Including the patient voice in the development and implementation of patient‐reported outcomes in cancer clinical trials

Abstract Context Patient‐reported outcomes (PROs) are used in parallel with clinical evidence to inform decisions made by industry, clinicians, regulators, health technology assessment bodies and other health‐care decision‐makers. In addition, PRO data can also guide shared decision making and individual patient choice. Yet, the quality of many PROs in cancer clinical trials is suboptimal and requires improvement to add value to health care and policy decision making. Objective To show how the integration of the patient and/or patient advocate at all stages of PRO development can help to realize the full potential of PROs. Methods We examined the literature to show that the patient voice is often absent from the planning and implementation of PROs in cancer clinical trials. Good practice examples from the literature were combined with guideline recommendations, training or educational resources, and our own experience to create detailed practical steps for the inclusion of patients and/or patient advocates throughout PRO development. Results Patient or patient advocates can play an active role in shaping PROs that are meaningful to the patient. They can contribute to content, choice of medium and implementation in a way that may support PRO completion and minimize missing data. Patients and their advocates can work to ensure PRO findings are disseminated appropriately in a way that is accessible to patients. Conclusion This practical guidance aims to optimize PRO development and implementation in clinical trials, resulting in robust, relevant data that reflect the patient experience and that support decisions made by all stakeholders involved in research and health care.


| INTRODUC TI ON
Patient-reported outcomes (PROs) have been used in cancer clinical trials for over two decades. 1 These evaluations, captured electronically or in person, in formats ranging from questionnaires to wearable devices, serve to provide a unique record of the patient's lived experience of a disease, its treatment and management, and the impact these may have on function and health-related quality of life (HRQoL). 2,3 The inclusion of PROs as trial endpoints aims to ensure a comprehensive assessment of burden of disease and the impact of an intervention. 4 In this way, PROs have established themselves as a central and indispensable component of the evidence evaluating medicines and can be used by clinicians, patients and policy makers to assess treatment choice, shape guidelines and enable regulatory and policy decisions based on the benefits and costs of treatment. 5,6 The inclusion of PRO data in clinical trials is supported by international guidelines, professional organizations and by regulatory and health technology assessment (HTA) bodies for measuring patient experiences that are not captured by conventional efficacy or adverse event data. 3,[6][7][8][9][10] In practice, however, few PROs in preapproval oncology clinical trials are reported to meet the Food and Drug Administration (FDA) standards. 11 Of the 40 treatments approved by the FDA Office of Hematology and Oncology Products between 2010 and 2014, only three (7.5%) were granted PRO labelling. 12 Collection of PROs has been shown to be inconsistent in cancer clinical trials with less than a third of the recommended PROrelated items included on average in study protocols in the UK. 13 PROs are not always accepted in HTA review processes: one recent study of submissions across several European countries found that PROs were mentioned in the final decision in less than half and were not reviewed by the HTA body in 21%. 14 Reasons cited for the poor quality of PROs in some cancer clinical trials include methodological, cultural and practical issues. 3,15 Choice and content of PRO measures are not always relevant to the aim of the study or may not reflect what is important to the patient. 4,16,17 PROs are all too often dismissed as subjective in contrast to laboratory findings and as less important than survival endpoints. 11,17 In its guidance on PROs in regulatory submissions, the European Medicines Agency (EMA) refers to a range of issues including bias, timing of assessments and missing data. 3 The practicalities of implementation may not be conducive to patients completing a PRO measure or to administrative staff supporting them in doing so. 4 In some cases, compliance with completion of PRO measures is so poor that the data are not even analysed. 18,19 PRO reporting can lack transparency, failing to provide details such as the rationale for a PRO measure or the approach to missing data, hampering interpretation. 20 Furthermore, PRO results may not be disseminated to the most relevant audiences or are not always reported in an accessible way. 4,21 As leaders of patient advocacy organizations, we believe that the failure of many PROs to fully reflect the patient perspective means that decision-makers such as clinicians, industry representatives, regulators, policy makers, HTA bodies and patients themselves are missing an important piece of evidence that could potentially resolve uncertainties about value and aid decision making. We urgently need to integrate the patient voice throughout PRO development and implementation, rather than the more usual practice of the patient acting as a consultant at certain stages of PRO development. 16 Incorporating the patient perspective will optimize the relevance of PROs in cancer clinical trials and support the delivery of patient-centred care. We are not alone in this view: patients, clinicians, industry and regulatory bodies have shown their support for a collaborative approach to the development of PROs and corresponding measures. 7,16,22,23 This does not yet happen routinely in cancer clinical trials. 16 How we might achieve such an approach and what benefits this may bring is the subject of this paper.

