Abstract
Objective
The study aimed to investigate the relationship between cognitive insight and treatment adherence among psychiatric patients.
Method
This descriptive, cross-sectional, and correlational study was conducted between February and April 2023 with 140 psychiatric patients in a public hospital in İstanbul. Data were collected using a personal ınformation form, the Morisky medication adherence scale, and the Beck cognitive insight scale. Statistical analyses, including confirmatory factor analysis and structural equation modeling, were performed using SPSS 25.0 and AMOS 24.0.
Results
The expressing oneself subdimension of cognitive insight had a statistically significant positive effect on treatment adherence (β=0.259; p<0.001), indicating that patients with higher expressive ability showed greater adherence. Conversely, the self-confidence subdimension did not have a significant impact (β=0.050; p>0.05). The model explained 10.5% of the variance in treatment adherence.
Conclusion
The ability to express oneself, an important subdimension of cognitive insight, is a significant predictor of treatment adherence in psychiatric patients. Enhancing expressive abilities may improve therapeutic engagement and adherence. However, self-confidence alone does not appear to directly influence adherence in this clinical context. These findings highlight the importance of incorporating insight-focused interventions into psychiatric care.
Introduction
Treatment adherence (TA) is a vital factor that directly impacts the progression of illness in psychiatric patients. It significantly influences treatment outcomes, quality of life, and the effectiveness of healthcare interventions provided to the patient. Non-adherence may result in higher hospitalization rates, increased symptom severity, and even a greater risk of suicide (1, 2). According to the World Health Organization (3), failure to stick to treatment in severe mental disorders such as bipolar disorder and schizophrenia substantially raises mortality, morbidity, and hospital readmission rates. TA in psychiatric patients has important implications not only for individuals but also for healthcare systems and society at large. Recognizing the impact of non-adherence across these domains promotes the development of strategies to improve adherence and, consequently, outcomes at multiple levels. For individuals, sticking to psychiatric medications is essential for effectively managing mental health conditions. Studies show that non-adherence is common, with nearly half of psychiatric patients not taking their medications as prescribed (4, 5). This non-adherence can worsen symptoms, aggravate psychiatric conditions, and heighten the risk of hospitalization and crises (1). Contributing factors include a lack of insight into one’s illness, negative attitudes toward treatment, and poor communication with healthcare providers (6). From a healthcare system perspective, treatment non-adherence increases the burden on hospitals and mental health services. Poor adherence frequently leads to more visits to emergency departments and psychiatric wards, boosting healthcare utilization and incurring substantial costs (1, 7). Moreover, ineffective management of psychiatric disorders due to non-adherence can cause repeated hospital stays and system instability, complicating mental health service delivery (8). On a societal level, the effects of TA in psychiatric patients are significant. Non-adherence not only affects individual health outcomes but also creates broader societal issues, such as higher unemployment, involvement of mentally ill individuals in the criminal justice system, and increased societal costs linked to untreated mental health conditions (7, 8). Patients with psychiatric disorders who do not adhere to treatment might also engage in behaviors that endanger public safety and health. For example, untreated mental illness is linked to higher rates of aggression or self-harm, which can pose risks to both the individual and others (9). Cognitive insight plays an important role in motivating patients to follow their treatment plans. Research indicates that patients with greater insight into their mental health are more likely to recognize the need for medication, which directly relates to better recovery outcomes (10, 11). Cognitive insight refers to a person’s metacognitive capacity to critically evaluate and reappraise their own distorted beliefs and misinterpretations, and to remain receptive to corrective feedback (12). It encompasses two main components: self-reflectiveness, which reflects openness to re-evaluating one’s beliefs, and self-certainty, which refers to overconfidence in one’s interpretations. Higher levels of cognitive insight have been linked to increased recognition of illness, stronger motivation to engage in treatment, and improved adherence outcomes (11). Understanding their illness often strongly influences adherence behavior. Conversely, poor insight into psychiatric symptoms tends to decrease motivation for treatment, leading to non-adherence (13). For instance, roughly 50% of patients with schizophrenia show poor insight, which is associated with lower medication adherence and worse recovery and treatment results (13). Cognitive insight involves a person’s ability to recognize and correct distorted thoughts and beliefs. It encompasses broader metacognitive skills, including awareness