Operational decisions directly affect patient outcomes, yet lack the scientific infrastructure supporting clinical care. From theatre scheduling to risk stratification tools, the NHS has adopted interventions without the rigorous evaluation expected of clinical treatments.
Patient outcomes are shaped by both clinical and operational excellence. Crucially, the two are not independent. Operational excellence often determines whether clinical excellence can be delivered in practice.
A surgeon may know the best treatment for a patient. But whether that treatment happens promptly depends on theatre capacity, staffing levels, diagnostic pathways, scheduling systems, equipment availability and organisational priorities. How healthcare is organised is therefore not peripheral to clinical care. It is one of its key determinants. Recent failures in NHS maternity services provide a stark reminder of this.
The evidence gap in operational decision-making
Since operational excellence shapes patient outcomes, one might reasonably expect it to be supported by the same scientific infrastructure as clinical excellence. It is not.
The NHS has spent decades building – however imperfectly – a coordinated ecosystem for generating, evaluating and disseminating clinical knowledge. Organisations such as the National Institute for Health and Care Excellence (NICE), the National Institute for Health and Care Research (NIHR), the Royal Colleges and national audit programmes perform different but complementary functions to support clinical decision making.
Operational management exists in a very different environment.
Decisions about staffing models, patient flow, discharge processes, theatre scheduling, waiting-list management, bed allocation and service configuration are often informed by local custom, managerial experience, policy fashion or consultancy advice. Tools frequently arrive in black boxes, without transparent methods, open documentation, reproducible code or independent peer review. Evaluation is variable. Evidence is fragmented. Learning, where it occurs at all, is often ad hoc rather than cumulative.
The consequences are often invisible because many operational decisions appear self-evident. Consider a surgical department making a strong case for a dedicated emergency theatre to improve access for emergency patients. Most managers and clinicians would support it. Yet an operations science study at Erasmus Medical Centre found that a dedicated emergency theatre performed worse than distributing emergency capacity across elective theatres. The hospital subsequently closed it. The counter-intuitive finding was not discovered through experience or intuition, but through operational science.
Now consider a clinical question: should a colorectal cancer patient receive radiotherapy before or after surgery? Clinicians are not expected to rely on intuition alone. They have access to guidance underpinned by trials, systematic reviews and independent appraisal.
The asymmetry is not between the clinical and operational importance of these decisions—both affect patient outcomes directly. It is between the infrastructures available to support them. If operational interventions are not held to the same standards as clinical interventions, we should not be surprised when ineffective, misleading or harmful approaches are adopted.
Examples are not difficult to find.
Predictive risk stratification tools were introduced in primary care with the aim of reducing emergency admissions. Yet when one such tool was evaluated in a randomised trial, emergency admissions increased, healthcare costs increased and there was no clear evidence of benefit for patients. A recent systematic review of population health management risk prediction models concluded: “The evidence does not support further integration of care pathways with costly population-level interventions based on risk prediction in unselected primary care cohorts.”
The NHS has often treated 85% utilisation as a general benchmark for operational efficiency. Yet the same target has been shown to be mathematically unachievable for operating theatres and, in one paediatric unit, would have produced a one-in-three probability of all beds being simultaneously full. The deeper error is to treat utilisation as a target at all. Utilisation is an output of system design, not an input – determined by the level of service, reliability and resilience required.
Hospital mortality statistics - which occupied over a decade of my academic research activities including being an expert witness to the Shipman inquiry- provide another striking example. Differences between observed and expected deaths were widely interpreted as estimates of avoidable mortality despite longstanding warnings that such interpretations were invalid. A prominent example was the claim that between 400 and 1,200 patients died because of poor care at Mid Staffordshire. Subsequent independent case-note review found perhaps one death plausibly attributable to poor care. David Spiegelhalter called such figures “zombie statistics” because they survive repeated methodological demolition – and then haunt hospitals.
Building a national capability for operational science
The issue is not a lack of expertise. Across the NHS, operational scientists have applied methods such as queueing theory, simulation modelling, optimisation and statistical process control to challenges ranging from waiting lists and patient flow to capacity planning and performance monitoring. Programmes such as GIRFT, the Elective Recovery Programme and specialist analytical teams across the NHS demonstrate the value of rigorous operational science.
The New Hospital Programme provides a recent illustration of what good looks like. Concerns about future capacity planning were addressed through the development of an open-source, peer-reviewed demand model supported by formal quality assurance and explicit treatment of uncertainty. The National Audit Office subsequently commended the approach for its transparency and rigour.
The NHS does not lack operational science capability. It lacks the infrastructure to coordinate and mobilise it at scale. As a result, operational knowledge remains fragmented: guidance is sometimes excellent, sometimes flawed and often absent.
NICE provides evidence-based guidance from which clinicians make context-specific decisions. An operational equivalent would serve a similar function.
My purpose, however, is not to prescribe a particular model, but to argue that the NHS should make a concerted effort to develop a coordinated national capability for operational science. The case for doing so is stronger today than at any point in the NHS's history. Advances in routinely collected data, computing power, modelling and artificial intelligence have transformed operational science from a specialist activity into a practical tool for everyday decision-making.
International experience shows what coordinated effort can achieve. Sweden's LAPS CARE programme applied operational science methods to workforce planning and service delivery across more than 200 organisations, improving efficiency and quality while reducing costs.
Operational excellence has never been a nice-to-have. It is a prerequisite for clinical excellence and better patient outcomes. The question is whether policymakers will finally recognise this and give operational excellence the same priority as clinical excellence.
This article originally featured in the Health Service journal: https://www.hsj.co.uk/service-redesign/we-need-a-nice-for-operational-excellence/8123691.article