Introduction

Well-designed feasibility studies can provide practical and logistical insights before undertaking a large-scale study [1]. Feasibility studies improve the likelihood of a successful future investigation and can save time and resources by minimizing the risk of undertaking large studies destined to fail [2]. Despite these potential benefits, feasibility studies are underutilized [3]. This editorial discusses the types of feasibility studies, the utility of feasibility studies, common pitfalls, and keys to success.

What are feasibility and pilot studies?

Eldridge et al. developed an initial framework to define pilot and feasibility studies in 2016 and the authors extended this framework in 2023 [4, 5]. They defined feasibility as an umbrella term encompassing any study that informs parameters of a future main study [4, 6]. Within the overarching concept of feasibility, studies are classified as either a) randomized pilot studies, b) non-randomized pilot studies or c) feasibility studies that are not pilot studies (Fig. 1). Supplementary Table 1 provides examples of the different study types [4].

Fig. 1
figure 1

Schematic representation of the types of feasibility studies proposed by Eldridge et al. [4].

Full size image

Randomized and non-randomized pilot studies are smaller versions of the main investigation evaluating all or part of the intervention [4, 6]. For instance, Airody et al.’s pilot trial randomized 44 patients to different treatment regimens for neovascular age-related macular degeneration to identify challenges before a large-scale RCT. Their investigation achieved pre-specified feasibility thresholds, and they concluded that with minor amendments, a large-scale RCT would be achievable [7].

Feasibility studies that are not pilot studies are designed to answer questions about an element of the future investigation. These studies focus on a component of the study without implementing the intervention or assessing the outcomes in the same manner as the future main study. For instance, Davara et al. assessed the effectiveness of a smartphone app for caregivers to assess patients’ visual acuity. This study was not a smaller version of a future larger investigation; rather, the objective was to explore a tool to facilitate future research [8].

The remainder of this editorial will focus on pilot studies as they are the most common feasibility study [4].

Why perform a pilot study?

Chalmers and Glasziou published a report in the Lancet estimating that more than 85% of research efforts are wasted due to inappropriate study designs and a lack of focus on correct research questions [9]. Trials are no exception; a review of nearly 30,000 abstracts from conference proceedings describing clinical trials reported that only 63% of abstracts were published in full [10].

Pilot studies are undertaken to improve efficiency, reduce wasted efforts in larger trials, and minimize the risk of committing greater resources to a trial likely to ‘fail’ [2]. The implications of pilot studies for large-scale trials can range from minor alterations to cancelling the trial. Publishing pilot studies, particularly pilots identifying methodologic issues, is informative in sharing the best research processes and practices [11].

Thabane et al. summarized the advantages of pilot trials into process, resource, management and scientific benefits [2, 3, 11, 12]. Process advantages include gaining insights into steps that will occur in the larger study, such as the recruitment and retention rates. Resource benefits arise from obtaining more accurate estimates of time, budget, and other requirements. Management benefits involve optimizing personnel and data management systems prior to the main study, and scientific benefits are derived from obtaining estimations of the variance of the effect and dose-response. A successful pilot can also demonstrate the feasibility of the trial to funding agencies and may even be required by certain agencies [2, 5, 13, 14]. Additionally, pilot and feasibility studies may facilitate assessments of surgical innovation, which may be more challenging than non-surgical studies to finalize details of the study, such as standardized techniques and the willingness of providers and patients to accept randomization [15].

Common misconceptions and caveats of pilot studies

While pilot studies should incorporate clinically important outcome measures, they are not designed to test hypotheses regarding these outcomes. Pilot studies are not powered to detect minimal clinically important differences, and using pilot studies to determine treatment effects is flawed and potentially misleading [2, 12, 14, 16]. In extreme cases in which repeated serious adverse events occur, they may be able to inform the safety of an intervention [14]. Instead, outcomes should be focused on feasibility, process, and describing initial results [2, 12, 17]. Pilot trials should have pre-specified criteria for progressing to the main study [17, 18]. These criteria include assessments of recruitment, adherence and outcome data [18]. Avery et al. suggest a traffic light system with green (go), amber (amend), and red (stop) to determine next steps. This approach facilitates less stringent thresholds and permits appropriate progression to the main study with amendments if required (Fig. 2) [18].

Fig. 2
figure 2

Summary of the pilot study process proposed by Avery et al. [18].

Full size image

A commonly cited argument against pilot trials is that they delay answering research questions [19]. To minimize this drawback, an internal rather than an external pilot can be performed [17, 18]. In internal pilots, the pilot phase is the first part of the trial, and the data generated can be used in the final analysis [17, 18]. In the 2023 extension of the pilot and feasibility conceptual framework, Bond et al. discussed the role of internal pilots and provided guidance, given their recent increase in popularity [5]. Internal pilots are preferred when feasibility issues have been primarily resolved, and the uncertainties are focused on recruitment and randomization [5, 20]. This approach, however, requires a pre-specified adaptive design and appropriate controlling for the risk of an increased type I error rate [14, 18].

Lastly, confusion may arise when differentiating between phase I-IV studies and feasibility studies. The primary objective of phase I-IV studies is fundamentally different from pilot and feasibility trials [4]. Nevertheless, they do share overlapping features, and phase 1 and 2 trials may provide information relating to the feasibility of a phase III trial. Supplemental Table 2 compares these study types [21].

Conclusion

Feasibility is a concept encapsulating any study that will help prepare for a future main study. Pilot studies are smaller-scale versions of the main study and are the most frequently conducted type of feasibility study. A pilot study can increase the likelihood of a successful main study, improve the methodology of the main study, and help avoid devoting resources to studies likely to fail. Feasibility studies are not designed to determine treatment effects; instead, established criteria should be used to interpret the data to determine the next steps depending on specified feasibility primary objectives.