The Role of Self- and Informant-Reports on Symptoms and Impairments in the Clinical Evaluation of Adult ADHD

: Little is known about which clinical features may aid the differentiation between attention deﬁcit hyperactivity disorder (ADHD) and other clinical conditions. This study seeks to determine the role of self- and informant reports on symptoms and impairments in the clinical evaluation of adult ADHD and explore their association with objective neuropsychological test performance by examining data of 169 outpatients referred for a diagnostic evaluation of adult ADHD. Participants were assigned either to an ADHD group (ADHD, n = 73) or one of two clinical comparison groups, depending on whether they show indications (Clinical Comparison Group, CCG, n = 53) or no indications (Clinical Comparison Group—Not Diagnosed, CCG-ND, n = 43) of psychiatric disorders other than ADHD. All participants and their informants completed a set of questionnaires. Compared to the CCG-ND, the ADHD group obtained signiﬁcantly higher scores on ADHD symptoms, impulsivity, cognitive deﬁcits, and anxiety. Compared to the CCG, the ADHD group scored signiﬁcantly higher on ADHD symptoms but lower on depression. Further regression analyses revealed that self- and informant reports failed to predict neuropsychological test performance. Self- and informant reported information may be distinct features and do not correspond to results of objective neuropsychological testing.


Introduction
Attention deficit hyperactivity disorder (ADHD) is a neuropsychiatric childhood disorder that persists into adulthood in a sizeable proportion of individuals and is characterized by symptoms of inattention, hyperactivity, and impulsivity [1][2][3][4]. ADHD in adulthood commonly occurs together with symptoms of other forms of psychopathology, which is underscored by research revealing a rate of up to 60-80% of patients with ADHD being diagnosed with one or more comorbid psychiatric disorders, with anxiety disorders (34%) and mood disorders (22%) being the most prevalent ones [5][6][7][8]. Additionally, ADHD symptoms can be observed in patients with anxiety (20%) and mood disorders (17%) [9][10][11]. It has been demonstrated that symptoms of ADHD and comorbid conditions contribute to functional impairments that are commonly observed in this population, such as lower academic achievement [12][13][14], lower employment rate [15,16], poorer financial the purpose of differential diagnosis but that measures for other clinical conditions with which ADHD is commonly confused may be more promising. In this context, Paucke and colleagues reported a large-sized difference between adults with ADHD and adults with major depressive disorder in self-reported symptoms of depression, as assessed with the Beck Depression Inventory-II (BDI-II) [40]. However, this effect was not found in an earlier study comparing patients with ADHD and bipolar disorder [47]. Similarly, Nelson and Gregg showed that college students with ADHD and dyslexia could not be differentiated from each other and, also, not from college students not having any diagnosis based on self-reported symptoms of depression and anxiety [48].
Considering no firm conclusions can be drawn so far about the role of self-and informant reports on symptoms and impairments in the differential diagnosis of adult ADHD, more research is needed, especially comparing clinical samples from the same referral context. Thus, this study aims to advance our understanding of the role of subjective reports in the clinical evaluation of adult ADHD and for differential diagnostic purposes specifically and to provide clinicians recommendations on how to use and interpret the standardized self-and informant reports. This study employs a large sample of 169 individuals clinically referred to an ADHD outpatient assessment. All individuals completed a comprehensive battery of self-and informant report rating scales for symptoms and impairments, including ADHD symptom domains in childhood and adulthood, cognitive functioning, depression, anxiety, and impulsivity. By using a large clinical sample of individuals who all completed a comprehensive battery of self-and informant reports, this study aims to determine whether individuals meeting the diagnostic criteria of ADHD can be differentiated from relevant clinical controls in the same referral context by reported levels of symptoms and impairments. We expect (1) individuals diagnosed with ADHD to show more pronounced ADHD symptoms and impairments but less pronounced symptoms of depression and anxiety when compared to individuals not reaching diagnostic criteria of ADHD but showing evidence for other psychiatric disorders. However, when compared to individuals who did not show evidence for any psychiatric disorders, we expected (2) individuals diagnosed with ADHD to endorse higher symptom levels and more impairments on all scales applied. Furthermore, because of reliable findings showing symptom under-reported [35,49,50] or over-reported [37,38,51] in the self-report of individuals with ADHD, we expected (3) individuals diagnosed with ADHD to show a more pronounced discrepancy to informant reports compared to clinically referred individuals not reaching diagnostic criteria of ADHD. Finally, (4) on the basis of previous research questioning the relationship between subjective reports and an objective neuropsychological test performance [52][53][54][55], this study seeks to further define the role of subjectively reported complaints by relating symptoms and impairments to test the scores of cognitive functions.

