Evaluation of Adherence to Pharmacological Treatment in a Large Sample of Patients with Personality Disorder

Abstract

Objective

Personality disorders (PDs) are chronic and pervasive mental health conditions associated with significant functional impairment and high psychiatric comorbidity. Although psychotherapeutic interventions are the primary treatment approach, pharmacotherapy is frequently prescribed to manage specific symptoms. However, adherence to pharmacological treatment in PDs remains a critical challenge, influenced by both personality traits and clinical factors. The present study aims to assess adherence rates in a large cohort of patients with PDs and explore potential sociodemographic and clinical factors associated with compliance.

Method

This observational study included 200 patients diagnosed with PDs according to DSM-5 criteria, recruited from different psychiatric services in Milan, Italy. Adherence was assessed using the Clinician Rating Scale (CRS), with positive adherence defined as CRS ≥ 5 and poor adherence as CRS < 5. Sociodemographic and clinical data were collected and analyzed across adherence groups and PD clusters.

Results

Positive adherence was observed in 64.5% of the sample. Cluster C PDs exhibited significantly higher adherence rates (83.3%) compared to Cluster B (61.3%), mixed-feature (60%), and Cluster A (73.3%; p<.05). A positive family history of psychiatric disorders was associated with greater adherence (60.3% vs. 45.5%, p<.05). A trend toward lower adherence was observed in patients with lifetime and current substance use.

Conclusions

Higher adherence in Cluster C PDs may be linked to anxiety-driven behavioral patterns, while lower adherence in Cluster B and mixed-feature PDs suggests impulsivity and mistrust contribute to non-compliance. Additionally, a positive family history of psychiatric disorders emerged as a potential protective factor, possibly enhancing treatment engagement through greater awareness and support networks. Future research should focus on developing tailored interventions to the specific needs of different PD clusters to improve long-term treatment outcomes.

Introduction

Personality disorders (PDs) are enduring and pervasive patterns of inner experience and behavior that significantly deviate from cultural expectations. These patterns are inflexible and manifest across multiple personal and social contexts, leading to substantial distress or impairment in social, occupational, and other key areas of functioning (American Psychiatric Association, 2022). PDs are prevalent in the general population, with estimates ranging between 3.9% and 15.5%, based on studies conducted across the United States and Europe (Torgersen, 2009). Epidemiological data suggest a relatively equal distribution between men and women, although some analyses indicate a higher prevalence in men (Busch et al., 2016). However, in clinical settings, PDs are more frequently diagnosed in women, likely due to a greater tendency among females to seek psychiatric care (Björkenstam et al., 2015).

The chronicity and stability of maladaptive personality traits contribute to the widespread perception that PDs are particularly challenging to treat (Bleidorn et al., 2022). Current guidelines emphasize psychotherapy as the first-line treatment for PDs (Bateman et al., 2015¸ Newlin & Weinstein, 2015bate,). Among the psychotherapeutic options available, cognitive-behavioral therapy (CBT), dialectical-behavioral therapy (DBT), mentalization-based therapy (MBT), and transference-focused therapy (TFT) are recognized as effective approaches (Bateman et al., 2004). However, pharmacotherapy can play a significant role in managing comorbid psychiatric symptoms, particularly when integrated with psychotherapeutic interventions (Ingenhoven & Duivenvoorden, 2011; Kuriakose, 2024; Links, 2007; Mercer et al., 2009; Rosenbluth & Sinyor, 2012;). Pharmacological treatments, including typical and atypical antipsychotics, antidepressants, and mood stabilizers, are often used as adjunctive strategies to address specific symptoms (Bateman & Tyrer, 2004; Hadjipavlou & Ogrodniczuk, 2010; Ingenhoven et al., 2010; Leichsenring & Leibing, 2003). However, there are currently no FDA-approved medications specifically indicated for the treatment of PDs, pharmacotherapy is often prescribed off-label and its efficacy remains limited. Evidence-based guidelines emphasize a combined treatment approach that integrates psychotherapy and pharmacotherapy (Bateman et al., 2015; Mazza et al., 2016). Nevertheless, adherence to pharmacological treatment remains a substantial challenge in this population.