and evaluation of one’s own thinking processes (14). For psychiatric patients especially those with schizophrenia or bipolar disorder cognitive insight is key to understanding the need for treatment. Without this insight, patients often deny their illness, resulting in lower adherence to medication and treatment plans (15, 16). Although previous studies have looked at the relationship between insight and TA in psychiatric groups, the concept of cognitive insight, characterized by self-reflectiveness and self-certainty, has received limited attention in this area. Self-reflectiveness refers to the individual’s capacity to question their own beliefs, consider alternative explanations, and remain open to feedback. For example, patients with higher self-reflectiveness may recognize that their symptoms such as unusual thoughts or perceptual experiences could be part of a mental health condition, leading them to seek clarification from healthcare providers and adhere more closely to treatment (17, 18). In contrast, self-certainty reflects a high level of confidence in the accuracy of one’s beliefs and interpretations, which may lead to resistance toward contradictory information and treatment recommendations (12). Although general clinical insight has been extensively studied in relation to TA, fewer studies have specifically addressed these two cognitive insight dimensions, particularly in bipolar disorder populations (19). Most research has focused on clinical insight or general symptom awareness, often relying on cross-sectional correlations without exploring underlying mechanisms or the direction of the relationship. Although several studies have investigated the association between clinical insight and TA (19, 20), most have conceptualized insight in broad terms, such as general symptom recognition or awareness of illness, and have predominantly relied on cross-sectional designs. This approach limits understanding of the temporal sequence and underlying mechanisms linking insight to adherence. Furthermore, relatively few studies have examined the specific cognitive insight dimensions self-reflectiveness and self-certainty as predictors of adherence, particularly in bipolar disorder populations (21). Addressing this gap is crucial for identifying targeted therapeutic strategies that can enhance adherence through improvements in these specific cognitive processes. Moreover, studies involving inpatient psychiatric populations, particularly those diagnosed with bipolar disorder, are rare, despite the importance of adherence in this group. Additionally, there is a notable lack of research using advanced statistical methods such as structural equation modeling (SEM) to investigate how aspects of cognitive insight might predict adherence behaviors through both direct and indirect routes. Although the study was open to all psychiatric patients meeting the criteria, all participants admitted during data collection happened to have bipolar disorder. This homogeneity reflects the patient profile of the hospital at that time rather than an initial focus on bipolar disorder inpatients.
The study aimed to investigate the relationship between cognitive insight and TA among psychiatric patients.
Materials and Methods
Study Design and Setting
This descriptive, cross-sectional, and correlational study was conducted from February 20 to April 27, 2023, at the psychiatry clinic of a public hospital, the University of Health Sciences Turkey, Başakşehir Çam and Sakura City Hospital, in İstanbul, Turkey.
Study Population and Sampling
The study population consisted of all psychiatric patients aged 18 years and older who received care in the aforementioned hospital during the study period (n=200). The final sample included 140 patients, selected through purposive sampling based on inclusion and exclusion criteria and voluntary participation.
The sample size was estimated using G*Power 3.1 software. A priori power analysis for a two-tailed correlation test was conducted using a medium effect size (r=0.30), an α level of 0.05, and a power of 0.80, resulting in a minimum required sample of 84 participants. This estimation was based on Cohen’s (22) guidelines. The final sample of 140 participants exceeded this minimum, enhancing statistical validity.
Inclusion and Exclusion Criteria
Inclusion criteria for the study were being 18 years of age or older, providing voluntary participation, having a psychiatric disorder diagnosed according to diagnostic and statistical manual of mental disorders-5, receiving oral or parenteral antipsychotic treatment, and having at least six months since diagnosis. Exclusion criteria were having significant visual or auditory impairments, a history of traumatic brain injury or central nervous system infection, or a substance use disorder within the past two months.
Research Hypotheses
H1: The “expressing oneself” (EO) subdimension of cognitive insight has a positive and statistically significant effect on TA.
H2: The “self-confidence” (SC) subdimension of cognitive insight has a positive and statistically significant effect on TA.
Data Collection Tools
Data were collected using the following instruments:
1. Personal information form: Created by the researcher based on literature, this form includes 8 items that assess demographic and clinical characteristics such as age, gender, diagnosis, family psychiatric history, and medication responsibility (21).
