Introduction

Personality is one of the main focuses of research on addiction, for various reasons. First, certain personality traits (primarily temperament traits) have a strong hereditary component1. Thus, personality is an important mediator of genetic factors in the development of substance use disorders (SUD)2. Second, the relative stability of personality traits, or at least the predictable course of their change3,4, allows researchers to track the inter- and intragroup variability in the personality traits of people with addiction in a meaningful way. Finally, the choice of personality traits as the main research topic in addiction studies resides in the fact that extreme levels of personality traits (too high or too low in value) can create a tendency to develop comorbidities, poor therapeutic response, recidivism, and chronicity5,6,7. The greatest number of studies on the relationship between personality and SUD are based on Cloninger’s psychobiological model and the five-factor model of personality8. Both models include the understanding of personality as a hierarchical, multidimensional construct that reflects relatively stable individual differences in affective, cognitive, and behavioral patterns, influenced by both biological and environmental factors9,10.

The seven-factor psychobiological model of personality, integrates findings from family and longitudinal developmental studies, psychometric research, and neurobiological investigations in both animals and humans9,11. The model conceptualizes personality as an interaction between four temperament dimensions—innate and biologically based—and three character dimensions, shaped through social learning and cognitive development. Temperament dimensions include:

  • Novelty Seeking: A dopaminergic system–based tendency to initiate behavior in response to novelty, reward signals, or cues for potential excitement. High NS individuals are curious, impulsive, enthusiastic, and disorganized.

  • Harm Avoidance: A serotonin-related predisposition to inhibit behavior in response to punishment or uncertainty. High HA is associated with shyness, pessimism, anxiety, and fatigue.

  • Reward Dependence: Reflecting noradrenergic sensitivity to social rewards, high RD is associated with warmth, attachment, sentimentality, and dependence on approval.

  • Persistence: The ability to persevere despite frustration or fatigue, now considered an independent trait. High P individuals are industrious, perfectionistic, and ambitious.

The three character traits represent aspects of self-concept:

  • Self-Directedness (SD): Reflects an individual’s capacity for self-regulation, responsibility, and goal orientation. High SD indicates maturity, self-determination, and the ability to adapt behavior to personal goals and values.

  • Cooperativeness (C): Describes the tendency to identify with and accept others. High C individuals are empathetic, tolerant, and helpful, in contrast to those who are self-centered and antagonistic.

  • Self-Transcendence (ST): Captures spiritual maturity and the capacity for transpersonal identification. High ST individuals experience themselves as integral parts of a greater whole and report feelings of connectedness and intuitive insight.

The Temperament and Character Inventory-Revised (TCI-R)12 operationalizes these seven personality dimensions with strong psychometric properties, including good internal consistency, factorial validity, and cross-cultural applicability13,14,15. It is widely used in research on substance use disorders and relevant studies repeatedly confirm Novelty seeking as a more pronounced personality trait of people with addiction in general15,16,17, while high Harm avoidance in combination with high Novelty seeking is a risk factor for alcoholism18,19.

The Five-Factor Model (FFM) of personality, also known as the Big Five, is the most extensively validated and widely accepted dimensional model of personality structure. It describes five broad, biologically influenced domains that capture stable patterns of thoughts, emotions, and behavior. Each of the five domains is composed of six more specific facets, allowing fine-grained personality profiling:

  • Neuroticism reflects emotional instability and sensitivity to negative stimuli. High scores are associated with a tendency toward anxiety, depression, and emotional reactivity.

  • Extraversion represents a proclivity for sociability, assertiveness, and positive affect. Individuals high in Extraversion are active, talkative, and seek stimulation.

  • Openness to Experience describes cognitive flexibility and openness to new ideas and experiences. High scorers tend to be imaginative, curious, and open-minded.

  • Agreeableness reflects prosocial tendencies such as empathy, trust, and cooperation. Highly agreeable individuals are altruistic, modest, and compassionate.

  • Conscientiousness captures self-discipline, goal orientation, and reliability. High Conscientiousness is associated with persistence, organization, and responsibility.