Previous studies have consistently reported a strong correlation between PDs and poor adherence to prescribed medications. For instance, Akerblad and colleagues (2008) showed that the presence of a PD significantly increases the risk of non-adherence, while Sirey and colleagues (2001) observed that patients without a PD diagnosis were more likely to exhibit good adherence to pharmacotherapy (Sirey et al., 2001). These findings underscore the importance of considering the presence of PDs as a critical factor in treatment adherence and emphasize the need for tailored strategies to improve patient engagement in pharmacological interventions. Additionally, clinical factors such as impulsivity and the quality of the therapeutic alliance have been shown to influence adherence. Impulsivity, commonly observed in certain PDs, can lead to inconsistent medication-taking behaviors, whereas a strong therapeutic alliance has been associated with improved adherence (Chang et al., 2019; Chapman & Horne, 2013).

Despite its clinical significance, research on medication adherence in patients with PDs is limited. Existing evidence suggests a general tendency toward low adherence in this population (Doesschate et al., 2009), along with a stronger preference for psychotherapy over pharmacological treatments (Mohammadabadi et al., 2022). Moreover, real-world data on adherence differences across PDs and their diagnostic clusters, especially within Italian clinical settings, are limited.

The present study aims to fill this gap by assessing potential differences in treatment adherence rates in a large cohort of patients with PDs, exploring sociodemographic and clinical factors, such as PD clusters, that may influence compliance. Understanding adherence patterns across specific PD diagnostic clusters is crucial for developing targeted interventions to enhance patient outcomes and optimize treatment efficacy.

Methods

A total of 200 patients, of any age and gender, with a diagnosis of PDs according to DSM-5 criteria, were consecutively recruited from different psychiatric services within the Department of Psychiatry and Mental Health of the ASST Fatebenefratelli Sacco, in Milan, Italy. Participants were eligible for inclusion if they met the following criteria: (1) a diagnosis of a PD according to DSM-5 criteria, (2) age ≥18 years at the time of enrollment, and (3) the ability to provide valid written informed consent. Exclusion criteria included the presence of intellectual disability and the inability to provide informed consent. Substance abuse, other pharmacological treatments (current or past), and medical comorbidities in the patients’ medical history were not considered exclusion criteria.

Diagnoses were obtained through the administration of the Structured clinical interview for DSM-5 Personality Disorders (SCID-5-PD; First et al., 2016). In the ‘mixed-feature PDs’ category, were included patients who met criteria for either ‘Other Specified Personality Disorder’ or ‘Unspecified Personality Disorder’ according to DSM-5. Symptom severity was assessed using the following psychometric questionnaires: Hamilton Anxiety Rating Scale (HAM-A; Hamilton, 1959), Montgomery Asberg Depression Rating Scale (MADRS; Montgomery & Asberg 1979), and Brief Psychiatric Rating Scale (BPRS; Overall & Gorham, 1988). Treatment adherence was evaluated using the Clinician Rating Scale (CRS), developed by Kemp and David (Kemp & David, 1996). The CRS quantifies the clinician’s assessment of adherence on a 7-point ordinal scale: 1 –complete refusal, 2 –partial refusal, 3 –accepts only because compulsory or very reluctant, persuasion or questions often needed, 4– occasional reluctance, 5 –passive acceptance, 6 –moderate participation, some knowledge and interest in medication and no promoting required, 7 –active participation. A CRS score of ≥ 5 was used to define positive adherence (A+), while scores <5 indicated poor adherence (A-), consistent with previous studies (Benatti et al., 2023; Holma et al., 2010; Kemp & David, 1996). The adherence evaluation covered the six months preceding study inclusion. Based on CRS scores, the sample was divided into two subgroups: A+ (positive adherence) and A- (negative adherence).

