisation during the same period. COVID-19 restrictions, staff shortages, and limited pre-assessment capacity contributed to underuse of theatres, increased waiting times, and financial losses. According to Mohammed et al (2022), theatres estimated cost during the restriction was £2,700 per hour compared to pre-pandemic estimated cost of £1,200 per hour as suggested by Briggs (2019). Even after restrictions were lifted, we faced large backlogs of patients awaiting surgery, while pre-assessment capacity remained unchanged. This mismatch resulted in underbooked theatres and sustained inefficiency, as noted in theatre utilisation reports and theatre scheduling meetings. Improving utilisation was therefore essential.
Capped utilisation time within scheduled sessions was chosen as the key measure of theatre efficiency. Capped utilisation counts only the scheduled session time, excluding overruns, whereas uncapped utilisation includes any extra time beyond the scheduled session. Using capped utilisation avoids distortion from overruns and provides a standardised measure of efficiency.
Nationally, the Royal College of Ophthalmologists (2022) adopted the GIRFT programme to introduce ‘high-flow’ cataract hubs to standardise surgical pathways and reduce backlogs. Locally, as suggested by Ahmed (2019), a Lean Six Sigma approach was adopted. Lean Six Sigma is a widely adopted quality improvement approach in health care and industry. According to George (2002), it focuses on streamlining processes, eliminating inefficiencies, and reducing errors to enhance outcomes through data-driven problem-solving. It integrates the speed of Lean with the quality and consistency of Six Sigma, utilising the DMAIC framework to guide improvements from initial definition to long-term control.
This article outlines the improvement journey and provides insights for other surgical services.
Methods
Detailed DMAIC framework for theatre optimisation
Define
The Define phase of Lean Six Sigma as Rosa et al (2024) have shown establishes the problem, objectives, and stakeholders. At WEH, theatre utilisation stood at 75.8% between January and September 2021. A SMART (Specific, Measurable, Achievable, Relevant and Time-bound) objective was set to raise this to 85% by December 2021.
A SIPOC (Suppliers, Inputs, Process, Outputs, Customers) analysis (Table 1) adopted from Improvement Academy of Public Health Wales (2023) and process mapping (Figure 1) adopted from Institute for Innovation and Improvement of NHS England (2017) clarified the theatre pathway and highlighted potential bottlenecks. These tools ensured that roles and responsibilities were clear and that areas for influence were identified.
WEH theatre utilisation SIPOC analysis
WEH theatre utilisation process mapping
Key stakeholders (Figure 2) included pre-assessment teams, booking coordinators, anaesthetists, surgeons, theatre staff, recovery teams, and managers. Their early engagement promoted collaboration and created shared ownership of the improvement process.
WEH theatre stakeholder analysis
Measure
The Measure phase gathered data to assess theatre utilisation and identify factors affecting performance as de Koning et al (2006) have shown. The primary output was capped touchtime utilisation, with input variables including late starts, early finishes, intercase downtime, on-the-day cancellations (OTDC), and unbooked capacity.
These measures were chosen for their known impact on efficiency. The data collection plan (Table 2) summarises how each was defined and used.
WEH theatre utilisation data collection plan
A process capability analysis, as suggested by Alatefi et al (2019), was conducted to assess whether current utilisation could meet the national benchmark of 85% (i.e. 210 min of operating time within a 240-min session). This analysis required the tested data to be stable and normally distributed. Therefore, the data distribution was assessed for normality and stability, and Box-Cox transformation methods were applied where necessary to correct skewness and enable valid statistical analysis.
The results (Figure 3) revealed that approximately 17% of theatre lists overran, with actual operating times exceeding 240 min, while 63.7% finished early, with actual operating times falling below the 210 min threshold. This highlights potential scheduling issues. In addition, the theatre utilisation Process Performance Capability Index (Ppk) was found to be 0.12, compared to the ideal Ppk range of 1.0–1.3, suggesting that the current process is not capable of meeting the 85% utilisation target.