The model has demonstrated robust cross-cultural validity in over 50 nations and populations and is supported by behavioral genetic studies showing heritability estimates between 50% and 79%. FFM traits show remarkable long-term stability, particularly after the age of 3010,20. Numerous studies, employing the FFM in addiction research, have identified high scores on the Neuroticism dimension and low scores on Conscientiousness as defining characteristics of people addicted to opioids and/or alcohol8,21,22,23,24,25,26,27,28,29. Furthermore, a study conducted in a Serbian clinical population reported that women with opioid use disorder scored significantly higher on Neuroticism and lower on Conscientiousness compared to women with alcohol use disorder30.

Sources of variability in research on the personalities of addicted individuals

Certain variables, outside the space of applied personality models, are important to consider when interpreting the results of research on the personalities of addicted individuals. For example, the temperament dimension of Novelty seeking has been repeatedly confirmed as a general risk factor for the development of SUD15,16,17,31. However, the expression of this trait is significantly affected by the age, type (clinical/nonclinical) of the studied population and severity of addiction16,32,33.

Similar sources of variability have been documented across the dimensions of the Five-Factor Model. Normative developmental research demonstrates that Conscientiousness and Agreeableness tend to increase throughout adulthood, while Neuroticism and Excitement-Seeking (a facet of Extraversion) tend to decline34,35. Furthermore, lower socioeconomic status has been associated with a personality profile marked by higher Neuroticism and lower levels of Conscientiousness and Openness to Experience36,37. Additionally, variations in personality profiles have been linked to addiction severity and type. For example, individuals with more severe SUDs often exhibit more pronounced elevations in Neuroticism and lower Conscientiousness, while different substances may be associated with distinct trait patterns23,38. This underscores the critical need to account for socio-demographic variables (e.g., age, socioeconomic status) and clinical characteristics (e.g., addiction severity and substance type) to accurately interpret Five-Factor Model personality traits in addiction research.

Female sex and addiction

Sex differences in SUD, according to Becker and associates39, reflect the complex interactions between neurobiologically based sex differences and the effects of sociocultural factors. Historically, these differences were largely overlooked, with most research conducted in male populations, predominantly among men with alcohol use disorders. Although contemporary epidemiological studies unequivocally indicate an increase in the incidence and prevalence of SUD in women40, especially during pregnancy41,42 the number of sex-sensitive studies in this field is disproportionately small.

This gap in research is problematic given the evidence that women differ from men in both their vulnerability to and experience of addiction. Women tend to escalate drug use more rapidly and are more susceptible to long-term health complications associated with substance use39,43. Withdrawal profiles also vary: during drug abstinence, women experience more intense symptoms than men, whereas alcohol withdrawal in women often presents with milder clinical features44.

In addition to biological vulnerability, women with SUD face considerable sociocultural burdens. Difficulties in fulfilling socially expected roles, particularly those related to motherhood and caregiving, increase the stigma surrounding SUD in women39,45. Many women lack supportive networks, often due to abusive relationships or partners who also misuse substances46. These conditions contribute to elevated stress sensitivity and a higher risk of relapse39. Economic inequalities—reflected in lower employment rates and income levels—further limit women’s access to treatment services47. As a result, women often seek professional help at more advanced stages of addiction, presenting with more severe physical and psychosocial consequences16,48. These intersecting challenges affect not only access to treatment but also adherence and long-term recovery46.

Research question

The primary objective of this study was to examine whether significant differences exist in personality profiles among three groups of women—those with alcohol use disorder (AUD), those with opioid use disorder (OUD), and healthy controls (HCs) — by employing two widely recognized theoretical frameworks: the Five-Factor Model and Cloninger’s psychobiological model of personality. A secondary objective was to investigate whether a composite of selected personality dimensions could reliably differentiate group membership (AUD, OUD, or HCs).

Hypotheses

  1. 1.

    Significant differences will be observed among the three groups (AUD, OUD, and HCs) across the investigated personality dimensions, specifically in the temperament and character traits outlined by Cloninger’s psychobiological model, as well as the basic personality dimensions defined by the Five-Factor Model.

