Data Range

The ratings assigned to the improvement work reported by the respondents ranged from 1121 – 5366/ 5554. Each integer within the rating is an independent variable, each relating to a separate dimension of the Healthcare Improvement Typology. The full range of improvement types identified is shown in Figure 12. For the purposes of illustrating the range of ratings, the data are presented here in numerical order. The higher each individual rating, the higher the complexity of that dimension. At the low end of this range, the improvement might, for instance, have involved an attempt to reduce mortality rates for elderly people undergoing a specific major surgical procedure in one department of a hospital. At the high end of this range, improvement work might have involved, for example, setting up from scratch a pan-city, community-

based multi-agency service for previously unidentified patients; or systematically improving nutrition to all patients across several trusts in a health community; or integrating children’s services across all related agencies within a health community.

ID ‘Type’ Rating (LFPI) ID ‘Type’ Rating (LFPI) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1121 1221 1234 1332 2242 2344 2244 2445 2446 3334 3335 3341 3343 3452 3535 3543 4346 4354 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 4444 4455 4533 4542 4542 4551 4554 4554 4664 5344 5354 5355 5366 5421 5444 5542 5554 5554

Figure 12. Data Range of Improvement ‘Types’

It is clear that the quality improvement work undertaken by the sample of THF Award Holders studied, did not encompass the whole spectrum covered by the Healthcare Improvement Typology, which extends up to rating 7 on each dimension. This is not surprising, given that the study population typically comprised clinicians and managers in middle to senior leadership positions in local healthcare organisations. Improvements where ratings extended into the realms of 7 would be likely to be led by

people in regional or national level roles and bodies, rather than within single local healthcare organisations.

On the Level dimension, the sample covered the range from 1-5. This illustrates that improvement work done by THF Award Holders tended to take place within a single organisation, a single health economy, or across a region, but did not tend to stretch to a national level or beyond.

In relation to the Focus dimension, the ratings ranged from 1-6, with just one example of ratings 1 and 6, but the majority of the sample falling in the range 2-5. This is likely to be explained by the fact that the Focus of the improvement work would correspond in broad terms to the level at which the improvement was taking place. Hence, as the level of improvements was limited to rating 5, it is unlikely that the Focus of the changes would fall into the realms of the descriptor ‘unlimited numbers of people’. Such indefinable numbers, with associated Focus ratings of 6 or 7, would relate more obviously to national and international levels of work, which are not represented within our sample.

The highest Focus rating of 6 related to a project to develop a self-management strategy for all people with long term conditions across a whole London borough, where the numbers of people involved and the range of conditions involved were inordinate but not unlimited. The lowest Focus rating of 1 related to a project to reduce the mortality rates of elderly patients admitted for emergency laparectomy. In this case, the Focus is clearly on a very specific condition, for a very specific, defined group of people.

When considering the Process dimension, the range covered by the sample was 2-6. The extremes at each end of the spectrum (ratings 1 and 7) were not relevant to the THF Award Holders studied, in that nobody was involved with improvement work affecting only one or two people (rating 1), nor was anybody involved with changes which were inordinately complex with indefinable numbers of people to influence (rating 7). Only 2 pieces of improvement work were rated 2 for Process and only 2 were rated 6, with the rest falling in the narrower range of 3-5. An example of a 2 rating for Process would be a project which aimed to reduce post-operative DVT (deep vein thrombosis), in which a small number of individuals were the main people to be influenced in changing their practice, to model the change to others. A 6 rating for Process was attributed to a project where stakeholders from a very wide range of agencies were being engaged in establishing a completely innovative mobile service for detecting and treating tuberculosis among the homeless and prison populations across London.

In relation to the Intended Impact dimension, the range of ratings for the sample of THF Award Holders was between 1 and 6. There were no examples of improvement work where the impact appeared to be sustainable indefinitely, which would have warranted a rating of 7. This was possibly partly due to the lack of time passing since the improvement work was done, meaning that this kind of assessment of sustainability was not yet appropriate. However, in those improvement projects with a 6 rating for Intended Impact, there were typically changes which had become largely embedded as a new way of doing things, with direct impact on patient experience and clinical outcomes. A new way of managing the transition of teenagers with diabetes

into adult services, would be such an example. The lowest rating of 1 for Intended Impact tended to be associated with attempted improvements where the goalposts changed mid-project, meaning that the work was never completed or where the post- holder moved jobs or responsibilities changed. Follow-through on the project was therefore rendered unrealistic or impossible.

The relevance of these results relating to improvement ‘type’ is considered in section 6.6.

