Perceived cognitive concerns are common among older adults, with 50–80% of adults over the age of 70 reporting some degree of decline compared to an earlier stage of life (Jessen et al., Reference Jessen2010; van Harten et al., Reference van Harten2018). While objective measures of cognitive functioning (i.e. neuropsychological tests) remain the gold standard for identifying pathological cognitive decline, subjective cognitive complaints (SCCs) are increasingly considered in comprehensive assessments of older adults. Numerous studies have demonstrated that self-reported SCCs are only weakly associated with current objective cognitive ability and are more strongly related to psychiatric factors (Burmester et al., Reference Burmester, Leathem and Merrick2016; Crumley et al., Reference Crumley, Stetler and Horhota2014). However, emerging research suggests that SCCs may, for some people, represent changes that are not yet detectable on objective testing (Mitchell et al., Reference Mitchell, Beaumont, Ferguson, Yadegarfar and Stubbs2014). Indeed, while many people with SCCs are cognitively healthy (i.e. “worried well”), SCCs in a subset of people may serve as a preclinical marker of neurodegenerative diseases, with implications for early detection and timely intervention (Hessen et al., Reference Hessen2017; Jessen et al., Reference Jessen2014, Reference Jessen2020; Mitchell et al., Reference Mitchell, Beaumont, Ferguson, Yadegarfar and Stubbs2014). Despite increased interest in the clinical utility of SCCs, there are differing opinions on how best to measure and interpret these reports.

The study by Truong et al. (Reference Truong2022) expands the existing body of SCC literature by evaluating dynamic and enduring patterns of SCCs reported by patients and informants over a 10-year span. In a large sample of community-dwelling older adults from Australia, the authors used generalizability theory to examine the overall reliability of two widely used measures of self- and informant-reported SCCs – the Memory Complaint Questionnaire (MAC-Q) and the Informant Questionnaire on Cognitive Decline in the Elderly (IQ-CODE). Specifically, they examined how a person (e.g. self vs informant), scale items, occasion, and the interaction of these factors influence the reliability of the two measures when repeatedly measured over 10 years. Outcome measures included whether patients met the criteria for dementia diagnosis, decided upon at a consensus meeting with clinicians from varying backgrounds (i.e. neuropsychiatrists, psychogeriatricians, neuropsychologists) using available clinical, neuropsychological, laboratory, and neuroimaging data. In addition, they measured objective cognitive performance using a battery of 10 tests that assessed attention/processing speed, language, executive functions, visuospatial ability, and learning/memory.

Truong and colleagues’ results suggest that both the informant-report (IQ-CODE) and a modified version of the self-report measure (MAC-Q) have at least adequate reliability in measuring enduring patterns of SCCs over time, and that the scores were likely to generalize well across time and populations. In reporting on the reliability of these measures, the authors contribute to a growing literature on improving assessment of cognitive impairment in older adults. Prior work has probed the reliability and diagnostic utility of performance-based cognitive tests, to identify those best suited to this purpose (McDonnell et al., Reference McDonnell2020; Nielsen and Jørgensen, Reference Nielsen and Jørgensen2020). Truong and colleagues’ approached the detection of cognitive impairment from the perspective of self and informant-reported cognitive decline, which is a complementary approach to objective testing.

In addition to the presented generalizability assessment, Truong and colleagues demonstrated that informant reports significantly predicted future cognitive test scores and dementia incidence across all occasions, while self-reported scores were less accurate in reflecting cognitive abilities and diagnosis at both the first and last assessments. In other words, collateral reports from people who knew the participants well were more consistent in the prediction of participants’ future cognitive decline than were participants’ reports of their own cognitive difficulties. The authors interpreted their findings as evidence that self-reported SCCs may reflect a tendency to ruminate about everyday memory errors (e.g. walking into a room and forgetting why they entered the room) rather than measuring their actual cognitive capacities (i.e. overall cognitive decline), possibly due to anxiety and stress. In contrast, the reliability and predictive accuracy of informant reports was interpreted as evidence that corroboration from knowledgeable informants is clinically important and useful. To expand upon these findings and further consider the clinical implications of this work, the current Commentary discusses issues related to the assessment and utility of both self-reported and informant-reported SCCs.

While both SCCs and performance-based cognitive data have implications for older adults’ daily functioning, it is important that clinicians resist the urge to map self-reported cognitive difficulties (e.g. the “memory” failure of driving past an exit on a highway) to current objective cognitive test scores (e.g., on a verbal list learning test) in a 1-to-1 fashion, due to the very low statistical association between SCCs and current cognitive test scores (Burmester et al., Reference Burmester, Leathem and Merrick2016; Crumley et al., Reference Crumley, Stetler and Horhota2014). Low correspondence between self-reported and performance-based cognitive functioning may be related to the influence of personality traits and psychiatric symptoms (Burmester et al., Reference Burmester, Leathem and Merrick2016, Koller et al., Reference Koller, Hill, Mogle and Bhang2019; Pavisic et al., Reference Pavisic2021). Indeed, both anxiety and depression can lead to cognitive difficulties such as inattention and memory impairment. Therefore, older adults experiencing either of these conditions may report SCCs due to alterations in mood and emotional functioning rather than from traditional “neurological” conditions such as stroke or Alzheimer’s disease. In addition, psychiatric disturbance may result in a heightened state of sensitivity and reactivity to (i.e., fear of) cognitive errors that would otherwise be interpreted as normal. For example, depression often includes negative mental states and cognitive distortions, where people perceive their own abilities to be worse than they are. This can lead to interpreting normative memory failures (e.g. briefly forgetting the name of an old friend or distant relative) as catastrophic errors (e.g., reflecting Alzheimer’s disease). However, the relationship between psychiatric symptoms and SCCs is not unidirectional and perceived cognitive decline can also lead to the onset or exacerbation of anxiety/depression in some cases (Hill et al., Reference Hill2016; Reid and MacLullich, Reference Reid and MacLullich2006). This makes the task of determining etiology and differential diagnosis particularly difficult.

