When a patient is too ill, too young, or cognitively impaired in a way that leaves them unable to complete a patient-reported outcome (PRO) report (or they are limited in some other way that precludes them from completing a PRO), those patient perspectives don't get collected. Jessica Roydhouse, PhD, Select Foundation Senior Research Fellow in Health Services Research of the Menzies Institute for Medical Research at the University of Tasmania in Australia, and Bryce B. Reeve, PhD, Professor of Population Health Sciences and also Pediatrics at Duke University School of Medicine discuss the possibility of using proxies to report PROs in these cases (J Clin Oncol 2023; doi: 10.1200/JCO.22.01049). In a joint interview, the two researchers shared their thoughts on the research and ongoing work on this topic.
1 What are the most important takeaways from this research?
Roydhouse: "Caregivers of children with cancer will have insight into how their children are feeling and functioning. This is not the same insight you would get from the children themselves. However, in recent research that looks at caregiver-clinician-child triads, caregivers do better than clinicians in terms of getting closer to what children say. If a child is not able to complete a PRO, then their caregiver can give useful insight that can help clinicians know how the child may be coping with treatment or how the treatment is impacting the child."
Reeve: "The information from the caregiver could help make sure the child is not suffering. If doctors can better manage the child's symptoms, then the child may stay on treatment longer and receive the maximum treatment benefit. There are multiple studies that provide evidence that children, their caregivers, and clinicians have related but different perspectives on what the children are experiencing in terms of symptoms and functional limitations. The caregivers' responses are more similar to the children than the clinicians' perspectives. However, we are not saying that one perspective is right and the other is wrong or flawed. Each perspective provides a complementary perspective about what is going on in the children's lives. We need to better learn how to use each perspective to provide more comprehensive care to the child. But care must begin with every attempt to listen to the children's voices."
2 What's the main barrier in wider use?
Roydhouse: "The main barrier is that proxy responses are in many ways a second-best option."
Reeve: "Ideally, patients will self-report all the time and investigators should consider every effort to include their voice. The challenge is to consider when to collect data from caregivers and/or clinicians. It could be only when the child is unable to complete a questionnaire (throughout the study or at an assessment point).
"However, this option is not ideal in that investigators may want to mix both patient self-report and proxy report in the same analyses, which would not be appropriate. Another option would be to always collect data from caregivers and/or clinicians at every data point where we collect the patient's self-report data. This would result in more comprehensive data, but also logistical and cost challenges to collect data from additional reporters. In addition, how do we make decisions on treatment benefit or safety when we have multiple reporters on the same outcome? There are potential answers, but this needs to be explored through further research."
Roydhouse: "There is also regulatory guidance discouraging 'proxy' report, although other guidance for registries does not provide similar discouragement. It's important to consider that the alternative in many cases would be either no information or trying to guess what the PRO would be from clinical or sociodemographic characteristics."
Reeve: "...using imputation methods."
Roydhouse: "This is likely far less predictive than a proxy report."
Reeve: "In addition, we need to: 1) provide better guidance about how to integrate proxy measures in research study to achieve high-quality data and limit survey fatigue and financial cost; and 2) identify best strategies to analyze symptom and function data when multiple reporters are used or mixed reporters."
3 What's the next step of your work?
Roydhouse: "Right now we are working on a project about methods for analyzing data from proxies."
Reeve: "We are interested in ways to enhance proxy measures to provide better information about patients, and this is an area where we are planning future work. In addition, we are looking at key decision points in clinical care and research and thinking of how to incorporate proxy data to make the most informed decisions."
Roydhouse: "Guidelines and standards will help, and getting more evidence about proxies beyond how proxy and patient responses compare when patients can self-report is needed."
Reeve: "Another consideration is how we can design better quality proxy measures with clear instructions to enable the proxy to provide more valid information about patients."
Roydhouse: "Caregivers know something about their children; it isn't necessarily the exact answer a child would give, but it can provide useful information. Working with caregivers to learn more when children or other patients can't provide this information can help."
Reeve: "Our discussion has been focused on children with cancer; however, these same important issues about using proxy data also apply to older cancer patients with cognitive limitations that reduce their ability to self-report."