Participants
A total of 248 participants were considered for inclusion in this study. All participants were recruited from the ADHD outpatient clinic of the Department of Psychiatry and Psychotherapy, LVR-Hospital Essen, University of Duisburg-Essen, Germany. Individuals were referred for a diagnostic assessment, because they were suspected of having ADHD by GPs, neurologists, psychiatrists, or by themselves. Qualified psychologists or psychiatrists performed a comprehensive assessment for all participants based on the criteria as outlined in the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition [2]. A semi-structured interview was conducted to evaluate ADHD psychopathology (i.e., the Wender-Reimherr-Interview and the Essen Interview-for-school-days-relatedbiography) [56,57]. Furthermore, all participants and their informants completed a battery of self-and informant report rating scales for symptoms and impairments commonly seen in ADHD, including ADHD symptoms in childhood, current ADHD symptoms, impulsiveness, anxiety, depression, and cognitive disturbances. Further, all individuals underwent cognitive testing using a battery of neuropsychological tests (see the Methods section for a detailed description of the applied measures). The diagnostic assessment also included objective measures of impairment (e.g. failure in academic and/or occupational achievement) and multiple informants (e.g. school reports, employer evaluation, partner or parent-reports) for all individuals.
Seventy-nine participants were excluded from data analysis for one of the following reasons, i.e., participants did not complete the diagnostic process, a formal diagnostic decision could not be established, or self-and informant report information was not assessed, resulting in a sample of 169 participants who were included in the final data analysis. All of those participants who were retained were assigned to one of three groups, i.e., the ADHD group (participants diagnosed with ADHD, n = 73), the Clinical Comparison Group (CCG; participants who did not fulfill diagnostic criteria for ADHD but showed evidence of one or more other psychiatric disorders; n = 53), and the Clinical Comparison Group-Not Diagnosed (CCG-ND; participants who did not fulfill diagnostic criteria for ADHD and were not diagnosed with any other psychiatric disorder; n = 43). Of those patients diagnosed with ADHD, 62 were diagnosed with the combined symptom presentation, and nine were diagnosed with the predominantly inattentive symptom presentation, while the symptom presentation of another two participants was not reported. Moreover, 27 patients diagnosed with ADHD showed evidence for one or more comorbid disorders, including mood disorders (n = 16), anxiety disorders (n = 3), addiction disorders (n = 6), personality disorders (n = 2), adjustment disorders (n = 3), obsessive-compulsive disorders (n = 2), intellectual development disorder (n = 1), mixed receptive expressive language disorder (n = 1), and autistic disorders (n = 1). Individuals in the CCG showed evidence for one or more psychiatric disorders other than ADHD, including mood disorders (n = 37), anxiety disorders (n = 4), addiction disorders (n = 15), personality disorders (n = 1), adjustment disorders (n = 1), obsessive-compulsive disorders (n = 1), eating disorders (n = 3), and schizoaffective disorders (n = 1). The characteristics of all the participants are presented in Table 1. A strict significance level of p < 0.01 was applied to control for alpha error inflation. Significant difference was observed between the groups in age, F (2) = 6.453, p = 0.002, but not in sex, χ 2 (2) = 4.500, p = 0.105, and not in educational level, χ 2 (8) = 17.268, p = 0.027. Compared to the CCG, patients with ADHD were on average significantly younger but did not differ significantly in sex and education level. The ADHD group did not differ significantly from the CCG-ND in either age, sex, or educational level.  3 Individuals were suffering from one or more psychiatric disorders other than ADHD: Mood disorders/anxiety disorders/addiction disorders/personality disorders/adjustment disorders/obsessive-compulsive disorders/intellectual development disorder/mixed receptive expressive language disorder/autistic disorders/eating disorders/schizoaffective disorders. * Statistically significant at p < 0.01.

