Subjective Executive Dysfunction in Patients with Primary Brain Tumors and their Informants: Relationships with Neurocognitive, Psychological, and Daily Functioning

Abstract

Objective:

We assessed agreement between patient- and informant-report on the Behavior Rating Inventory of Executive Function – Adult (BRIEF-A) in patients with primary brain tumors (PBT) and differences on BRIEF-A in neurocognitive (intact v. impaired), psychological (asymptomatic v. distressed), and functional (independent v. needing assistance) categories using both patient- and informant-report.

Method:

patients with PBT (n=102) completed neuropsychological evaluations including the BRIEF-A, clinical interview, neurocognitive tests, and mood questionnaires. Correlations between the BRIEF-A and Informant (n=39) were conducted. Differences in patient and informant BRIEF-A indices were investigated across five classifications: neurocognitive functioning, psychological functioning, medication management, appointment management, and finance management.

Results:

Patient and informant BRIEF indices were correlated. There was no difference on BRIEF-A or Informant indices for intact v. impaired neurocognitive status. Higher BRIEF-A and Informant indices were observed among psychologically distressed v. asymptomatic patients. Results showed higher BRIEF indices among those requiring assistance with medication, appointments, and finances.

Conclusions:

Patients and informants agreed in their reports of executive function (EF). These reports, while not different in neurocognitive classification, were different in psychological functioning and in those needing assistance with instrumental activities of daily living (IADL). Patient- and informant-reported EF may provide important data regarding psychological and IADL functioning in this population.

Introduction

The prevalence of cognitive dysfunction and its propensity to affect daily functioning are major challenges among those with primary brain tumors (PBT) (1-3). Impairments are caused in part by tumor burden, as well as the effects of standard treatments including surgery, radiation, and chemotherapy (4-6). Executive function (EF)—generally agreed to comprise higher-order, goal directed, and problem solving abilities that supervise the self-regulation of cognitive, emotional, and behavioral outcomes (7,8)—is a particularly common cognitive concern among those with PBT (9). Impairments in EF can negatively affect one’s ability to complete instrumental activities of daily living (IADLs) (10) such as financial planning, medication management, and appointment scheduling. EF impairments can also disrupt adjustment and coping, leading to frustration, sadness, fear, and reduced quality of life for individuals and their families (11,12)

Perceived executive dysfunction may provide a unique perspective into individual concerns and daily challenges (2,13). The Behavior Rating Inventory of Executive Function (BRIEF) (14); is a subjective measure of EF, with self- and informant-versions (BRIEF-A and BRIEF-Informant, respectively). Based on behavioral and neural correlates (15-17), the BRIEF was developed to capture perceived deficits in daily functioning and inform treatment of cognitive impairment across a wide range of populations (18). BRIEF-A profiles of individuals with PBT are reported to be similar to those of individuals with mild cognitive impairment related to traumatic brain injury, with the highest elevations in the domains of metacognition, i.e., working memory, planning and organization, and task monitoring (19).

Previous findings suggest that subjective cognitive function is more strongly related to self-reported mental and physical health rather than to specific objective neurocognitive tests in brain tumor samples (2,20,21). This is consistent with research in other medical populations, demonstrating poor agreement between measures of subjective and objective cognitive function, yet a consistent relationship between subjective cognitive function and psychological distress (22-25). This pattern raises the possibility that the discrepancy between objective and subjective cognitive dysfunction reflects poor insight on the part of the respondent; thus, the consideration of collateral informant reports adds significant value in understanding the needs of individuals with brain tumor. Generally, across studies in neuro-oncology, a significant agreement between patient- and informant-report has been found, even when patient cognition was impaired (20,26-28). Taken together, these studies suggest that neuro-oncology patients may have preserved insight, and the differences between objective and subjective cognitive dysfunction are attributable to some other cause. Thus, researchers urge continued consideration of both patient- and informant-reported EF as reliable and valid data points in neuropsychological examinations (20,26,27).