| THE PATIENT AND PATIENT ADVO C ATE CONTRIBUTI ON
In order for the patient voice to be integrated throughout the development, implementation and dissemination of a PRO, it needs to be included from the outset. 24,25 This enables a truly collaborative approach with patients, patient advocates and in some cases caregivers as patient partners, moving beyond traditional roles as consultants or providers of information and into roles as advisers, co-creators and even drivers of the process, as shown in the model of the participation ladder ( Figure 1). 24,25

| PRO design and selection
Patients and patient advocates can play a central role in the design and selection of PROs. They bring to the process knowledge of the disease, symptoms and attributes of care with the greatest impact Choosing a study endpoint takes place early in the planning process and patients and patient advocates can contribute to the identification of PRO-assessed endpoints that are meaningful to patients, such as pain or unmet medical needs. In a disease where quality of life can be a priority for patients, patient experiences and outcomes may well be considered at least as important as improvements in survival in shaping research priorities. 11,26 Cancer patient psychosocial well-being and physical well-being are not mutually exclusive: psychosocial factors, such as depression, have been shown to predict recovery and health status after surgery. 27 Subsequently, the patient or patient advocate voice can be combined with the findings from a literature review and expert input in order to define a conceptual framework that will detail the rationale of the PRO and the specific concepts it aims to measure; for example, experience of pain, fatigue or HRQoL. 6,7,28 It should also detail the intended population who will complete the measures. 12,18 Patients and patient advocates can ensure that the PRO concept is meaningful and relevant to the intended population and can contribute further by selecting domains for inclusion in the PRO that reflect the concept and ensure that the patient experience is captured fully. 6,7,29 This may involve identifying domains that are missing from existing PROs: the social and emotional experiences of cancer patients, for example, are not always assessed. 30,31 Domains not relevant to the intended population may also be identified.
Using their familiarity with the patient experience, patients and their advocates can provide input into item selection for any existing or new measures, giving careful consideration to the symptoms and effects associated with a specific type of cancer. They should also take into account whether the study participants are starting treatment, receiving treatment or living with or beyond the disease. 21

| Patient-reported events
Patients and patient advocates might consider calling for the use of a PRO measure in a clinical trial to document both positive and negative effects experienced by the patient that may otherwise not be identified as disease-or treatment-related. Discrepancies can exist between health-care professional and patient perception of adverse events (AEs) associated with treatment 32 and clinicians, unlike patients, are required to judge how likely an AE is to be treatmentrelated. It is important that clinicians' judgement on the relevance of an AE does not include only AEs that can be clinically managed while AEs such as fatigue are not taken into account. A measure that records patient-reported events has the potential to provide a comparison with clinician reports. Industry, HTA bodies, regulatory bodies or payers can assess PRO AE findings rather than relying on interpretation or selection by trial investigators or clinical experts.
One such tool has been developed. Based on the Common Terminology Criteria for Adverse Events (CTCAE), the PRO-CTCAE allows patients to record their experience of AEs that do not require laboratory assessment, without clinician selection and interpretation. 33,34 Patient representatives were active participants throughout the development of this measure and a wider group of patients was involved in providing detailed feedback at several points in the process. 34 Basch and colleagues have shown that patients are willing and able to complete the PRO-CTCAE for the collection of adverse event recording and that minimal additional resources are required. 35 Making the use of a PRO on AEs mandatory may provide a better assessment of the balance between the efficacy and toxicity of cancer treatments. 36 Patients and patient advocates recognize the negative impact that a grade 1 or 2 AE, such as nausea or fatigue, can have on an individual's quality of life if it is persistent rather than transient and may suggest that PRO measures include duration as well as frequency. 37,38 This is particularly relevant in the cancer patient population where extended survival may mean that treatment is given over longer periods of time. Additionally, all grades of AE data (from mild AEs at grade 1 to severe AEs at grade 4) should be included in trial reports, not only AE data of grade 3 or above which can result in under-reporting of toxicity. 36,39 Use of patient education tools provided by patient advocates is beneficial in supporting patient understanding of the differences between AE grades. 40 Patients and patient advocates may also choose to add their voice to the call for standardized AE grading decision criteria. 40 Finally, in discussion with their research partners on AEs, patients and patient advocates are able to support the necessity of disseminating full AE data so that individual patients can make informed choices. Some patients may be prepared to tolerate greater AEs than others based on age, stage in their patient journey and other factors. 41 Patient-reported AE data are increasingly accepted or solicited by regulatory and HTA bodies and can play a key role in their decision making. 42 In the case of the non-small-cell lung cancer treatment crizotinib, for example, post-marketing PRO data on symptom control and improved HRQoL resulted in a reversal of an earlier negative reimbursement decision by the German HTA body which was based on data only from the initial study endpoints. 42