Measures
The current study is part of a larger project on clinical and neuropsychological functioning of adults with ADHD in an outpatient referral context. Since the present study focuses on self-and informant-reported symptoms and impairments, it describes these instruments in detail. The role of objective neuropsychological test performance has been addressed in a previous study of our group on an overlapping sample [32]. The current manuscript is therefore restricted to a brief description of neuropsychological tests.
WURS-K. The German version of the Wender Utah Rating Scale (WURS-K) was administered to all participants to quantify self-reported retrospective ADHD symptoms [58][59][60]. The scale includes 25 items on a 5-point Likert scale ranging from 0 (not at all or very slightly) to 4 (very much). Participants were asked to rate each item based on their recall of experiences in childhood. Internal consistency (Cronbach's alpha) of this scale was excellent and reported to be 0.91. A sum score was calculated for the severity of ADHD symptoms in childhood.
ADHD-SR. The German version of the ADHD Self-Report Scale (ADHD-SR) is a selfreport scale used to assess the severity of current ADHD symptoms [61][62][63]. The ADHD-SR comprises 18 items corresponding to the diagnostic criteria of DSM-IV. Participants are asked to rate each item based on how often an ADHD symptom occurred over the past six months on a scale ranging from 0 (never) to 3 (very often). The internal consistency (Cronbach's alpha) of this scale was high and reported to be 0.90. A sum score was calculated for the severity of current ADHD symptoms.
CAARS. The Conners' Adult ADHD Rating Scales (CAARS) is a self-and informant report instrument that was developed to assist in the assessment of ADHD in adulthood [64]. The present study includes both the self-report (CAARS-S:L) and observer report form (CAARS-O:L). Each scale includes 66 items, which are rated on a 4-point Likert scale ranging from 0 (not at all/never) to 3 (very much/very frequently). Item scores are summed up to derive eight subscale scores, including inattention (CAARS_SR_IA and CAARS_OR_IA for the self-and other report, respectively), hyperactivity (CAARS_SR_HA and CAARS_OR_HA), impulsivity (CAARS_SR_IM and CAARS_OR_IM), problems with self-concept (CAARS_SR_SC and CAARS_OR_SC), DSM-IV: inattentive symptoms (CAARS_SR_DSM and, CAARS_OR_DSMI), DSM-IV: hyperactive-impulsive (CAARS_SR_DSMH and CAARS_OR_DSMH), DSM-IV: total ADHD symptoms (CAARS_SR_DSMT and CAARS_OR_DSMT), and the ADHD index (CAARS_SR_Index and CAARS_OR_Index). Internal consistency (Cronbach's alpha) of the CAARS was excellent and ranged from 0.74 to 0.95 [44].
STAI. The State-Trait Anxiety Inventory (STAI) is a self-report scale designed to measure the presence and severity of anxiety symptoms in adults. The inventory consists of 40 items, each rated on a 4-point Likert scale. Twenty items assess the presence of anxiety as an emotional state (state anxiety, STAI-S), whereas other 20 items assess individual differences in anxiety proneness as a personality trait (trait anxiety, STAI-T) [65,66]. Internal consistency (Cronbach's alpha) was reported to be 0.93 and 0.90 for state anxiety and trait anxiety, respectively [67]. The sum scores were calculated for both state and trait anxiety, with higher sum scores indicating higher levels of anxiety.
BDI-II. The Beck Depression Inventory-II (BDI-II) is a self-rated scale assessing the presence and severity of depressive symptoms in individuals aged 13 years and older. The BDI includes 21 items, each rated on a 4-point Likert scale [68]. To each item, participants are asked to select the statement that best characterizes their emotions and functioning in the past two weeks. The internal consistency (Cronbach's alpha) of the BDI-II was reported to be high (alpha ≥ 0.84) [69]. Scoring of the BDI includes the calculation of a total score, with high scores indicating more severe depressive symptoms.
