Adult ADHD and emerging models of maladaptive personality: a meta-analytic review

Background ADHD is a highly consequential disorder that is estimated to affect 2.5% of the adult population. Emerging models of psychopathology posit that disorders like ADHD can be usefully situated within general models of individual differences in personality, such as those recently implemented in the DSM and ICD for the diagnosis of personality disorder. Previous research and systematic reviews have linked adult ADHD to the personality traits Conscientious Inhibition and Negative Emotionality. However, there have been some inconsistencies in the literature and research embedding ADHD-personality connections in the DSM-5 and ICD-11 personality disorder models has been limited. The goal of this paper was to systematically review associations between adult ADHD and personality traits, organized within a maladaptive five factor framework. Method A comprehensive literature search yielded 13 papers whose effects were meta-analyzed. Results Results supported associations between ADHD and low Conscientious Inhibition and high Negative Emotionality. However, interesting patterns of variability were observed, potentially related to issues such as instrumentation and facet variation. Conclusion Results support the clinical application of personality assessment for suggesting risk for ADHD symptoms, and point to important directions for further research.

functioning in a wide range of areas, including dysfunction related to inattention, impulsivity and hyperactivity. Associations between personality and psychopathology are particularly strong when maladaptive trait measures, or instruments that focus on personality-related problems, are used [16].
Several previous reviews have documented associations between FFM personality traits and ADHD [17][18][19]. These reviews indicate that ADHD as a unified construct is most consistently related to low Conscientiousness, low Agreeableness, and high Neuroticism. This research also suggests some specificity in associations between inattentive and hyperactive/impulsive aspects of ADHD and certain traits [20,21]. Specifically, Inattention is positively correlated to Neuroticism and negatively to Conscientiousness; whereas Hyperactivity/Impulsivity relates negatively to Agreeableness and positively to Extraversion [17]. As such, there is considerable potential in using basic traits like those of the FFM, RDoC, HiTOP, AMPD, and ICD-11 to understand how individuals diagnosed with ADHD differ from individuals with other diagnoses in terms of underlying personality traits. The goal of the present study was to use meta-analysis to summarize associations between adult ADHD and create a bridge from non-clinical to maladaptive personality traits.

The integrated five-factor model
The FFM consists of the following five factors: Neuroticism which refers to the extent of negative emotions, i.e., sadness, fear, hostility, and emotional lability that the individual experiences [22]. Individuals high in this domain may be at risk for many different psychiatric and physical disorders, high comorbidity, lower quality of life, shorter lifespan and more extensive use of health care [23]. Extraversion concerns how outgoing or talkative the individual is in most situations. Core traits are sociability, assertiveness, positive affect, and activity level [22]. Low Extraversion, or introversion, may include social withdrawal, social detachment, intimacy avoidance, restricted affectivity and anhedonia [24][25][26] and extremely high Extraversion may represent personality pathology in sexual promiscuity, emotional intrusiveness, excessive self-disclosure and thrill-seeking behavior [27]. Openness to experience describes the depth and breadth of an individual's intellectual, artistic, and experiential life, with key facets such as aesthetic sensitivity, intellectual interests, and imagination [22]. Although open individuals generally tend to have greater psychological wellbeing, maladaptive Openness can be found in distinguishing between major depression and bipolar disorder as well as different variants of schizotypy [28]. Agreeableness concerns the extent of the individual's motivation for prosocial behavior and pleasant interpersonal relationships. Important traits are compassion, trust, and politeness [22]. Individuals low in Agreeableness tend to be critical, skeptical, try to push limits, express hostility and being condescending [28] . Agreeableness is also inversely linked to psychopathy and aggressive behavior [29,30]. Maladaptive variants of extreme Agreeableness can result in gullibility, submissiveness, clinging, subservience, servility etc. [28]. Conscientiousness refers to the individual's level of organization, ability to complete tasks, and persistence in achieving long-term goals. Key concepts are orderliness, self-discipline, and reliability [22]. However, both high and low Conscientiousness is associated with decreased functioning. Low Conscientiousness is characterized by disinhibition, irresponsibility, negligence, and rashness [31,32]; and inflexible high Conscientiousness can result in perfectionism, fastidiousness, punctiliousness, workaholism and other facets of compulsivity [33][34][35].
Research has shown that multiple models of both normal-range and maladaptive personality traits can be integrated into the structure of the FFM [36][37][38]. For instance, the trait Neuroticism is empirically and conceptually similar to the trait Negative Affectivity in the AMPD and ICD-11, Emotion Dysregulation in Livesley's Dimensional Assessment of Personality Pathology (DAPP [39];, and the traits Harm Avoidance and (low) Self-Transcendence in Cloninger's Temperament and Character Inventory (TCI [40];. Similarly, FFM Conscientiousness is similar to Disinhibition in the AMPD/ ICD-11, Compulsivity on the DAPP-BQ, and low Novelty Seeking and Persistence in the TCI. As such, in this study, we will conceptualize personality traits in terms of an Integrative Five Factor Model (IFFM) inclusive of instruments that are explicitly conceptualized as measures of the FFM, as well as those, such as the DAPP and TCI, that have commensurate scales (See Table 1). We will use domain labels that integrate normal-range and maladaptive aspects of each trait [37]: Negative Emotionality, Positive Emotionality, Openness, Conscientious Inhibition, and Agreeable Inhibition.

