Proxy reporting in health: a scoping review of instructions, perspectives, and reporting experiences

In this review, we explore the instructions provided to informal carers (‘proxies’) completing questionnaires for individuals who are unable to complete questionnaires. By understanding these instructions, informal carers can provide useful information about the heath-related quality of life and symptoms of others. This will help to improve the quality of data collected from proxies. This review identified a lack of consistency in how proxies are instructed to report on the health of people who are unable to complete questionnaires. More consistent instructions and understanding how well proxies follow instructions may be helpful.

Individuals with advanced illnesses or cognitive impairments can encounter challenges in reporting their own health status [1‐3]. In such instances, proxies, often caregivers or family members, may be asked to provide this information. The use of proxy respondents is common in epidemiological research [4, 5]. Proxies are often used in specific populations such as children [6, 7], older adults [3, 8, 9],or in specific clinical contexts such as dementia [2, 10], hearing loss [11] or critical illness [12]. They are typically asked to provide their perspectives in situations where patients have cognitive impairments, severe illnesses, or are otherwise incapacitated [13].

This scoping review focuses on informal caregivers, who are typically family members, friends, or neighbours providing unpaid care. In contrast, formal caregivers are paid professionals who have received training in caregiving and may be employed at healthcare institutions or private agencies in which they provide care to individuals. We chose to focus on informal caregivers because they often lack the professional training that formal caregivers receive, which can lead to unique challenges in reporting the health-related quality of life and symptoms of the person they care for due to limited healthcare knowledge. Greater understanding of the instructions given to informal caregivers, as well as their experience of proxy reporting, can inform better questionnaire design and, ultimately, potentially better reports from informal caregivers acting as proxies. In this review we focused on quality-of-life measures including health-related quality of life and symptom measures.

Reporting on behalf of someone else presents several challenges, primarily due to proxies’ limited access to complete firsthand information compared to individuals reporting on their own health [14, 15]. Proxies may lack complete insight into an individual’s health history and symptoms, leading to difficulties in reporting. Proxies acquire information differently than individuals who self-report. While self-reporting individuals draw information from direct experiences, proxies rely on observation or communication fundamentally shaping their reporting approaches [16]. This difference can make it challenging for proxies to provide a response that closely aligns with what the individual would have reported [17]. Proxy reports are affected by a number of factors, including proxies’ perception, memory, motivation to respond, the type of information to be reported [15, 18], and how the proxy engages with the individual, including the type of relationship, and the frequency of contact [18, 19].

Proxies may be asked to report about others from their own (“proxy–proxy”) or the individual’s (“proxy–patient”) viewpoints [20, 21]. The former is often defined as proxies responding from their own experience based on their observation and interpretation, while the latter entails proxies attempting to report as if they are the patient themselves and reconstruct the patient’s internal mental state when answering questions [22‐24]. In the proxy–proxy perspective, proxies are asked to respond based on their own observations and interpretations of the patient’s condition. They provide information from their viewpoint, reflecting what they see and understand about the patient’s health-related quality of life or symptoms [21, 25]. To illustrate these concepts, one might imagine a proxy being asked in the proxy–proxy perspective, as “How much pain do you think the patient is experiencing?” In contrast, from the proxy–patient perspective, a question might be phrased as, “If you were the patient, how much pain would you say you are experiencing?” The US Food and Drug Administration (FDA) [26] and the European Medicines Agency (EMA) [27] definition of proxy report uses the proxy–patient perspective. Research suggests that instructions provided to proxies regarding perspective-taking influence their responses to questionnaires [28‐30]. However, whether proxies follow these instructions is not usually examined [28, 30]. There is no clear evidence indicating that one perspective may better enable proxies to provide responses similar to patient responses, with inconsistent results that vary across different health domains [30‐32]. Neither proxy–proxy nor proxy–patient perspectives have been shown to have consistently better concordance with self-reports for all domains or conditions. Rather, there is some evidence of better concordance for one perspective with self-report for some domains, and for the other perspective with self-report for other domains. For example, a study in an orthopaedic outpatient clinic assessing agreement between self-completed EQ-5D-5L scores by patients and the proxy reports found substantial agreement between proxy and patient reports while using the proxy–patient perspective and moderate agreement when the proxy–proxy perspective was used [32]. Lobchuk et al. reported that asking the proxies to imagine themselves as the patients was effective in reducing discrepancies across symptoms and underlying symptom dimensions (e.g. severity, distress and frequency) [22]. However, Gundy and Aaronson found that concordance between patient and proxy reports was better for functional scales on a cancer-specific tool when using the proxy–proxy perspective [30]. A review on the effect of perspective on the report difference between the self-reported and proxy-reported quality of life of people living with dementia has indicated that adopting a proxy–patient perspective reduces the difference between proxy and patient report compared to the proxy–proxy perspective [33].