BIS-11. The Barratt Impulsiveness Scale (BIS-11) is a self-report questionnaire designed to measure impulsiveness [70,71]. The scale consists of 30 items, each rated on a 4-point Likert scale ranging from 1 (rarely/never) to 4 (almost always/always). The internal consistency (Cronbach's alpha) of the BIS-11 was high (0.83) [71]. A total score was calculated for the BIS-11, with larger scores indicating higher levels of impulsiveness.
FLEI. The Questionnaire on Mental Ability (FLEI) was administered as a measure of subjectively experienced cognitive deficits [72]. In this scale, participants are asked to rate 35 statements regarding the presence of problems in attention, executive functioning, and memory in everyday life. The FLEI includes 35 items scored on a 5-point Likert scale ranging from 0 (never) to 4 (very often). The internal consistency (Cronbach's alpha) of the FLEI was high (0.94) [72]. A sum score is computed to indicate the severity of the cognitive deficits.
Neuropsychological test battery. A battery of neuropsychological tests was administered to all participants to assess several aspects of cognition, including selective function (Perceptual and Attention Functions-Selective Attention, WAFS), vigilance (Perceptual and Attention Functions-Vigilance, WAFV), working memory (N-back Task), interference (Stroop Interference Test), inhibition (Go/No-Go Test), figural fluency (5-Point Test-Langensteinbach Version), flexibility (Trail Making Test-Langensteinbach Version, TMT-L), planning ability (Tower of London-Freiburg Version, TOL-F), and task switching (SWITCH Task). All tests were retrieved from the test set Cognitive Functions ADHD (CFADHD), which is a computerized test battery assessing cognitive functions in which adults with ADHD commonly show difficulties [73,74]. The test variables recorded include the speed of responses (mean reaction time); variability of response times (SD of reaction time); and accuracy measures (i.e., number of omission errors, commission errors, correct responses, correctly produced patterns, or number of solved items). Based on the test variables, compound Z-scores per domain are computed, e.g., basic attention, working memory, inhibition/interference control, cognitive flexibility, and convergent/divergent thinking (for details, see Guo et al. 2020 [32]).

Procedure
The assessment of symptoms, impairments, and neuropsychological functions using the various approaches was part of the standard clinical procedure for all participants referred to the ADHD outpatient clinic of the department of psychiatry and psychotherapy, LVR-Hospital Essen, University of Duisburg-Essen, Germany. All participants signed a written informed consent that declares their agreement for their data being used for scientific purposes. Furthermore, approval for this study was obtained from the ethical review board of the medical faculty of the University of Duisburg-Essen, Germany (20-9380-BO). It was stressed to all individuals that agreeing to take part in this study was voluntary, unpaid, and would not affect their clinical assessment or treatment. All participants were asked to complete the battery of self-and informant report questionnaires and to perform the neuropsychological assessment around the date of their clinical interview. The clinical assessment, including all measures, took about four hours in total.