Conceptualizing ADHD within an integrative individual differences framework
Conceptualizing ADHD as a dimension or set of dimensions that can be placed within a general, evidencebased model of maladaptive personality traits has a number of potential clinical advantages [41][42][43][44]. For instance, this model may help provide a principled and efficient means for conceptualizing heterogeneity in the presentation of individuals with ADHD diagnosis. Specifically, previous research suggests that Negative Emotionality and low Conscientious Inhibition are most strongly related to inattention, whereas Agreeable Inhibition and Positive Emotionality maybe more strongly related to hyperactivity and impulsivity [17,[45][46][47]. Evidence-based models of individual differences also help explain comorbidity. Insofar as many disorders share underlying propensities for high Negative emotionality and low Conscientious and Agreeable Inhibition [48], these disorders should be expected to co-occur with inattentive, impulsive and hyperactive symptoms. Conceptualizing ADHD within a multidimensional personality model that can help provide explanations and order to patterns of heterogeneity and comorbidity may thus pave the way for more targeted treatments and research. As such, individual differences models such as the IFFM have the potential to improve the efficiency, validity, and utility of ADHD diagnoses [43].
Four previous studies summarize associations between personality traits and adult ADHD [17][18][19]47]. This study provides additional information not included in those studies in the following ways. First, this unlike some reviews [18], uses formal meta-analytic procedures to quantify associations between personality traits and ADHD. Second, six not previously reviewed studies are included in the current meta-analytic review [46,[49][50][51][52][53]. Third, whereas the cited earlier reviews have included both child and adult studies with varying approaches to the assessment of ADHD [17][18][19]47], the present review has only included papers with adult individuals and excluded studies without a clear description of a formal ADHD diagnosis using a stringent clinical assessment. Fourth, this study conceptualizes personality using various models that can be integrated within an IFFM framework.
In summary, the aim of this study was to quantify associations between ADHD and the traits of an Integrated FFM to better understand how ADHD fits into evidence-based models of personality and psychopathology, and ultimately enable improved clinical practice and research. Given the existing body of research in adult populations, we focused on traits from the FFM, DAPP, HEXACO [54], and TCI. Given that relatively few studies have distinguished between ADHD subtypes, we conceptualized ADHD as unitary construct.

