Crafting Tempo and Timeframes in Qualitative Longitudinal Research: Case Studies From Health Research

The Nature of the Phenomena: Intense, Acute and Time-Bound

Thus, it makes sense that phenomena or processes which are time-bound, acute, rapid, and multiple can particularly benefit from a more intensive orientation towards data gathering. Being able to capture these experiential features as they unfold will likely require more frequent data collection waves and shorter timeframes (Hermanowicz, 2016). This was particularly evident in our eight-month study involving three waves of interviews with asthma patients in the context of the later stages of the COVID-19 pandemic (study 1). The backdrop of external events potentially shaping individual experiences was intense and acute. Specifically, during the period of data collection, the UK had experienced multiple changes to public health restrictions, while in primary care we witnessed the acceleration of the COVID-19 vaccine programme and various changes to the provision of care for acute and long-term conditions. These events were important as they influenced how patients perceived their condition and engaged with healthcare providers. Thus, the modest timeframes and intense tempos allowed for the mirroring of context and the highlighting of the rapidity and fleeting nature of changes, as experienced by the participants.

The Benefit of Detail

A more intensive orientation also provides opportunities to focus on detail and nuance in studying the processes of continuity and change (Hermanowicz, 2016; Neale, 2021). A more intense tempo, for instance, can enable a greater depth of engagement and a more nuanced examination of change (Howell et al., 2012). In fact, the more intensive the tempo, the more opportunities to create a “movie” which can zoom in on the intricacies of participants’ lives, described as “walking alongside” participants as “fellow travellers” (Neale & Flowerdew, 2003). Such an intensive tempo was particularly suitable in study 1 on asthma and study 2 on health workers’ experiences of the COVID-19 pandemic. The intensive tempo allowed us to capture rapidly changing contexts (e.g., policies and restrictions) and – consequently – rapidly changing experiences and views, such as perceptions of risk. Thus, frequent data collection allowed a close examination of how participants’ views changed or remained the same. On the extreme end of the intensive orientation, the intensive tempo can take the form of almost continuous visits (Epstein, 2002). This approach can facilitate even more access to the immediacy of views, experiences, and “epiphanies in peoples’ lives” (Saldaña, 2003, p. 33). Study 2 illustrates this approach: at the start, participants were interviewed every two weeks, while over the course of the year, the frequency of interviews was adjusted to fit the pace of changes in participants’ perceived experiences and situations.

Challenges With Intense Timeframes

However, it is worth noting that a more intensive orientation may not always allow enough time to “detect” change (Neale, 2021). As highlighted by Neale, it may prove difficult to align the study tempo with the tempo of the phenomena under investigation (Neale, 2021). It is also possible that different tempos will be suitable for different aspects of the phenomena. For example, while the intensive tempo in our asthma study lent itself to detecting rapid changes in some areas, as described above, other aspects (such as how patients utilised their self-management strategies) could have benefited from a less intensive tempo or longer timeframe as changes simply did not occur dynamically and at the same rate. However, it is worth noting that the slow pace of change became a finding in itself, showing that patients did not make dramatic changes to their self-management strategies in ways that had originally been anticipated. This highlights the need for also allowing enough time to explore change and continuity in sense-making rather than prioritising a focus on observable change. This applies to both sense-making done by the participants as well as researchers analysing the data in shorter periods of time and thus not being able to take the “long view” perspective.