Statistical Analysis
Descriptive statistics of all measures of self-and other reports are presented per group. Furthermore, self-and other reports of the CAARS subscales for inattention, hyperactivity, impulsivity, self-concept, and total ADHD index were contrasted by calculating discrepancy scores per person (i.e., computing scores of self-reports minus other reports) and presenting the mean and standard deviation of the absolute discrepancy scores per group. After checking for assumptions of parametric testing, groups were compared on all measures employing ANCOVA and post-hoc pairwise comparisons between the ADHD group and the CCG and CCG-ND, respectively. As the assumption check for STAI variables indicated a violation of the normality assumption (presumably because of the small sample size resulting from missing values), results of ANCOVA were confirmed by nonparametric testing (Kruskal-Wallis Tests). Age was taken as a covariate to control for age differences between groups. A strict significance level of p < 0.01 was applied to control for alpha error inflation. Effect sizes (Cohen's d) were calculated for significant pairwise group differences to indicate the magnitude of findings. Age-adjusted mean scores were used for the calculation of effect sizes to control for age differences. Based on the interpreting guidelines for Cohen's d, d < 0.2 indicates a negligible effect, 0.2 ≤ d < 0.5 indicates a small effect, 0.5 ≤ d < 0.8 indicates a medium effect, and d ≥ 0.8 indicates a large effect [75]. Furthermore, to determine the validity of subjective reports in predicting an individual's diagnostic status, two binary logistic regression models, using backward elimination of predictor variables, were calculated for distinguishing the ADHD group from both the CCG and CCG-ND, respectively. Only scales that showed significant effects on the group were included as predictors in the models. To reduce the influence of multicollinearity between STAI-T and STAI-S, a new anxiety variable (STAI) was calculated by adding for each individual the scores of STAI-T and STAI-S and was included in a binary logistic regression analysis. Missing values (about 12% of the data) were replaced by group means for regression analyses to obtain a sufficiently large sample size.
Finally, the objective neuropsychological test performance per domain (i.e., basic attention, working memory, inhibition/interference control, cognitive flexibility, and convergent/divergent thinking) were presented per group and were compared using ANCOVA. In order to explore the predictive value of self-and other reports for objective neuropsychological test performance (following a clinically oriented order of prediction but not necessarily causal relationship), multiple linear regression models were computed for each group separately and the total group. To improve the power of regression models, binary correlation analyses were performed between all subjective reports and different aspects of objective neuropsychological test performance prior to the regression analyses. Only variables that were significantly correlated with the outcome in bivariate analyses were eventually included in multiple regression models. Additionally, backward elimination was used for potential predictor variables. All statistical analyses reported so far were performed using IBM SPSS 25 (IBM Corp. Released 2017. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp). Statistical power for the group comparisons was calculated with G-Power. These calculations indicate, based on a significance level of 0.05, large power to reveal medium (83.0%) and large effects (99.8%). However, in order to minimize the risk of running into a type-1 error, we decided to reduce the significance level to 0.01, which resulted in lower power for medium effects (63.0%) but retained high power for large effects (98.7%).

Results
Descriptive statistics and group comparisons (ANCOVA controlling for age, with post-hoc pairwise comparisons) of subjective reports of symptoms and impairments, as well as objective neuropsychological test performances, are presented in Table 2. Significant differences were found in symptom domains of ADHD, anxiety, depression, impulsiveness, and subjective experience of cognitive deficits, whereas no significant effects were observed in any of the discrepancy scores or in the neuropsychological test performance. Compared to the CCG, the ADHD group endorsed significantly higher symptom scores in the WURS-K, ADHD-SR, and CAARS-SR-IM (medium to large effects) and significantly less symptoms of depression (medium effect). The ADHD group did not differ significantly from the CCG in any variables of the informant report, anxiety, impulsiveness, and subjective experiences of the cognitive deficits (effects up to medium size). Compared to the CCG-ND, the ADHD group (and their informants) indicated significantly more symptoms in the WURS-K, ADHD-SR, CAARS-SR-HA, CAARS-SR-IM, CAARS-SR-Index, and CAARS-OR-HA (medium to large effects), significantly higher symptom scores in the STAI-T and STAI-S (large effects), BIS-11 (large effect), and FLEI (medium effect). Group differences between ADHD and CCG-ND on depression did not reach significance (negligible effect). Furthermore, two significant binary logistic regression models were obtained for the differential diagnosis between the ADHD group and both the CCG and CCG-ND, respectively. Regarding the differentiation between the ADHD and CCG groups, subjective reports had a significant predictive value for an individual's diagnostic status, χ 2 (4) = 47.54, p < 0.001, with 31.4% explained variance (Cox and Snell). This model correctly classified 75.4% of the individuals. The contribution of each scale to the model is presented in Table 3. Significant effects are observed for WURS-K, CAARS_SR_IM, and BDI-II. Regarding the differentiation between the ADHD and CCG-ND group, subjective reports again had a significant predictive value for an individual's diagnostic status, χ 2 (4) = 69.79, p < 0.001, with 45.2% explained variance (Cox and Snell) and 84.5% of the individuals correctly classified. Table 3 shows that WURS-K, CAARS_SR_HA, BDI, and STAI had significant effects on predicting the group membership.

