A framework for conducting policy-relevant primary care research: a COVID-19 case study in Australia

Keywords: action research, collaborative research, COVID-19, design thinking, general practice, pandemic, primary care, research methods.

The onset of the coronavirus disease 2019 (COVID-19) pandemic, caused by SARS-CoV-2, triggered a global response across all research domains, most prominently in the fields of health and medical research. Within Australia, COVID-19 pandemic research has been undertaken within all states and territories across a broad range of healthcare and medical domains (Research Australia 2020), including primary care (Halcomb et al. 2022).

The Australian primary care landscape experienced changes during the COVID-19 pandemic period, commencing with the Australian Government’s ‘$2.4 Billion Health Plan to Fight COVID-19’ announced on 11 March 2020, which initiated the introduction of telehealth services, home medicine services, electronic prescribing and general practitioner (GP)-led respiratory clinics (Australian Government Department of Health 2020; Desborough et al. 2020). As the COVID-19 pandemic progressed, additional changes were introduced to address emerging issues, and initiatives were modified, extended or made permanent (Australian Government Department of Health 2021, 2022). The introduction of pandemic response measures directly impacted the delivery of primary care services. Understanding how these changes affected practice activity is important for gauging both their effectiveness and their implications for future service planning.

For primary care practitioners and healthcare stakeholders seeking to understand the impact of the COVID-19 pandemic, or any policy/procedural change on their practice activity and care delivery, there are a range of quantitative and qualitative research approaches available. Herein, we detail a research project approach that was undertaken during the first 2 years of the COVID-19 pandemic to understand the pandemic’s impact on general practice activity as part of a project titled ‘COVID-19 – utilising near real-time electronic General Practice data to establish effective care and best-practice policy’ (Georgiou et al. 2021; Australian Institute of Health Innovation 2022). The collaborative approach, research outputs and lessons learned during the research process are transferrable to any primary care service and can assist in planning future research engagements between primary care practitioners and researchers.

Research methodologies that promote collaborative undertakings between researchers and healthcare stakeholders and/or aim to bridge the gap between research and practice (Ovretveit et al. 2014), are increasingly being used in health care. There are various approaches available to collaborative research teams (Ovretveit et al. 2014). Herein, we provide an overview of the two approaches utilised in the current project, namely ‘design thinking’ (DT) and ‘action research’ (AR).

DT, which has been used in healthcare settings (Altman et al. 2018), has been described as ‘…a human-centred, iterative, team-based mindset that leads to breakthrough innovations’ (Hasso-Plattner-Institut 2022) using a phased approach. The understanding of DT as a concept and its practical application can be context specific (Carlgren et al. 2016), hence, the approach used in our project (Georgiou et al. 2021) was based on a framework proposed for use in healthcare settings. The framework (Roberts et al. 2016) details three methods most suited to use in health care, which emphasise empathy (to gain an understanding of end-user needs), collaboration, and rapid prototyping (of multiple potential solutions). As an iterative process, the literature suggests DT is best introduced from the outset of an initiative (Roberts et al. 2016), and following initial engagement, then proceeds through phases of synthesis, review, iterative prototyping and testing until an implementable solution is achieved (Roberts et al. 2016).

AR includes participatory involvement of key stakeholders as co-researchers (Given 2008) throughout all iterations of the research cycle, leading to the production of knowledge (Given 2008; Vallenga et al. 2009), which informs action (Dick 2014). The AR cycle involves iterations of phases during which researchers and stakeholders participate in planning the research, acting on the knowledge generated, observing and reflecting (Altrichter et al. 2002; Vallenga et al. 2009). AR has been used in primary care settings (Hampshire 2000). It has been advocated as an approach to reducing the ‘research-practice’ gap as the process incorporates practitioners’ experience and knowledge to identify research problems, thereby producing knowledge that is directly relevant (Meyer 2000; Vallenga et al. 2009).