Data collection took place from May 2022 to October 2024. The patients provided written informed consent to participate in this study and for the use of their anonymized data for research purposes. After obtaining consent, sociodemographic and clinical variables were collected and entered a common database. The study was conducted in accordance with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008 (PMC2566407).

Sociodemographic and clinical variables were collected for the entire sample and stratified by PD clusters. Data were examined to assess differences between treatment adherence groups (A+ vs. A-) and to explore potential associations between adherence rates and specific clinical and sociodemographic features. Statistical analyses were performed using Student’s t-test and one-way ANOVA for continuous variables, while the χ²test was used for dichotomous variables to compare sociodemographic and clinical characteristics between treatment adherence subgroups. Additionally, an exploratory multivariate logistic regression analysis was performed to assess the independent contribution of potential predictors to adherence (A+ vs. A-). The variables included in the model were age, gender, service type, lifetime substance use, and psychometric scores. The Omnibus Test of Model Coefficients was used to assess the overall significance of the model. Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS) 26.0 software for Windows (SPSS Inc, Chicago, IL, USA), and statistical significance was set at p< .05.

Results

The study included 200 patients diagnosed with PDs according to DSM-5 criteria, recruited from different psychiatric services: 44.9% from community mental health services, 40.3% from day hospital services, 10.7% from psychiatric wards, and 4.1% from outpatient tertiary clinics. Regarding PD clusters, the distribution within the sample was as follows: Cluster A accounted for 7.5%, Cluster B for 53%, Cluster C for 12%, and mixed-feature (MF) for 27.5%. The sample comprised 65% females and 35% males with a mean age of 43.4 ± 17.0 years. The mean age of illness onset was 26.3 ± 12.6 years, with an average of untreated illness duration (DUI) of 6.2 ± 22.6 years.

More than half of the sample (55.2%) reported a positive family history of psychiatric disorders, and 52.4% showed lifetime suicidal ideation, although only 23% had attempted suicide at least once. Regarding psychopharmacological therapy, the majority of patients were receiving at least one medication: antidepressants (77%), mood stabilizers (32.1%), antipsychotics (56.3%), and benzodiazepines (58.4%). At the time of inclusion in the study, 95.3% of the sample reported no significant adverse effects related to pharmacological treatment and 63.5% of patients were receiving psychotherapy. The mean age at first pharmacological therapy was 30.2 ± 12.9 years, with a mean daily administration frequency of 0.7 ± 0.4. Lifetime substance use was reported by 32.6%, while 12.5% were current users. At the time of inclusion, psychometric assessment showed mild-to-moderate symptom severity, with mean scores of 15.3 ± 6.59 on the HAM-A, 22.77 ± 11.31 on the MADRS, and 20.81 ± 9.25 on the BPRS. The main sociodemographic and clinical variables of the study participants are also provided in table 1.

Table 1.