Criteria of Neuropsychological Test Performance
Regression Models

Discussion
This study aimed to explore the role of self-and informant reports on symptoms and impairments in the clinical evaluation of adult ADHD and examine the predictive value of self-and informant reports for objective neuropsychological test performance.
Group comparisons revealed various significant effects of medium to large sizes between the ADHD group and both the CCG-ND and CCG. Compared to the CCG-ND, the ADHD group reported, consistent to our expectations, significantly more pronounced symptoms and impairments in most of the self-report scales. In line with previous findings, both the scales for current (ADHD-SR; CAARS) and retrospective ADHD symptoms (WURS-K) unfold to have discriminative value to distinguish individuals with and without ADHD [44,[76][77][78][79]. Further, in line with the core features of ADHD and commonly seen comorbidity, patients with ADHD indicated more pronounced symptoms of impulsivity [80,81], cognitive deficits [24,26,27], and higher levels of anxiety [82,83]. Against our expectations, however, depressive symptoms did not differ between the ADHD group and CCG-ND. This is surprising, considering the indications of a large number of comorbid mood disorders in the ADHD sample (16 of 73). The pattern of group differences between the ADHD group and the CCG-ND is also reflected in the logistic regression analysis, revealing that subjective reports significantly predicted an individual's diagnostic status with almost 45% explained variance. The strongest predictors were self-reported ADHD symptoms in childhood and current anxiety. The observed predictive value of anxiety is consistent with the widely recognized knowledge that anxiety is one of the major comorbidities of ADHD [82,83]. Retrospective ADHD symptoms in childhood seem to stand out in the diagnostic process and appear to be most informative to identify ADHD in adulthood in a psychiatric outpatient referral context [76]. We conclude that the exploration of childhood onset and continuity of ADHD symptoms through adolescence and adulthood should be done with care and deserves sufficient time and resources in the assessment of first-time ADHD diagnosis in adulthood.
Comparing individuals with ADHD to the group of individuals showing indications of other psychiatric disorders (CCG), fewer differences were observed, i.e., the ADHD group reported higher scores on their current (ADHD-SR; CAARS) and retrospective (WURS-K) ADHD symptoms but lower scores on depression (BDI). Hence, also in differentiation to patients having other psychiatric disorders, ADHD symptomatology seems to have a key role and appears to have clinical value [38,40,42]. Considering the fact that the majority of individuals in the CCG showed indications of mood disorders (37 of 53), higher depression scores appear logical and may serve as useful information to differentiate ADHD from depressive disorders. The discriminative value of self-reported ADHD symptoms in childhood and adulthood, as well as depressive symptoms, are underscored in the logistic regression analysis, with no other score adding significantly explained variance for group differentiation. Compared to other ADHD scales used in this study, the CAARS appeared to play a minor role in distinguishing the ADHD group from the CCG, which adds evidence to previous works stressing the limited value of the CAARS for a differential diagnostic purpose [84][85][86][87]. In contrast to differences in depression, no significant effects were found for the symptoms of anxiety, impulsivity, and subjective reports of cognitive deficits. As an explanation, one may consider that whether or not group differences in psychopathology can be observed may depend on the composition of samples (comorbid disorders to ADHD; diagnostic status of the clinical comparison group) and may be difficult to generalize. Additionally, nonsignificant differences on a variety of scale scores support the view that it is difficult to differentiate ADHD from other psychiatric disorders based on self-reported information only [88][89][90].
