Development of decision-making considerations to support equitable patient selection in paediatric haemophilia trials

In order to identify gaps in our knowledge of equitable enrolment processes, a preliminary literature search was conducted using the Google Scholar and MEDLINE databases for information on competitive participant selection using keywords such as ‘participant selection’, ‘competitive selection’, ‘fair subject selection’, ‘resource allocation’, and ‘selection bias’ in combination with the term ‘clinical trial’ for articles published between January 1960 and July 2021. Inclusion criteria consisted of any English language article that provided trial participant selection guidelines or examined challenges in maintaining equity of opportunity when distributing limited resources, whether in the allocation of clinical trials or other therapies. Articles about clinical trial recruitment were excluded as our concern was not with increasing enrolment in trials, but rather selecting between participants who were already deemed eligible for enrolment. In our experience, recruitment has typically not been a challenge amongst paediatric haemophilia patients, and there are limited avenues for increasing enrolment within this population because certain essential qualifying criteria (such as body weight or previously untreated status) cannot be altered by clinicians.

After screening, we found 11 relevant articles (Appendix), many of which outlined guidelines for fair subject selection, but few of which had attempted to implement the recommendations and discuss the outcomes. No articles concerning enrolment in haemophilia trials were found, but two articles were identified that discussed participant selection in cystic fibrosis clinical trials with competitive enrolment ^[6,10]. One article specifically examined challenges in equitable subject selection for COVID-19 therapy trials ^[3].

The literature review revealed that current practices for formulating enrolment processes are highly heterogenous, although there are some overarching themes. Recent literature on COVID-19 therapy allocation proved to be especially helpful in our search, as the pandemic provided a prominent example of how to distribute limited medical resources amongst an extremely large group of eligible people with competing and diverse interests.

Of the strategies proposed and reviewed in the literature search, ‘first come, first served’ was a tactic that was generally not recommended, as it unfairly advantages those who can access or learn about treatments more quickly, and neither attempts to maximise benefits nor minimise harms with what limited resources are available ^[3,12,14]. Several of the articles reviewed explicitly stated the need to prioritise vulnerable patients or those with a medical need, indicating that protecting at-risk patients is a common priority ^{[3,11,12,13,14,15]}. However, this may not necessarily be a consideration that carries over to clinical trials, as clinical trials have scientific aims to fulfil that scenarios involving resource allocation do not. Prioritising only those with the greatest medical need for trial enrolment may therefore adversely influence the generalisability and quality of the results one can obtain from the trial – a factor that needs to be taken into consideration during the decision-making process.

Another frequent concern was the importance of reducing healthcare disparities and overcoming barriers such as personal biases or prejudices, geographic limitations, or language in equitable subject selection ^{[3,7,8,9,10,11,12]}. Amongst articles that did not limit their scope to the distribution of a particular resource or patient population, preventing the unjustified inclusion or exclusion of certain populations was a popular aim, except in cases where the participant or their community would be put at unacceptable risk, suggesting that patient safety was critical ^[7,8,9]. A small number of articles concerning resource allocation during the COVID-19 pandemic encouraged the prioritisation of participants based on ‘instrumental value’ to some degree – although the value of any individual is difficult to determine or translate to a clinical trial setting for a disease such as haemophilia, and invites its own ethical questions ^[12,13,14].

Regarding the major realms of disagreement, random selection was a somewhat controversial method of decision-making. While random selection was lauded for its impartiality and received positive feedback for its usage in Dobra et al.'s algorithm for patient selection in competitive cystic fibrosis trials ^[6], it has also been said to not meaningfully reduce morbidity or mortality, provide any economic or social benefit, or reduce health disparities ^[12]. In addition, many articles seem to lay out criteria or recommend selection processes that random chance is inadequate in fulfilling alone ^{[3,8,9,12,13,14,15]}. Furthermore, although a considerable number of articles examined in the literature search made an effort to take into account the conflicts that may arise when attempting to satisfy multiple criteria during decision-making, each recommended a different order of priority ^{[6,9,11,12,14,15]}. One article did not recommend that the interplay between criteria be taken into account at all, arguing that attempting to consider such complex interactions would make the processes too inconsistent or difficult to implement, and instead suggested following a linear, rigid sequence of priority ^[13]. Again, we see that any effort to weigh these ethical principles against each other proves to be troublesome, providing another example of a pertinent ethical challenge that researchers face.

Lastly, in attempting to develop and implement a decision-making protocol that balances these competing priorities, four articles specifically recommended assembling a diverse team of stakeholders and using iterative deliberations to create a shared set of values and ethical principles to inform patient selection ^[3,6,14,15]. Of these articles, one recommended using a blinded scoring system to reduce bias and introduce a quantitative aspect to the selection process ^[3]; one team came to the conclusion of using randomisation ^[6]; one used a scoring system to develop solutions but left the relative weightings of each criterion up to individual judgement ^[15]; and one suggested that multiple approaches should be used and that researchers’ personal ethical and clinical reasoning would be required to evaluate trade-offs between conflicting priorities ^[14].