A Focus on Sense-Making

Examples such as those described above in which experiences are explored in near-real time, or during a specific temporal context such as a health emergency, are perhaps instinctively ‘intensive’. However, processes such as those in study 3, which explored the lived experience of chronic illness, illustrate how intensive approaches can indeed be used to explore sense-making over time. In this study exploring how patients with COPD used a digital remote monitoring platform following an exacerbation in their symptoms, the overall timeframe was bounded conceptually by the principles underpinning the Year of Care approach (Roberts et al., 2019) which reconsiders the role of the individual in the daily management of living with a long-term condition. Patients make multiple daily decisions and carry out activities which affect their health and quality of life, with often limited interaction with healthcare professionals. By focusing on individuals’ stories and sense-making over time, we can explore how these interact with established clinical pathways or novel digital tools. However, as might be expected in the context of long-term chronic illness, time may take on a different quality, with lived experiences described as static and stagnant. Significant physical recovery is unlikely, and sense-making focuses predominantly on how patients experience the mundane reality of managing a condition that severely limits overall quality of life. This approach to intensive QLR feels quite different to the more ‘urgent’ examples described above and impacts on the approach taken to data collection. For example, interviews are still semi-structured and guided by research aims but are longer, with time and space given for articulation and probing of whatever seems meaningful to the participant. These kinds of practical adjustments are important factors to consider, reminding us that we are not just trying to capture a holistic understanding of participant experiences, quickly.

The Nature of the Phenomena: Protracted, and Less Bounded

A more extensive orientation is apt for gaining a sense of change and continuity when the processes or phenomena of interest are known to be protracted or less clearly bounded. Two of our example studies – a 32-month study of older people’s decision-making regarding knee surgery in the context of multimorbidity (study 4) and a 60-month study of patient journeys through genomics (study 5) - had more extensive orientations. In each case, the timeframe and tempo were determined by the nature of the inquiry, and, in particular, an interest in exploring how participants (re)make decisions about and manage the process over time. For study 4, participants were recruited at the time of referral to an orthopaedic consultant, with the intention of capturing decision-making regarding surgery and, if applicable, experiences of recovery in the subsequent months. Similarly, in study 5, the emphasis was on understanding the experiences of patients and families who were participating in genomic testing for a rare disease or cancer. The study was designed to capture, over time, their experiences from querying a potential heritable tendency, through to making decisions for, about and/or with relatives, receiving (certain or uncertain) findings, and living with a result(s) or uncertainties. This process often spans several years. For both studies, whilst a more intensive timeframe and tempo may have permitted exploration of one or two aspects of the process, an extensive orientation was considered vital to gaining a holistic (as far as possible) sense of each journey.

A Focus on Sense-Making

An extensive orientation is often, although not exclusively, characterized by more ‘punctuated’ gaps between encounters with participants (Neale, 2021). A longer period between waves can allow for greater reflection on the part of both the researcher and participant, potentially offering new insights or re-interpretations of past accounts. In study 5, participants were encouraged during each new wave to reflect on aspects of their previous interviews to explore how they (re)made sense of different decisions and encounters. More ‘punctuated’ visits can, however, make it more challenging to detect the intricacies of the dynamic processes at play. Adaptations can also be made to the study tempo as the research progresses. The original intention in study 4 was to interview participants on two occasions; at the point of referral to orthopaedics and six months after surgery as clinically this is considered a reasonable length of time by which surgery success can be evaluated (e.g., in terms of pain, mobility, functionality). If they were not recommended or did not opt for surgery, the interview was scheduled for six months after their consultation. Changes to patient’s pathways due to the pandemic (detailed below) necessitated the decision to add an additional wave for those who had been on the waiting list for 6 months. This provided a better understanding of if/how their experiences and perceptions were affected by these delays.

Incorporating a Retrospective Focus

Even with a more extensive timeframe, it is not always possible to capture a process in its entirety. A longer timeframe can, however, permit greater opportunity to extend the temporal reach of a study by obtaining in-depth retrospective accounts, inviting participants to look back over their lives. In both studies 4 and 5, taking a retrospective approach, as well as following participants prospectively through the process, was essential as their present circumstances and imagined futures could not be understood without gaining in-depth insights into their backstories. For example, the focus in study 4 on how patients with multiple long-term health conditions assess the relative importance of their knee problem necessitated an emphasis on their other pre-existing conditions to understand their priorities and expectations regarding the potential knee surgery. Similarly, in study 5, most participants had already experienced multiple investigations and tests, and encountered a range of healthcare specialists prior to being referred for genomic testing. These experiences shaped their hopes, expectations and anxieties. Understanding their backstories was essential to interpreting how they conceive of and contend with the process of testing.

