Organisational culture: problem-sensing and comfort seeking


THIS Institute logo

You can view this chapter as as PDF here.

Publication date: 17 April 2023
Last reviewed: 17 April 2023

In this chapter:


Back in 2004, Bill Moyes, who was at the time executive chair of the foundation trust regulator Monitor, said: "There is no such thing as a perfect organisation. The best we can ever hope for is that an organisation is self-aware, recognises its issues, and deals with them effectively".  This remains as true today as it was nearly 20 years ago. 

The complexity of large organisations and even greater complexity of working within systems challenges organisations to be self-aware. The evidence of what is going well, what can be improved and what requires urgent attention is readily available in organisations that use soft intelligence effectively. There will always be the temptation to take comfort in getting most things right rather than being disconcerted by the significant minority of things that go wrong. There will always be a danger in regarding compliance as an end in itself rather than as useful, but limited, intelligence on organisational performance.

In this section Mary Dixon-Woods and Graham Martin contrast problem-sensing with comfort- seeking, confront structural complacency and a lack of eagerness to use hard and soft intelligence, and discuss the crucial importance of openness.

It would be surprising if there were many who would disagree with the content of this chapter, but the challenge is to embrace openness and make real a learning health system in which structural secrecy is identified and challenged at all levels.


Key messages:



Unwarranted variations and deficits in patient safety and quality of care continue to present a major source of harm and distress for patients, families and staff, as well as consuming a growing proportion of the NHS budget on payments for negligence claims. 1 2 3 At their most extreme, these problems result, with depressing frequency, in organisational catastrophes – including, but not limited to, the Bristol Royal Infirmary and Gosport War Memorial Hospital failures in the 1990s, Mid Staffordshire NHS Foundation Trust in the 2000s, and, in the recent past, the Morecambe Bay, Cwm Taf, Shrewsbury and Telford, and East Kent maternity units. Some vulnerable groups are especially badly impacted by poor quality care. The Learning Disabilities Mortality Review Programme,4 for example, has revealed a grim pattern of premature death for people with learning disability. Though each of these organisational degradations and failures in care is distinctive, they demonstrate many shared features. They include discounting of warning signs, poor management systems, failure to listen or act on patient and staff concerns, fragmentation of knowledge about problems and vacuums of responsibility for addressing them, cultures of secrecy and protectionism, and fragmentation of knowledge about problems and responsibility for addressing them.5

It is clear that healthcare systems are vulnerable to what Diane Vaughan, in her account of the Challenger disaster, terms "structural secrecy". This describes how "patterns of information, organisational structure, processes, transactions, and the structure of regulatory relations systematically undermine the attempt to know and interpret situations."6  In this chapter, we discuss a crucial contribution to structural secrecy: behaviours that can undermine the capacity to recognise and act on sub-optimal care. These behaviours may be found at all levels of the health system – from individuals and teams all the way upwards through organisational leadership and the regulatory and policy level.  Based on a very large study of culture and behaviour in the English NHS, these behaviours can be characterised on a spectrum from "comfort-seeking" to "problem-sensing".5 Problem-sensing involves actively seeking out weaknesses in systems relating to quality and safety, typically using multiple techniques and sources of organisation intelligence. Comfort-seeking, on the other hand, is characterised by seeking reassurance, by taking undue confidence from the data available, and by the inability or unwillingness to seek out information that might challenge the sense that all is well.


Reluctance and eagerness to know 

Problem-sensing and comfort-seeking behaviours to some extent reflect variations in dispositions towards wanting to know. Tendencies towards complacency, over-optimism, and even self-deception reflect patterns found in disasters outside healthcare – including, memorably, the financial crisis of the late 2000s.7  Within healthcare, these kinds of orientations are highly consequential. At Mid Staffordshire, for example, many of the behaviours described by Sir Robert Francis 8 fell firmly in the category of comfort-seeking. Senior leadership at the trust demonstrated comfort-seeking behaviours that appeared to be rooted in hubris, including the belief that it was compliant with quality and service standards despite numerous internal indicators that it was not. As in the Challenger disaster6 and other failures in aerospace and the oil industry,9 staff at the trust with specialised technical expertise were marginalised from the trust’s decision-making structures. Over 900 incident reports submitted by staff in Mid Staffordshire on understaffing and other safety concerns were neglected. Experienced doctors and nurses were pressurised to collude in creating favourable accounts of the Trust, for example by getting patients through the emergency department on time—or making it appear that they did—regardless of the consequences for quality of care. More recently, the Kirkup report into maternity and neonatal services at East Kent found that "the Trust wrongly took comfort from the fact that the great majority of births in East Kent ended with no damage to either mother or baby."10