2. Morisky medication adherence scale (MMAS-4): A 4-item scale developed by Morisky et al. (23) to measure adherence using “yes/no” responses. Adherence is classified as good, moderate, or poor. The scale’s Turkish version was utilized, with reported internal consistency (Cronbach’s α=0.61). The Turkish adaptation and psychometric assessment for use in individuals with bipolar disorder was performed by Bahar et al. (24), showing satisfactory validity and reliability.
3. Beck Cognitive Insight Scale (BCIS): A 15-item self-report tool developed by Beck et al. (25) to assess cognitive insight through two subscales: Self-reflectiveness and self-certainty. A composite score is created by subtracting self-certainty from self-reflectiveness. The Turkish version and psychometric validation were carried out by Aslan et al. (26).
Data Collection Procedure
Participants were recruited by the researcher during routine ward visits. After obtaining informed consent, the questionnaires were administered face-to-face in a quiet and private setting. The researcher remained present to ensure that questions were understood and answered appropriately.
Statistical Analysis
Data obtained from 140 participants were analyzed using IBM SPSS Statistics version 25.0 and AMOS version 24.0. Descriptive statistics including mean, standard deviation, frequency, and percentage distributions were used to summarize the demographic and clinical characteristics of the sample. Prior to model testing, the normality of data distribution was assessed using Skewness and Kurtosis coefficients. All variables exhibited values within the acceptable range (-1.5 to +1.5), indicating normal distribution. Accordingly, the maximum likelihood (ML) method was selected for parameter estimation in SEM.
A two-tailed significance level of p<0.05 was considered statistically significant for all analyses.
Ethical Considerations
The study was approved by the Ethics Committee of İstanbul Arel University, İstanbul, Turkey (approval date: 10 February 2023; approval no: 2023/03). The research was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. All participants were informed about the purpose, procedures, and voluntary nature of the study, and signed informed consent was obtained. Anonymity and confidentiality were ensured throughout the research process. Data were securely stored and used exclusively for research purposes.
Results
The mean age of the participants was 31.47±7.12 years. Of the total sample, 66.4% (n=93) were female, and 66.4% (n=93) were single. High school graduation was the most common educational level (42.9%, n=60), and 71.4% (n=100) lived in a nuclear family structure. Regarding living arrangements, 51.4% (n=72) lived with parents and siblings. All participants (100%, n=140) were diagnosed with bipolar disorder, and 69.3% (n=97) reported regular medication use. The majority (71.4%, n=100) reported no family history of mental illness. According to the categorical distribution of the medication adherence scale, 47.6% (n=67) demonstrated high adherence, 23.6% (n=33) moderate adherence, and 28.6% (n=40) low adherence.
Confirmatory Factor Analysis (CFA)
Cognitive Insight Scale
The BCIS, consisting of 15 items and two subscales (self-reflectiveness and self-certainty), was subjected to CFA to evaluate construct validity. All item factor loadings were above 0.50, ranging between 0.54 and 0.80, indicating no item removal was required (Figure 1).
Medication Adherence Scale
The MMAS-4, composed of 4 items and a unidimensional structure, was also evaluated using CFA. All factor loadings exceeded 0.50, ranging from 0.52 to 0.86 (Figure 2).
Convergent Validity, Discriminant Validity, and Normality Analysis of the Scale Subdimensions
Composite reliability (CR) values are calculated based on the factor loadings obtained from CFA. A CR value of 0.70 or higher (CR ≥0.70) indicates that the condition for CR has been met. Convergent validity is indicated by the average variance extracted (AVE). For convergent validity to be confirmed, AVE should be at least 0.50 (AVE ≥0.50). However, if CR is above 0.70 for all dimensions, an AVE value of 0.40 or higher is also considered acceptable. To establish discriminant validity, the square root of the AVE for each construct (√AVE), presented in parentheses on the diagonal in Table 1, must be greater than the inter-construct correlation values in the same row and column.
Table 1 presents the reliability and validity values of the scales used in the study. All scales and their subdimensions demonstrated adequate internal consistency (Cronbach’s alpha ≥0.70) and met the thresholds for CR and convergent validity. The discriminant validity results confirmed that each construct measured distinct concepts. These results indicate that the instruments were psychometrically sound and suitable for examining the relationship between cognitive insight and TA in the present study.