  2. 2.

    A multivariate combination of personality dimensions will significantly and consistently differentiate women with AUD, women with OUD, and healthy controls.

Methods

The research was conducted as a cross-sectional study. The sample consisted of women addicted to alcohol and opioids, and treated at the Special Hospital for Addiction Disorders in Belgrade (a specialized inpatient facility offering both detoxification and initial rehabilitation) from June 2014 - February 2016. The inclusion criteria for women treated for SUD in this study were 26–50 years of age and a confirmed diagnosis of alcohol and opioid use disorder (based on the ICD-10 classification criteria). Furthermore, the inclusion criteria also comprised an established abstinence from alcohol i.e. opioids for at least 10 days before entering the research study. This duration was selected based on clinical guidelines indicating that acute withdrawal symptoms typically subside within the first 7–10 days, allowing for a more stable emotional and cognitive state at the time of psychometric assessment. Abstinence was verified through a combination of toxicological urine screening and continuous clinical monitoring by trained psychiatric staff during inpatient treatment.

The study did not include subjects with unstable somatic or neurological conditions that may impair cognitive functioning and the validity of psychological assessments, such as epilepsy, traumatic brain injury, neurodegenerative diseases, or decompensated hepatic insufficiency, a comorbid SUD, or psychiatric comorbidities (schizophrenia-spectrum disorders, unipolar or bipolar affective disorders, anxiety disorders, and organic mental disorders). This applied to both the subjects from the clinical sample and the subjects from the control group.

The control group consisted of women from the general population of [withheld for review], selected via random sampling from the database of all residential subscribers of fixed telephony, provided that they have no history of abuse/addiction to alcohol and drugs, nor do they currently abuse these substances. All participation was voluntary and conducted under institutional ethical approval without financial compensation. Based on these inclusion and exclusion criteria, three samples of equal size (N = 80) were formed - women with (AUD), women with (OUD), and a sample of women from the nonclinical population with no history of abuse/addiction, i.e. healthy controls (HCs). To detect a statistically significant difference among the three groups at an alpha level of 0.05 and a statistical power of 0.80, a sample size of 76 participants per group was required to identify significant differences in at least five out of the seven measured dimensions49.

Data were collected on the respondents’ age, educational level, professional status, marital status, and number of children via semi-structured interviews. This questionnaire included questions about somatic, neurological, and psychiatric status; personal and family history of drug and alcohol abuse/addiction.

To assess the personalities of the respondents, the revised Temperament and Character Inventory (The Temperament and Character Inventory-5-Revisited; TCI-5-R)12, a self-report questionnaire based on Cloninger’s psychobiological personality model50, and the revised Neuroticism Extraversion Openness Personality Inventory (Neuroticism, Extraversion, Openness Personality Inventory-Revised; NEO PI-R)51, which is based on the Five-factor model of personality52, were used.

The TCI-R evaluates seven dimensions of personality, comprising four temperament dimensions—Novelty Seeking (NS), Harm Avoidance (HA), Reward Dependence (RD), and Persistence (P)—and three character dimensions—Self-Directedness (SD), Cooperativeness (C), and Self-Transcendence (ST).

Each broad dimension consists of several narrower facets:

  • NS: Exploratory Excitability-Rigidity, Impulsiveness-Reflection, Extravagance-Reserve, Disorderliness-Determination.

  • HA: Anticipatory Worry-Optimism, Fear of Uncertainty-Confidence, Shyness-Inhibition, Fatigability-Vigor.

  • RD: Sentimentality-Indifference, Attachment-Detachment, Dependence-Independence.

  • P: Industriousness-Laziness, Work Hardened-Susceptible, Ambitious-Unmotivated, Perfectionism-Pragmatism.

  • SD: Responsibility-Blaming, Purposefulness-Helplessness, Resourcefulness-Passivity, Self-Acceptance-Self-Striving, Congruent Second Nature-Incongruence.

  • C: Social Acceptance-Intolerance, Empathy-Indifference, Helpfulness-Unwillingness, Compassion-Cruelty, Principles-Selfishness.