Touchtime process capability histogram
Through this thorough measurement process, a clear understanding of current performance was obtained, which laid the foundation for identifying root causes in the subsequent Analyse phase.
Analyse
The Analyse phase focused on identifying the root causes of underutilisation, as Driesen et al (2022) have shown, to explore the relationships between key performance indicators and overall theatre efficiency. Root cause analysis (Figure 4) pointed to efficiency variation between theatre sessions; while some achieved 97% utilisation, others performed as low as 55%. The analysis also indicated that inconsistent scheduling processes and limited pre-assessment capacity contributed to significant underutilisation.
WEH theatre utilisation root cause analysis (Fishbone Diagram)
Regression analysis as suggested by Field (2024) tested late starts, early finishes, intercase downtime, OTDC, and unbooked capacity (input variables) against the output utilisation rates. Results (Figure 5) demonstrated that early finishes were the strongest predictor of underperformance, contradicting the historical assumption that late starts were the primary cause. Further analysis of the reasons for early finishes (Figure 6) showed that under-booking of cases, rather than OTDC, strongly correlated with early finishes, highlighting the importance of scheduling.
Impact of input factors on touchtime output (correlation analysis)
Impact of OTDC and case capacity used on early finish
Predicted response calculator (Figure 7) was used to develop a data-driven model for forecasting theatre utilisation based on varying levels of input conditions. This tool enabled the team to simulate different operational scenarios and determine the combination of inputs most likely to achieve the target of 85% utilisation.
Regression analysis predicted response calculator
Overall, the Analyse phase enabled a clear understanding of the specific variables hindering performance. By shifting the focus from assumed causes to evidence-based contributors, the project laid the groundwork for designing precise and impactful improvements in the next phase.
Improve
McDermott et al (2022) suggest Lean Six Sigma in healthcare can aid the improvement by identifying the type and nature of problems. The Improve phase addressed root causes through targeted interventions. Stakeholders, guided by the Improvement Lead, used structured brainstorming and prioritisation tools to generate solutions and rank them based on their potential benefit and required resources.
The main bottleneck was pre-assessment capacity, which limited the supply of patients to fill theatre lists. COVID-19 swabbing and infection control measures contributed to this restriction, alongside patient reluctance to attend appointments.
The agreed interventions included:
Increase pre-assessment capacity to ensure enough patients are ready to be booked.
Reduce unbooked theatre capacity to zero and early finish to a maximum of 13 min through robust theatre scheduling process under direct supervision by service manager and theatre manager.
The theatre team reviews case-level theatre lists one day in advance to confirm the availability of required implants, equipment, and instruments, aiming to reduce OTDC to zero.
The theatre manager closely monitors late starts and intercase downtime, with targets to reduce it to 13 and 10 min, respectively.
As per the regression analysis and formulated predicted response calculator (Figure 7), these interventions should enable us to achieve 85% capped theatre utilisation.
A Plan–Do–Study–Act (PDSA) cycle (Figure 8), based on the Institute for Healthcare Improvement (2017) framework, was used to ensure changes were trialled and measured before wider adoption
PDSA cycle
During the Improve phase, some stakeholders assumed that cataract cases involving ASA 3 patients, those scheduled for afternoon sessions, or those performed by trainees and non-consultants take longer to complete. Consequently, they suggested these factors be considered when scheduling theatre lists. Advanced statistical methods were used to validate this assumption. A factorial Design of Experiments (DoE) tested the impact of three variables: surgeon grade (consultant vs non-consultant), session time (AM vs PM), and ASA classification (⩽ 2 vs ⩾ 3). Initial plots (Figure 9) suggested longer touchtimes for consultants, afternoon sessions, and higher ASA scores. However, interaction analysis (Figure 10) refined these observations. A minor interaction was identified between surgeon grade and ASA classification, with consultants spending slightly more time on ASA ⩾ 3 patients. Similarly, session timing interacted with ASA status; cases involving ASA ⩾ 3 patients were longer when performed in afternoon sessions. Notably, these effects were inconsistent across groups. Furthermore, case duration was not influenced by an interaction between surgeon grade and session time, which remained statistically insignificant.
