Talking to the people that really matter about their participation in pandemic clinical research: A qualitative study in four European countries

2.1 Design and setting

This study was conducted as part of the European Commission funded Platform foR European Preparedness Against (Re-)emerging Epidemics consortium (PREPARE; www.prepare.eu). It aims to inform study optimization of multisite, pan European clinical studies in primary and critical care for over 5000 patients during interpandemic periods and in preparedness for a research response during a future pandemic. We used qualitative research methodology because we aimed to identify important issues from the perspectives of the public before attempting to quantify them. We therefore prioritized quality and richness of information gathering over quantity or prevalence of views. This work has informed construction of a survey instrument that is being applied to a large, representative sample.

2.2 Recruitment and sampling

We selected one country each from northern, southern, eastern and western Europe, as defined by the United Nations macro-geographical regions.6 Focus groups and interviews took place between July and November 2015 in Cardiff, UK; Barcelona, Spain; Lodz, Poland; Antwerp and Bruges, Belgium. These locations also had teams that were members of the PREPARE consortium recruiting patients into pandemic research. Additionally, their health services differ in important, contrasting ways (eg fee for service and free at the point of delivery). Local, experienced qualitative researchers with good command of the English language were recruited through pre-existing research networks. Informants (>18 years) within the four countries were purposively sampled to include parents, carers and informants with experience of hospital admission. Participant recruitment was by local advertisement. Interested respondents completed a short demographic questionnaire before being invited to attend the group.

2.3 Data collection

Focus groups and semi-structured individual interviews were used. Our rationale for using focus groups was to stimulate discussion between those involved, anticipating that this would give rise to varied ideas and perspectives beyond those of the research team.7, 8 We held 10 two-hour focus groups. Topic guides were developed iteratively in two pilot focus groups (UK). Two focus groups were subsequently held in each of the four countries in accessible community venues. Pilot data were included in the data corpus as there were no substantive changes to data collection procedures that invalidated their use. We conducted 16 follow-up telephone interviews (15-30 minutes in length) within 2-7 days of focus groups with two informants from each group. We selected interviewees for these individual interviews who were less vocal in the group or those who had held discrepant or contrasting views.7 Focus groups and interviews were conducted in the local language. A face-to-face training meeting of all focus group moderators was held prior to data collection to ensure consistency, cultural sensitivity and to inform reflexivity. All study materials were translated into local languages. Focus groups were conducted in the local language by two local researchers (one moderator and one observer). Local researchers came from a range of different backgrounds (sociology, primary care, public health, culture science). Post-group debriefing and the use of field notes further enhanced reflexivity.

Focus group questions were rooted in evidence on pandemic clinical research and standards for good clinical practice in research.4, 9, 10 We used images and hypothetical scenarios of a moderate-risk influenza pandemic to stimulate discussion of participating in the following three research scenarios: a point of care test evaluation (primary care), evaluation of a routinely used antiviral medication with known safety profile (primary care) and evaluation of a newer antiviral medication with known safety profile (hospital intensive care unit) (Box 1). We also asked about the acceptability of conducting research that uses excess from routinely collected clinical samples, and enrolment to adaptive clinical trials (Box 1). We selected a moderate-risk pandemic influenza because an airborne virus, such as a novel influenza, is a likely candidate for an epidemic or pandemic spread that might affect European countries. Scenarios were aligned with clinical studies being conducted in PREPARE that aim to recruit patients from primary care and critical illness settings and that aim to generate evidence to inform clinical management of outbreak-related illness. We chose to avoid an outbreak with very high mortality, such as Ebola, which might polarize decisions and limit transferability of findings to a European context. We also asked about reasonable adjustments to research enrolment processes, that is from the point when a patient is identified as eligible to be invited into a study, through the process of information exchange to support their decision about participation, and, if selected, the provision of consent. Follow-up interviews aimed to capture post-group reflections.

2.4 Analysis

Focus groups and interviews were audio-recorded, transcribed verbatim and translated into English. Local researchers checked each stage for accuracy. Data were analysed using the framework approach, which follows a systematic process moving from data management to description to interpretation.11 We charted the focus group (NG) and interview (HS) data by identifying text that related to our study objectives as evidence of a particular theme. In this way, a visual overview of the data was produced from which patterns, including similarities and differences, could be identified. In practice, this involved charting data related to each specific objective on multiple Excel spreadsheets, which then formed the basis for team discussions regarding emergent, explanatory themes. Charting the data enabled recognition of contradictory or opposing views. The process of charting gave rise to the initial coding frame, which was then sent to local researchers (MG, SA, MPV), who double coded one of their focus group transcripts, and gave feedback on the analytic frame. At this point, the analytic process moved further into an interpretative stage where links between themes and cases were explored to expand the explanatory frame. In this higher-level analysis, we aimed to identify the range and diversity of factors influencing decision making and retrospectively applied decision-making theory12 in our analyses. NVivo 10 software (QSR International, Melbourne, Australia) was used to facilitate analysis.

2.5 Rigour and quality criteria

We used the COREQ checklist when designing and reporting this study13 and adhered to the following rigour criteria14: description of context, of informants and of the research process, methodological adequacy and reflexivity of the multidisciplinary research team. Throughout the research process, we actively sought opportunities to consider how multiple perspectives influenced the research process.

2.6 Ethical considerations

Researchers obtained local ethical approvals in each country. Informants gave voluntary, written informed consent, and researchers guaranteed anonymity, confidentiality and data protection.