  • ST: Self-Forgetfulness-Self-Awareness, Transpersonal Identification-Isolation, Spiritual Acceptance-Materialism.

The instrument contains 240 items rated on a 5-point Likert scale, ranging from 1 (definitely false) to 5 (definitely true). The Serbian standardized version of the TCI-R was employed, validated on our population14.

The TCI-R demonstrates satisfactory reliability and validity in both international and local contexts. Internal consistency (Cronbach’s alpha) for the broad temperament and character dimensions typically ranges from 0.70 to 0.85, with facet-level alphas varying owing to fewer items per scale but generally acceptable for research purposes. Test-retest reliability coefficients over several weeks to months commonly exceed 0.70, indicating good temporal stability12,13.

The NEOPI-R51 evaluates five broad personality domains: Neuroticism (N), Extraversion (E), Openness to Experience (O), Agreeableness (A), and Conscientiousness (C). Each domain consists of six distinct facets, with each facet represented by eight items, enabling a comprehensive assessment of personality traits.

The facets under each domain include:

  • N: Anxiety, Hostility, Depression, Self-Consciousness, Impulsiveness, Vulnerability.

  • E: Warmth, Gregariousness, Assertiveness, Activity, Excitement-Seeking, Positive Emotions.

  • O: Fantasy, Aesthetics, Feelings, Actions, Ideas, Values.

  • A: Trust, Straightforwardness, Altruism, Compliance, Modesty, Tender-Mindedness.

  • C: Competence, Order, Dutifulness, Achievement Striving, Self-Discipline, Deliberation.

It comprises 240 items rated on a 5-point Likert scale (ranging from 1 = strongly disagree to 5 = strongly agree). For example, items include statements such as “I often feel tense or jittery” (Neuroticism: anxiety) and “I am organized and thorough” (Conscientiousness: competence), which participants respond to according to their degree of agreement.

The NEO PI-R has demonstrated excellent psychometric qualities in both international and local contexts. In this sample, internal consistency coefficients (Cronbach’s alpha) for the five domains ranged from 0.86 to 0.92, indicating high reliability, while facet scales showed acceptable reliability between 0.54 and 0.83, consistent with previous research. Test-retest reliabilities for domain scores typically exceed 0.80 over multiple years, supporting the temporal stability of measured traits. Validity evidence includes strong convergent and discriminant validity confirmed across diverse populations. The standardized Serbian version applied here53.

After the goal and procedure of the study were clarified to the respondents and written consent for participation was obtained, semi-structured interview and the stated questionnaires were applied.

Group differences in sociodemographic characteristics were assessed with chi-square tests, and differences in personality dimensions scores were evaluated using one-way ANOVA with post-hoc Šídák tests for pairwise comparisons. The influence of specific sociodemographic confounders was further examined using one-way analysis of covariance (ANCOVA). To comprehensively adjust for multiple potential confounders, we employed multivariable linear regression models controlling for age, education, marital status, employment status, number of children, and family psychiatric history. Lastly, the ability of the personality dimensions to predict group affiliation was tested via canonical discriminant analysis (CDA).

Results

Sample structure

As shown in Table 1, the three groups of respondents differed significantly across all the examined sociodemographic characteristics.

Table 1 Socio-demographic characteristics.
Full size table

ANOVA revealed significant differences (F = 35.85, p < 0.001) in mean age: AUD group had the highest (43.17 ± 6.03), HCs followed (37.50 ± 9.21), and OUD group had the lowest (33.99 ± 4.76), with post hoc Tukey tests confirming significant pairwise differences (Supplementary Table S1). Pairwise comparisons further showed that women with OUD differed significantly from both AUD and HC groups in education level and employment status, having lower education levels and higher unemployment (Supplementary Table S2). Marital status and number of children also differed significantly between the AUD group and both OUD and HC groups, with AUD women being less often single and less frequently without children compared to the other groups. Differences between the AUD group and HCs were significant for education level (with more university degrees in the AUD group), but not significant for employment status. A positive family history of SUD was significantly more common in both clinical samples than in the HCs (20%) with no significant difference between OUD (69%) and AUD (66%).