Method
Literature searches were conducted in November 2019 via Cochrane, PubMed, PsychInfo and SCOPUS using the following keywords: (ADHD OR ADD OR ADDH OR Attention Deficit Disorder with Hyperactivity) AND (personality traits OR personality dimensions OR NEO OR five-factor model OR FFM OR BFI OR Temperament and Character Inventory OR TCI). The reference sections of each article retrieved was searched manually for any potentially relevant studies. Included studies were a) written in English; b) published after January 1, 2000; c) contained adult cases with a diagnosis of ADHD according to DSM-IV or DSM-5; d) used validated measures to assess adult ADHD; e) used established dimensional personality measures (FFM, DAPP, HEXACO, or TCI); and f) presented results in a format that allowed for straightforward conversion to the FFM i.e. into the Integrated Five-Factor Model (IFFM [37]; (Fig. 1).

Openness
Openness to experience Self-transcendence Psychoticism (AMPD only) Emerging personality models AMPD and ICD-11 factors are not represented in included studies of the current review

Conscientious Inhibition
As explained above, and illustrated in Table 1 All formal ADHD diagnoses were based on the DSM criteria, although different instruments were used to assess current and historical symptoms and impairments. All data included in this analysis examined group differences as they relate to personality traits. Control groups included community samples, non-clinical parents of children with ADHD, blood donors, clinical groups without ADHD, and population norms extracted from manuals ( Table 2). In five of the papers [17,46,51,58,62] correlations with the ADHD symptom clusters: Combined (C), Inattentive (I), Hyperactive/Impulsive (HI), were described. Only the Jacob [51] 2016 study presented individual data according to specific subtype of ADHD. Thus, results on individual ADHD type or presentation could not be aggregated and analyzed separately. However, indicative heuristic patterns are presented and discussed.

M (SD)
Correlation trait -ADHD dx TCI Self-transcendence 55,1 (12,9) 50,0 (10, 0) .22 TCI Novelty seeking 59,2 (10,9) 50,0 (10, 0) . metric, weighted by sample size, and analyzed separately for each personality dimension. Effect sizes, 95% confidence intervals and prediction intervals with a random effects model were calculated and illustrated in forest plots. Heterogeneity of effects was also calculated to estimate the robustness of results [65]. I 2 was used to evaluate the heterogeneity of mean effect sizes. An I 2 value above 50% can be interpreted as meaningful heterogeneity around the mean effect size [66].

Quality assessment of papers
The 14 papers were ranked for scientific quality according to an adaptation of QUADAS [67]. QUADAS was developed as a tool to systematically assess scientific papers that concern diagnostic tests. The following QUA-DAS items: a) design, b) sample size, c) relevant population, d) assessment method, e) relevant assessment instrument, f) sources of recruited cases, g) blinding, h) conflict of interest, and i) drop-out, were assessed by the first author (PJ) and two co-authors (TN and BS). The papers were first assessed independently and then discussed to reach consensus. Each aspect was scored from 0 to 2: unsatisfactory = 0; fair = 1; good = 2. The scores for the aspects were summarized for each paper to generate total scores between 0 and 18. The 14 papers ranged from 5 to 13. The paper with lowest score [5] was removed after quality assessment and subsequent preliminary analyses due to deviating both in quality assessment and being identified as an outlier in subsequent analyses, thus resulting in 13 papers that were included in the final analysis. The excluded paper did not meet the quality criteria (unsatisfactory = 0) for five of the nine assessed criteria: small sample, limited number of sources from which participants were recruited, no blinding, no description of conflict of interest, no description of drop-out. High drop-out rate or not specifying drop-out and insufficient blinding were the most common quality deficits in the other papers. Assessment methods and relevant populations were generally study strengths. The range of the included papers was 7 to 13 (median 10).