Proxy reports may influence clinical decisions, treatment plans, and interventions. If proxies provide inaccurate information and if the information is used in clinical decisions, it can impact patient care. A study indicated that it is easier for proxies to report on observable factual information than private subjective behaviour [18]. This is due to the fact that proxies may not have the same level of insight to the patient’s experience as the patients themselves. Recognising proxies’ confidence in reporting may help to judge the quality of evidence prior to using the information for decision making. Factors such as proxies’ knowledge and understanding of the health situation of the care recipient may be associated with their confidence. When proxies are confident, their reports may more likely reflect the patient’s true condition and preferences. Conversely, a lack of confidence may indicate potential uncertainties in the information provided [34]. When developing proxy-reported instruments, cognitive interviews may be used to identify potential issues in understanding and interpreting survey questions. This can help refine questions to ensure they are clear and comprehensible for proxies, thereby improving the accuracy of their responses [35]. During these cognitive interviews, proxies may be asked additional questions such as how difficult they find it to respond on behalf of someone else [36], and how confident they are that their responses align with those of the target individuals [18].

In the current scoping review, the term ‘proxy reporting experience’ refers to any question asked to proxies about their experience while completing a questionnaire for someone else including their level of confidence in their report. There is a knowledge gap regarding the specific instructions provided to proxies about perspective-taking [25, 37], the perspectives they adopt, and proxies’ experience when reporting about others [14]. Understanding whether specific instructions or prompts regarding proxy perspectives are provided to proxy respondents may be helpful. By identifying this, we can assess whether researchers and practitioners guide proxies on how to approach reporting (e.g., proxy–proxy or proxy–patient perspective). Knowing which perspective (if any) proxies adopt while completing health questionnaires may provide valuable insight into which perspectives they may find more useful. Examining how proxies’ reporting experiences are captured in published articles allows us to explore the experience of proxy reporting and understand real-world challenges. This scoping review had three specific objectives:

This scoping review aims to shed light on the current practices in proxy reporting, including the instructions provided to proxies, the perspectives they adopt, and their experiences in reporting. By examining these aspects, we anticipate uncovering patterns in proxy use across different clinical populations and identifying potential areas for improvement in proxy assessment.

A scoping review was used to address the research questions. It was chosen over a systematic review as it focuses on mapping the evidence landscape and discerning key knowledge gaps in the field [38]. Unlike systematic reviews, a scoping review allows for the exploration of broad and less specific research questions [38‐40]. The approach taken in this scoping review adhered to the methodological guidelines proposed by the Joanna Briggs Institute (JBI) [41]. We employed the framework outlined by Arksey and O’Malley [42], encompassing a structured process involving five key stages: (1) formulating the review questions, (2) identifying relevant studies, (3) selecting eligible studies, (4) charting and collating data, and (5) summarising and reporting results. We did not undertake the sixth optional stage of the scoping review, which is stakeholder consultation [43]. We used the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA- ScR) for scoping reviews to present the results [44].

A total of 5816 articles were identified, from which 2104 duplicate records were removed. Following title and abstract screening, 3652 articles were excluded, and 60 studies were retained for full-text review. Based on a full-text review, 44 studies were excluded: most of them (23 out of 44) do not address the concepts of the research questions. Initially, 16 studies were considered for data extraction. Four of the 16 studies were found to be ineligible during data extraction.

We conducted a scoping review of the perspective instructions provided to proxies, the extent to which these perspectives were adopted by proxies, and the proxies’ reporting experiences. In this review we identified variation in the instructions given to proxies assessing patient health, and did not find any assessment of whether proxies followed these instructions.