Dimensions of psychobiological model

The mean values of the TCI-R dimensions for the three groups of respondents, as well as the significance of differences between groups, are shown in Tables 2 and 3.

Table 2 Descriptive characteristics of study groups on temperament and character dimensions.
Full size table
Table 3 Sidak post hoc pairwise comparisons of TCI-R dimensions: effect sizes and Power.
Full size table

Concerning temperament traits, a significant difference was observed only in Novelty Seeking. Compared to women with AUD and HCs, women with OUD showed significantly higher scores in this domain. Although women with AUD, compared to HCs, showed a higher mean score for Novelty Seeking, the difference was not statistically significant.

At the level of character traits, significant differences were detected in the dimensions of Self-Directedness, Cooperativeness, and Self-Transcendence. Women with OUD attained significantly lower scores on Self-Directedness and Cooperativeness, and higher scores on Self-Transcendence compared to women with AUD and HCs. Women with AUD scored similarlyto the control group on these dimensions, with no statistically significant differences observed. Post hoc power analysis was conducted for all TCI-R personality traits, including those without statistically significant group differences. Large effect sizes and high statistical power were observed for comparisons involving Novelty Seeking (OUD vs. HC: d = 0.87, power = 1.00; OUD vs. AUD: d = 0.71, power = 0.99) and Self-Directedness (OUD vs. HC: d = 1.37, power = 1.00; OUD vs. AUD: d = 1.03, power = 1.00). Moderate effects were found for Cooperativeness (OUD vs. AUD: d = 0.61, power = 0.97) and Self-Transcendence (OUD vs. HC: d = 1.00, power = 1.00). In contrast, traits such as Harm Avoidance, Reward Dependence, and Persistence showed small effect sizes (d < 0.30) and correspondingly low statistical power, indicating limited ability to detect differences in these domains given the present sample size.

Additionally, a one-way analysis of covariance (ANCOVA) was conducted to examine differences in TCI-R personality traits among the three participant groups, using year of birth as a covariate. After controlling for age, all previously observed between-group differences in personality dimensions remained significant. In multivariable linear regression models, other sociodemographic factors —marital status, education, employment, number of children, and family history of mental illness—were not significant predictors of personality dimensions after adjustment for substance-use group (Supplementary Table S3).

A CDA was used to test the possibility of predicting group affiliation based on temperament and character traits (Supplementary Table S4). A single significant function emerged (Λ = 0.629, χ²(14) = 108.43, p < 0.001), explaining 93% of the discriminant variance. This function was mainly defined by lower Self-Directedness (r = − 0.785) and higher Self-Transcendence (r = 0.631) and Novelty Seeking (r = 0.522) (Supplementary Table S5). Group centroids showed that women with OUD (1.007) were clearly separated from HCs (–0.655) and women with AUD (–0.352), with distances of approximately 1.66 and 1.36 standard deviations, respectively, whereas the AUD group was located nearer to the HCs (≈ 0.30 SD) (Supplementary Table S6). Based on this function, 57.1% of the cases were successfully classified (Supplementary Table S7).

Dimensions of FFM

One-way ANOVA revealed significant differences in mean scores across all five FFM dimensions among the three groups: women with OUD, AUD, and HC (Table 4). Post hoc pairwise comparisons showed that women with OUD scored significantly higher (p < 0.001) on Neuroticism compared to both the AUD and HC groups, with medium-to-large (d = 0.61) and large effect sizes (d = 0.91), respectively. Similarly, mean scores for Agreeableness and Conscientiousness were significantly lower (p < 0.001) in the OUD group compared to both the AUD group and HCs, also medium-to-large (d = 0.71 and 0,65) and large effect sizes (d = 0.87 and 0.83), respectively.