Despite this documented relationship with psychological distress, research on the BRIEF-A and other measures of subjective cognitive dysfunction have found these self-report instruments to be clinically useful in identifying those with cognitive and functional impairment. For example, the BRIEF-A independently contributed to distinguishing between healthy controls and those with attention deficit / hyperactive disorder (ADHD), underscoring its importance in assessing features of ADHD which were non-redundant to performance-based tasks (29). Among patients with Parkinson’s Disease, perceived EF explained variance in IADLS—such as managing medications and completing housework—that was unaccounted for by performance-based measures (30). One recent study, interestingly found that, among individuals with PBT and impaired cognition, a general measure of subjective cognitive concerns correlated significantly with neurocognitive tests (31). Nevertheless, most previous research in those with PBTs has failed to find a relationship between domain-specific neurocognitive tests of EF (e.g., Trail Making Test B) and patient- and informant-reported EF (2,20). But scales of these kinds may still be helpful in identifying those classified as impaired based on a composite of objective neurocognitive tests, rather than deficits observed on domain-specific tests, as the BRIEF-A focuses on difficulties in daily functioning that may be better captured by an overall composite of neuropsychological tests. This is especially relevant when considered alongside the evidence demonstrating the potential for preserved insight in neuro-oncology patients.

Thus, the aims of the present study were to: 1) replicate the findings of agreement between patient- and informant-report on the BRIEF-A in a heterogenous sample of patients with PBT and 2) build on previous work and investigate whether self- and/or informant-report on the BRIEF-A indices differed between patients with and without neurocognitive dysfunction, psychological distress, and/or functional dependence. Due to the consistent lack of relationship between the BRIEF-A and neurocognitive tests of EF (2,20), we did not attempt to replicate this null finding, rather, we sought to examine whether the BRIEF-A differed between patients classified as neurocognitively intact v. impaired. Based on previous literature, we hypothesized that patient and informant reports on the BRIEF-A indices would be congruent. Given potentially preserved insight among individuals with PBT, it was hypothesized that both patient- and informant-reported EF would differ between patients with intact v. impaired neurocognitive functioning. We also hypothesized that both patient- and informant-reported executive dysfunction would differ between those with psychological distress v. asymptomatic and between those independent v. in need of assistance with IADLs.

Methods

Design and Procedures

This cross-sectional retrospective study used archival data from baseline neuropsychological assessments conducted with patients with PBT who were referred by their medical team for cognitive testing. Relevant demographic and medical characteristics were gathered from patient medical records following the process of informed consent. If patients elected to participate, their data was included in a clinic-specific registry, which was approved by the ethical review board at Virginia Commonwealth University (HM 2005129). The present study reports on data collected as part of patients’ initial neuropsychological assessment spanning January 2015 to March 2020.

Patients and Informants

Eligibility criteria consisted of 1) confirmed PBT diagnosis via histopathological findings, 2) a minimum of 1-week post-surgical resection or biopsy (if applicable), 3) English speaking, and 4) available age-adjusted BRIEF-A normative data (ages 18-90). If patients were unable to complete the associated study measures (see measures below) due to severe language or visual impairment, they were excluded from this data set. All efforts were made for patients to be included in the study (e.g., reading questionnaires aloud or increasing font). Potential bias was reduced in this convenience sample by a high rate of referral to neuropsychology (assessments are recommended as clinic standard of care) and a high registry consent rate (98%). Adult individuals who accompanied the patient during their neuropsychological evaluation, were present during the clinical interview, and were aware of the patient’s daily functioning were identified as informants.