Average cataract theatre touchtime plots for the tested variables
Interaction plots of the tested variables
To validate these initial findings, a full-factorial Design of Experiments (DoE) was conducted over eight runs to produce a predictive model. Through iterative refinement guided by Pareto chart visualisations (Figure 11) and confirmed by p-values, non-significant terms were removed step-by-step as follow:
Removed AC (Surgeon Grade × Session Time).
Dropped AB (Surgeon Grade × ASA Classification).
Excluded ABC (three-way interaction).
Removed A (Surgeon Grade) once no interactions remained.
Dropped BC (ASA Classification × Session Time).
Final model retained B (ASA Classification) and C (Session Time), though neither was statistically significant (p = 0.2134).
This confirmed that, in the data reviewed, surgeon grade, session timing, and ASA classification did not have a meaningful impact on cataract touchtime. This is different from the actual surgical time, which is measured from knife-to-skin to closure. These findings could change with a much larger sample size or if there is significant change in the workforce.
Pareto chart of coefficients for case duration (min)
These findings are important. They show that theatre scheduling should not rely on surgeon or patient factors alone. Instead, efficient list booking and sufficient pre-assessment capacity drive utilisation more reliably than attempts to stratify cases by ASA grade or surgeon seniority in isolation from other contributing factors.
Through this comprehensive Improve phase, the project moves from diagnostic insight to practical implementation, ensuring that interventions are grounded in data and designed for sustainable impact on theatre efficiency and patient care outcomes.
Control
The Control phase in Lean Six Sigma ensures theatre utilisation improvements are sustained by embedding monitoring systems and preventing regression. Its focus is on keeping processes stable, predictable, and aligned with the touchtime benchmark.
The following data were collected and reported weekly:
Late start minutes (input metric),
Early finish minutes (input metric),
Intercase downtime minutes (input metric),
OTDC number (input metric),
Unbooked capacity number (input metric),
Touchtime utilisation percentage (output metric).
Regression analysis, as explained above, was used monthly to determine the impact of each input metric on theatre utilisation and guide our improvement focus. In addition, Process Capability (PC) analysis was conducted monthly to assess whether our control processes could consistently deliver a controlled outcome
The Improvement Lead supports this by coaching stakeholders to adopt control measures. These include weekly utilisation reviews, clear ownership of metrics such as underbooked capacity and early finishes, which are managed by the scheduling team with interventions to fully utilise pre-assessment capacity and align sessions to the known average surgery time per surgeon per procedure. OTDC had to be authorised by the general manager on a case-by-case basis. The theatre manager monitors intercase downtime and late starts, intervening to address issues when needed. In addition, sharing anonymised surgeon utilisation data in monthly operational meetings and weekly on the theatre bulletin board has fostered a culture of healthy competition. One of the lowest-utilisation surgeons even volunteered to be named as a case study to demonstrate the improvement they made in utilising their theatre list. Continuous data collection and feedback loops reinforce these processes and promote a proactive improvement culture.
Furthermore, Statistical Process Control (SPC), using control charts, helps distinguish between natural (common cause) variation and special causes. For example, monitoring late theatre starts over time can highlight when performance drifts beyond control limits, triggering timely root cause investigations and corrective actions. In the SPC chart (Figure 12), special causes were identified on two occasions. The investigation found that patient non-attendance (DNA) led to an unusual delay in the start time on those two occasions. As a result, an intervention was agreed to introduce a prompt call to all listed patients one day before surgery to confirm their attendance.
Late start SPC chart, highlighting instances of variation for special causes
Embedding SPC within routine practice supported resilience and a culture of continuous improvement, ensuring utilisation gains were maintained and refined over time.