Table 4 Descriptive characteristics of NEO PI-R domains of study groups.
Full size table

In the Extraversion domain, women with OUD scored higher than those with AUD (d = 0.57; p = 0.001), but did not differ significantly from HCs. For Openness to Experience, significant differences favored the OUD group compared to HCs (medium effect sizes: d = 0.53 ; p = 0.003), but did not differ significantly from women with AUD (Table 5). At the dimensional level, AUD patients did not differ significantly from healthy controls. In contrast, post-hoc facet analyses within Neuroticism indicated significant differences for Depression (p = 0.001, d = 0.39, 95% CI [0.09, 0.69]) and Self-Consciousness (p = 0.004, d = 0.26, 95% CI [–0.04, 0.56]). Across most dimensions, women with AUD occupied an intermediate position between women with OUD and HCs, with the exception of Extraversion, where they scored lowest among the three groups. Statistical power exceeded 90% for most significant comparisons, ensuring robust detection of large effects, whereas lower power was observed for non-significant differences. After controlling for age, all previously observed group differences in personality dimensions remained statistically significant, except for Extraversion. Multivariate linear regression analyses indicated that other sociodemographic factors, including marital status, education, employment, number of children, and family history of mental illness, exerted non-significant effects on NEO-PI-R dimensions when substance use group was accounted for in the model (Supplementary Tables S8).

Table 5 Sidak post hoc pairwise comparisons of NEO PI-R domains: effect sizes and power.
Full size table

A CDA was utilized to test the possibility of predicting group affiliation based on the dimensions of the basic structure of personality (Supplementary Tables S9 and S10).

Two discriminant functions were obtained, of which only the first (Function 1) was statistically significant (Λ = 0.709, χ² (10) = 80.98, p < 0.001). This function was defined by higher Neuroticism (r = 0.615) and lower Agreeableness (r = − 0.615) and Conscientiousness (r = − 0.571), with a smaller positive contribution of Openness (r = 0.349). Women with OUD were located at the positive pole of Function 1 (centroid = 0.841), whereas healthy controls (HCs) were located at the negative pole (centroid = − 0.520), representing a distance of approximately 1.3 standard deviations. Women with AUD (centroid = − 0.321) were positioned closer to HCs than to women with OUD (Supplementary Table S11). Based on this function, 56.7% of cases were correctly classified (Supplementary Table S12).

Discussion

Women with opioid use disorder

Women with OUD are predominantly younger, unmarried, childless, less educated, and unemployed. The earlier onset of OUD is an expected finding. In fact, unlike men, women, owing to greater biological sensitivity, tend to progress faster from initial experiences with opioids to the development of an addiction39. Earlier onset of this form of addiction may also suggest a greater genetic burden, which is a common finding in opioid-related disorders54. This finding is supported by our research, which detected a significantly higher rate of prevalence of family history of SUDsin this group of respondents.

Women with alcohol use disorder

Compared to women with OUD and HCs, women with AUD were characterized by greater average age, higher levels of education, and higher employment rates. In addition, this group had the highest number of married respondents, while simultaneously they showed greater marital instability; i.e., in this group the tendency towards separation and divorce was by far the most common. This may be one of the reasons why they have fewer children than women do in HCs. Instability in partnerships or marital relationships can be a source of intense stress, where alcohol use disorders have a symptomatic function44,46. However, due to the cross-sectional design of this study, it is not possible to determine whether such instability is a cause or a consequence of the disorder.

In terms of personality dimensions, the success of the classification of respondents (AUD, OUD, or HC) based on applied personality inventories was similar – 57.1% for the TCI-5-R and 56.7% for the NEO PI-R. This gives researchers and practitioners a certain degree of freedom in choosing instruments or models of personality in research, evaluation, and treatment.

Cloninger’s psychobiological model of personality

According to the results of our study, the three groups of respondents differ in the dimensions of Novelty Seeking (temperament dimension) as well as all three character dimensions: Self-Directedness, Cooperativeness, and Self-Transcendence. Moreover, women with OUD have significantly higher scores on the Novelty seeking and Self-Transcendence dimensions, and significantly lower scores on the Self-Directedness and Cooperativeness dimensions, compared to the other two groups. The results of the CDA suggest that women with OUD can be clearly distinguished from both healthy controls and women with AUD on the basis of specific personality profile characterized by lower Self-Directedness in combination with higher Self-Transcendence and higher Novelty Seeking. In contrast, women with AUD were positioned closer to healthy controls, with substantially smaller differences, suggesting that their personality structure may not deviate as markedly from non-clinical populations.