3.1 Public understanding of pandemics

Informants referred to historical pandemics such as the Black Death, Cholera in London, Spanish flu and, more recently, Ebola. Awareness of these outbreaks contributed to a perception that pandemics, by definition, are highly infectious, spread rapidly and result in high mortality. In contrast, the 2009 swine flu (“H1N1”) pandemic was associated with expectations of milder illness. Public health information distributed through official channels was seen as trustworthy and as able to contribute to a more balanced understanding of what pandemics involve. However, informants described particular news stories or events as having introduced new perspectives. For example, the social activism of a Catalan physician during the swine flu epidemic, who heavily criticised the pharmaceutical industry, was cited as a reason to be cautious about the motives of those conducting research. Informants also noted criticisms of public health agencies and in one group, in Wales, informants reflected on the political dimension to pandemics, where political parties are under pressure to respond effectively to pandemic outbreaks.

3.2 Attitudes towards pandemic research and willingness to participate

In endorsing these models of research enrolment, informants referred to organ donation initiatives. Thus, in Belgium and the UK (Wales), opt-out policies for organ donation were referenced. In Poland, parallels with opt-in to organ donation registers were made. For ICU research, informants recognized the need for a third-party to consent on their behalf should they lack decision-making capacity. Preference was expressed for a close family member who could represent their wishes and values to decide. However, if family members were unavailable (for example, ill with the pandemic), some informants, particularly those with prior ICU experience, considered it acceptable for a doctor to decide on their behalf. High levels of trust in health-care professionals and the relatively low risks of harm with involvement in the study determined acceptability.

3.3 Factors influencing decisions about research participation

Decisions about participation were influenced by a complex interplay between personal attributes, research factors and contextual cues. Personal factors, such as age, life stage and previous ICU experience, were influential. For example, older informants felt their age made them more “philosophical” and described feeling more socially conscious, while parents of younger children described needing to be especially cautious about exposing themselves to risk. Research factors, particularly the benefits and risks or burdens of participation, were deliberated. These were judged contextually. For example, when deliberating about participation in the ICU scenario, risks associated with acute illness and the pandemic were considered.

In the ICU scenario (scenario 3), informants perceived that they would receive more effective treatment if in a trial than if not. Some informants said they would only participate if guaranteed to receive an experimental trial intervention. Fear, driven by a perception that that no known effective treatment may be available for a life-threatening illness meant that informants placed greater hope in the intervention's superiority. Informants were also motivated by altruism, social responsibility and by contributing to medical knowledge in the “long-term.” They wanted to receive feedback about the outcome of research they supported and felt feedback of results was a neglected aspect of research participation, which reduced the chances of subsequent research involvement.

In the ICU scenario (scenario 3), risks and burdens were judged proportionate to the context of critical illness in which the decision about research participation was being made. Informants expressed less concern about potential side-effects of new interventions, describing the critical illness scenario as being “serious” and “you're in dire straits” (Poland).

In primary care, informants described a high level of trust in clinicians that had developed over time and with continuity of care. Several informants spoke of actively seeking to consult with primary care clinicians they trust; care pathways in Poland and Belgium allow patients to choose their primary care physician and this is possible to some extent in Spain and the UK. Informants with experience of ICU spoke of trust developing very rapidly with treating clinicians by virtue of the intensity of acute illness and vulnerability coupled with perceived expertize of the clinical team, including the quality of their communication. Informants felt they would seek advice from these clinicians about research participation. A clear, minority view was also expressed suggesting that informants may feel unable to refuse participation in the context of an existing relationship with their treating clinician.

3.4 Acceptability of using routinely collected clinical samples for research

However, a minority of respondents from all countries wanted upfront information about how and for what their sample might be used. They also sought assurance that their samples would not be used for commercial purposes and were generally less trusting of research funded by industry sponsors, or where financial gain was the main ultimate motive. A counter-argument was also raised, highlighting the need to support investment required to develop new, more effective products.

3.5 Views on Participating in Adaptive Clinical Trials

4.1 Strengths and weaknesses

Sample sufficiency was attained with the use of complementary techniques and the richness of data collected across four countries. Richness of data was achieved, in part, by skilful group facilitation through which informants felt comfortable to change their opinions or express discrepant views; researchers actively elicited individual and group deliberation. Similarity of findings in the different countries and with findings from other studies, using follow-up interviews with participants holding discrepant views and reflexive discussions on analysis among researchers, supports the credibility of our research. Our sample is not reflective of the age group most affected during the H1N1 pandemic; however, we have no way of predicting in advance which age groups will be affected in future pandemics. Our groups were largely mono-cultural, and the views of minority ethnic groups are not represented. As respondents volunteered to participate, we assume some responder bias. We used hypothetical scenarios and views may have been different had we used different scenarios, or following actual research participation during a pandemic. Process evaluations conducted as part of pandemic clinical research are required to offer insight into participants’ actual experience. Despite training, there was some variation in the way informants were recruited and data were collected in different countries, which may have impacted our ability to compare across countries. The median age differences in different country data also limits our ability to compare directly across countries. Appropriate to the hypothesis generating nature of this work, we worked with a purposive sample which is designed to provide transferrable but not generalizable data.25 We have used these findings to develop a cross-sectional survey of public opinion that will allow us to quantify views and consider generalizability in nationally representative samples in eight countries within the Organisation for Economic Co-operation and Development (OECD).