Sociodemographic and clinical features stratified by personality cluster

	Cluster A PDs n=15 (7.5%)	Cluster B PDs n=106 (53%)	Cluster C PDs n=24 (12%)	Other Specified PDs and Unspeci- fied PDs n=55 (27.5%)	Total n=200
Gender Male Female	13 (86.7%) 2 (13.3 %)	30 (28.3%) 76 (71.7%)	8 (33.3%) 16 (66.7%)	19 (34.5%) 36 (65.5%)	70 (35%) 130 (65%)
Age	46.8±15.7	41.32±17.9	44.5±16.3	46.07±15.6	43.44±17.04
Service Day hospital service Psychiatric ward CMHS Outpatients tertiary clinics	5 (33.3%)** 1 (6.7%) 9 (60.0%)** 0 (0.0%)	41 (40.2%) 16 (15.7%) 41 (40.2%) 4 (3.9%)	16 (66.7%) 1 (4.2%) 4 (16.7%) 3 (12.5%)	17 (30.9%) 3 (5.5%) 34 (38.6%) 1 (1.8%)	79 (40.3%) 21 (10.7%) 88 (44.9%) 8 (4.1%)
Age at first pharmacological therapy	32.2±11.3	28.4±13.3	30.9±11.1	32.5±12.8	30.18±12.88
Psychiatric family history No Yes	5 (35.7%) 9 (64.3%)	46 (45.5%) 55 (54.5%)	9 (37.5%) 15 (62.5%)	26 (49.1%) 27 (50.9%)	86 (44.8%) 106 (55.2%)
Pharmacotherapy Taking ADs Taking MSs Taking APs Taking BDZs	7 (46.7%) 3 (21.4%) 10 (71.4%) 5 (35.7%)	78 (78.8%) 42 (42.4%) 61 (61.6%) 58 (58.6%)	22 (91.7%) 3 (12.5%) 7 (29.2%) 13 (54.2%)	40 (75.5%) 13 (24.5%) 29 (54.7%) 35 (66.0%)	147 (77%) 61 (32.1%) 107 (56.3%) 111 (58.4%)
Suicidal thoughts lifetime No Yes	11 (78.6%) 3 (21.4%)*	37 (37.0%) 63 (63.0%)*	15 (62.5%) 9 (37.5%)*	28 (52.8%) 25 (47.2%)*	91 (47.6%) 100 (52.4%)
Suicide attempts lifetime None Just one More than one	14 (100%) 0 (0.0%) 0 (0.0%)	69 (69.0%) 30 (30.0%) 1 (1.0%)	22 (91.7%) 2 (8.3%) 0 (0.0%)	42 (79.2%) 11 (20.8%) 0 (0.0%)	147 (77%) 43 (22.5%) 1 (0.5%)
N. of hospitalization None From 1 to 3 From 3 to 6 From 7 to 10 Over 10	10 (66.7%) 4 (26.7%) 0 (0.0%) 0 (0.0%) 0 (0.0%)	51 (51.0%) 19 (19.0%) 22 (22.0%) 6 (6.0%) 2 (2.0%)	15 (62.5%) 5 (20.8%) 2 (8.3%) 2 (8.3%) 0 (0.0%)	30 (56.6%) 12 (22.6%) 10 (18.9%) 1 (1.9%) 0 (0.0%)	106 (55.2%) 40 (20.8%) 34 (17.7%) 9 (4.7%) 3 (1.6 %)
N. of involuntary commitments None From 1 to 3 From 4 to 6 Over 10	12 (85.7%) 2 (14.3%) 0 (0.0%) 0 (0.0%)	92 (92.0%) 4 (4.0%) 3 (3.0%) 1 (1.0%)	23 (95.8%) 0 (0.0%) 1 (4.2%) 0 (0.0%)	47 (88.7%) 6 (11.3%) 0 (0.0%) 0 (0.0%)	174 (91.1 %) 12 (6.3 %) 4 (2.1 %) 1 (0.5%)
Current medication side effects No Yes	13 (92.9%) 1 (7.1%)	93 (93.9%) 6 (6.1%)	24 (100%) 0 (0.0%)	51 (96.2%) 2 (3.8%)	181 (95.3%) 9 (4.7%)
Lifetime substance use No Yes	12 (85.7%) 2 (14.3%)	59 (59.0%) 41 (41.0%)	20 (87.0%) 3 (13.0%)	37 (69.8%) 16 (30.2%)	128 (67.4%) 62 (32.6%)
Current substance use No Yes	13 (92.9%) 1 (7.1%)	84 (83.2%) 17 (16.8%)	24 (100%) 0 (0.0%)	47 (87.0%) 7 (13.0%)	168 (87%) 25 (13%)
Psychotherapy lifetime No Yes	9 (64.3%) 5 (35.7%)	38 (37.6%) 63 (62.4%)	9 (37.5%) 15 (62.5%)	14 (26.4%) 39 (73.6%)	70 (36.5%) 122 (63.5 %)