Measures

Behavior Rating Inventory of Executive Function – Adult and Informant versions. The BRIEF is a standardized questionnaire for patients and informants aged 18 to 90 years. It was designed to capture an individual’s view of their own EF in the everyday environment. The validity, reliability, and clinical utility of the BRIEF-A and Informant have been established in several medical populations with cognitive difficulties including mild cognitive impairment, dementia, Parkinson’s Disease, and traumatic brain injury (14,18,30,32). Both the BRIEF-A and Informant are 75-items and take approximately 10 minutes to complete. They yield an overall summary score, the Global Executive Composite (GEC), which is comprised of two indices: Behavioral Regulation Index (BRI) and Metacognition Index (MI). The BRI includes four scales: Inhibit, Shift, Emotional Control, and Self-Monitor. The MI includes five scales: Initiate, Working Memory, Plan / Organize, Task Monitor, and Organization of Materials. Each clinical scale includes 6-10 items. Reponses use a three-point scale, scored as follows: never = 1; sometimes = 2; and often = 3. Raw scores are converted to T-scores using age-corrected norms. Higher scores reflect greater difficulties experiences by the individual reporter. T-scores ≥65 indicate clinical significance.

Beck Anxiety Inventory (BAI) (33) is a 21-item self-report inventory of anxiety symptoms that is widely used in clinical settings and has demonstrated utility in diagnosing anxiety disorders as well as monitoring treatment progress (34,35). Each symptom is rated on a 4-point scale ranging from 0 to 3: 0 (not at all), 1 (mildly, It did not bother me much), 2 (moderately, It was very unpleasant, but I could stand it), and 3 (severely, I could barely stand it). The score for the 21 items is based on the patients’ responses in regard to each symptom over the past week, which is then summed to yield a single anxiety score. The following guidelines have been proposed to interpret the scores: 0-7 = minimal anxiety, 8-15 = mild anxiety, 16-25 = moderate anxiety, and 26-63 = severe anxiety.

Beck Depression Inventory – Second Edition (BDI-II) (36) is a 21-item self-report inventory that evaluates the severity of depressive symptoms in adolescents and adults. Similar to the BAI, the BDI-II is a widely used measure to assist in diagnosis of mood disorders, monitor depressive symptoms, and track treatment progress (37). Validity and reliability of the BDI-II have been well-established in previous work (38-40). Each symptom is rated on an item-specific 4-point scale ranging from 0 to 3, with total scores ranging from 0 to 63. The score for the 21 items is based on the patients’ responses in regard to each symptom over the past two weeks, which is then summed to yield a single score to determine the depression severity. The following guidelines have been proposed to interpret the scores: 0-13 = minimal depression, 14-19 = mild depression, 20-28 = moderate depression, and 29-36 = severe depression.

Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) (41) is a performance-based assessment of cognitive functioning composed of 12 subtests used to assess cognitive impairment across five domains: immediate memory, delayed memory, attention, language, and visuospatial-construction. The RBANS yields age-normed scores for each subtest and domain in addition to a total score based on a normal distribution curve (i.e., mean of 100 and standard deviation of 15). The RBANS is a well-validated tool (42) that has been used with patients with PBT in previous research (43,44). The RBANS has shown good sensitivity in identifying patients with dementia and characterizing unique cognitive profiles (45,46), it also shows clinical utility in identifying those with mild cognitive impairment from healthy controls (47,48). Further, it has four alternative forms to allow for retesting, making it clinically useful when tracking neurocognitive functions with serial testing and has been found to reliably predict cognitive change over time (49,50).

Trail Making Test B (TMT A & B) (51) was administered as a brief measure of EF. TMT A is a measure of processing speed and visual scanning. TMT B is a measure of processing speed, visual scanning, and task switching. In TMT A, participants are asked to connect numbers as quickly as possible, then in TMT B participants are asked to connect numbered and lettered bubbles, switching between numbers and letters. The raw score is the time to completion. When participants make errors the administrator corrects them and this is reflected in longer completion times. Raw scores are converted to normed scores, accounting for participant age, race, level of education, and gender using previously published normative data (53). Given that previous research has established the BRIEF-A is not related to TMT-B in neuro-oncology and other brain injured samples (2,20), we did not investigate this specific test’s relationship with the BRIEF-A. Instead, the TMT-B was one of several domains of cognitive function that determined neurocognitive classification.

Classification

Cognitive.