The results of relevant studies consistently show that temperament dimension of Novelty Seeking - is a general risk factor for the development of addiction15. Our findings partially align with with previous research reportingsignificantly higher scores in Novelty Seeking and Self-Transcendence and significantly lower scores in Self-Directedness among individuals with OUD compared to both alcohol-dependent individuals and healthy controls19. However, unlike Milivojević et al., who also identified a significant difference in mean scores of Reward Dependence and Harm Avoidance between the groups, which we did not observe. Additionally, while they found significantly higher Novelty Seeking in individuals with alcohol use disorder compared to controls, in our sample this difference was not statistically significant. A post-hoc power analysis suggests that the absence of statistically significant differences for these traits may be partly attributable to insufficient sample size rather than the absence of a true effect. However, compared to earlier similar study30, with an increase in sample size in this study, a convergence of personality characteristics of women with AUD and HCs was noted.

In our sample, women with AUD predominantly displayed Type 1 characteristics as defined by Cloninger’s typology of alcohol use disorder55, including older age, dependent personality traits, and heightened exposure to social stressors.

The five-factor model of personality

Personality assessment of respondents, obtained by applying the FFM revealed statistically significant differences between women with OUD and women with AUD in four domains (all but Openness), and in four domains (all but Extraversion) between OUD patients and HCs. CDA further supported these findings, identifying one significant discriminant function characterized by high Neuroticism and low Agreeableness and Conscientiousness (with a smaller positive contribution of Openness), which clearly separated women with OUD from HCs, while AUD patients clustered closer to the control group.

These results are consistent with the findings of foreign studies on people with SUD21,25,55,56,57,58. After controlling for age, all previously observed differences in personality dimensions remained significant, except for Extraversion. Considering that additional analyses showed a significant, negative correlation between the scores on the Extraversion domain and the age of the respondents the lowest mean score of women with AUD were, in all likelihood, a reflection of the fact that the respondents were the oldest in this group. Previous studies have shown that Extraversion tends to be more pronounced in younger individuals with addiction, as well as in those with OUD, compared to individuals with AUD21,36. Moreover, within the CDA model, this domain did not contribute significantly to the discriminatory function. This suggests that the variation in Extraversion is more likely attributable to age differences among participants than to their group membership. Women with AUD exhibited significantly higher levels of Agreeableness and Conscientiousness compared to those with OUD, while demonstrating less pronounced Openness to Experience. In this respect, their personality profile more closely resembled that of HCs. Moreover, no significant differences were observed between women with AUD and HCs across any of the five personality domains. This finding contrasts with prior research30, which - using a smaller sample - reported significantly higher Neuroticism and lower Conscientiousness in women with AUD. This discrepancy may be attributable to differences in sample size and participant age across studies. However, women with AUD show some potential for psychological destabilization (an increase in the Neuroticism facets - Depression and Self-Consciousness). As this is a cross-sectional study, it is not possible to distinguish whether depression is a personality trait or a state, or if it is a state, whether it is a primary or secondary depressive disorder.

Clinical implications

Finally, it is important to consider the expected therapeutic responses in light of the prominence of Neuroticism and Novelty Seeking, as these are the dimensions that most distinguished women with OUD in our research. An anxious, vulnerable, self-conscious, moody, depressed person who is high in the domain of Neuroticism, is inclined to see situations as stressful and is more vulnerable to stress and thus at greater risk of relapse59. Additionally, there is a positive correlation between Novelty Seeking and the tendency toward treatment dropout ahead of time or relapse17,32 probably due to high impulsivity, risky behavior, poor self-control, and psychological vulnerability16.

A combined, psychopharmacological and psychotherapeutic approach, supplemented with social skills training, is recommended for women with OUD, with close monitoring for dropout and relapse. Being generally younger and exhibiting traits of immature, maladaptive personality organization, they require therapy focused on fostering developmental growth, self-esteem, and adaptive coping.