Notes. Values for categorical and continuous variables are expressed in percentages and mean ± SD, respectively. Reported variables had a percentage of missing data ranging from 0 to 14%. CMHS: Community mental health services; ADs: antidepressants; MSs: mood stabilizers; APs: antipsychotics; BDZs: benzodiazepines.*p<.05; **p<.005

Consistent with cluster-based analysis, significant differences emerged regarding psychiatric service setting: the majority of patients in Cluster A (60%, n = 9, p < 0.05) were followed at community mental health services, whereas those in Cluster C (66.7%, n = 16, p < 0.05) were linked to day hospital services. Cluster B showed a balanced distribution, with 40.2% of individuals recruited from both community mental health services and day hospital services (see figure 1). No other significant differences were observed across the PD clusters.

Figure 1.

Distribution of personality disorder clusters (Clusters A, B, C, and Mixed-feature) across different psychiatric service settings

To assess treatment adherence, the sample was stratified into two subgroups based on CRS scores: A+ (positive adherence) with a score ≥ 5, and A- (poor adherence) with a score < 5. A comprehensive summary of clinical features classified by adherence groups is provided in table 2. A+ was observed in 64.5% of the sample, composed of 35.7% males and 64.3% females, while A- accounted for 35.5%, with 33.8% males and 66.2% females. A significantly higher rate of A+ were observed in patients with a positive family history of psychiatric disorders (60.3% vs 45.5%, p<.05). When examining adherence across PD clusters, the lowest adherence rates were observed in patients with mixed-feature and Cluster B PDs (A-: MF: 40%, B: 38.7%, A: 26.7%, C: 16.7%). Patients diagnosed with Cluster C showed significantly higher rates of A+ compared to other groups (83.3% vs A: 73.3%, B: 61.3%, MF: 60%; p<.05, see figure 2). Lifetime substance use was reported in 39.4% of A- patients compared to 29% of A+ patients (p = .223), while current substance use was higher in A- patients (19.1%) than in A+ patients (9.6%), showing a trend toward statistical significance (p =.060). No significant differences emerged between adherence groups regarding the type of pharmacological treatment prescribed (p=.328) or in the number of daily medication administrations (A+: 0.77 ± 0.42 vs. A-: 0.68 ± 0.47, p=.200). A higher percentage of A+ patients reported lifetime psychotherapy (67.2%) compared to A- patients (56.7%), though this difference was not statistically significant (p=.150). No other significant differences were found in treatment adherence groups. Finally, in the exploratory multivariate logistic regression analysis, the overall model was statistically significant (Omnibus Test, p=.017), indicating that the combined variables explained a portion of the variance in adherence status. However, none of the individual predictors achieved statistical significance after adjusting for the other variables (p<.005).

Table 2.

Socio-demographic and clinical characteristics of the sample related to treatment adherence