The Interplay Between Micro and Macro

Finally, a key advantage of a more extensive timeframe is the opportunity to follow participants over a significant period of time and understand how wider economic, social and political change shapes their individual or collective biographies. Whilst during periods of intensive change this might be feasible through a more intensive orientation (as discussed above), a more extensive timeframe is more conducive to capturing the (often relatively slow) pace of such change. QLR with an extensive orientation is also valuable for capturing how individuals or collectives experience the implementation of a new policy initiative or practice, the tempo of which could be shaped by key moments in the process. For example, participants in study 5 were interviewed as the new national genomic medicine service started to become established within the National Health Service (NHS) offering the possibility of capturing their views on a significant change in healthcare policy.

Identifying a Clear Baseline and Closure Points

To define a study timeframe, it is important to consider the baseline and closure points. The baseline for a QLR study should be a temporal marker, which could be biographical or historical (Neale, 2021). Biographical events can be understood as closely linked to the individual’s experiences, such as receiving a diagnosis of a health condition or becoming a parent. In contrast, historical events can be defined as something external to, and independent of, an individual, such as the introduction of a national policy. Baseline and closure points can be a useful way of checking whether the scope of the project remains the same in order to ensure the research aim(s) and questions remain central (Yates & McLeod, 1996). They can be considered bookends in the studies we conduct.

In some studies, a historical baseline might be most relevant and can be a clear, discrete starting point (Neale, 2021). For example, the beginning of the COVID-19 pandemic was a baseline for study 2 on health workers’ experiences during the pandemic as the study was designed specifically to explore the impact of this event on people from as early as possible. Similarly, in the asthma study, while the starting point was not the beginning of the COVID-19 pandemic, there was a common baseline: namely, at the first interview, all participants shared the same context which was the implementation of restrictions by the UK government.

Difficulties in Conceptually Defining Baseline or Closure Points

Alternatively, studies with biographical baselines may provide a similarly discrete ‘start’ point in terms of study design. However, they can at times be conceptually more difficult in terms of explanatory power or meaning-making. Indeed, for studies primarily interested in experiences around an ‘active phase’ of a lived experience, this idea of a conceptual boundary is certainly useful in terms of framing the study design but in reality, it is not always clear-cut. For example, in the studies on experiences of COPD (study 3) and knee joint replacement surgery (study 4), the participants’ starting point was essentially conflated with their eligibility to participate in the research. Patients in study 3 were eligible to participate in the year-long study if a recent exacerbation in their symptoms had led to hospitalisation. However, this invariably was not the first exacerbation they had experienced. Thus, although the ‘baseline event’ was clinically meaningful, it may or may not have been as significant to the patient in terms of sense-making. For some people, this exacerbation was anticipated and relatively routine in the context of their condition’s overall trajectory, while for others the exacerbation was an intense and frightening experience.

Similarly, in some projects, the starting point may not always coincide with the beginning of the individual participants’ journeys. For example, in study 5, patients referred for genomic testing were likely to have had many other tests and encountered a range of clinical specialists beforehand. The point at which their journey began was not clear-cut. To allow for this, Neale’s (2021) suggestion of sampling comparatively across different cohorts, focusing on those who are at different points in their journeys, proved fruitful. Accordingly, two cohorts were recruited with different temporal baselines. The first cohort included those who participated in genome sequencing via their involvement in the 100,000 Genomes Project (100kGP); a UK-based initiative that included 85,000 patients affected by rare disease or cancer (Peplow, 2016). Participants in this cohort were first interviewed 1–2 years after joining the 100kGP. Retrospective accounts of their journey prior to the test were captured, thereby extending the study’s longitudinal reach. Participants were then followed prospectively to see how their journey evolved. The second cohort comprised those accessing testing via the new NHS genomic medicine service and to which recruitment is ongoing. For these participants, the baseline is generally denoted by initial contact with the service. Retrospective accounts about their journey to the point of testing are combined with prospective interviews aimed at capturing participant’s experiences of the process in real-time.