Normality values of the variables are presented in Table 2. Prior to testing the structural model, the normality of the subdimensions’ distribution was assessed through Skewness and Kurtosis coefficients. The values for all variables were found to lie within the acceptable range of -1.5 to +1.5, suggesting that the data followed a normal distribution. Consequently, the ML estimation method was considered suitable for parameter estimation within the model.
Path Analysis of the Structural Equation Model Using Observed Values
In the proposed research model, the effects of the subdimensions of the cognitive insight scale—EO and SC—on the TA score were examined.
In the path analysis model using observed variables, the model fit indices indicated that the model was statistically significant. The chi-square value (χ²) was 215.55 with 146 degrees of freedom, resulting in a χ²/df ratio of 1.476, which falls within the acceptable threshold (less than 3), suggesting a good model fit. Furthermore, additional fit indices reinforced the adequacy of the model: The goodness-of-fit index was calculated as 0.920, and the comparative fit index was 0.953—both exceeding the commonly accepted threshold of 0.90, indicating good fit. The standardized root mean square residual was 0.071, and the root mean square error of approximation was 0.059, both within acceptable limits, confirming that the structural model fits the data well (Figure 3).
Regression results are presented in Table 3. The following findings were obtained from the model examining the effects of the cognitive insight scale subdimensions on TA:
The EO subdimension had a positive and statistically significant effect on TA (β=0.259; p<0.05). This indicates that an increase in EO scores leads to higher levels of TA.
The SC subdimension had a positive but non-significant effect on TA (β=0.050; p>0.05), suggesting no statistically meaningful relationship between SC and TA.
The model explained 10.5% of the variance in TA through the EO subdimension (R²=0.105).
Discussion
This study investigated the effect of cognitive insight—specifically its subdimensions, EO and SC on TA in psychiatric patients diagnosed with bipolar disorder. The findings revealed that only the EO dimension significantly predicted adherence, whereas the SC dimension did not show a meaningful effect.
In psychiatric patients, the way individuals express themselves directly impacts their awareness of the illness, emotional regulation, and social interaction, thereby significantly influencing TA. Expression style helps individuals make sense of their inner experiences (emotions, thoughts, fears) and communicate them to others. As a subcomponent of cognitive insight, the ability to express oneself plays a crucial role in TA. This interaction is influenced by various factors such as communication skills, emotional expression, and the ability to understand and articulate mental health issues. Being able to effectively express emotions and thoughts is essential for establishing a therapeutic alliance between patients and healthcare providers. When patients can communicate their concerns and experiences clearly, healthcare providers gain a better understanding of their perspectives and can develop personalized treatment plans (27). Open communication also helps reduce stigma and encourages TA (28).
Patients who are better at expressing themselves are more likely to cooperate with treatment. Conversely, a lack of insight, including difficulties in expressing emotions and thoughts about their condition and treatment, can negatively affect adherence (28). The inability to recognize and articulate personal experiences related to mental health challenges may hinder engagement in treatment. Interventions aimed at increasing insight and enhancing patients’ expressive abilities can promote adherence (20). Coping strategies also impact patients’ expression styles and adherence. Patients with ineffective coping mechanisms may avoid confronting mental health issues, struggle to express themselves, and thus weaken adherence (6).
Patients who can communicate their needs are more likely to participate in care decisions and follow treatment plans (29). Strong social support systems also support better expression and adherence. Studies show that patients with limited social support and low expressive skills face more difficulties sticking with treatment, highlighting the importance of social support in maintaining participation (30, 31). A patient’s level of insight the ability to recognize their condition and treatment needs—may also reflect their expressive style. Patients with high self-awareness often clearly express their treatment goals and concerns, promoting adherence (32). Conversely, those with low insight may find it hard to see the value and necessity of treatment, complicating adherence (11, 33). As a subcomponent of cognitive insight, expressive ability greatly influences adherence in psychiatric patients. Effective communication fosters a stronger therapeutic relationship with healthcare providers and enhances adherence. Emotional expression, coping strategies, and social support are also believed to affect TA outcomes.