If patients performed ≥ 1.5 SD below the normative sample on any of the 5 indices of the RBANS, on TMT A, or on TMT B, they were categorized as cognitively impaired. This method aligns with previous work in brain tumor samples (54).

Psychological.

Psychological distress was based on the established severity groups for the BAI and BDI-II (cutoffs and ranges described above). Patients falling in the mild to severe range on either or both the BAI and BDI-II were considered psychologically distressed, those falling in the minimal range for both scales were considered psychologically asymptomatic (55-57).

Instrumental Activities of Daily Living (IADL).

IADLs were assessed during the clinical interview with patients, and when present, informants. Specifically, they were asked whether aspects of daily living were conducted independently within the home setting. Tasks included managing medication, schedule, and finances. Based on response, each task was coded as independent or assisted. If discrepancies were present between patient and caregiver report, or when there was a question about patient capability, a clinical neuropsychologist determined classification on each task using available information, behavioral observations, and clinical judgment. This method of determining functional status is standard practice for many neuropsychologists.

Statistical Analysis

Data met assumptions for analyses conducted here. Pearson correlations were conducted to determine the association between BRIEF-A and BRIEF-Informant indices. Paired-sample t-tests were run to compare the BRIEF-A and BRIEF-Informant indices. Neurocognitive scores were converted to z-scores. To determine covariates to be included in the models, chi-square and t-tests compared the five categorical variables of interest (cognitive functioning, psychological functioning, and medicine, appointment, and financial management) by demographic (age, gender, and education level) and disease-related variables (tumor grade, time since diagnosis, and history of resection, radiotherapy, and chemotherapy). Any covariates found to be significant were included in the following analyses. Next, multivariate analysis of covariances (MANCOVAs) compared differences in BRIEF-A indices (BRI and MI) across the five functional classifications: neurocognitive functioning, psychological functioning, as well as medication, appointment, and finance management. Then, the same MANCOVAs were run with BRIEF-Informant indices across the five functional classifications. There were no issues of multicollinearity between BRIEF indices. SPSS version 27 was used for all analyses; p < .05 indicated statistical significance.

Results

Sample Characteristics

Data from baseline neuropsychological assessments were available for 102 patients and 39 informant-reports. For analyses of IADL classification, there was data available for all 102 patients with PBT; due to variations in clinical assessments, 96 patients had full neurocognitive data and were included in analyses of cognitive classification, and 98 had full data on psychological measures and were included in analyses of psychological classification. Demographic and disease characteristics for the full sample of 102 patients are presented in Table 1. Group frequencies for each classification method are presented in Table 2.

Table 1.

Demographics of Patients with PBT (N = 102)

	Demographic	n (%)
Age		49.0 yrs (range 21-81 yrs; SD=14.8)
Gender
	Male	54 (52.9%)
	Female	48 (47.1%)
Race
	European American/White	87 (85.3%)
	African America/Black	13 (12.7%)
	Other	2 (2.0%)
Ethnicity	Hispanic	4 (3.9%)
Level of Education		15.2 yrs (range 8-20 yrs; SD=2.8)
Time Since Diagnosis		33.6 months (range 0 – 218 months; SD=52.9 months)
Tumor Type
	Atypical Meningioma	3 (2.9%)
	Oligodendroglioma	19 (18.6%)
	Astrocytoma	24 (23.5%)
	Glioblastoma Multiforme	45 (44.1%)
	CNS Lymphoma	5 (4.9%)
	Craniopharyngioma	2 (2.0%)
	Hemangioma	2 (2.0%)
	Ganglioglioma	1 (1.0%)
	Germinoma	1 (1.0%)
Tumor Grade
	Low	33 (32.4%)
	High	69 (67.6%)
Tumor Hemisphere
	Left	47 (46.1%)
	Right	41 (40.2%)
	Bilateral	11 (10.8%)
Tumor Lobe*
	Frontal	52
	Temporal	23
	Parietal	22
	Occipital	8
	Cerebellum	1
	Brain Stem	3
	Other	1
Seizure Activity
	History	47 (46.1%)
	Ongoing	19 (18.6%)
	Current Anticonvulsant	42 (41.2%)
Surgery
	Biopsy	24 (23.5%)
	Resection	78 (76.5%)
Hx of Radiation Therapy		50 (49.0%)
Hx of Chemotherapy		50 (49.0%)

Note. yrs = years, hx = history

^*tumor lobe is not necessarily mutually exclusive

Table 2.

