Cancer strategy - monitoring and evaluation: participatory systems mapping

Cancer Strategy Systems Map insights

Applying network analysis principles and clustering

PRSM provides some network centrality measures in the data view of the system map, which are described below. These measures can be used to sense check and understand the importance of certain factors. Please note that for this qualitative system map these measures should only be used as a guide and not as a quantitative fact, in particular because the map is based on the subjective views of the 11 workshop participants. See Caveats and limitations for more details. The network centrality measures include betweenness, and degree centrality.

Degree centrality is a measure of the number of links going in and/or out of a factor. If a factor has one link arriving, and three links leaving then it would have an in-degree of 1, an out-degree of 3 and a total degree of 4 (Figure 1). This highlights factors that are well connected. If a factor has a lot of links going in then it could hold a lot of information about the wider system, or be more susceptible to changes in different parts of the system. If a factor has a high out-degree then it can quickly connect with a wider network and potentially influence the system more rapidly.

Example network/system map to explain the network centrality measures.

Betweenness counts the number of times that a factor appears on the shortest path between other factors (see Figure 1). Factors with a higher betweenness can be seen as bridges between factors in the system, these factors will have more influence on the flow of change around the system.

For example, the network centrality measures highlight the cross-cutting nature of the policy ambition to tackle inequalities as it has the highest betweenness value of all the Cancer Strategy ambitions. The betweenness value compared to another ambition and a focal factor in the map can be seen in Table 2. The high betweenness value implies that “tackling inequalities” appears on many paths between other factors, and if this ambition was successful this would improve flow of change across the system. This gives some assurance that the system map is a useful representation of the system.

The network centrality measures have been used to identify the illustrative factors which may have high leverage in the system i.e. strong influence. For these examples the chosen focus is on the betweenness measure. The examples chosen include the three factors with the highest betweenness – these are listed in Table 2.

The role of the illustrative factors (Table 2) in the system was explored in more detail. A summary of the insights that were explored in the system map can be found in the following section: Insights from the visualisation tool.

Clustering to sense check the map and identify limitations

One of the approaches used to quality assure and sense check the system map was to cluster the factors into different categories and examine the links between the clusters. This was done within the PRSM software.

Factors were clustered into:

• Infrastructure
• Healthcare workforce
• Financial
• Digital and data
• Strategic
• Patient focused
• System goals
• Policy ambitions

This map exploration highlighted that the policy ambitions are most connected with the healthcare workforce cluster (38 links) and the patient cluster (27 links). The system goal cluster is most connected with the healthcare workforce cluster (8 links). This clustering generally gives an overview of where participants were focused when linking factors. There are understandably more links to healthcare workforce related factors than patient related factors. This is a limitation that arises from not involving patients directly – while the focus of the workshops centred on patient experiences, views of the system necessarily emerged from participants’ own experiences as members of the workforce.

System goals and focal factors

The system goals were chosen by participants during the qualitative analysis stage of the PSM process. The outcome of this iterative discussion was three system goals: psychological well-being, better cancer survival outcomes, and minimal time to diagnosis. These system goals were considered appropriate as they align well with the strategic aim of the Cancer Strategy to “improve cancer survival and provide excellent, equitably accessible, care”, nevertheless they were decided by only a small number of participants. The system goals were mostly used to sense check the cause and effect in the system map for these participants; other factors could be classed as system goals by a group with different perspectives.

Focal factors were also identified by all participants throughout the mapping process. These are the actions considered by participants as essential for achieving the system goals. The final map had four focal factors: coordinated care, patient feeling supported, access to care and treatment, and primary care capability.

System goals and focal factors may be useful to consider when developing indicators to measure the impact of the Cancer Strategy. Consideration of system goals and focal factors are used in this report as a way of assessing the impacts in the illustrative examples.

Insights from the visualisation tool

The visualisation tool was used to explore the influential factors identified with the network centrality measures (see Table 2). The visualisation tool functionality to increase a factor was employed and the map visually explored to assess the impact. In the visualisation tool change is propagated through the system by increasing and decreasing the size of the factors. For example, if factor X was increased and as a result one of the system goal factors significantly increased in size – this would be described as factor X having a strong positive impact on the system goal.

When looking at impacts across the whole system it is important to remember that the links are all qualitative. So, while there can be strong positive impacts on the participant-identified system goals, the possibility cannot be ruled out that in reality some of these positive impacts could be reduced, or fully counteracted, due to other interactions in the system. Therefore the conclusions here are not definitive.

The visualisation tool also has a built-in theory of change functionality which helps to identify potential levers for the factors. A theory of change considers the key steps that are anticipated between an intervention and its outcome. A factor is described as a “lever” if it has a strong influence on another factor – this is usually because there are multiple paths connecting the two factors.