6.3

Analysis of Interview Data

As detailed in section 5.3, semi-structured interviews were conducted with 36 THF Leadership Award Holders, all of whom were middle- or senior-level clinicians or managers. The aim of these interviews was to elicit descriptions of the improvement work each interviewee was undertaking, to enable the improvement ‘type’ to be determined, and to gather data about the leadership behaviours used to enact the improvements. This section explains in detail how the interview data were analysed. The 36 semi-structured interviews were fully transcribed from audio recordings. This resulted in over 60 hours’ worth of transcribed data. Each transcription containing data relating to the nature and aims of improvement work undertaken by the interviewee and the leadership behaviours reported by the interviewee in effecting that improvement work.

As part of the semi-structured interviews, detailed descriptive accounts were obtained from all respondents about the leadership behaviours they had used to lead the improvements. These accounts consisted of verbal explanations from each participant

about how he or she acted or behaved to lead the work. For all the interviews, the researcher used the full transcripts to code and analyse the data, in order to identify the leadership behaviours reported, and the frequency with which different behaviours were mentioned. As detailed in Section 5.4.2, the framework used for analysing reported leadership behaviours was the Indicators of Quality Leadership (IQL©). A reminder of the structure of this framework is provided in Figure 13 for reference.

Figure 13. Structure of the IQL Framework

There were three different levels at which the data could have been coded. One option was to code data according to the Competency Areas. However, there are only 3 Competency Areas within the IQL and this would not have differentiated the data sufficiently, as the analysis would be at too high a level to be useful. At the other extreme, there was the option of coding the data according to the 120 Behavioural Indicators within the IQL. This approach would have allowed the detail of the data and the subtle differences between the meaning of data to be most accurately captured and reflected, which was deemed important for identifying patterns in the data. However, the main drawback of this approach was that in a 90 minute interview, it

was unlikely that data associated with 120 different Behavioural Indicators would have been reported. This detailed level of analysis therefore risked yielding no data for many of the 120 Behavioural Indicators. Analysis at the mid-level, according to the 24 Key Competencies, offered an approach which differentiated the subtleties in the data sufficiently while also allowing similar data to be grouped into categories which would show any emerging patterns.

Given the options, an approach was adopted which intended to combine rigour with pragmatism. The author decided to code the data according to the most detailed level, using the 120 Behavioural Indicators. This meant that if, at a later stage, it became necessary to interrogate the data at this level of detail in order to identify or explore data patterns, it would not involve a repeat of the coding process. Having coded at this level, the author then aggregated the Behavioural Indicator data into the 24 Key Competences, for reporting purposes.

For each interviewee, the behavioural data were therefore coded according to the 120 IQL Behavioural Indicators. To illustrate the coding process, an example is given in Appendix 14. Within this example, one extract of the verbatim interview data extract is shown in Figure 14:

“They sat and listened to me in great appreciation but it still didn’t happen. Then, when we are going on digging on, you know, which person is actually capable of doing it, we found a lady who was one of our secretaries in the past — for me and my boss — and then we rang

her and said, you know, this is what we want to do and she was like ‘yeah, what’s the problem, I’ll do it.’ ’’

Figure 14. Extract of verbatim interview data (ID 06)

The statement contained in this excerpt was coded as 1dv, which relates to the Behavioural Indicator ‘Engages the support and allegiance of informal networks in formal situations’ (IQL dimension 1dv, comprising competency area 1; key competence d; behavioural indicator v), as marked in red in Figure 15.

COMPETENCY AREA 1: INTERACTS AUTHENTICALLY

---

a) Seeks, understands and values the viewpoint of others

i. Solicits all points of view and uses these perspectives to build consensus ii. Regularly initiates discussion and facilitates open sharing of opinions iii. Harnesses different opinions and capitalises on the benefits of diversity iv. Takes other people’s perceptions seriously and empathises with their feelings

v. Encourages the differing and preferred working styles of individuals

--- ----

b) Understands personal impact and influence on others

i. Anticipates how other parties may react to the content of personal communication ii. Makes convincing and balanced arguments, tailored to others’ needs and expectations iii. Takes account of others’ reactions re: tones of voice, gestures and facial expressions iv. Monitors others’ understanding of what is discussed and corrects misunderstandings

v. Interprets the face-to-face impact of own conduct on others’ behaviour and responses ---

c) Values the skills and expertise of others

i. Capitalises on the range of skills and talents present in the organisation ii. Identifies and nurtures talent to build capacity and capability

iii. Offers support, rewards achievements and celebrates success

iv. Gives clear constructive feedback, timely praise and focused recognition v. Delegates work to provide challenge and opportunities to learn and develop

--- d) Creates networks for the creation and sharing of ideas

i. Identifies and consults with key stakeholders to obtain buy-in for ideas ii. Build and enthuses a wide base of support for innovation and change iii. Develops and sustains a diverse range of internal and external relationships

iv. Invests time to establish, sustain and broaden information and intelligence networks

v. Engages the support and allegiance of informal networks in formal situations

Indicators of Quality Leadership (IQL ©)