External Factors Dictating Closure Points

Some closure points may also be of biographical or historical nature, as explained earlier. This was the case in the asthma study; similarly to baseline, historical events shaped the closure point of the study. The study’s end was not aligned completely with the end of the pandemic, as what that meant has been contested and widely discussed. However, at the time of the third and final wave of data collection, the healthcare system was re-introducing some services and this provided a natural, yet not planned, endpoint.

It is important to acknowledge more pragmatic approaches to determining the closure point. These can include funding coming to an end, or researchers coming to the end of their employment contracts. This was the case for study 4. The planned endpoints were either six months after the knee surgery to explore the process of recovery (for participants who had surgery) or if they were not recommended for, or did not agree to have surgery, six months after their consultation. However, due to disruptions to orthopaedic surgery, the clinical journeys for those recommended for a knee replacement were much more protracted than the original study design had anticipated. Consequently, the end of data collection was pragmatic, based on the maximum window to which the funding arrangements could span. This meant that some participants at the end of data collection continued to be waiting for surgery and some had a surgery only recently. Similarly, in study 3, closure was simply one year of usage of the telemonitoring system - not necessarily clinically or personally meaningful. Indeed, some people found this arbitrary and difficult, and the removal of the monitoring system and its associated ‘care’ was a concern for some.

The Nature of Phenomena Requiring Flexibility in Design

Despite the necessity for creating baseline and end points for studies, there is also a need to remain flexible (Elliott et al., 2008). Flexibility can be a real strength of a well-resourced QLR study and intentional flexibility is an important aspect that should be embedded in the design from the beginning (Holland et al., 2006). As highlighted in a review by Audulv, only a minority of studies using QLR design, tend to build in flexibility, alluding to the difficulty of doing so (Audulv et al., 2022). Studies focusing on the phenomena of more unpredictable length or tempo can particularly benefit from more flexible design. Study 5, for example, sought to capture patients’ experiences as genomic medicine started to feature as part of routine healthcare in the NHS. As a new service, the patient pathway was elusive, making it difficult to anticipate the potential timeframe and appropriate tempo. Participants were also on different clinical pathways, seeking genomic tests for a range of conditions and diseases. At times, there was much to discuss over a relatively short period of time, and at others, the journey was marked by much waiting. The tempo for each case was not prescribed and, instead, discussions were scheduled to coincide with aspects of individual’s journeys. A flexible approach and keeping abreast of the circumstances of different participants offered many benefits, allowing the team to capture ‘the dynamic process under investigation’ (Neale, 2021, p. 109). It was also labour intensive and maintaining engagement relied heavily on building long-term relationships and being clear about the fluid nature of the study from the start.

The Need to Respond to Unplanned and Unpredictable Changes

The second aspect of intentional flexibility is the need to respond to unplanned and unpredictable changes as uncertainty can also be born from events external to the substantive focus of the study. For example, it was impossible to predict how the COVID-19 pandemic was going to unfold and what the most appropriate tempo would be. Thus, in study 2 on health workers’ experiences during the pandemic, we incorporated a flexible tempo, which changed over the study period. Such flexibility was intended to accommodate the uncertainties of delivering this research during a public health emergency. The protocol outlined that participants would be interviewed repeatedly (e.g., every 2–4 weeks) over an agreed period. The interviews initially occurred every two weeks. However, as the speed of changes varied for participants in different settings and roles, over time more flexibility was introduced. Interviews were agreed upon with each interviewee individually. This meant that some participants were interviewed more frequently (as their contexts/settings and availability changed more) than others, resulting in between 4 and 10 interviews per participant. This flexibility allowed the participants themselves to decide on the tempo, which was appropriate to their circumstances. This also represented a combination of researcher-led contact and participant-led decisions on whether and when each interview should take place. Thus, the tempo was the result of this approach, rather than being specifically pre-planned.