Example 1: Tackling inequalities

In the system map increasing the “Tackling inequalities” factor can be interpreted as making progress with this policy ambition to tackle inequalities. As previously mentioned, this factor has a high betweenness value which suggests that it acts as a bridge for many paths across the system. This bridge concept can be explored by looking at the short paths that include this factor. For example, the system map has a factor for “understanding how to cope with geographical challenges”. Participants thought that an increase in this understanding would positively benefit the ambition to tackle inequalities, which would then be the policy lever to improve travel infrastructure and accessibility.

When this factor is increased in the visualisation tool there are a wide range of positive outcomes, reflecting participants’ view that tackling inequalities is a good thing to do, and necessary for improving outcomes for all people affected by cancer. Even when considering the qualitative, subjective nature of the system map, this policy ambition does appear to influence positive change across the system. If progress is made on this ambition this is likely to amplify positive change resulting from other Cancer Strategy ambitions.

Another way to draw insights from the system map is to look at the neighbouring factors. Factors that are one link downstream from “Tackling inequalities” demonstrate participants’ opinions of immediate outcomes of tackling inequalities but also indirectly suggest some of the ways in which participants thought that inequalities could be tackled. These include improvements to travel infrastructure and accessibility, more public communications around cancer symptoms, clearly communicated information about the patient journey, and improved support services such as prehabilitation, rehabilitation, mental health and social care. The immediate positive impacts on patients in the system can be summarised as an improved patient experience, with patients feeling supported and heard.

Looking for neighbouring factors upstream allows us to explore potential system impacts on the ambition. The integrated theory of change function of the visualisation tool suggests that the factor in the map with the strongest influence on “Tackling inequalities” is the “person centred care” ambition of the Cancer Strategy (Figure 2A). The suggested synergy between these two policy ambitions relies on the participants’ perspective that the person centred care ambition would ensure that the patient voice is present and would improve the availability of single points of contact (SPOC). Participants generally held the view that SPOCs were a good way of supporting people affected by cancer and a core component of person centred care. This in turn translated into the view that widespread, consistent availability of SPOCs would reduce inequalities. The system map highlights this view as SPOC availability is found on multiple theory of change paths for tackling inequalities (Figure 2B).

Figure 2

Example theory of change paths from the visualisation tool. Panel A illustrates the influence that “Person centred care” has on “Tackling inequalities” via two paths. Panel B highlights some of the paths influenced by single point of contact (SPOC)/cancer support worker (CSW) availability.

Panel A.

On the left hand side there is a yellow circle labelled “person centred care”. Two arrows, labelled “S” for same, leave this circle going to the right. One goes to a blue circle labelled “patient voice is present”, the other goes to a blue circle labelled “SPOC/CSW availability”. Both blue circles have arrows, labelled “S”, leaving them and arriving together at a yellow circle, to the right, labelled “tackling inequalities”.

Panel B.

On the left hand side there is a blue circle labelled “SPOC/CSW availability”. Two arrows, labelled “S” for same, leave this circle going to the right. One goes to a yellow circle labelled “person centred care”, the other goes to a yellow circle labelled “tackling inequalities. Both yellow circles have arrows, labelled “S”, leaving them and arriving together at a blue circle, to the right, labelled “improved travel accessibility”.

A second factor identified as a potential lever for “Tackling inequalities” is the alignment of priorities of NHS Education for Scotland, the Centre for Sustainable Delivery, and other professional organisations. The participant perspective was that a disjointed approach and conflicting priorities reduced the impact of current work to tackle inequalities.

Example 2: Person centred care for all

In the system map increasing the “Person centred care” factor can be interpreted as making progress with this policy ambition to achieve person centred care for all. This factor has a high betweenness value and a high out-degree reflecting participants’ “enthusiasm” for this ambition and the shared perspective that this will be integral to improving care for people affected by cancer. This is further emphasised by the wide range of positive outcomes when this factor is increased in the visualisation tool (Figure 3).

As for Example 1, the qualitative, subjective nature of the map has to be continually considered, however this policy ambition does appear to influence positive change across the system – in particular by appearing to exert a varying degree of positive influence on all of the other Cancer Strategy ambitions. Of note, despite being a patient-focused ambition on the surface, the participants held the view that this could have positive impacts on staff morale and workforce capacity. The system map suggests that staff morale would improve when staff feel like they have more time as patients become empowered to have more responsibility over their own care.

Figure 3

Screenshots of the system map showing the impact when the “Person centred care” factor is increased. Panel A contains an overview of the system map when there are no changes. The key point is that all the factors (represented by circles) are the same size. Panel B shows the entire system map when the factor is increased, where some factors grow in size and some become smaller. The factors that increase most in size are “Psychological wellbeing”, “Patient feeling supported” and “Better cancer survival outcomes”.

Panel A.

A zoomed out view of the entire system map using the visualisation tool. It is not possible to see much detail. There are lots of circles representing factors, and lines going from factor to factor. All of the circles are the same size.

