Kotter argues that 70% of change initiatives fail because organisations do not put in the necessary preparation [Kotter 1995]. His 8-Step Change Model, described in his book Leading Change focuses on “unfreezing” the organisation from its current state, (4 stages), then making the change (3 stages), then refreezing in the new state.

Kotter’s 8-step change model

Step One: Create Urgency

The idea of the change being necessary for the success of the organisation is very powerful, so the advice is to help individuals become aware of an existing problem (eg. polypharmacy in the elderly) and offer them a possible solution (eg. a DSS designed to reduce over prescribing). To generate a conversation about what is happening and what direction the organisation could go, the change manager can create a forum where issues and potential solutions are discussed. Kotter estimates that roughly three quarters of an organisation’s management need to be behind a change for it to be successful, so it is important to prepare well before jumping into the change process. This step creates a 'need' for change, rather than just a 'want' for change. This difference is very important when it comes to the likely support and eventual success of the change.

Step Two: Form a Powerful Coalition

It is very hard for one individual to lead a whole change process, so it is important to build a coalition made up of a range of skills, experience and people who come from different areas. This coalition can help spread messages throughout the organisation, delegate tasks and ensure there is support for the change organisation wide. Team members that are chosen to complement each other and can drive each other to work harder will make life easier and the change process more likely to succeed.

Step Three: Create a Vision for Change

Some change initiatives (such as introducing the RDS as a DSS platform with both local and national content, covering a wide range of clinical and social care areas) can be very complicated. This makes it hard for all levels of staff to understand. So, to generate support from the whole organisation it is essential to create a clear, persuasive vision that encapsulates the overall aim. This vision should be simple, understandable and inspirational to have maximum effect.

Step Four: Communicate the Vision

The vision then needs to be communicated throughout the organisation using the coalition already built up. Bs between them they are likely to have networks in every area of the organisation. It is important to communicate this message clearly and continuously, as aften competing messages arise and are also being spread.

Step Five: Remove Obstacles

The first four steps build the strength of the change initiative, but it is also important to look for barriers that are likely to reduce its chances of success. These can be individuals, traditions, IT or resource barriers, legislation, or other obstacles. Identify these as early as possible and rely on available resources to break them down, without disrupting other areas of activity.

Step Six: Create Short-Term Wins

Since change processes take time to bring benefits, individuals may think their effort has been wasted, causing support to fall. It is therefore important to demonstrate the advantages of the new process by creating short-term wins. Shorter term targets are also useful tools for building motivation and direction. Using these wins to justify investment and effort can help to re-motivate staff to continue backing the change.

Step Seven: Build on the Change

Many change processes fail if staff develop complacency towards the end and projects are not finished. Therefore, Kotter argues it is important to sustain and cement the change for long after it has been accomplished. This can be done by continued quality improvement cycles, repeated goal setting, evaluation and analysing what could have been done better.

Step Eight: Anchor the Changes in Corporate Culture

Changing the habits and processes of employees may not be enough to achieve culture change across the organisation. To have a lasting effect, the changes should become part of the core of the organisation. This can be done by keeping senior stakeholders on board, encouraging new employees to adopt the changes and celebrating individuals who adopt the change.

Critics of Kotter argue that his model may not apply well in healthcare, where there are distinctive professional groups with specific concerns – e.g. about patient safety and quality of care - who do not readily conform to standard organisational management practices.

Adapted from Accipio

The main theory underlying NASSS [Greenhalgh 2017] is that implementation is complex, and this complexity should be embraced rather than being ignored. NASSS comprises six specific domains that need to be considered in any healthcare implementation programme:

• the illness or condition
• the technology
• the value proposition
• the individuals intended to adopt the technology
• the organisational environment
• the wider system

A seventh domain is included that considers how all these domains evolve over time.