A thorough clinical assessment of both SCCs and current psychiatric functioning can help to distinguish people with deficits reflecting early markers of a neurodegenerative disease from those whose self-reported cognitive deficits are best explained by psychiatric factors (historically called “pseudodementia”). Indeed, accounting for psychiatric functioning when interpreting SCCs may improve clinical accuracy, even in the preclinical stage of a neurodegenerative condition (Pavisic et al., Reference Pavisic2021). With this in mind, providers are encouraged to consider and assess not only SCCs but also current psychiatric functioning when evaluating older adults. High rates of psychiatric distress at the time of reported SCCs may warrant cautious interpretation of concern and inclusion of collateral report. Ultimately, we recommend thorough cognitive and psychiatric evaluations that are repeated across time, with appropriate interventions for potentially reversible sources of cognitive impairment such as depression/anxiety and neuromedical syndromes (e.g., hypothyroidism, sleep apnea). Once medical and psychiatric factors are addressed and serial cognitive testing batteries have been completed (ideally, over the course of several years), diagnostic confidence in the presence/absence of a neurodegenerative condition is greatly increased.

Whereas the presence of self-reported SCCs in the preclinical and early stages of neurodegeneration may, in some cases, signify declines not yet detectable on cognitive testing, lack of self-reported cognitive concerns at later stages of decline should be interpreted cautiously. Indeed, neurodegenerative conditions are often marked by insidious onset and gradual progression that, in later stages, can interfere with metacognition (understanding one’s own thinking) and result in diminished insight and anosognosia (Jack et al., Reference Jack2010; Rosen, Reference Rosen2011). In other words, moderate to severe cognitive impairments can result in misperception of difficulties, and this may partially explain Truong and colleagues’ finding that self-reported SCCs had decreased predictive validity at the last measured time point, 10 years after the initial evaluation. Although initially unimpaired, declines in cognitive ability may have reduced the accuracy of participants’ self-report to some degree. This finding, juxtaposed with the persistent reliability of informant reports across time points, highlights the importance of including informant report, particularly in later stages of decline or disease. Consistently, evidence suggests that informants’ endorsements of SCCs correlate more closely with objective cognitive impairment and disease status than self-report (Buelow et al., Reference Buelow, Tremont, Frakey, Grace and Ott2014; Van Harten et al., Reference van Harten2018).

However, one cannot consider the utility of including informants without also acknowledging the limits of their reports, as evidence suggests that, while better at times, informant reports are also only weakly associated with clinical outcomes (Holland et al., Reference Hollands2015; Slavin et al., Reference Slavin2015). There are many interacting variables that predict complex outcomes such as pathological cognitive decline and skilled clinicians do not put too much emphasis on any one single indicator, including informant-based SCCs. Indeed, informant reports should always be considered in context and with acknowledgement that they are also susceptible to the influence of outside biasing factors. In other words, “informants” represents a heterogenous group of people with varying characteristics (e.g., age, health literacy, cognitive functioning, mental health, personality) that can impact a collateral report of a patient’s cognitive functioning. When evaluating an informant’s report, consideration should be given to several factors. First, the nature of the relationship between the informant and the patient will impact the report of SCCs, as some may give incorrect or incomplete assessments due to limited opportunities to observe the patient in daily life. Friends, children, or other relatives who have irregular contact with the patient may misrepresent the situation based on lack of recent experience with the patient. In contrast, while spouses/partners may have greater access to the patient and, therefore, more opportunities to observe them, they may be susceptible to their own cognitive limitations. Second, relational problems can occur between informant/caregiver and patient (e.g. caregiver burden, marital discord), which can color an informant’s perceptions of the patient’s cognitive functioning. Third, informants, like anyone, can experience their own life stressors, psychological symptoms, and personality disorders, which might impact their report of the patient’s cognitive symptoms. Fourth, in some cases, informants can be motivated by external incentives (e.g. financial gain) to present the patient in a negative cognitive light and will skew their report accordingly. All these factors should be considered for each patient.

In conclusion, there is growing recognition of the utility and limitations of self- and informant-reported SCCs in identifying those who may be at risk of developing objective declines in the future. The work by Truong et al. (Reference Truong2022) adds support for the potential utility of SCCs when measured by reliable and generalizable measures. Their findings and the extant literature suggest that both self- and other-reported SCCs can be important parts of a comprehensive, integrated assessment of an older adult’s cognitive and mental health status. In this Commentary, we suggest that consideration of the role and influence of affective states is vitally important when interpreting self-reported SCCs, while inclusion of corroboratory evidence from knowledgeable informants is particularly valuable in later stages of disease, where one’s ability to accurately report on their own difficulties may be diminished. Taken together, this work and the broader literature related to SCCs suggest that understanding each patient from multiple perspectives allows the clinician to paint a more complete picture of their functioning, to identify contributing medical and psychosocial factors, and, importantly, to tailor treatment recommendations to a particular patient’s view of themselves and their unique interpersonal context.