Routine research practice can also face unpredictability and challenges. This can be amplified in the context of QLR and requires an intentionally flexible approach. While it is important to plan the overall timeframe of the study, external factors, such as recruitment challenges, can make following these timeframes difficult. In study 3, which intended to follow an intensive approach to longitudinal data collection, significant recruitment challenges led to a revision of the planned research. It was hypothesised that recruiting patients at the point of hospitalisation might be an optimal time at which to provide support. However, very low recruitment numbers together with high dropout rates (Whelan et al., 2021) necessitated delays to accommodate a significant protocol amendment. The adjusted strategy improved overall recruitment numbers somewhat, but the adjustment resulted in alternative challenges around clinical and research staff capacity and workload. These factors changed the overall study design and the delay in commencing and conducting intensive longitudinal work served in particular to magnify the burden of workload around data collection, management and analysis. Similarly, the challenges in the NHS affected the clinical pathways of patients and altered both the timeframe and tempos of studies 4 and 5. For example, in study 4, pre-pandemic it was realistic (in most NHS trusts) for patients to wait around 6 months before undergoing knee joint replacement surgery. But with elective surgeries ceasing during the pandemic and waiting list backlogs growing, at some orthopaedic surgery sites, patients were advised to expect waits of up to 36 months. The study protocol had to be revised to accommodate these new clinical realities, which meant a subgroup of participants was interviewed on three occasions to accommodate reflections whilst on the waiting list. The longer timeframe and variable tempo did, however, allow shifting attitudes in relation to policy and zeitgeists of the pandemic to be captured; for example, views about NHS staff and expectations of the NHS; perspectives on the role of the UK government in public health; and preferences over modes of health service delivery (including remote care).

A Lack of Flexibility Brings Some Limitations

In contrast, in some studies, there was little pre-designed flexibility. Study 6 was developed as part of a larger (pre-COVID-19) research programme on implementing interventions to optimise antibiotic prescribing in general practice. It aimed to opportunistically explore the impact of the COVID-19 pandemic on antibiotic prescribing and stewardship. The original study involved two rounds of interviews with clinicians implementing the ‘Antibiotic Optimisation’ intervention at pre-set time-points (6- and 12-month post-implementation to coincide with follow-up surveys). Study 6 was designed to add two further rounds of interviews 6 and 12 months later. We selected the time-points at the study design stage assuming that 6-month intervals between the interviews would constitute sufficient time for changes in context, practice and experience to occur and for participants to be able to reflect on them. As is often the case, studies evaluating interventions typically involve pre-designed data collection points, with qualitative data collection usually matching ‘standard’ time-points (e.g., 3, 6, 12 months) for quantitative outcome measures. Such design allows answering research questions (to evaluate the intervention) and generating qualitative and quantitative data that can be triangulated. Similarly, in study 1 about asthma management, there was arguably little flexibility in the longitudinal design: the timing and number of waves of data collection were based on the three-month regular interval schedule pre-defined by the research team. Given a modest timeframe (8 months), largely driven by the funding constraints, there was also little flexibility in introducing more frequent waves of data collection. However, the intensive tempo allowed for capturing detailed accounts. In addition, relatively short intervals between interviews enabled reflections on events, which happened in-between the data collection waves with less reliance on retrospective accounts.