Although this literature search revealed a great deal of knowledge on how clinicians may go about developing more equitable trial selection processes, there is a general dearth of information and recommendations for resolving this ethical issue, the importance of which researchers have only recently begun to acknowledge.

In light of this literature search, our objective was to develop a list of preliminary decision-making criteria to describe a more objective process for selecting paediatric patients for participation in therapeutic haemophilia trials with competitive enrolment, in hopes of making participation opportunities more equitable.

The criteria for these considerations were developed through discussions amongst our team of interdisciplinary clinical and allied health professionals, followed by iterative sessions of revision and deliberation, drawing primarily upon the ethical foundations outlined by the Canadian Tri-Council Policy Statement on Ethical Conduct for Research Involving Humans ^[16].

Creating a selection process for competitive clinical trial enrolment is undoubtedly a difficult undertaking that involves balancing many diverse, competing factors. We sought to create a list of considerations that could be used as a guideline when there were many eligible candidates that could be offered the opportunity to participate in a competitive clinical trial for paediatric patients with haemophilia. Although the principles of these considerations may be applied to other situations in clinical settings where equitable resource distribution is an issue, each therapeutic area will differ, and users are therefore advised to tailor the criteria to best suit their specific requirements. Haemophilia is a unique therapeutic area in that there may be more potentially eligible patients who express interest in participation in clinical trials than there are opportunities to enrol. Larger studies in other therapeutic areas may find other methods of participant selection more appropriate to mitigate bias and ensure fair opportunity.

The criteria presented provide a potentially more objective process for facilitating equitable patient selection in clinical trials with limited enrolment availability by urging clinicians to intentionally consider factors such as medical need, systemic barriers, and patient engagement during the selection process. By developing a comprehensive list of the considerations involved in competitive participant selection, we hope that this tool will enable clinicians to recognise and discuss potential sources of bias more readily. Traditionally, factors such as language barriers or a therapeutic team's lack of cultural competence have proven to be obstacles to clinical trial recruitment. These may emerge as issues if clinicians preemptively and unjustly exclude patients on this basis without awareness of the possible biases involved. Encouraging deliberate conversations around the factors involved in enrolment processes may therefore enhance equitable decision-making ^[23].

In our experience, these considerations have proven to be an effective framework for discussion amongst a multidisciplinary team through encouraging a variety of professionals – including social workers, physiotherapists, nurses, and physicians – to provide insight into their areas of specialty and facilitating a more holistic decision-making process overall. However, as the criteria are still being trialled, we encourage other clinicians to modify them to meet their needs as they see fit.

In light of these criteria, our team also recognises that clinical trials serve not only a therapeutic purpose, but a scientific one. As such, one final factor to consider when selecting amongst patients is whether their participation will be conducive to the study's investigational aims. Our main concern is that the patients selected should be appropriately representative of the therapy's target population to ensure generalisability. In selecting patients who may meet the considerations described here, a clinician may skew the trial population towards those who they believe may be more ‘deserving’ of the opportunity or more successful in a trial – such as those with greater medical need, or those who require less support. As such, patient selection must also take into consideration which demographics may be systematically excluded in trying to optimise these factors. It is imperative that the scientific goals of the study are assessed before considering any other criteria for competitive selection.

There are situations where these considerations may come into conflict with each other. For instance, Iterative Step 1 may lead clinicians to conclude that patients with poor adherence be recommended for trial enrolment due to their greater medical need, while Iterative Steps 2 and 3 suggest that patients with fewer support needs and greater adherence could have a greater chance of success. As evidenced by the literature search, this area of ethical consideration is rich in disagreements on how to weigh such criteria against each other. As a guideline, our team recommends weighing the considerations developed in each step equally, unless there is a need to adjust them based on the scientific aims of the study. Although this reliance on individual judgement could result in some decision-making discrepancies between each team, it permits these considerations to be flexible enough to meet a wide variety of needs. For this reason, we elected to avoid assigning any scoring system to the criteria to avoid inhibiting decision-making.

Issues involved with negotiating the intersectionality of aspects such as race and gender have not been considered here. Our team recognises the role that inherent biases and systemic barriers play in clinical decision making and initial consultations were held with equity and diversity specialists at McMaster Children's Hospital to help clarify the ways in which these circumstances may affect patient selection. Although this information was not included in our initial set of matrix criteria, future steps should involve incorporating the voices of patients and communities, as well as consultations with greater numbers of stakeholders, in order to better understand the role that social determinants of health play in influencing participant selection.

Finally, although historically we have not found enrolment to be an issue within our population of interest, we recognise that the task of increasing enrolment in clinical trials as a whole and selecting between eligible patients who are already enrolled share common ethical concepts. As such, our exclusion of articles discussing enrolment introduces limitations to our literature search. These articles could possibly provide insight into additional barriers towards trial participation or reveal patient perspectives on the burden of clinical research, and are recommended as an area of future investigation. Ultimately, we hope that these criteria serve as a starting point for stimulating team discussions around how to make the competitive clinical trial enrolment process more representative, equitable, and rigorous.