Over a prolonged period, scholars have attempted to define and measure quality in physiotherapy clinical education. This has been a complicated process because of the complex, multidimensional nature of clinical education and the paucity of research (Higgs 1993; Jette et al. 2014; McCallum et al. 2013; Stachura, Garven & Reed 2000; Strohschein, Hagler & May 2002). A consensus has yet to be reached on what may be alluded to as the quality of clinical physiotherapy education. The quality of clinical education appears to be affected by the processes involved in clinical education, namely models of teaching and learning that lead to the development of a competent graduate, the enhancement of technical and non-technical skills, the attitudes and behaviour of students and the roles and relationships between students and clinical educators (Dunfee 2008; Jette et al. 2014; McCallum et al. 2013; Moghadam, Kashfe & Abdi 2017; Stachura et al. 2000; Strohschein et al. 2002).

Because of the challenge experienced in measuring quality in clinical education, Strohschein et al. (2002) advocated for clear clinical education goals and objectives, as had Higgs (1993) and Stachura et al. (2000) in earlier papers. These clear clinical education goals and objectives should facilitate appropriate processes and models of clinical education. They recommended conducting more research to identify prominent features of current clinical education processes using qualitative and quantitative research methods. They suggested that a tool(s) must be developed to measure the important aspects of physiotherapy clinical education.

The quality of clinical education can only be defined and measured if standardisation in structure, processes and outcomes exists in physiotherapy programmes; structure – standardisation of the number and duration of clinical education experiences; standardised forms for communication of important information with clinical sites or standardised supervisory models that allow for student autonomy over time and the use of a standardised national student evaluation form and standardised tools to assess stakeholder satisfaction with clinical education (Jette et al. 2014).

It seems that a lack of standardisation in the structure and processes of clinical education, more so than outcomes, impacts the ability to define quality and evaluate physiotherapy clinical education. Furthermore, the need for more consensus regarding what to measure and how to measure components of structure, process and outcome and the lack of a reliable measuring tool(s) amplifies this problem. Holistic evaluation of clinical education must examine the structure, process and outcome using a valid and reliable tool (standardised measure). Thus, our study aimed to identify items that could be included in a tool to evaluate clinical physiotherapy education in South Africa.

This qualitative study involving focus group discussions (FGDs) was undertaken nationwide in South Africa. Permission to undertake the study at the universities where physiotherapists are trained was obtained from various clinical sites. All participants signed informed consent before being included in our study.

Purposive sampling of participants included academics from seven of the eight South African universities involved in educating physiotherapy students, physiotherapy clinicians, physiotherapy clinical managers, physiotherapy clinical educators and community service physiotherapists from various clinical sites. The FGDs included either mixed groups of academics and clinicians or separate academic and clinician groups, depending on the availability of participants.

Before conducting the FGDs, the first author undertook a pilot study of two FGDs to ensure that she ran the FGDs in a manner that would provide participants discussed issues in a safe environment and that she asked appropriate open-ended questions like ‘tell me about your clinical education experience’ with suitable prompts as needed. This pilot study was conducted with colleagues in the Physiotherapy Department of University of the Witwatersrand and a senior academic observer who discussed the FGD procedure with the first author after each pilot FGD. She then conducted all the FGDs countrywide.

The FGDs included 6–8 participants, and each discussion lasted approximately 2 h with appropriate comfort breaks and refreshments. The FGDs were audio-recorded, and the collected demographic information was kept separately from the audio recordings. The first author kept field notes during the FGDs.

The data were analysed using Tesch’s (1992) method of data analysis for qualitative studies (Vuso & James 2017). The audio recordings were transcribed verbatim by a professional transcriber. The first author then cross-checked the transcriptions with the audio recordings to minimise any errors that may have occurred during the transcription. The transcripts were then sent to the FGD participants for verification and were subsequently confirmed as an accurate account of the discussions.

An inductive thematic content analysis using MaxQda version 2018.2 was undertaken; the data were coded, and similar foci were categorised and subcategorised. This process led to the identification of themes (Greenwood et al. 2017; Vuso & James 2017). Two authors (A.S. and D.M.) and an independent researcher checked the thematic content analysis on a selection of transcripts. There was a high degree of agreement, and where there were minor differences, these were discussed and resolved leading to an agreed list of codes, categories and subcategories. They then identified themes from the codes and categories discussed them with the first author and reached a consensus on any minor differences.

Thus, data triangulation was assured through member checks, field note comparison, observer reflection and the above data verification. This triangulation aimed to ‘overcome the inherent biases in a single investigator, single theory or single method approach’ (Halcomb & Andrew 2005).

To ensure trustworthiness that is the credibility, dependability, confirmability and transferability of the data, the following steps were undertaken.

Prolonged interaction with participants during the FGDs; audio recordings and transcriptions; conferring with the observer of the pilot FGDs; using an interview guide during the FGDs; checking the transcripts and comparing them to the field notes taken during the FGDs; member checking and checking of codes, categories and themes by three researchers. Transferability was assured by purposively choosing the sample, namely the academics in the training universities, clinicians and clinical educators involved in the clinical training of undergraduate physiotherapy students (Connelly 2016; Krefting 1991).

The first author considered her background and perceptions in collecting the data during a reflective process (Krefting 1991). She currently coordinates the clinical training programme at the University of the Witwatersrand. Before this, she was a physiotherapy clinician.

Fourteen FGDs were held countrywide, including seven of the eight universities and three clinical physiotherapy departments. Eighty-one participants were included with approximately eight participants per group, either in groups including academics and clinicians or just academics and clinicians. Data saturation was obtained after eight FGDs. However, the first author continued collecting data for 14 FGDs given that the participants had been invited to participate in the study 6–12 months before the data collection to accommodate academic timetables and clinician responsibilities (see Table 1).

Governance, structure and clinical experience emerged as themes from the inductive thematic content analysis. These themes are macro-, meso- and microstructures (see Figure 1, Figure 2 and Figure 3). The macro level includes the policies, procedures and systems; the meso level is the organisational structure of the academic department and clinical site and the micro level relates to the student’s clinical learning experience. The items of the tool were generated from the resultant codes of each section. A preliminary tool of 131 questions and three sections was developed: governance contained 17 questions; academic structure contained 55 questions and operational structure contained 59 questions.

The 14 FGDs yielded three themes related to clinical physiotherapy education namely governance (macro), structure (meso) and clinical experience (micro). The complexity and diversity of physiotherapy clinical education both in South Africa and in the global community can be seen in these themes developed from the collected data (Higgs 1993; Jette et al. 2014; McCallum et al. 2013; Patton, Higgs & Smith 2018; Stachura et al. 2000; Strohschein et al. 2002). To identify items to be included in a tool to evaluate physiotherapy clinical education, all the categories and subcategories making up these broad themes need to be considered to ensure that a tool developed to evaluate physiotherapy clinical education is comprehensive and sufficiently detailed.

Items were generated from codes under each theme, resulting in the preliminary tool. The themes and the extensive number of items generated (131 items) reflect the complexity and diversity of physiotherapy clinical education currently navigated. This first step in developing a tool that pulls together and encapsulates the breadth and depth of physiotherapy clinical education challenged several scholars. The items in these themes can now be considered for the next stage of the development of a tool to evaluate clinical physiotherapy education.