The ‘COVID-19 – utilising near real-time electronic General Practice data to establish effective care and best-practice policy’ project commenced in July 2020, aiming to ‘examine the feasibility of using near real-time electronic general practice data to promote effective care and best-practice policy’ (Georgiou et al. 2021, p. 2; Australian Institute of Health Innovation 2022). The project was approved and funded by the Digital Health Cooperative Research Centre (DHCRC) as a collaborative undertaking between Macquarie University, Outcome Health (the Primary Health Network (PHN) contracted data custodians/digital developers), the Royal College of Pathologists of Australasia Quality Assurance Programs (RCPAQAP), and PHNs in Eastern Melbourne, South Eastern Melbourne and Gippsland, with data also provided by two NSW PHN’s (South Western Sydney, Central and Eastern Sydney). Detailed aims, methods and ethics have been published in a protocol paper (Georgiou et al. 2021). Knowledge and outcomes achieved throughout the project were enabled through the combination of the key elements within the project framework; namely, collaboration, iterative cycles of AR, the availability of general practice data via the POpulation Level Analysis and Reporting (POLAR) platform (Pearce et al. 2019) and an overarching project governance. The project’s framework is summarised in Fig. 1.

Fig. 1.  The methodological framework used for the project. Collaboration was the core element, which served as the foundation for the initial design thinking approach, which was subsequently supplanted by iterative cycles of action research. The project was overseen by an engaged governance structure, comprising key representatives from partner organisations. The knowledge and outcomes element of the framework is presented in greater detail in Fig. 2.

A two-tiered governance structure was established comprising a Project Control Group (PCG) responsible for overseeing the project and its deliverables, and a Project Management Team (PMT) responsible for the everyday operational aspects of the project. Key decision-makers from each organisation were invited to participate in the PCG, which met twice per year to review project progress and quarterly reports. The PMT met monthly throughout the project duration to provide updates and review priorities and progress. PMT meetings provided a forum for key decision-makers to lead and input into strategic discussions and feedback on research findings as the project progressed. The project governance structure served as the conduit for stakeholder engagement, knowledge dissemination and served as a regular forum for stakeholder organisations to contribute to the project direction.

DT was instituted from the commencement of the project. The initial DT session introduced the research process through a matrix detailing the phases, infrastructure, process, and outcomes anticipated from each phase. DT sessions resulted in the identification of the initial priority areas for research. However, after the completion of three DT sessions, the decision was made to redefine the research approach to place greater emphasis on timely information generation. The research methodology was subsequently amended, and AR was adopted with the aim of generating rapid knowledge-orientated outcomes.

The AR cycle involved four iterative phases commencing with an exploration of the initial research priority areas identified during the DT sessions. AR sessions were attended by the research team, PHN representatives (including representatives/GPs with a particular interest in the topics being addressed), the DHCRC and Outcome Health. During the second phase of AR, the research team explored the research priority areas and performed quantitative analyses of POLAR data to generate preliminary findings for further discussion. During the third phase, the research team presented the preliminary research findings for open discussion and reflection on the results, their context, and relevance. This phase added contextual understanding to the results and provided clarification for final analyses, while also identifying additional areas for further research. The contextual knowledge and insights gained were subsequently incorporated into the fourth AR phase, where analyses were finalised and published in the form of succinct evidence ‘Snapshots’. Ideas for further research were then explored in the next AR cycle and, as the project progressed, more in-depth analyses were published in relevant journals. Through cyclic iterations of AR, seven priority research areas were investigated, resulting in more than 20 quantitative research undertakings. In addition to quantitative analyses, the research team also collected focus group data in the latter stages of the project (reflective phase of the AR cycles), for qualitative analysis. Qualitative data provided a more profound understanding of stakeholder perspectives/experiences related to focus group topics.

The core component of the project framework was collaboration. Throughout the DT and AR cycles, participation from PHN representatives and GPs, Outcome Health (data custodians) and the research team were integral to achieving knowledge outcomes. PHN/GP representatives set the research agenda by identifying the areas of general practice most impacted by COVID-19 pandemic circumstances and provided context and understanding to the data analyses. The data custodians, with permission from each participating PHNs (PHNs are the owners of the de-identified general practice POLAR data used within this project), provided the research team with the data necessary for quantitative analysis. The research team analysed the quantitative data and planned focus groups to collect qualitative data for analysis. While producing the research outputs, the RCPAQAP provided expert knowledge for pathology-related analyses.