Women with AUD display greater maturity and better impulse control, which supports sustained engagement in therapy; however, their vulnerability and depressive traits, together with their older age and potentially greater physical health problems, make the therapeutic process not necessarily less complex.

Limitations of the study

The clinical sample included women in specialized inpatient treatment, who may differ from those not seeking treatment in their level of insight into the disorder and motivation to address it, potentially reflecting a different constellation of personality traits. Consequently, findings may not generalize to broader populations, including untreated women with substance use disorders or at-risk women in the general population. Additionally, larger samples would enhance the reliability of results, particularly for finer-grained analyses of TCI and NEO PI-R facets. A post-hoc power analysis indicated that the study had adequate statistical power (≥ 0.80) to detect medium-to-large effects for significant results. However, small differences may have been underestimated due to limited statistical power, suggesting that some non-significant findings could reflect insufficient sample size rather than the absence of a true effect.

This cross-sectional design precludes causal inference and limits the feasibility of mediation analyses. Temporal ordering among key variables cannot be established; thus, we cannot determine whether being without a partner among women with OUD is a cause or consequence of addiction, whether partner/marital instability predisposes to or results from alcohol misuse, or whether professional/educational stress precipitates or follows addictive patterns. In addition, moderation (interaction) tests were not preregistered, and were not part of the primary analysis.

One notable limitation of this study is the relatively short abstinence period (7–10 days) prior to psychometric testing. While major guidelines (e.g., UNODC/WHO, NICE) emphasize the importance of conducting assessments during abstinence and clinical stability, they do not specify a precise minimum duration for personality testing60,61,62. Although acute withdrawal symptoms are typically resolved within 1–2 weeks, post-acute withdrawal syndrome (PAWS)—characterized by persistent emotional and cognitive disturbances—can last for several months63,64,65. Consequently, it is highly likely that some of the observed elevations in Neuroticism and impulsivity scores among women with OUD reflect transient withdrawal-related affective states rather than stable personality traits. Indeed, with prolonged abstinence, lower scores might be expected, as suggested by prior research66,67. This limitation should be carefully considered when interpreting the results, and future studies with longer abstinence periods are warranted to validate these findings.

Furthermore, while formal comorbid psychiatric diagnoses were an exclusion criterion, the study did not employ screening tools to assess for subclinical symptoms of mood or anxiety disorders. Such subclinical symptoms are common in SUD populations and could potentially influence personality test scores. We recommend that future research include standardized symptom measures to better control for this potential confounding factor.

Finally, the generalizability of our findings is limited by the focus on only two SUDs (alcohol and opioids). We did not collect systematic pharmacotherapy data (medication class, dose, and timing relative to assessment) or detailed treatment histories, nor did we assess psychiatric comorbidity with standardized diagnostic tools. As a result, potential effects of medication exposure and clinical history on personality scores remain unaccounted for. Future studies should include a broader range of SUDs and incorporate standardized diagnostic assessments and detailed treatment-history and pharmacotherapy measures to better control clinical confounding.

Conclusion

The results from this research confirm the expected differences in the personality dimensions of women with opioid and alcohol use disorders, revealing that the dissimilarities between those two groups are much greater than those between women with alcohol use disorders and healthy women. Independent of the personality inventory used, consistent results were obtained, and the TCI-5-R and NEO PI-R were similarly successful in the classification of respondents (AUD, OUD, or HCs). Women with ОUD are younger, with a constellation of character and temperament traits that point to immature and maladaptive personality organization. Women with AUD are more mature, but still vulnerable due to their neurotic personality structure, with dependent traits including traits characteristic of dependent personality - such as submissiveness, need for reassurance, and interpersonal dependency- accompanied by depressive symptoms, and multiple sources of social distress. These findings highlight the importance of considering substance-specific personality profiles in clinical assessment and treatment planning. Future research should further explore targeted interventions addressing distinct personality traits associated with different substance use disorders.