Group frequencies for each classification method

		Patient-Report			Informant-Report
		N	BRI M(SD)	MI M(SD)	N	BRI M(SD)	MI M(SD)
Neurocognitive Function	Intact	66	53.4 (10.8)	57.1 (12.2)	23	48.7 (8.1)	52.0 (8.7)
	Impaired	29	56.8 (13.4)	60.3 (15.0)	15	50.7 (10.0)	56.4 (12.6)
Psychological Function	Asymptomatic	44	48.6 (8.8)	52.1 (11.2)	23	46.5 (6.9)	48.0 (7.9)
	Distressed	54	58.7 (11.2)	63.2 (12.2)	16	51.2 (9.6)	58.1 (10.7)
Medication Management	Independent	71	52.7 (11.3)	56.5 (13.1)	21	50.2 (10.3)	52.6 (12.3)
	Needs Assistance	31	58.7 (10.8)	62.7 (11.6)	18	49.2 (7.2)	56.0 (8.3)
Appt. Management	Independent	63	52.3 (11.7)	55.6 (13.2)	19	50.4 (10.5)	52.2 (12.7)
	Needs Assistance	37	58.8 (10.1)	63.6 (10.9)	19	48.5 (6.8)	55.2 (7.7)
Financial Management	Independent	60	53.0 (11.1)	56.2 (12.3)	18	47.4 (8.3)	48.9 (8.9)
	Needs Assistance	42	56.7 (11.8)	61.5 (13.2)	21	51.8 (9.1)	58.6 (10.2)

Note. N = number of patients / informants, BRI = Behavioral Regulation Index, MI = Metacognition Index, Appt. = Appointment, BRI and MI scores are normed per the published manual.

Patient- and Informant-Report Agreement

A total of 39 informant reports were gathered, resulting in 39 dyads available for analysis. Patient and informant report were significantly correlated on the BRI (r = .34, p = .04) and MI (r = .39, p = .02). Results of t-tests revealed no differences between patient and informant reports on the BRI or MI (see Table 3).

Table 3.

Differences on BRIEF-A indices between patients and informants (n = 39 dyads)

BRIEF-A Index	Patients M(SD)	Informant M(SD)	Mean Difference (SD)	p-value
Behavioral Regulation	52.6 (12.2)	49.8 (8.9)	2.9 (12.4)	.161
Metacognition	55.3 (11.9)	54.2 (10.7)	1.1 (12.5)	.577

Note. BRIEF-A = Behavior Rating Inventory of Executive Function – Adult

Patient- and Informant-Reported EF Differences on Functional Classifications

Neurocognitive Function.

Radiation [χ(1) = 5.70, p = .02] was the only significant covariate for neurocognitive classification, with irradiated patients more likely to be cognitively impaired. There were no differences on BRIEF-A or BRIEF-Informant indices between cognitively impaired v. intact patients.

Psychological Function.

Age [t (96) = 2.81, p < .01] was significant, with younger patients more likely to be in the psychologically distressed group. Psychologically distressed patients had significantly higher scores on the BRI [F (2, 95) = 12.82, p < .001] and MI indices [F (2, 95) = 13.32, p < .001] Based on informant reports, psychologically distressed patients had significantly higher scores on MI [F (2, 34) = 5.20, p = .01], but not on the BRI.

Medication Management.

Age [t (100) = −5.24, p <.001] was significant, with older patients more likely to need assistance. Patients who needed assistance with medications had significantly higher scores on the BRI [F (2, 99) = 4.33, p = .02] but not the MI. Based on informant reports, there were no differences between those who needed assistance and those who did not.