Data analyses were performed for multiple aspects of each of the seven research priority areas (Fig. 2), including analyses in progress at the time of writing. Three tiers of knowledge eventuated to meet short-term, long-term, and future information needs, including: (1) succinct ‘Snapshots’ of rapid evidence for immediate priority areas; (2) in-depth analysis of dimensions of the priority areas, with findings published in relevant journals; and (3) the translation of evidence to inform the development of a suite of impact measures for use beyond the COVID-19 pandemic and to inform ongoing development of the data custodian’s COVID-19 dashboard.

Fig. 2.  The knowledge and outcomes resulting from the action research cycles. Seven priority research areas were identified by stakeholders, resulting in knowledge generation across more than 20 quantitative and qualitative research undertakings. The knowledge generated was translated into a suite of impact measures for determining the impact of the COVID-19 pandemic on general practice activity and served to inform the ongoing development of a COVID-19 visual dashboard by Outcome Health (data custodians), using wisdom based on clinical judgement and intuition. The outcomes of the project were disseminated across practices and to practitioners within primary health networks, as well as being made available to broader audiences.

The three tiers of knowledge generated provided evidence of the impact of the COVID-19 pandemic on general practice service provision including, for example, face-to-face versus telehealth consultations for diagnostic testing (Hardie et al. 2022) and medication prescribing (Wabe et al. 2022), decreases in GP visits to residential aged care facilities (Dai et al. 2021), and impacts on breast and cervical cancer screening activities (Imai et al. 2021). One priority area, telehealth, permeated across many ensuing priority areas, emphasising the need for evidence and understanding of the impact of this policy change on a wide range of aspects of general practice. The knowledge produced also served as the basis for the development of a suite of evidence-based measures that could be used to monitor the ongoing impact of the COVID-19 pandemic on general practice activity. A suite of 10 impact measures, documented in detail elsewhere (Thomas et al. 2022), were developed for the purpose of measuring the ongoing impact on, for example, medication prescribing volumes, non-acute respiratory illness testing volumes, HbA1c testing volumes, and HbA1c testing frequency and levels in diabetes patients, representing a translational output that could be used into the future.

Fig. 2 also shows the approaches employed to disseminate research findings, with PHNs playing a central role in communicating to general practice networks/GPs. Research evidence was communicated via newsletters, bulletins, email, websites, and included in annual reports or PHN meetings, with particular emphasis on better informing their commissioning programs and activities designed to improve patient healthcare outcomes. General practice/GPs and PHNs were the primary recipients of all research outputs (Fig. 2), with traditional research dissemination (journal publications/conference presentations) and web publications available to the broader community.

Outcome Health received ethical approval from the Royal Australian College of General Practitioners National Research and Evaluation Ethics Committee (NREEC) 17-008 for the POLAR GP data warehouse. The Macquarie University Human Research Ethics Committee (HREC) Medical Sciences Committee granted approval to the project (Reference no. 5202067517176).

Adaptability was a key learning during the project, including acknowledging the initial DT approach was not optimal for achieving short-term information needs, hence the decision to adopt AR as the subsequent approach. Adaptability was also required in responding to the need for different tiers of evidence to meet short- and longer-term knowledge needs. Throughout the project, many of the cited benefits of AR (Meyer 2000; Given 2008; Vallenga et al. 2009) were realised, including the participatory role of key stakeholders as co-researchers, through GPs providing researchers with a clinical perspective of the pandemic impact, GPs identifying immediate research priority areas based on their experiences and GPs providing clinical insight into factors impacting data sources (e.g. understanding service item numbers). Strong collaboration between PHNs/GPs, an existing research team and the data custodians was integral to the success of the AR cycles encompassing clinical, data and research perspectives to produce clinically relevant knowledge outcomes pertinent to general practice activities impacted by COVID-19 pandemic circumstances. Collaboration with PHNs also facilitated the dissemination of knowledge outputs to general practices/GPs, supporting existing literature on the role of professional networks in increasing research impact (Reed et al. 2017; Sibbald et al. 2019) and the role of AR in bridging the research-knowledge gap.