Results
Thirteen original articles consisting of 2023 unique adult subjects diagnosed with ADHD and 16,835 control participants, were included in this meta-analysis. Mean age ranged from 21.6 to 42.8 years; 54% of the included population identified as male. Most of the clinical cases of ADHD were recruited through psychiatric clinics, either general or specialized centers (see Table 2). A few clinical cases were recruited through advertisements. The meta-analytic analyses used in this review indicated a range of combined effect sizes from d = 0.15 (negligible) for Openness to d = 1.11 (large) for Negative Emotionality [68,69] (Fig. 2). The effect size for Conscientious Inhibition was also large (d = − 0.89). Effect sizes for Agreeable Inhibition and Positive Emotionality were more modest, d = − 0.39 and − 0.43, respectively. All associations with the exception of Openness were significant. Negative Emotionality and Conscientious Inhibition are thus the personality dimensions that were consistently and most strongly elevated in the ADHD samples in this review.
Even though the combined effect sizes for Negative Emotionality and Conscientious Inhibition were large, there was significant variability in the ADHD populations within these personality domains. Effect sizes for Negative Emotionality ranged from d = 0.50 (medium to large) to d = 2.06 (very large) (Fig. 3). An even greater range was observed for Conscientious Inhibition: d = 0.06 (negligible) for the Cloninger TCI facet Persistence The I 2 is a measure for the proportion of observed variance that reflects real differences in effect size. It is expressed as a percentage with a range from 0 to 100% [70]. All effect sizes for the various personality domains in the ADHD samples ranged from 85 to 97%. Heterogeneity decreased when FFM measures were used exclusively. The I 2 for FFM Negative Emotionality was 47% and I 2 for FFM Agreeable Inhibition was 64%.

Discussion
The results of this systematic, quantitative review suggest a robust connection between adult ADHD and the IFFM personality domains Negative Emotionality and low Conscientious Inhibition. Positive Emotionality and Agreeable Inhibition have somewhat smaller, although significant, associations with adult ADHD. These results are consistent with the review from Gomez et al. [47], with the exception that Gomez et al. observed a larger effect size for Agreeable Inhibition (d = −.64) than we did (d = −.39).
Low Conscientious Inhibition consists of features that are also implicated in the core symptoms of ADHD. These features are conceptualized in ADHD as difficulties in executive functions or self-regulatory processes [71,72]. Barkley [73] proposed that behavioral inhibition, consisting of: a) inhibition of pre-potent responses, b) stopping ongoing responses, and c) interference control, can be seen as a superordinate factor that affects several cognitive and behavioral modalities, eventually manifesting themselves in specific symptoms commonly found in ADHD. This construct is largely commensurate with Conscientious Inhibition as conceptualized in trait measures, thus providing a bridge between neuropsychiatric and quantitative trait approaches to conceptualizing underlying deficits in ADHD [14,48,74].
Negative Emotionality is elevated across many psychiatric disorders and is highly predictive of psychological and physical comorbidity as well as general quality of life [23,48]. Although symptoms of Negative Emotionality are not central to ADHD criteria, this review and other evidence clearly shows that emotional distress and dysregulation are common among individuals with ADHD  [75]. There is also evidence that high Negative Emotionality also predicts worse outcomes in individuals with ADHD [76][77][78].
Significant variability was observed in effect sizes between and within samples in the current review, as well as between instruments. For example, the high percentage in the I 2 measure, described above, indicates that there is a high degree of heterogeneity between studied populations within the different personality domains in the meta-analysis. Ideally, subgroup or moderator analyses would be performed to explain this variation. However, the number of studies available for this metaanalysis was too small to test moderation. Nevertheless, some potentially reliable patterns were observed.
First, when specific measures of FFM were extracted for separate analysis, heterogeneity decreased. This suggests more consistency when personality traits were measured with FFM instruments. One particularly important source of variability may have to do with differing facet models across those instruments. Facet variability is complicated by the interstitial nature of some traits, which leads different models to place similar facets on different domains. For instance, we conceptualized TCI Self-transcendence as an aspect of Negative Emotionality, although it also shares some features with Conscientious Inhibition. Future work exploring a broader range of traits, and conceptualizing personality at the level of facets, would shed significant light on this issue.
Second, stronger effects with psychopathology can generally be expected with maladaptive as opposed to normal range personality measures [16]. As maladaptive range measures are more likely to be used in clinical settings, associations with these measures are potentially both stronger, and more informative for clinical practice. As such, future research examining links between adult ADHD and maladaptive range personality measures would be particularly useful.
A third source of variability involves ADHD presentation, given that traits may differentially relate to inattentive, impulsive and hyperactive features [17,45,47,51,58]. Unfortunately, our relatively small sample size precluded moderator analyses of ADHD subtypes. Future studies with larger samples and symptom or subtypelevel assessments of ADHD are needed to further explore the heterogeneity of effects we observed here.