The relationship between self-confidence and lack of insight in psychiatric patients is a complex interaction that significantly impacts adherence and overall mental health outcomes. Self-confidence, defined as an unwavering belief in one’s own understanding, beliefs, and perspectives, is generally negatively associated with insight, which refers to recognizing and accepting one’s mental illness. When measured with tools like the BCIS, self-confidence typically correlates with lower cognitive insight. BCIS assesses cognitive insight through two main components: Self-reflectiveness (the ability to evaluate oneself) and self-certainty (overconfidence in one’s beliefs). Self-confidence involves excessive trust in one’s beliefs and resistance to questioning them. Research shows that especially high levels of self-confidence reduce individuals’ awareness of their mental state and lower their level of insight (19, 34, 35). Self-confidence relates to a rigid stance on one’s thoughts and perceptions and often leads to resistance against external feedback and corrections. Cooke and colleagues discovered that higher self-confidence scores in patients with schizophrenia correlated with lower cognitive insight, indicating a tendency to ignore evidence that challenges their beliefs (17).
This resistance can hinder therapeutic engagement, as patients might reject treatment suggestions or not recognize the need for adherence. High self-confidence can also suppress the expressive component of cognitive insight, which enables individuals to critically evaluate their thoughts and behaviors. Studies in clinical contexts like first- episode psychosis show that individuals with high self-confidence demonstrate significantly lower expressive ability (18, 36, 37). This cognitive dynamic can make it harder for individuals to recognize and accept psychotic symptoms, leading to symptom worsening and clinical deterioration (24, 38). Research indicates that self-confidence not only weakens cognitive insight but also reinforces maladaptive thought patterns that contribute to emotional dysregulation and are linked to worsening psychiatric symptoms (25). In summary, the relationship between self-confidence and lack of insight in psychiatric patients especially those within the psychotic disorder spectrum is a crucial interaction that complicates treatment. High levels of SC may limit expressive capacity and the ability to recognize mental states, reinforcing cognitive biases that worsen psychiatric symptoms. Addressing these cognitive processes through therapeutic interventions that improve insight and cognitive flexibility is vital for better clinical outcomes and care quality.
In this study, the finding that the “self-confidence” subdimension had no statistically significant effect on TA is noteworthy. This result suggests that individuals’ excessive confidence in their thoughts and beliefs may not directly hinder TA. While the literature often links high self-confidence to poor insight and lower adherence (18, 39), such a relationship was not observed here. This could be because the sample exclusively included individuals diagnosed with bipolar disorder, most of whom were receiving inpatient treatment. Additionally, the specific context of this study may further explain why the self-confidence subdimension did not significantly affect TA (12, 24). These findings highlight the importance of incorporating communication- focused interventions and psychoeducation programs into psychiatric care. For example, structured psychoeducation sessions that help patients articulate their symptoms, treatment concerns, and emotional experiences may strengthen the therapeutic relationship and improve adherence.
Communication focused therapies, like social skills training and cognitive- behavioral strategies, can further improve patients’ expressive abilities by providing a safe environment for practicing emotional expression and perspective- taking (12, 39). Given Turkey’s cultural context, where family involvement in care is common, including family members in psychoeducation sessions could also reinforce communication skills and promote adherence by fostering a shared understanding of illness management. Overall, tailoring interventions to improve self-expression may not only boost TA but also positively influence broader psychosocial functioning. In inpatient psychiatric settings, where treatment plans are structured and closely monitored, high self-confidence might have less impact, as adherence relies more on external regulation than on individual motivation. Studies show that poor insight is linked to lower adherence and weaker therapeutic alliances in schizophrenia spectrum disorders, contexts similar to inpatient environments.
Additionally, cultural factors may influence insight dimensions; in collectivist societies like Turkey, adherence to medical authority and group- oriented values can reduce the negative effects of high self-confidence on treatment compliance. Therefore, the intensive clinical support available in inpatient care and cultural norms around authority and obedience may lessen the impact of self-confidence as a factor in TA. In these environments, patients tend to follow medication and treatment protocols more consistently due to the structured setting and ongoing professional oversight. Moreover, the influence of self-confidence may also depend on the stage of illness acceptance or the severity of psychotic symptoms. In this context, high self-confidence does not necessarily indicate treatment resistance; some patients may remain adherent to their treatment without critically questioning their choices.