In producing knowledge outputs (Fig. 2), AR also served as an enabling framework to actualise the components of the data, information, knowledge, and wisdom (DIKW) theory. Hebda et al. (2019) provide detailed descriptions/explanations of the components of DIKW theory, which include data as ‘…discrete observations with little interpretation’ and information as ‘…a continuum of progressively developing and clustered data…’ (Hebda et al. 2019, p. 23). In the current project, electronic data sourced via POLAR contained the elements of patient/clinical data and information collected by GPs. Knowledge, as ‘…information that has been processed and organised…’ (Hebda et al. 2019, p. 23), was generated through quantitative analyses of POLAR data. During the reflective AR phases, wisdom in the form of ‘…an appropriate use of knowledge…’ (Hebda et al. 2019, p. 23) was contributed through clinical judgement/intuition from PHN representatives, GPs and data custodians to add context and interpretation to the knowledge translation process, culminating in research outcomes.

Key learnings also emerged from the research analyses (e.g. various uses of telehealth), which generated evidence that could be used to inform both general practice and policymakers for directing policy and resources both now and for future disruptions (whether negative, e.g. COVID-19, or positive, e.g. introduction of telehealth) to the healthcare system. The research knowledge also identified missed opportunities/care during the COVID-19 pandemic (e.g. cancer screening, residential aged care onsite GP consultations), which serves as substantial evidence for the future provision of services/care to vulnerable groups, as missed care may lead to future morbidity/mortality. Analyses of quantitative data also identified gaps in the availability and usability of general practice patient data, which, if addressed, would be beneficial for future research and evidence to guide clinical practice.

Towards the end of the project, qualitative data were also collected to gain stakeholder feedback on their involvement across the entire project, including suggestions for future directions. The methodology, analysis and findings from this process will be reported in a future research publication.

The strength of the methodological approach used in the current project stems from the interrelationship between the four key elements of collaboration, AR cycles with the availability of an existing general practice data source, engaged governance, and the production and dissemination of apposite knowledge outcomes. The methodology served to reduce the research–knowledge gap through: GPs defining research priorities directly relevant to clinical practice based on their experiences during the COVID-19 pandemic; data custodians providing (with permission) the PHN owned de-identified general practice POLAR data; researchers producing timely evidence with context from PHN/GP feedback; the manifold dissemination of research findings through PHNs; and translation of research outputs to inform clinical practice, including helping PHNs to inform their commissioning programs and activities designed to improve patient healthcare outcomes. Consolidation of key research evidence into a suite of impact measures also provides an avenue to inform future policy and resource management beyond the current project to any disruption to the healthcare system, enabling continuing translational impact. The research approach used herein can be adapted to future research undertakings across any primary care setting and highlights the utility of AR and collaboration in primary care research to produce knowledge directly relevant to clinical practice.

The data that support this study are available in the article.

The authors have no conflicts of interest to declare.

This project was funded and supported by Digital Health Cooperative Research Centre Limited (‘DHCRC’) (project ID: DHCRC-0118). DHCRC is funded under the Commonwealth’s Cooperative Research Centres (CRC) Program.

All authors made substantial contributions to this project. All authors were involved in the conception and design of knowledge outputs throughout the project. Research team members (AG, GS, CI, ZD, GF, NW) were involved in the acquisition of the quantitative data from Outcome Health (the PHN contracted Data Custodians), with permission from each participating PHN (PHNs are the owners of the de-identified general practice POLAR data used within this project). The research team were involved in the collection of qualitative data. All authors were involved in the interpretation of either the qualitative or quantitative data analysis during the course of the project. The manuscript was drafted by JT and AG, and all authors critically reviewed the manuscript for important intellectual content. All authors approve the final version of the manuscript for submission. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.