2 METHODS
We used the same methodological approach for this study as in a recently published study,9 because both studies arose from one research project based on the same methodological approach.
We used semi-structured individual interviews and focus groups to collect qualitative data from participants in Lincolnshire in the East Midlands region of England. The School of Health and Social Care Ethics Committee at the University of Lincoln granted ethics approval for the study.
We recruited a convenience sample of service users (adults without active cancer) and primary care practitioners (GPs and nurses) who agreed to participate in the study. Cancer patients were not recruited because they were not likely to use the tool for a cancer which was already known, and we felt it was not worth the stress, even if some cancer patients had some information to offer.
The interview schedule seeking barriers and facilitators to use of the tool was informed by an implementation theoretical framework, the Consolidated Framework for Implementation Research [CFIR],10 which seeks to understand human and other factors involved in the deployment of innovations such as cancer risk assessment tools for symptomatic individuals. The data analysis and interpretation were also informed by relevant constructs within the CFIR10: relative advantage (for the facilitators); patient needs and resources, compatibility, knowledge and beliefs of individuals involved, and reflecting and monitoring the implementation process (for the barriers). This theoretical framework was selected because some of its constructs (those relating to barriers), helped to explain how the different barriers may make the implementation of the tool difficult, while other constructs of the framework (those relating to facilitators) also helped to explain how the different facilitators identified by participants may aid the implementation of the tool in patient consultations.
Some service users were recruited using flyers in local public places (e.g. the public library, advertisements on notice boards) while others were recruited through members of the patient and public involvement group who helped to identify and invite potential participants. Service users were offered individual interviews because these were considered more feasible for them than focus groups. Service users who were willing to participate in the interviews contacted the researcher (JNA) for more information and an appointment for individual face-to-face interviews. These interviews were subsequently conducted in the preferred place of interviewees – in their own homes or at the university.
Following invitation letters to practitioners through their general practices, interested practitioners contacted the researcher for more information and an appointment to meet the researcher (JNA) for either individual interviews or focus groups. Some practitioners were unable to find time to be interviewed individually, and were focus groups instead, enriching the data through participant interactions. Other willing practitioners were interviewed individually.
Before asking for the views of participants on the barriers and facilitators to the implementation of the tool in general practice consultations, a vignette of the QCancer tool was shown, explained, and demonstrated to participants where there was Internet access.
Prior to commencing the individual interviews and focus groups, participants gave a written consent by signing a consent form, and they were assured of their freedom to discontinue with at any point. All participants gave permission for audio-recording of the interviews and focus groups, and notes were taken to complement the audio recorded data.
After transcribing the data verbatim, the Framework approach11 was used to analyse which was facilitated by NVivo version 10. A priori codes, which informed the interview guide, formed an initial coding framework, and further inductive codes were identified as analysis of the interview data progressed. Two investigators (JA, ANS) read the transcripts thoroughly and derived an initial coding framework which was discussed and agreed by the research team. Through further interpretation and discussion, the initial themes were developed iteratively into a smaller number of overarching themes. There was no need for further collection of data from bother service user and practitioner participants as the data analysis achieved saturation, which meant that no new codes or themes were generated.12 The service user and practitioner data were analysed separately and then compared for similarities and differences.