Análisis Externo

The participant sample was purposive. A snowball sampling technique was also used to support this. The sampling frame was identified from staff contact lists at each of the hospital sites. This included clinical and non-clinical

administrative staff and hospital managers (see Table 6).

Table 6 A matrix representation of the three vertical levels of participants within each case

Ho sp it al lo ca ti on

Case A Case B Case C

Managerial Managerial Managerial

Non-clinical Non-clinical Non-clinical

Clinical Clinical Clinical

Site lead contacts were the Trauma and Orthopaedic Leads/Managers who represented each hospital Trust and assisted in participant identification. I invited approximately equal numbers of people from each group (clinical, non- clinical, managerial) and from all three organisations to be interviewed. Total final numbers interviewed were 30 clinical staff, 15 non-clinical staff and 11 managers. As planned, I did not interview patient participants. Individuals who participated in key informant interviews (KII) were selected from local and national stakeholder organisations.

3.3.2.4.1 Interviews

Initial invitations were sent to each identified participant. Emails detailed the nature of the study and what would be expected of them. The study participant information sheet was attached to the email (Appendix 4). If interest was indicated, I sent a follow up email outlining how the interviews were structured and the topics which might be discussed, and arranged a time and place to meet the interviewee. In snowball sampling, participants would suggest colleagues with whom I should make contact to try to arrange an interview. Often these

individuals played a key role in delivering hip replacement services but were not easily identifiable through the organisation’s website.

Interviews took place between December 2014 and December 2015. Each interview was digitally recorded and transcribed professionally. A hard copy of the consent form was signed before the planned interview commenced, or verbal consent was obtained for ad hoc interviews. During each interview, I followed my topic guide (Appendix 5). The guide was divided into two parts. The first included structured questions aimed at obtaining general information about the individual’s role, responsibilities, time in post, position within the hospital, academic links and external positions, for example if they sat on a hospital board or national organisational committee. The second included unstructured

questions about the participant’s overall experiences, beliefs and attitudes towards evidence and clinical practice. Each interviewee was immediately given an ID label to maintain confidentiality. Labels were divided into three groups within each site, hence each of the three sites A, B and C contained interviews from ‘C’ (clinical), ‘A’ (non-clinical administrators) and ‘M’ (managers).

Of the 102 formal requests for interview that were sent (purposive and snowball sampled), 50 participants agreed to be interviewed and 48 interviews actually took place. Of the two drop-outs, one failed to attend for the interview and the second was unable to find a suitable time for interview during the three-month period when I was located at the Trust. Of the further 52 formal requests which did not result in interviews, most people did not respond and a small number declined due to lack of time available. These figures do not include the 16 ad hoc interview requests that were made during observations at each site. Often, these were completed there and then, such as after observations of meetings or during coffee breaks.

As planned, there was a relatively balanced number of participants from each of the sites and across the three professional groups. Please see Table 7 for a breakdown of these participants.

Table 7. Participant numbers detailed by site and by professional group

Professional group Site A Site B Site C KII (from local and national stakeholder organisations) Clinical (C) 12 10 8 4 Non-clinical (A) 4 5 6 2 Managers (M) 2 4 5 2 Sub-total 18 19 19 8 Total 64

Interviews ranged from 28 minutes to 1 hour 35 minutes (one hour on average). Most interviews were conducted in offices, personal and shared, located within the hospital Trusts. Some interviews were undertaken in the hospital coffee shop or canteen at the request of the participant (for example if they shared an office and did not want to be overheard or disturb colleagues). All of the eight KII interviews were conducted off NHS Trust sites, such as at the offices of NICE, the CCG offices and at national conferences located across the country.

3.3.2.4.2 Observations and document analysis

At each site, approximately three months of observation took place. This time did not include weekends, time dedicated to other academic commitments or periods of annual leave, hence overall observation time represented just over one full calendar year from 1st _{December 2014 to 11}th _{December 2015.}

Observations ranged from 7.00am to 8.00pm, to include morning and evening meetings that were often scheduled outside traditional core working hours. Site A observation was conducted first, between 1st _{December 2014 and 1}st _February

2015; site B took place between 1st _{March 2015 and 30}th _{June 2015 and site C}

between 1st _{July 2015 and 11}th _{December 2015. Observations in Site C were for}

an extended period of calendar time to compensate for my other academic commitments taking place during this time. However, my actual time dedicated to data collection was similar in all three sites. This was to ensure consistency across data collection.

An outline of the key documents included in the analysis is provided in Table 8. Reading and referring to these documents throughout my time at the

the wider context of the hospitals and to help frame the decisions that were made. For example, board meeting reports stated the financial situation of the hospital and/or where planned cuts would be made, which allowed me to understand the financial pressures facing management teams responsible for purchasing orthopaedic implants.

Table 8. Document type and quantity by case study site

Document type Site A Site B Site C

Clinical pathways 5 3 6

Protocols 17 2 4

Meeting notes 7 5 11

Strategy documents 2 1 0

Quarterly and annual reports 14 18 17

Internal presentations 2 5 2

Sub-total 47 34 40

Total 121