There is an arrow going from panel A to panel B labelled “increase person centred care”.

Panel B.

A zoomed out view of the system when the “person centred care” factor has been increased in the visualisation tool. Increasing the “person centred care” factor causes change to flow through the system based on the causal relationships. Some factors grow in size while others shrink. The circles representing factors are now a range of different sizes. Some factors are greyed out because they are not impacted by the change.

Downstream neighbouring factors from “Person centred care” highlight ways in which participants felt this ambition could be achieved in practice. For example, by ensuring that patients have sufficient information such as guidance for their particular pathway, a symptom booklet, or financial planning support. Some of the other factors cover more operational considerations such as the presence of multidisciplinary teams, single points of contact, and improved support services such as prehabilitation, rehabilitation, mental health and social care. The synergy between tackling inequalities and person centred care is mentioned in Example 1 and is further emphasised by both of these ambitions being identified as key leverage points for the factor “Patient feeling supported”. In the system map patients feeling supported is directly linked to better cancer survival outcomes. It is not clear in the system map why participants felt that there was a direct relationship between patients feeling supported and better cancer survival outcomes – this could be a causal relationship to explore further in future work.

The potential system impacts on the person centred care ambition were also explored. The integrated theory of change function of the visualisation tool suggests that the lever variables for “Person centred care” are generally workforce factors such as ensuring that staff are supported and provided with training to deliver person centred care. Similarly to Example 1, a common factor appearing on the theories of change is the availability of single points of contact (SPOC). Participants generally held the view that SPOCs were a good way of supporting people affected by cancer and a core component of person centred care. Participants also thought that patients had a role to play in person centred care, particularly in the form of self-monitoring which could be enabled by having the right IT infrastructure in place and building patient knowledge via comprehensive symptom booklets. To summarise, participants generally held the view that finding ways to support the people on both sides of the healthcare relationship would result in improved person centred care.

Example 3: Coordinated care

In the system map increasing the “Coordinated care” focal factor can be interpreted as increasing the amount and effectiveness of coordinated care. Participants described coordinated care as “fast access between people” with streamlined requests, while also covering money and data flows between health boards. “Coordinated care” is the factor in the map with the highest betweenness suggesting that this factor will likely play an important role in the success of the Cancer Strategy.

Exploring the role of this factor in more detail, the visualisation tool was used to identify that “Coordinated care” is connected to “Earlier and faster diagnosis” by multiple paths and positive reinforcing loops. This reflects participants’ strongly held view that improved coordination of care is necessary to streamline the patient journey and result in earlier and faster diagnosis. This point is further emphasised by looking at the downstream neighbouring factors of coordinated care. Participants thought that coordinated care would directly improve access to care, treatment and diagnosis; enable staff to deliver the right care in the right place; and make it easier to share clear information on the patient journey because the journey itself is clearer.

Changing perspective to consider system impacts on “Coordinated care”, the visualisation tool was used to identify potential levers of influence. Two factors that may have a high influence on coordinated care are “Ability to talk to relevant colleagues” and “Effective IT infrastructure”, as demonstrated in Figure 4. These factors are likely a reflection of the particular frustrations that the participants experience when trying to provide coordinated care. The system map highlights that the ability to talk to relevant colleagues can improve coordinated care by building good relationships, saving staff time (participants shared experiences of having to contact multiple people before finding the right person), by improving the care from multidisciplinary teams (MDT) and simply by being a foundational aspect of coordinated care. Having more effective IT infrastructure also appears to have a strong, positive influence on coordinated care by enabling improved access to primary care (e.g. by video or e-consultations), by improving the quality of data so that staff do not have to repeat tests, for example, and by saving staff time (participants shared various frustrations with IT being slow or cumbersome to use).

Figure 4

A screenshot of the visualisation tool output when theories of change for “Coordinated care” has been selected. It highlights that “Ability to talk to relevant colleagues” and “Effective IT infrastructure” have a high degree of influence on “Coordinated care” because these factors are linked by multiple paths made up of positive links (indicated by the letter “S” which stands for “same”).

Six out of the 11 Cancer Strategy ambitions were identified by the tool as potential levers of influence for “Coordinated care”: person centred care; sustainable and skilled workforce; safe, realistic and effective treatment; cancer information and intelligence lead services; earlier and faster diagnosis; and preventing more cancers. This may reflect a participant view that coordinated care is closely linked to the Cancer Strategy ambitions.

The strongest lever ambitions for coordinated care have two linking paths. The first is the ambition for safe, realistic and effective treatment which could act as a positive lever for coordinated care by supporting successful hub and spoke models and ensuring that MDTs are competent and confident. The second lever ambition is person centred care which participants thought would enable coordinated care by reducing geographical challenges and also promoting and supporting MDTs.