Problem-sensing behaviours, on the other hand, involve caution about being self-congratulatory.5 They also involve actively seeking out data that offer challenge, disrupting any incipient risk of complacency.11 Equally importantly, when they do uncover problems, problem-sensing behaviours involve the use of strategies that go beyond merely sanctioning staff at the sharp end of care delivery, 12 instead making more systemic and holistic efforts to strengthen their organisations and teams.


Variations in ability to use routine data to monitor quality and safety and make improvements 

As well as cultural dispositions, problem-sensing and comfort-seeking behaviours also vary according to team, organisational, and institutional capacity to use routine data as the basis of monitoring safety and quality of systems and ability that data as the basis of action. High quality data collection and analysis is needed to support managers and clinicians in their work and to facilitate co-production of health with patients.13 Good measurement provides the information that organisations need to monitor quality and safety and take action where needed,5  and it supports innovation and evaluation of service change, which improves the allocation of finite resources and flow through the system.14   Despite a long history of performance measurement in the NHS,15 and the many data sources available, several challenges remain in monitoring quality and safety in health systems and translating insights from measurement into beneficial use.

One issue is that, despite their ubiquity, the results of many performance management schemes are mixed: they are susceptible to multiple unintended consequences.16 17 18  Even when launched with an explicit emphasis on improvement, they may become regarded by staff at the sharp end more as blame allocation devices than supports for practice.19 For organisations with a tendency towards comfort-seeking, measurement aimed performance management may too easily incentivise exactly the behaviours that contribute to structural secrecy. Serious blind spots can arise when organisations are preoccupied with demonstrating compliance with external expectations. 5 Future performance management efforts should be subject to careful design and evaluation. At a minimum, they should not be regarded as providing a fast-track to improvement, and the threats they may pose to learning should be acknowledged more fully.

A second set of problems in monitoring safety and quality is to some extent located in an institutional context that is still maturing. Despite recent efforts to improve monitoring and measurement 20 that have sought to go beyond crude indicators, and to emphasise the complexities and multidimensionality of safety, much remains to be done to support the NHS in high quality measurement for the multiple purposes for which it is needed. Safety in particular has remained difficult to measure in part because of the absence of a unifying construct and associated valid indicators,21 and because methods of effective, reliable surveillance have been slow to develop.22

Though the number of quality indicators currently available is now enormous, many are inconsistently defined, poorly operationalised, or may not address the priorities of patients or staff.23 24 Data collection endeavours for many monitoring and improvement efforts are often very time-intensive and may involve long delays so that valuable information is received too late. 25  Even for those with a problem-sensing disposition, these are key challenges. For those inclined more towards comfort-seeking, these more technical problems provide too many opportunities to avoid confronting discomfiting knowledge. The future may involve more automated data collection, collation and processing techniques that take advantage of emerging artificial intelligence techniques to learn adaptively and offer real-time predictive analytics. But for the moment, the hype exceeds the reality.26 Much more needs to be done to recognise the value of high quality data collection, analysis and interpretation at every level of the NHS.14

Also important is the recognition that better data by itself is only part of the solution – and indeed the quest for more data can sometimes thwart the goals of improvement. 27 The ability to use data to make sound diagnoses of problems and use those diagnoses as the basis of action requires more attention. For example, organisations and systems need to be able to distinguish between what quality issues can be attributed to the individual performance of healthcare staff, what can be achieved through process improvement, and what represents defects in the design and resourcing of systems.