	A - n= 71 (35.5%)	A + n= 129 (64.5%)	Significance
Cluster of personality Cluster A Cluster B Cluster C Other Specified PDs and Unspecified PDs	4 (26.7 %) 41 (38.7 %) 4 (16.7 %)* 22 (40%)	11 (73.3 %) 65 (61.3%) 20 (83.3 %) 33 (60 %)	.015
Gender Male Female	24 (33.8%) 47 (66.2%)	46 (35.7%) 83 (64.3%)	.792
Age	43.43±17.01	43.45±17.09	.933
Age at onset	25.53±12.68	26.65±12.63	.556
Service Day Hospital Services Psychiatric ward Community mental health services Outpatients tertiary clinics	33 (47.1%) 6 (8.6%) 29 (41.4%) 2 (2.9%)	46 (36.5%) 15 (11.9%) 59 (46.8%) 6 (4.8%)	.490
Duration of illness (years)	26.96±58.89	20.77±47.35	.845
DUI (years)	7.71±30.75	5.42±16.98	.516
Age at first pharmacological therapy	29.29±12.34	30.64±13.17	.495
Number of daily administrations	0.68±0.47	0.77±0.42	.200
Positive psychiatric family history Pharmacotherapy Taking ADs Taking MS Taking APs Taking BDZ	30 (45.5%)* 47 (72.3%) 20 (30.8%) 39 (60.0%) 40 (61.5%)	76 (60.3%) 100 (79.4%) 41 (32.8%) 68 (54.4%) 71 (56.8%)	.029 .328
Suicidal thoughts lifetime	31 (16.2%)	69 (36.1%)	.279
Suicidal attempts lifetime	11 (5.8%)	33 (17.3%)	.274
Current medication side effects No Yes	62 (34.3%) 3 (33.3%)	119 (65.7%) 6 (66.7%)	.955
Age at onset substance use	19.76±5.60	22.88±9.22	.141
Lifetime substance use	26 (39.4%)	36 (29%)	.223
Current substance use	13 (19.1%)	12 (9.6%)	.060
Psychotherapy lifetime	38 (56.7 %)	84 (67.2 %)	.150
N° of Hospitalizations None From 1 to 3 From 3 to 6 From 7 to 10 Over 10 N° of Involuntary commitments None From 1 to 3 From 4 to 6 Over 10	36 (54.5%) 16 (24.2%) 12 (18.2%) 1 (1.5%) 1 (1.5%) 61 (31.9 %) 2 (3.0 %) 3 (4.5%) 0 (0%)	70 (55.6%) 24 (19%) 22 (17.5%) 8 (6.3%) 2 (1.6%) 113 (90.4 %) 10 (8.0 %) 1 (0.8 %) 1 (0.8%)	.602 .162
HAM-A at inclusion	11.4± 6.15	16.3 ± 6.46	.141
MADRS at inclusion	14.75 ± 6.60	24.56 ± 11.49	.119
BPRS at inclusion	28.33±1.53	19.08±9.44	.121

Notes: Values for categorical and continuous variables are expressed in percentages and mean ± SD, respectively. Reported variables had a percentage of missing data ranging from 0 to 14%. A+: positive adherence to treatment, A-: negative adherence to treatment, DUI: duration of untreated illness, ADs: antidepressants; MSs: mood stabilizers; APs: antipsychotics; BDZs: benzodiazepines. *p<.05; **p<.005

Figure 2.

Distribution of positive (A+) and poor (A-) adherence rates across different personality disorder clusters (Clusters A, B, C, and Mixed-feature)

Discussion

In our study, approximately two-thirds of patients with PD exhibited positive adherence to their medication regimen. When analyzing adherence rates across different PD clusters, patients with Cluster C showed the highest adherence, aligning with previous literature. Research on treatment adherence in individuals with Avoidant, Dependent, and Obsessive-Compulsive Personality Disorders (OCPD) suggests that these patients exhibit higher compliance with medical treatments, likely driven by heightened anxiety, fear of disapproval, and a strong need for reassurance (Emilsson et al., 2020; Mohammadabadi et al., 2022). Holma and colleagues (2010) found a significant association between good adherence and Cluster C PDs in patients with major depressive disorder, reinforcing the notion that pronounced anxiety traits contribute to treatment engagement (Holma et al., 2010). Moreover, Gudjonsson and Main (2008) reported that anxiety-driven personality traits, particularly those in Cluster C, correlated with higher adherence rates (Gudjonsson & Main, 2008). Notably, OCPD has been linked to better adherence likely due to its rigid cognitive flexibility and heightened concern for rule-following compared to those with Cluster B disorders, who are more prone to impulsivity and emotional instability (Pinto et al., 2022). These findings suggest that, while anxiety-related characteristics can be impairing, they may also serve as facilitators for adherence through reinforced behavioral patterns and a greater tendency to comply with medical advice (Ingenhoven & Abraham, 2010; Livesley & Jang, 2000). Furthermore, our study found that the lowest adherence rates were observed in patients with Cluster B and mixed-feature PDs, which is consistent with existing evidence indicating that Cluster B traits, particularly in borderline personality disorder (BPD), are associated with poor adherence (Hancock-Johnson et al., 2017; Mohammadabadi et al., 2022; Timäus et al., 2019). BPD is characterized by impulsivity, emotional dysregulation, and an unstable self-identity, factors that compromise treatment engagement and lead to inconsistent medication use (Zimmerman & Mattia, 2001). Additionally, the limited efficacy of pharmacotherapy in BPD may contribute to lower adherence rates, as patients may perceive medications as ineffective or unnecessary (Chiesa et al., 2017). Cluster A PDs, including schizoid, schizotypal, and paranoid PDs, have been historically associated with poor adherence due to their personality traits. In particular, Cluster A patients often exhibit high levels of mistrust and social detachment, pervasive suspiciousness, cognitive-perceptual distortions, emotional detachment, and isolation, which can lead to negative attitudes towards medication. Patients may doubt the necessity or safety of their treatment regimens and may not perceive the same urgency or benefit in prescribed medications (Chauhan et al., 2020).