In the present study, the Cronbach’s alpha for the MMAS-4 was 0.61, which falls below the generally accepted threshold of 0.70 for internal consistency. This relatively low reliability may partly result from the brevity of the scale, as shorter instruments often produce lower alpha coefficients. Additionally, the MMAS-4 emphasizes a limited range of adherence behaviors, which might not fully represent the multidimensional nature of TA in psychiatric populations. While the MMAS-4 has been widely used in clinical research and its brevity provides practical advantages, the lower reliability suggests that any observed associations between TA and cognitive insight should be interpreted with caution.
An additional strength of this study is its use of SEM, which provides advantages over traditional correlational analyses. Unlike simple correlation methods that look at relationships between two variables at a time, SEM models multiple relationships simultaneously and accounts for measurement error. This approach offers a more accurate estimation of how different subdimensions of cognitive insight contribute to TA and allows for testing theoretical models in a single, comprehensive framework. Using SEM in this context introduces a new perspective to psychiatric research and reveals insights that might be missed with less advanced analytical techniques.
Strengths of the Study
This study addresses essential gaps in the literature by examining the impact of cognitive insight on TA in a sample of psychiatric patients using SEM. The integration of metacognitive constructs within a structural framework allows for a more nuanced understanding of how reflective thinking and belief certainty influence adherence behaviors. By analyzing a sample that predominantly consisted of individuals diagnosed with bipolar disorder, the study offers novel insights into the determinants of TA in this population and contributes to the development of targeted, insight-informed interventions.
Study Limitations
This study has several limitations. First, the cross-sectional design precludes any causal inferences regarding the relationship between cognitive insight and TA. Second, the sample was drawn from a single public hospital in Istanbul, which may limit the generalizability of the findings to other psychiatric populations or clinical settings. Third, the study focused on the relationship between cognitive insight subdimensions and TA within a SEM framework, and did not perform subgroup analyses based on variables such as gender, educational level, employment status, treatment duration, or presence of comorbidities. Although these factors may influence TA, they were beyond the scope of the current research questions and dataset. Future studies should incorporate these variables into their analyses to provide a more comprehensive understanding of the factors affecting adherence. Finally, self-report measures were used to assess TA and cognitive insight, which may be subject to response biases. Another limitation of this study is the relatively low internal consistency of the MMAS-4 (Cronbach’s alpha =0.61). The brevity of the scale and its narrow behavioral focus may limit the accuracy and stability of the adherence measurement. Consequently, the findings related to TA should be interpreted with caution, and future research may benefit from using longer or more comprehensive adherence instruments to improve reliability. The face-to-face administration of the questionnaires by the researcher in a psychiatric inpatient setting may have introduced response bias, as participants might have tended to provide socially desirable responses. This should be taken into account when interpreting the self-reported measures of cognitive insight and TA. Additionally, the structural equation model explained only 10.5% of the variance in TA. This indicates that a large proportion of the variability in adherence is likely explained by other psychosocial, clinical, and contextual factors not included in the model. Future research should incorporate additional variables such as illness severity, treatment history, social support, and comorbidities to improve the model’s explanatory power and provide a more comprehensive understanding of the determinants of TA.
Conclusion
This study demonstrates that cognitive insight particularly the EO dimension is a critical factor influencing TA in psychiatric patients. The findings indicate that patients with a greater ability to express themselves tend to adhere more closely to treatment plans, likely due to their enhanced capacity to communicate personal experiences and engage more effectively in therapeutic interactions. In contrast, the SC dimension did not show a statistically significant impact on TA, suggesting that excessive confidence in one’s beliefs may not directly hinder adherence behavior in this particular sample.
These results underscore the importance of integrating interventions that enhance communication and self-reflective abilities into psychiatric care. Improving patients’ capacity to articulate their emotions and thoughts may strengthen therapeutic alliances with healthcare professionals, thereby facilitating the delivery of individualized care and promoting higher levels of adherence.
From a clinical standpoint, mental health professionals can translate these findings into practice by incorporating structured psychoeducation sessions, role-playing exercises to improve emotional articulation, and communication-focused therapies into routine care. Interventions should also be adapted to the patient’s cultural and clinical context, with special consideration given to those with limited expressive abilities.
Future research should employ longitudinal and experimental designs to explore the role of cognitive insight subdimensions across diverse psychiatric populations. Testing targeted intervention programs that explicitly focus on enhancing expression and insight while assessing their impact on adherence could provide actionable evidence for improving treatment outcomes.