Differences in the ability to access and use "soft intelligence" 

In addition to being able to gather, analyse and act on metrics, achieving quality and safety requires attention to what is not measured: forms of soft intelligence 28 that may not be easily surfaced and are often highly fugitive in character. Comfort-seeking behaviours may result in organisations neglecting or being highly selective in how they access and use soft intelligence. Problem-sensing behaviours, on the other hand, are characterised by going beyond mandated measures, and using multiple techniques for gaining access to softer forms of intelligence. Such techniques should always include active and participatory forms of listening to patients and staff, but may include a range of other methods, such as informal, perhaps unannounced visits to clinical areas; use "mystery shopper" style data gathering, shadowing of staff, swapping roles for a short period, and use of clinical simulation as a diagnostic method. Culturally, problem-sensing encourages staff to engage in active noticing of where there might be defects, speaking up about them, and ensuring that systems are in place to make improvements.29

While sometimes discomfiting, this less routinely gathered knowledge enables fresh and penetrating insights.  Accessing softer forms of intelligence requires care and sensitivity, including intentional efforts to hear and make sense of the views and concerns of people at the sharp end of care. Often these are nascent, partial and may not be fully formed as safety concerns.30 But they also represent the signals that, experience in healthcare and other industries suggests, can be crucial in identifying and heading off emerging disasters.31 As with the collection of "harder" data, though, it is important not to mistake activity for action. Simply undertaking listening activities or unannounced visits is no substitute for the hard work of analysing and responding to the issues they unearth, and it can even undermine efforts to learn and improve if perceived as an exercise in inspection or a meaningless performative gesture.32 33


Orientations towards openness

The ability to access and make use of intelligence, soft or hard, is not dependent solely on the attitudes and behaviours of leaders and board members. Rather, it is also crucially reliant on the broader culture and systems of the organisation, particularly the extent to which their values, norms, behaviours and institutional capacities are oriented towards openness and learning.34 Past tragedies in healthcare have been attributed in part to a reluctance on the part of a wide range of people to raise concerns about quality and safety. At Mid Staffordshire, for example, many staff at the sharp end of care felt unable to speak up about the issues they saw. It was due to the tenacity of a determined few – including advocacy by patient groups – that the problems were finally recognised. 8

The willingness and ability of those at the sharp end to speak and the willingness and ability of those at the "blunt end" of senior leadership to listen exist in a somewhat reciprocal relationship.35 The efforts of boards and other senior leaders to foster openness, and the extent to which they model good behaviour, make proactive use of data for improvement, and listen and act on soft intelligence, are crucial in setting the organisational tone, for better or worse. But we should not overestimate the power of leaders or of "transformational leadership" in influencing behaviour across complex, disparate and dispersed organisations. The cultures of healthcare organisations are often multiple, and may be patterned along specialty, occupational groupings, professional hierarchies, and service lines.36 Consequently, the most important role of boards and senior leaders to nurturing positive cultures may be in authentically seeking and collating knowledge about variations in performance, behaviour and culture across their organisations, and supporting local leaders, located within units with their own subcultures, in their efforts to improve openness.



The complexity of healthcare organisations makes them highly prone to structural secrecy, and the consequences have been seen, tragically, in multiple disasters of quality and safety of care over the last two decades, in the UK and elsewhere. Organisations vary in their ability to capture intelligence and use it as a basis for improvement, and the contribution of comfort-seeking behaviours to organisational malaise need to be recognised and addressed.

One way of addressing these challenges may in the opportunities presented by the concept of a learning health system:13 one that constantly integrates quality monitoring, quality improvement, research, operations, staff and patient engagement.37 Embracing this concept at organisation and at system level may help to routinise the use of data and insight from a variety of sources in the improvement of care – not least patients and carers themselves. Again, however, capacity to become a learning organisation will vary, and so institutional action at the level of the whole NHS system is also important.

The principles of positive deviance are likely to be especially helpful in the move towards learning at system level. Effective learning in health systems requires both learning from failures 38 and from success. 39 40 Encouraging examples are now beginning to appear, 41 42 suggesting the potential value of the approach – but how to replicate and scale positive deviance remains an important question.39 System-level actors – including national bodies, integrated care systems, regulators and improvement agencies – are likely to have an important contribution to make – not least because coordinating system-level activity in a way that enables individual organisations to realise problem-sensing behaviours. At system level, it is important that  "priority thickets" of goals that may conflict, compete or fail to cohere 5 are avoided if the potential of learning health systems is to be fulfilled, and the safety of patients advanced.



Available in the PDF version


Return to guide.