Intended for healthcare professionals


Applying quality improvement approaches to health care

BMJ 2009; 339 doi: (Published 02 September 2009) Cite this as: BMJ 2009;339:b3411
  1. Martin Marshall, clinical director
  1. 1Health Foundation, London WC2E 9RL
  1. martin.marshall{at}

    The health sector could learn much from industry

    People who work in the health system can be resistant to learning from other sectors. They claim that health care is different, and it is hard to argue otherwise when you look at the range and diversity of the stakeholders and the ways in which authority is distributed between them. But some challenges are common to all sectors, and it would be wasteful to ignore transferable solutions.

    The manufacturing sector’s approach to improving quality is a case in point. More than half a century has passed since pioneers transformed Japanese industry by applying simple statistical and behavioural principles, by applying simple statistical methods like statistical process control and behavioural practices such as effective teamwork.1 Companies like Toyota and Unipart have adapted these approaches to produce levels of quality and reliability that the health sector can only dream of.2

    It is not that the health sector hasn’t tried to apply these techniques. Twenty years ago Berwick described how healthcare providers might learn from industry.3 These ideas were first taken up by a small band of enthusiasts,4 5 and they are now being used by a large number of providers internationally. But the so called “industrial” approaches to improvement are not yet embedded in all organisations, and they seem to have largely bypassed most clinicians.

    Recently published reports from the NHS Institute for Innovation and Improvement and NHS Quality Improvement Scotland attempt to shed light on why this might be the case.6 7 The reports provide a brief overview of the most common methods that have been transferred from industry to health, describe where and how they have been used, and summarise the evidence of their effectiveness.

    Reassuringly, the two reports reach similar conclusions. Firstly, despite the weird and wonderful names given to the different approaches—Lean, Six Sigma, Business Process Re-engineering, and Theory of Constraints to name but a few—they are based on simple common principles. These core principles are that the customer must be central to everything; work processes should be categorised, redesigned if necessary, and understood as components of a wider system; measuring components of the process and understanding the importance of variation in these measures is fundamental; and the expertise of people who work in the front line should be recognised and valued.

    The approaches vary in their focus on these core principles and the extent to which they advocate incremental or radical change. Nevertheless, although a small number of purists will stick to their preferred technique (sometimes for commercial reasons), most practitioners are highly realistic in their application, drawing on the best of different techniques and adapting the methods to suit local needs.

    Secondly, agreement exists about the conditions—sometimes described as the “organisational context”—under which the approaches are most likely to be effective. Leaders of organisations need to show sustained personal commitment and ensure alignment between the improvement method and the wider organisational strategy. Health professionals, especially doctors, need to be involved. Organisations that have invested in training and supporting their staff, and in the development of high quality information systems, are more likely to achieve success.

    Thirdly, evidence shows that the approaches can engage and enthuse many of the staff who use them, and that they often lead to changes in working practices. However, there is little rigorous evidence that they have a significant or sustained effect on clinical or patient outcomes. The authors suggest that these findings reflect an absence of appropriate evidence of effect, rather than evidence that the interventions are ineffective, and they highlight the ongoing and often heated debate about the most appropriate methods to evaluate complex interventions.8 Most importantly, they advocate evaluative approaches that attempt to understand and explain the implementation environment, rather than treating it as a variable that needs to be excluded. Context, they say, is all.

    These findings help to challenge current thinking and practice about how to improve quality. For example, experience and evidence show that the health sector can learn from industry’s approach to improving quality. However, we should avoid exaggerated claims or overly high expectations of the impact of such approaches. The extent to which they engage frontline staff makes these approaches compelling, but we do not yet know how to optimise their effectiveness. Most importantly, we know little about the cost effectiveness of large scale improvement interventions,9 although we do know that implementing them requires considerable human and financial resources.

    Next, it is time to unpack the black box called “context,” which risks becoming a catch-all excuse for our ignorance about why an intervention succeeds or fails. The two recent reports highlight some of the contextual factors that seem to be important.6 7 We need to understand these factors and the ways that they interact more fully, building on the insights gained from high quality qualitative research.10 We also need to understand the relation between the largely organisationally focused approaches to improvement described in the reports and the more established interventions aimed at individual clinicians (for example, guidelines, clinical audit, or peer review) or high level health system interventions (such as regulation or funding mechanisms). These approaches are not incompatible, and it is reasonable to assume that when used synergistically they are more likely to reap benefits.11 12

    In recent decades we have found several partially effective solutions to the challenge of building a high quality, responsive, and efficient service for patients. We still have a long way to go, however, and we need to draw on, test, and rigorously evaluate a wider set of methods to achieve substantive and sustained change.


    Cite this as: BMJ 2009;339:b3411


    • Competing interests: None declared.

    • Provenance and peer review: Commissioned; not externally peer reviewed.