Result
This quality improvement project successfully applied a structured Lean Six Sigma framework to address a Covid impacted under-utilisation of theatre time at the Western Eye Hospital, Imperial College Healthcare NHS Trust. By focusing on data-driven insights rather than commonly assumed causes, such as late starts, we achieved a notable improvement in capped theatre utilisation, increasing from average baseline of 75.8%–81%. This improvement was gradually achieved alongside the incremental increase in pre-assessment capacity and enhancements to the scheduling process. Progress was monitored, and data were collected weekly over 3 months (October, November, and December), consistently moving towards our target of 85%. However, due to estate-related and fire safety issues, the unit had to decant to another hospital, and services were provided by a different team. This demonstrates that targeted, evidence-based interventions can yield rapid operational gains.
Discussion
A key strength of this study is its robust, data-driven methodology. The use of advanced statistical tools, including multiple regression and factorial Design of Experiments (DoE), allowed for a detailed understanding of process variation that simple observation would have missed. Our analysis revealed that early finishes and under-booking were the primary contributors to low utilisation, rather than late starts. This challenges conventional assumptions and highlights the importance of optimising patient scheduling and pre-assessment throughput. This finding aligns with the national GIRFT guidance and supports the move towards high-flow cataract hubs.
However, the study has several limitations. Being a single-centre ophthalmology project, the findings may not be fully generalisable to other specialties or operational settings. Although the DMAIC methodology provides a replicable framework, the specific root causes and interventions identified may require adaptation elsewhere. Furthermore, while the improvement achieved is commendable, the final utilisation rate of 81% remains below the national benchmark of 85%, indicating that additional strategies are needed to close the remaining gap. The sustainability of the initial gains could not be assessed due to theatre decanting to another site; ongoing monitoring will be essential to ensure that improvements are maintained.
Our DoE analysis found no statistically significant effect of surgeon grade, session time, or ASA classification on cataract case duration. This finding has important implications for surgical list planning, suggesting that fluid scheduling based on a consistent supply of pre-assessed patients may be more effective than rigid categorisation by surgeon or patient factors. This aligns with the principles of high-flow cataract hubs and reinforces the value of standardising processes over relying on individual variables.
Future research should explore the application of this model to other specialties, as well as the long-term impact on patient outcomes, staff satisfaction, and financial performance. Further work is also needed to address the remaining gap to the 85% utilisation benchmark and to reduce on-the-day cancellations.
Conclusion
This quality improvement project successfully applied Lean Six Sigma methodology to increase theatre utilisation from 75.8% to 81% at the WEH theatres during the challenging COVID-19 recovery period. By focusing on capped utilisation as a robust performance measure, the initiative uncovered key drivers of underperformance, including early theatre finishes, under-booking due to pre-assessment limitations, and process variation across the surgical pathway.
The project’s structured use of the DMAIC framework enabled a deep exploration of these issues, supported by advanced statistical tools such as regression analysis, Design of Experiments (DoE), and Statistical Process Control (SPC). These tools facilitated not only the identification of root causes but also the development and refinement of evidence-based interventions. The successful deployment of PDSA cycles further ensured that changes were rigorously tested before full-scale implementation.
Perhaps most critically, this initiative demonstrated the importance of stakeholder engagement, data transparency, and iterative learning in driving sustainable improvement. Although statistical models suggested that certain clinical variables had limited impact on touchtime duration, the project created a platform for challenging assumptions and fostering a culture of continuous improvement.
Moving forward, the principles and learning from this work can serve as a blueprint for other healthcare service providers seeking to improve theatre efficiency. Sustaining these gains will depend on maintaining robust data monitoring systems, leadership support, and a shared commitment to high-value care delivery.
The model presented here offers a recommended, replicable, and scalable approach to theatre optimisation across the wider healthcare system, particularly in units of comparable size.
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