An additional significant finding in our study was the higher adherence rate observed in patients with a positive family history of psychiatric disorders. One possible explanation is that these individuals may have greater awareness of mental illness and its trajectory, leading to a stronger understanding of the importance of adherence to treatment in preventing relapses and maintaining long-term stability. This supports prior research indicating that familial exposure to psychiatric illness may foster increased mental health awareness, leading to greater recognition of treatment benefits and a stronger commitment to pharmacotherapy (Zygmunt et al., 2002). Moreover, family members may serve as key facilitators in medication management, reinforcing adherence behaviors and reducing the likelihood of discontinuation (Náfrádi et al., 2017). Given these findings, future research could explore whether family history serves as a direct predictor of adherence or represents a proxy for other factors, such as increased healthcare engagement or early psychiatric interventions.

Although our study did not find significant differences in adherence based on medication class, this result aligns with previous literature suggesting that adherence in PDs is more influenced by personality traits, therapeutic relationships, and perceived treatment effectiveness than by the specific pharmacological agent prescribed (Chapman & Horne, 2013). Finally, while substance use was not significantly associated with adherence, a trend toward lower adherence was observed in patients with lifetime and current substance use, a finding consistent with previous studies indicating that comorbid substance use disorders represent a sgnificant barrier to adherence in psychiatric populations (Chapman & Horne, 2013; Girone et al., 2024).

Limitations

The abovementioned results should be interpreted in light of some methodological limitations. First, the cross-sectional design precludes any causal inferences regarding the relationships between PD clusters and adherence. Second, although the sample size was relatively large, some PD clusters and service types were underrepresented, potentially limiting the statistical power and affecting the generalizability of the findings. Third, although the multivariate regression model was statistically significant overall, none of the individual predictors reached significance; this may reflect the complex and multifactorial nature of adherence behavior as well as the sample size and variable distribution. Finally, the “mixed-feature PDs” group included patients classified as “Other Specified” or “Unspecified” according to DSM-5 criteria, representing a heterogeneous group with variable clinical presentations. Future research should address these limitations by employing longitudinal designs and larger, more balanced samples to confirm and extend our findings.

Conclusion

Our findings emphasize the heterogeneity of adherence patterns in PDs, reinforcing the need for personalized interventions. Cluster C patients may benefit from structured reinforcement strategies that leverage their anxiety-driven compliance, whereas patients with Cluster B or A traits may require more intensive psychoeducation, motivational enhancement techniques, and structured therapeutic alliances to improve adherence to treatment. Moreover, involving family members in treatment planning could be beneficial, as caregiver support may contribute to better adherence in psychiatric conditions.

Overall, our study contributes to the growing evidence that adherence in PDs is a complex and multidimensional phenomenon influenced by personality traits, family dynamics, and service-related factors. Further research is necessary to investigate the complex interplay of factors influencing adherence in PDs, with a focus on identifying specific strategies that could enhance treatment engagement and efectiveness.