Given this focus on complexity, NASSS does not produce specific recommendations about how to solve implementation problems. Instead, it is used to “generate a rich and situated narrative of the multiple influences on a complex project; to identify parts of the project where complexity might be reduced; and to consider how individuals and organisations might be supported to handle the remaining complexities better” [Greenhalgh 2019].

NASSS has not yet been fully validated, but the authors of a comparison of NASSS domains with 7 articles describing implementation frameworks for DSS concluded that our understanding of how to implement AI-driven DSS in healthcare practice is still in its early stages of development [Gama 2022]. New domains that are not covered by NASSS were identified in these articles, including: dependency on data input and existing processes, shared decision-making, the role of human oversight, and the ethics of population impact and inequality.

In strict contrast to the organisational and sociological theories underlying the Kotter and NASSS models, a psychology perspective emphasises that implementation of innovation depends on individual decisions, actions, and establishing new behaviour using behaviour change techniques.

The COM-B model builds on the premise that establishing a new behaviour (for example, using DSS routinely in health and care practice) is generally a result of three factors: capability, opportunity and motivation [Michie 2011]:

• Capability is about a person’s physical and psychological ability to use the new technology, and could be influenced by education, physical limitations (eg. sensory, cognitive or physical disability) or by their perception of what is possible in their working environment, due to the expectation of others such as their peers or patients.
• Opportunity means that the technology should be readily available in the workplace, there should be time for the person to use it and the data needed by the DSS should be available in coded format.
• Motivation is clearly important for behaviour change, as it is always easier to continue with an existing behaviour than adopt a new behaviour that may bring risks or carry a burden of learning. Positive motivation can come through immediate benefits, such as the DSS reminding a user to carry out a preventive care action they would have forgotten, or avoiding a drug interaction they were unaware of. Later benefits could include the learning benefits from using RDS over time or the opportunity to contribute suggestions to RDS authors that are acted on – and potentially peer recognition for this.

The ‘B’ in COM-B stands for behaviour, which is what we are trying to predict or modify.

The following section describes effective behavioural change techniques. Michie (2013) has also produced a short list of 26 behaviour change techniques that are used most frequently and which should satisfy most behaviour change needs.

The COM-B theory and the catalogue of behavioural change techniques can be used by implementers to help individuals to overcome barriers to changing their behaviour and to focus on supporting reluctant and slow adopters.

This organisational model builds on the insights from Everett Roger’s 1950s classic work on the diffusion of innovations [Rogers, 1959]. The journey through implementation and spread to a business as usual (BAU) service can be represented in a graphical format (Figure 5):

The spread and embedding of a new technology or service can falter after the initial uptake by agile and ambitious leaders as the ‘trough of disillusionment’ is populated with a mix of disciplined and reluctant followers, along with conflicted laggards – See Figure 6 below.

Groups of staff can be categorised as:

• Agile Leaders, who are aggressive in their pursuit of advantage through technology. They are the fastest adopters of new technologies and either find a way to get value or abandon and move on to the next new things. In the case of RDS, these may be represented by people who both see the need (and benefit) of RDS and have the resources available at their disposal to promote uptake.
• Ambitious Leaders will move once a clear business use cases is identified. They will be on the forefront of the fall into the trough, often adopting without a full view of implications and disappointment is common. This can be the case for organisations who see the clear need for RDS and have a handful of possible ‘easy wins’ but have not considered the recurring resources required to continue support beyond initial projects.
• Fast Followers come online slightly later because they are waiting to see evidence that it’s worthwhile. They want to see both what has been achieved and the problems that have been encountered.
• Disciplined Followers will start to move when they see that the risks are either under control, or that not moving now brings a higher risk than the risk of moving. They may also be in the situation that they have had to make both strategic and operational changes to be in a position for uptake.
• Reluctant Followers and Conflicted Laggards will only get onboard with the innovation towards the end of the journey.
• Disinterested Laggards are not shown in Figure 6 as they will never adopt RDS. It is important that local teams quickly identify individuals or groups who fall into this category so that efforts are not wasted by trying to include them in adoption work.