Clinical implications
Overall, these findings support the emerging view that personality, personality disorders, and neurodevelopmental disorders can be conceptualized within an integrated and evidence-based dimensional trait system. This supports the clinical application of personality trait assessment models according to the DSM and ICD Categorical psychiatric diagnoses are fraught with high comorbidity, complicating the clinical picture and reducing treatment specificity. Dimensional models provide an opportunity to integrate comorbidity into a more comprehensive and individualized clinical description, which makes treatment planning more in line with current clinical praxis. The current review suggests that the features of ADHD can be economically situated within evidence-based models of personality and psychopathology. As such, the clinical use of personality trait measures and the emerging dimensional personality models included in the AMPD and the ICD-11 proposal provide opportunities to implement a more individualized approach for patients with problems related to hyperactivity, impulsivity, and inattention. Specifically, multidimensional trait profiles may help clinicians estimate long-term risk for psychosocial consequences of ADHD, as well as the likelihood of other comorbid conditions. They also help provide explanations for heterogeneous clinical presentations and point to appropriate pathways for intervention [41][42][43]79].
In clinical practice it is useful to rely on heuristics to make efficient clinical decisions [80]. For example, these findings indicate that individuals with scores on personality measures that are more than one standard deviation higher on Negative Emotionality (NE) constructs or lower on Conscientious Inhibition (CI), in relation to non-clinical populations, may be at risk for ADHD or other disorders. Such heuristics can be helpful in screening because they both indicate risk for specific kinds of problems, while also efficiently providing a comprehensive portrait of the individual's strengths and weakness, which has significant value for making prognostic predictions and guiding treatment [17,[43][44][45].

Limitations
The literature on associations between adult ADHD and personality traits is scarce, thus limiting the degree to which the current results can provide robust information about these links. The existing literature, while sufficient to quantitatively summarize connections between adult ADHD and IFFM trait domains, limited our ability to test moderators such as ADHD subtype, personality measure, or various sample features, and it necessitated inclusion of measures with varying connection to the FFM or DSM/ICD personality disorder proposals. Some of our conclusions were based on assumptions about how to organize facets into maladaptive five factor domain [37,47] that have not been fully tested empirically. All personality measures were self-report. This is particularly important because some aspects of personality (developmental, variability, neurocognitive, reflective functioning) may be better assessed using other methods (e.g., informant reports, repeated reporting, implicit tests, neuropsychological testing). Also, the initial screening was performed only by the primary author, which might have risked possible exclusion of relevant studies. Subsequent multiple rater inclusion process was based on consensus; however, reliability statistics for how consensus was reached, were not calculated. Finally, the samples used in this study were WEIRD (i.e. White Educated Industrialized Rich and Democratic). Greater diversity in samples is needed to add confidence to these findings and enhance generalizability.

Conclusion
Psychiatric nosology is in a transition from legacy diagnostic categories to evidence-based models of individual differences that closely resemble personality traits [14,81]. Research on how ADHD fits into these models is relatively nascent, particularly in adult samples. This study shows that existing research on personality associations with ADHD suggests that ADHD is strongly related to Negative Emotionality and low Conscientious Inhibition, and moderately related to low Positive Emotionality and low Agreeable Inhibition. These results support the clinical assessment of personality traits in ADHD diagnosis, clinical care, and research, while also pointing to the need for further research to more specifically delineate how ADHD can be fit into the personality/psychopathology hierarchy.