Most of the included studies in the current review involved informal carers of people living with dementia. This is expected, as proxy reporting is prevalent in this population [55] due to the under-development of self-report measures that work for this population [56‐58]. The instructions given to proxies in these studies were heterogeneous. In eight of the twelve studies, proxies were asked to report from both perspectives, while in four studies, they were requested to report only from the proxy–patient perspective. Even within the same proxy perspective, variation in the emphasis on perspective taking instructions provided was noted. For example, in a study [22] detailed instruction was provided as follows: “Sometimes when people try to understand what the other person is feeling or thinking, they imagine how they themselves would feel in the person’s situation. We would like you to try the same…While you are doing so, please try to imagine how you would feel if you had the patient’s diagnosis and how this would affect your life. In your mind’s eye, try to picture how you yourself would feel if you were experiencing the same symptoms. Focus on yourself. As you imagine how you would feel if you were diagnosed with the patient’s disease, please tell us…” In contrast, in another study [47] the instruction for proxy–patient perspective was: “We are interested in how you think the patient would assess their health. Answer these questions as if you are the patient.” Providing consistent instructions on perspective-taking may help proxies better understand how to approach the questions they are asked. This, in turn, could result in more interpretable and useful data. Specifically, consistent instructions can enhance the reliability of data collected by proxies, making it easier to observe changes over successive appointments.

Additionally, consistency between different proxy reporters can ensure that the data is comparable and interpretable, leading to more accurate assessments of quality of life [25]. The heterogeneity of proxy instructions observed in our study aligns with a previous systematic review on proxy ratings for patients with primary brain tumors [59]. Studies have emphasised the importance of clearly defining and documenting the proxy perspective [60, 61]. Reviews have also indicated that clear instructions should be provided to proxy respondents [7, 62]. This clarity may be helpful for healthcare providers, significant others, and proxy respondents when evaluating patients’ quality of life [63]. Clear instructions help proxies respond from the intended perspective [61]. Providing detailed guidance in instructions helps clarify whether proxies should respond based on their own views or those of the care recipient [14]. These studies collectively reinforce the importance of providing proxies with clear, well-defined instructions to collect better quality data.

In addition to providing perspective instructions, it might be important to check whether proxies understand and follow instructions as proxies might not always adhere to the instructions provided. Researchers should ensure that proxies understand their role and the intended perspective (their own or the care recipient’s), as this understanding is crucial for accurate data interpretation and reliable results. Evaluating content validity of the instructions, clear definition of intended perspective as suggested by previous guidance [61]. Asking proxies about their reporting experience and any challenges they encountered at the end of the survey may provide helpful context to responses.

The included studies did not specifically examine proxies’ reporting experiences. However, insights from validation studies using cognitive interviews and open-ended questions have shed light on proxies’ comfort, confidence, and difficulties in responding. During these cognitive interviews, proxies were asked about their interpretations of the domains in the proxy measures. By incorporating closed-ended probing questions that assess the ease or difficulty and confidence level in answering specific survey questions, researchers can evaluate whether proxies’ responses to these questions are linked to data quality. A higher perceived ease or confidence may suggest a greater likelihood of accurate matches between proxy and self-responses [36]. Researchers may wish to include debriefing questions at the end of the surveys to assess respondents’ experience of completing the survey. The debriefing questions may  help in assessing  respondents' percieved difficulty while  completing questionnaires [64]. This information can provide valuable context for interpreting the findings. When proxies are not given clear instructions regarding the perspective they need to adopt, it may lead to inconsistent data and make interpretation of results difficult as it will not be clear whether the respondents follow their own or the patients’ perspective.

One limitation of this review is the exclusion of grey literature, which may have led to relevant studies being overlooked. For example, relevant reports or theses could contribute to a greater understanding of the topic. Additionally, as most studies in the review were in dementia and cancer populations, it is not clear how generalisable these findings are to other conditions. The small number of included studies potentially limits the strength of the evidence but also underscores an important research gap. Specifically, relatively little is known about proxies, despite their important role in providing data in many situations. This lack of information makes it difficult to improve questionnaire design. Additional work can help address this gap.

Despite these limitations, the use of a systematic search strategy, comprehensive literature review, double screening, and rigorous methodology—including predefined eligibility criteria and adherence to established JBI guidelines—enhances the credibility and robustness of the review’s findings. These methodological strengths provide a comprehensive overview of the current state of research on proxy reporting instructions for informal carers, ensuring that the conclusions drawn are well-supported and reliable.

The review highlights the need for more consistency in instructions given to proxies, as well as information on how well proxies follow these instructions and their experience of reporting. This can help improve questionnaire design and the quality of proxy-reported information about the health of others. Providing clear instructions helps the researchers and clinicians to collect useful and interpretable information from informal carers. To this end, clear instructions and an understanding how and if instructions are followed and interpreted can be useful in interpreting proxy-reported information.