| Good practice examples of design and selection
Integration of the patient voice in the development of PROs used in cancer clinical trials does not appear to be commonplace to date. 15 Emerging examples do exist, however, of patient and patient advocate involvement in the design or selection of PROs in the routine care setting and in patient surveys and registries that are engaging with key questions that are relevant to PRO use in clinical trials. [43][44][45][46] Patients and patient advocates can offer valuable advice on the challenging question of whether to use standardized pre-existing PRO instruments, or to adapt or devise a new measure. Standardized PRO sets are recognized as enabling population comparisons, which are of value to comparative processes such as HTA assessment, in addition to making efficient use of resources (Table 1). 17,43 However, experience of a specific patient population may not be reflected in a standardized measure. 47 Where a PRO measure does not reflect the patient experience in question, the choice is whether to adapt an existing measure with the same concept, to combine validated generic and disease-specific measures in a trial or to develop a new measure. 47,48 An example of a PRO measure developed for a specific population is outlined in Table 2 health-care professionals, industry and policy makers. 45 Registry patients were involved in this way in one recent study which examined patient-reported toxicities and quality of life in lung cancer patients treated with immune checkpoint inhibitors. 46 Crucially, patients from this group along with oncology providers and informal caregivers were interviewed to develop an item bank on toxicities, which formed part of the PRO measure used. 46 These examples, along with the elements of PRO design and selection outlined above, show the potential of the contribution of patients and patient advocates to developing or choosing a PRO for use in a clinical trial that fully reflects the patient experience.

| PRO implementation and administration
The patient and their advocates can also bring insights based on experience to guide practical decisions around PRO implementation and administration. They are well placed to collaborate on the production of a trial protocol which details how, when and by whom the PRO should be implemented, along with a record of the conceptual framework. 6 Lack of such detail in the protocol may result in methodological inconsistencies between centres and how individuals implement or document findings. This can result in suboptimal data quality. 19  interpretation; compromising the validity of study findings; or resulting in treatments being denied PRO labelling by regulatory bodies. 12,19 The completion of PRO measures must not be burdensome for either well or unwell patients. 21 Shorter and more reliable measures can improve response rates. 48  and reassuring them that their responses will help them receive better care on the research study or in the clinic. Advocates can also prompt study participants to remind their health-care professional if a PRO questionnaire they are expecting is not received.

| Linguistic and cultural input
Patients and patient advocates can make an informed contribution to the language used in questionnaires in order to ensure the use of plain language, the lack of which has been noted in some existing cancer PRO measures. 50 Similarly, patients and patient advocates from each region or community where the measure will be used can help validate linguistic and cultural aspects a PRO measure in order to make it understandable and accessible to all members of the study population. 51

| Achieving inclusivity
Importantly, patients and patient advocates from low-or middleincome backgrounds or countries appear to be less frequently

| Medium, setting and technologies
Recognition is increasing that in order to improve data quality and adherence to data collection, PRO data need to be collected in ways beyond the standardized questionnaire format on cancer clinic days. 4 Table 2

| A FR AME WORK FOR PATIENT AND PATIENT ADVO C ATE INVOLVEMENT IN PRO DE VELOPMENT
This paper and its references contain numerous practical suggestions and resources which can further support the involvement of patients and patient advocates in PRO development, implementation and dissemination. In addition, a framework has been developed to guide a collaborative partnership between health-care professionals and patients and patient advocates, which enables the inclusion in clinical trials of PROs that truly reflect what matters to patients. The criteria used for the selection of patient partner (patients, patient advocates and caregivers in some instances) will be dependent on best fit between the partner, the focus of the research and their anticipated roles in developing the PRO. 24,25,56 Best-fit patient advocacy organizations may be assessed by con-

| Optimizing communication in the development team
Traditionally, the patient and patient advocate have been involved only at the later stages of giving feedback on what has already been conceived and is in development. When they are fully engaged in F I G U R E 2 Checklist for involving patient partners when initiating a PRO development project the development process from the very beginning, the patient and patient advocate can actively influence decisions, pose questions and even drive the process. This more comprehensive and complex approach brings with it the need for communication and facilitation skills in order to achieve meaningful and productive engagement. 24 An open dialogue between all stakeholders that allows discussion of both the patient experience and formal care systems is crucial for mutual understanding. 22 Also important is the support required for all team members, including regular and direct communication, use of a partnering system between clinical researchers and patients and patient advocates where appropriate and opportunities for reflection on the collaborative process. 22

| A framework for moving forward
To  The patient and patient advocacy community should be working collaboratively with researchers, industry, clinicians and investigators.

| CON CLUS ION
Patient and patient advocate input has great potential to improve design, reduce missing data and impact regulatory processes, policy decisions, shared decision making, and ultimately patient outcomes.
As leaders of patient advocacy organizations, we are committed to working towards this goal and we invite all stakeholders to join us.

ACK N OWLED G EM ENTS
We would like to thank Claire Montell, Medical Writer (UK), for assistance in writing and editing the manuscript with financial support from Bristol-Myers Squibb. Bristol-Myers Squibb had no role in drafting or approving the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analysed in this study.