Appointment Management.

Age [t (100) = −5.00, p <.001] and education [t (100) = 2.02, p = .046] were significant covariates for appointment management, with older and less educated patients more likely to need assistance. Patients who needed assistance with appointments had significantly higher scores on the BRI [F (3, 96) = 4.13, p < .01] and MI indices [F (3, 96) = 3.81, p = .01). Based on informant reports, there were no differences between those who needed assistance and those who did not.

Financial Management.

Age [t (100) = −3.53, p = .001] was the only significant covariate for financial management, with older patients more likely to need assistance. There was no difference on the patient-reported BRI and MI indices between those who needed assistance and those who did not. Based on informant reports, patients who needed assistance had significantly higher scores on the MI [F (2, 36) = 5.61, p < .01] but not on the BRI.

Discussion

On a measure of subjective daily EF, there was agreement between patient-informant dyads with respect to both behavioral regulation and metacognitive abilities. The strength of this relationship was modest, but builds upon several other neuro-oncology studies demonstrating concordance between dyads on measures of subjective cognition, even when cognitive impairment was present (20,26-28). Together, these studies suggest that patients with PBT may exhibit preserved insight into their daily EF, despite the negative effects of brain tumors and their treatments on cognition.

Next, we sought to extend the current literature by testing whether patient- and informant-reported EF differed as a function of patient cognitive and/or psychological classification. As predicted, with respect to psychological functioning, both patient and informant reports on the BRI differed between psychologically distressed and psychologically asymptomatic patients. This is consistent with the subscales that compose the BRI, which include abilities critical to adaptive emotion regulation: Inhibit, Shift, and Emotional Control.

Given the possibility of preserved insight, we hypothesized that patient-reported EF would differ between those who were neurocognitively intact v. impaired. However, we found no evidence to support this hypothesis. This could reflect that, while the classification method was based on previously established criteria for grouping patients with brain tumors by cognitive status (54) and included several indices of neurocognitive function, it remained dependent on neurocognitive tests which have been consistently found to be unrelated to reported EF across neuro-oncology patients (2,20). We did not investigate the relationship between the BRIEF-A and TMT-B, a specific test of EF, due to the bounty of published findings that demonstrated no correlation between the two; instead, we hypothesized that a composite of neurocognitive functioning may be more representative of daily functioning, like that measured by the BRIEF-A. Previous work in neuropsychology has found that neurocognitive tests capture abilities that may not accurately reflect the cognitive demands of daily life functioning (58-60). Our findings concur that subjective cognitive measures, like the BRIEF-A, might be capturing daily EF impairments in a manner different than that captured by objective neurocognitive tests (2,20). When we consider this alongside the finding that both patient- and informant-reported EF differed between those classified as psychologically distressed v. asymptomatic, it may be that subjective EF is more related to psychological functioning than neurocognitive functioning.

Lastly, our findings demonstrated inconsistent patterns by which patient- and informant-reported EF distinguished between those needing assistance and those independent in their IADLs. Patient-reported EF on the BRI distinguished between those who needed assistance with medication management and those who did not. Both indices of patient-reported EF distinguished between those who needed assistance with appointment management and those who did not. For both medication and appointment management, patients who needed assistance reported higher subjective executive dysfunction. Patient-reported EF did not identify dependence in financial management; which, interestingly was the only IADL that informant-reported EF, on the MI, identified those needing assistance from those who did not.

These inconsistent results may reflect a unique characteristic of financial management relative to other IADLs measured in this study, which is reflected in previous research (61). One study found that financial management dependence, and not other IADLs, predicted the development of dementia 10 years later (62). Research has also demonstrated the multi-domain cognitive requirements for independent financial management including preserved EF and delayed memory, versus other IADLs, which may become routinized and thus require less demanding cognitive capability (63,64). Further, informant-reported cognition has been shown to predict task inefficiencies and errors on an ecologically-valid task of functional independence in those with mild cognitive impairment (65). Considered within this body of literature, our findings may reflect the more nuanced knowledge informants have of highly demanding tasks, like financial management, even when patient insight for other levels of functioning is largely preserved. For example, an informant may notice and correct errors in a patient’s financial management, unbeknownst to the patient; or an informant may point out that a patient is only able to manage independently due to auto-pay and more complex financial management would require assistance.