Minimising Attrition Through Relationship Building

Longitudinal studies are at risk of participant attrition (Hermanowicz, 2016). Thus, establishing and managing relationships with participants is crucial. Rapport is essential to all qualitative endeavours and is regarded as a marker of quality (Prior, 2018). It involves building rapport with the participant as well as trust which can be facilitated by clear communication about the study objectives and expectations. Studies with shorter timeframes may suffer from less attrition as participants may consider the length of engagement as less demanding (Neale, 2021). However, the frequency of contact, even in the context of shorter timeframes, needs to be taken into account when building, maintaining and exiting the research relationship (Calman et al., 2013). Walking intensively alongside participants may call for “the ethics of walking alongside” (Neale, 2021, p. 142). In our studies, involved between two and 10 waves of data collection, we have experienced low attrition. In studies with shorter timeframes, this was perhaps due to engagement with the study being less demanding for the participants. In addition, we have also witnessed how even modest timeframes could lead to strong relationships with the participants, opening dilemmas of how to maintain professional boundaries. This was perhaps heightened by the fact that some of the studies were conducted during the pandemic when some participants had limited contact with other people, which to some extent might have led to research relationships taking on a different meaning (Pilbeam et al., 2022). Indeed, pandemic notwithstanding, isolation and loneliness are common for many people living with chronic illness and this is an important consideration in the negotiation of boundaries within the researcher-participant relationship and the management of expectations. This was the case in the context of study 3, which spanned a year of life but which could be considered relatively brief in the context of wider QLR: participants were elderly and/or living alone, and repeated contact with the same researcher over the course of a year allowed for the formation of connection and rapport with a ‘curious stranger’ (Denzin & Lincoln, 2011).

In studies involving longer timeframes, the importance of rapport and sustaining long-term research relationships comes to the fore. The burden of participating in a longitudinal study, especially if it involves repetitious activities, can lead to participant fatigue. It is essential to balance the depth of data collection with the participant’s capacity and willingness to engage over time. Trends in UK Higher Education, particularly the propensity for early career researchers to be on short-term employment contracts, means that it is not always feasible to maintain researcher continuity over time (for reflections see Shirani, 2013). Yet, when participants engage with the same researcher, this can help foster an enduring connection. In study 4, the potentially long gap between interviews generally was not seen as a concern, nor was the renewal or maintenance of rapport required. Some participants remarked that their participation had felt like a useful process allowing them to ‘close the chapter’, suggesting they perceived the timings of the interviews as appropriately mapping their clinical journeys and outcomes.

An extensive orientation can also open up the possibility for incorporating a range of interim activities that help sustain participant engagement, and minimise attrition whilst also generating interim data. This could also include short catch-up phone/video calls, online/postal activities or working with participants to co-produce different aspects of the research. In our studies with longer timeframes, we have also experienced limited attrition. This can potentially be attributed to our engagement activities. In study 5, for instance, a sub-sample of participants worked intensively with the team and an artist to co-produce visual representations of their journeys. This deepened further the team’s understanding of participants’ experiences, as well as leading to the production of resources designed to support new patients and healthcare professionals.

Finally, while maintaining relationships is important, ending them also needs attention (Morrison et al., 2012; Reiss, 2005; Smith, 2003). Ending relationships needs to facilitate closure for the participants who have invested a considerable amount of time and energy in sharing their experiences (Warin, 2011). Thus, ending the relationship requires careful preparation, starting earlier than during the last interview (Warin, 2011). Strategies, such as clearly communicating the end of the study to participants, explaining the outcomes and how their participation has contributed to the research, and offering a debriefing session, are crucial (Morrison et al., 2012; Reiss, 2005; Smith, 2003). Also, acknowledging that a relationship has been formed is important. We view incorporating these components in our studies as part of our ethical practice, even in studies with shorter or less intensive contact with the participants. We have also reflected on our own emotional responses to endpoints: as others have highlighted previously, the degree of rapport can equally affect researchers who are likely to have been heavily invested in the research project and relationships fostered with participants (Morrison et al., 2012; Reiss, 2005; Smith, 2003).