Furthermore, against our expectations, neither the informant report scores nor any of the discrepancy scores between the patients and their informants contributed to the differentiation between the ADHD group, CCG, or CCG-ND. Against previous evidence questioning, the reliability of patients' self-reports due to symptom under-or over-reporting of patients with ADHD [35,37,38,[49][50][51]91,92], the present study strengthens the role of patients' self-reports in the clinical evaluation of adult ADHD [33,34,81]. This must not be confused with the conclusion that informant reports are of no added value to the clinical evaluation of adult ADHD, as it has been repeatedly demonstrated that ratings from multiple informants on symptoms and impairments increase the accuracy of clinical evaluations [91,93]. Besides, even though no significant effects of discrepancy were observed on group levels, disagreement between different sources of information (e.g., including self-and informant reports) is believed to provide unique and crucial information in clinical evaluations [94,95]. On an individual basis, this means that, independent from the diagnostic group, disagreements between self-and other reports may, for instance, give an indication for noncredible symptom reporting and may represent a risk factor for an adverse outcome [96][97][98]. Such a disagreement between self-and other reports, however, may also indicate that patients lack awareness of their symptoms or that they may apply efficient coping strategies so that their significant others do not experience the full degree of their difficulties.
Finally, a number of significant regression models of subjective reports on objective neuropsychological test performances have been obtained in this study. However, considering the negligible to small proportions of explained variance, and no reoccurring patterns of prediction across the models, it can be concluded that subjective reports of symptoms and impairments have no meaningful predictive value for objective neuropsychological test performances [99,100]. While both subjective reports and objective test performance have been advocated to provide the clinician with valuable information in a clinical evaluation and treatment planning, they seem to be distinct and nonredundant sources of information and should not be treated interchangeably [53,55,91,[101][102][103][104]. It remains a challenge for future research to determine the value of any of the information derived in a clinical assessment to predict the sustainable change, long-term outcome, improvement in symptoms, impairments, and general well-being.

Limitations
Several limitations of this study must be considered. First, the representativeness of group differences in the various scales of psychopathology is difficult to determine, because this largely depends on the composition of groups regarding ADHD symptom presentation, comorbidity (ADHD group), and psychopathology (clinical control groups). Second, it must be stressed that the various comorbid disorders of patients with ADHD and psychiatric conditions of individuals in the CCG were suggested based on the clinical assessment for ADHD but were not confirmed by subsequent psychiatric evaluations per specific disorders. Third, this study is largely based on self-reports, which are known to be vulnerable to bias, including over-reporting, under-reporting, or careless responding [35,51,91]. Fourth, because data were collected in a clinical setting and not under strict controlled experimental conditions, administrative challenges resulted in the occurrence of some missing values depending on the variable type, and individuals were not selected based on matching group criteria. Finally, the subjective reports and neuropsychological test results were accessible to clinicians establishing a diagnosis; thus, these data were not completely independent of diagnostic decision-making.

Conclusions
This study highlighted the role of self-reports of ADHD symptoms in the clinical evaluation of adult ADHD in an outpatient referral context, giving particular emphasis on the role of ADHD symptoms assessed retrospectively from childhood. Compared to selfreports, informant reports, and, also, the discrepancy between self-and informant-reported information unfolds to have limited values for the differential diagnosis of ADHD. While a comprehensive assessment of individual strengths and weaknesses, clinical diagnosis, treatment planning, and outcome assessments typically benefit from a comprehensive approach, including various types of information, self-reports still seem to be most informative for a differential diagnosis. This study also demonstrates that the present inventory of symptoms and impairments differentiated ADHD from a clinical comparison group without psychiatric disorders more successfully than from a clinical control group with indications of other psychiatric disorders. Comorbidities and shared features between ADHD and other psychiatric disorders may presumably be the main reasons for this effect. Finally, our data confirm previous evidence showing that subjective reports of symptoms and impairments on the one side and objective neuropsychological test performances on the other side are distinct and nonredundant information. The roles of various types of information for clinical evaluations, the prediction of outcomes, and sustainable improvement still need to be determined.  Informed Consent Statement: All participants agreed and signed a written informed consent for their data being used for scientific purposes.

Data Availability Statement:
The data that supported the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to privacy and ethical concerns.