While the BRIEF-A and BRIEF-Informant are consistently related to one another in neuro-oncology samples, it appears that the two diverge when examining functional outcomes, particularly IADLs. The measurement of IADLs via the clinical interview may capture day-to-day task management that maps differently onto patient- versus informant-perceived EF. One could argue that these results might be due to low insight in patients with PBT or one member of the dyad underreporting on the BRIEF. However, neither of these explanations were supported by the rest of our findings or previous studies in neuro-oncology (20,26-28). The lack of congruence between patient- and informant-reported EF in distinguishing between functional classifications could be affected by relationship dynamics, unmeasured in the current retrospective study. For example, whether the informant aided the patient in managing their IADLs could affect how they responded on the BRIEF and thus its relationship to functional classifications. Depression has also been found to predict financial management and may affect a patient’s insight into their cognitive functional status (63,64), thus explaining these discrepant findings in IADL management and subjective EF between patient and informant report.

Overall, it appears that both patient- and informant-reported BRIEF indices have utility in providing additional information on functional dependence for patients with PBT. Given that these findings are novel in the literature base, there is a need to build upon these results. Future research should continue these investigations while adjusting for relationship type and length between patient and informant, as well as the use of ecologically-valid measures of IADLs.

It is also important to note that, in our sample, age was a consistent covariate in identifying those who need assistance with medication, appointment, and financial management. Research has shown older adults with cancer are at an increased risk for decline due to functional impairments (66,67). While not surprising, these results underscore the need for functional assessment in older neuro-oncology patients and encourage resource allocation to improve quality of life for patients and their families.

This study is limited by a small sample of informant-reported EF, resulting in limited statistical power to detect differences in analyses using the BRIEF-Informant. Further, IADL classification was based on the clinical interview, and, while it can yield useful and clinically relevant information regarding functional dependence, the analyses could be enhanced with an ecologically valid measure such as the Texas Functional Living Scale (68). Importantly, this was a retrospective research study and thus specific details regarding informants were unavailable. However, allowing informal caregivers to provide informant-report mirrors previous studies using the BRIEF-Informant in neuro-oncology (20) and ultimately does not negate the findings. Finally, studies of patients with PBT are not without important medical considerations, including effects of tumor grade and treatment history, covariates adjusted for in the present study. However, we did not have information on tumor size and we were underpowered to adjust for tumor type, given the heterogeneity, limiting the ability to control for these medical characteristics.

In conclusion, patient- and informant-reported EF on the BRIEF were significantly related to one another and both identified patients with psychological distress but failed to identify those with neurocognitive impairment. It appears that patient- and informant-reported EF are more closely related to psychological functioning, rather than neurocognitive functioning, which is consistent with previous work. Building on these findings, we also demonstrated that patient- and informant-reported EF differentially identified those needing assistance on functional outcomes. Inconsistencies between which IADLs were identified by the patient- and informant-report point to the high cognitive demands of financial management and potential ability for informants to identify nuances in daily functioning. These findings demonstrate that patient- and informant-reported EF contribute important information in the assessment of quality-of-life and functional status in patients with PBT. Measuring patient- and informant-reported EF may not distinguish between those with and without cognitive impairment, but instead provide supplemental neuropsychological data on psychological and daily activity impairment necessary for making treatment recommendations. Our findings add to a growing body of research suggesting the possibility of preserved insight in this population. As such, we caution against excluding patients with PBT from quality-of-life research based on assumptions regarding insight, as these practices have led to a significant underrepresentation of patients with PBT in behavioral and cognitive intervention studies (69-72) and have concerning implications for the development of behavioral interventions in a population needing support.