Previously, severe mental illness was predominantly regarded as chronic and deteriorating. However, over recent years there has been a shift towards more recovery-oriented perspectives and approaches (1, 2) Numerous studies demonstrate that many individuals with severe mental illness achieve recovery in terms of symptomatology, (social) functioning or/and personal growth (3–5). Despite the increasing acceptance and comprehension of these recovery-oriented approaches, the implementation and delivery of services continues to pose significant challenges (2, 6).

The prevalence of stigmatization towards patients with severe mental health problems contributes to the challenges in implementing recovery-oriented approaches (7). Even though mental healthcare professionals (MHCPs) themselves do not always recognize having stigmatizing attitudes (8), stigmatization within mental healthcare remains common (9, 10). Disturbingly, over twenty percent of reported instances of stigmatization are attributed to encounters with MHCPs (10) even though attitudes of MHCPs seem overall more positive compared to the general public (11). One of the underlying causes can be a binary perspective held by some MHCPs, where mental health and mental illness are perceived as fundamentally distinct, leading to stereotypes and a separation of ‘us’ versus ‘them’; seeing ‘them’ as different from their own ‘normal’ experiences and behavior. Stereotypes and beliefs can be implicit (i.e., beliefs that reside outside of conscious control and/or awareness; associations that one would not explicitly endorse or reveal) or explicit (beliefs that are self-reported and occur within conscious control and/or awareness). There is an increasing understanding that explicit measures may underestimate levels of stigmatization (12, 13). Further, explicit and implicit measures may differentially predict behavioral outcomes due to operating through reflective (e.g., basing decisions on knowledge about facts and values) versus impulsive (e.g., basing decisions on associative links and motivational orientations) systems (14), or based on whether outcomes are controllable or spontaneous (15). Both implicit and explicit stereotypes and beliefs are associated with labeling, and can result in the marginalization and discrimination of a specific group. This categorical distinction plays a significant role in the stigmatization process (16). Stigmatization of individuals or groups with mental health problems involves perceiving, experiencing, or labeling mental conditions or issues as deviant or negative, leading to the detrimental consequences of discrimination, prejudice, exclusion, stereotyping, and status loss (17).

Potential consequences of stigmatization are a decrease in life satisfaction, lesser likelihood of help seeking behavior, less employment chances, less housing opportunities, fewer social opportunities and received healthcare is of inferior quality (18–23). Stigmatization does not affect all individuals with mental health problems in the same way. Some respond with indifference towards stigmatizing behavior and others show high levels of psychological resilience, which may even fuel their motivation to challenge negative beliefs (24). However, the majority of individuals tend to endorse and thereby suffer from stigmatizing attitudes (25). When someone considers stigmatizing beliefs to be true, these beliefs can be internalized and cause self-stigmatization. Self-stigmatization is the process in which an individual thinks that he or she is socially less acceptable and/or valuable, it involves identification with a stigmatized group, and the internalization of the negative attributes and stereotypes associated with this group (26). Self-stigmatization is prevalent and is related to demoralization, reduced physical and/or mental health and a diminished self-efficiency, self-esteem and quality of life (27) (25, 28) (29–31).

Several meta-analyses on the effectiveness of interventions to diminish stigmatization (16, 32, 33) and self-stigmatization (24, 31, 34, 35) have been published. These include psycho-education programs, trainings, individual- and group sessions. The results vary substantially; some studies show no effects, while others report high effectiveness in reducing stigma. One of the challenges in providing stigma-reduction trainings is that people often resist acknowledging or recognizing stigmatizing beliefs in themselves because this threatens their self-image. This can cause social desirability and consequently unreliable outcomes on stigma measurements (36, 37). Recently there has been a shift towards recommending promoting continuum beliefs (e.g. understanding that mental health problems can be seen on a continuum from mild to severe; to some degree all humans suffer from them) rather than explicitly targeting the reduction of stigmatizing attitudes (16, 37). Additionally, an ethical question arose while reviewing self-stigma intervention studies: shouldn’t the first step be to offer interventions to mental health care workers (MHCWs) before addressing self-stigmatization in patients? This is important because, for many patients, self-stigma often begins within mental health care and is driven by labelling and stigmatization. It seems reasonable to tackle self-stigma by focusing on stigma reduction among mental health care workers (38). Given that categorical beliefs play a pivotal role in the stigma process, continuum beliefs can be regarded as a countermeasure against stigmatization (16). We conducted a pilot study to understand the effects of an intervention enhancing continuum beliefs, based on acceptance and commitment therapy (ACT), offered to MHCPs to reduce self-stigmatization in their patients. A continuum concept of mental health and mental illness assumes one dimension from severe psychiatric symptoms to subclinical, light, or non-existent symptoms. Since every person is likely experiencing symptoms of mental illness at some points during their life, a person with severe mental illness might be seen as someone with similar, but with more severe experiences, thus remaining ‘someone like us’ (16). In ACT, mental health problems are explained using continuum beliefs, understanding that we all experience emotional difficulties to some extent and in everyone the intensity of the struggle fluctuates over time. Furthermore, ACT enhances psychological flexibility, which is defined as the ability to maintain a gentle and kind awareness of thoughts, emotions, and the present moment, and take action, based on personal values and goals. Psychological inflexibility is found to be positively related to stigmatization (39, 40). ACT interventions are promising in the reduction of (self)stigma of mental health problems (38, 39, 41). The current study focusses on whether self-stigmatization in patients decreases after their MHCP received ACT, with a focus on enhancing continuum beliefs. It is hypothesized that after the intervention for their MCHPs, patients’ agreement with, beliefs about and application of common mental health stereotypes as well as the ‘hurt-self’ (self-esteem or self-efficacy as a consequence of internalizing these stereotypes) decreases compared to the control group.

No significant differences were found between characteristics of patients and MHCPs in both conditions (the age of MHCPs, the amount of years working in mental health care, the age of patients, the amount of years of patients having mental health problems, gender, having mental health problems in loved ones or personal mental health problems in MHCPs, the function of MHCPs, gender of patients, marital status of patients, and living situation of patients).

No significant differences between conditions on the SSMIS-SF were found, t(39) = -0.43, p = .67, indicating that pre-test self-stigmatization scores were not different over conditions (see Table 2 for descriptive statistics).

Considering Application, no significant differences between conditions on pre-test scores were found, t(39) = -1.71, p = .10, indicating that pre-test Application scores were not different over conditions. No significant differences between conditions on Agreement, t(39) = 0.74, p = .46, Awareness, t(1, 33.52) = 1.29, p = .21, and Hurt-self, t(1, 33.11) = -1.98, p = .06, were found, indicating that pre-test scores were not different over conditions.

No effect of time was found on the subscale Awareness, F(1,39) = .04, p = .85 and no effect of condition was found for Awareness, F(1,39) = .85, p = .36, also no significant interaction effects were found F(1,39) = 2.02, p = .16 suggesting that the ACT-training had no effect on the ‘Awareness’ subscale, compared to the control condition.

For the subscale Agreement, a significant effect of time was found ( Figure 3 ), F(1,39) = 9.86, p <.01 η_p ² = .87, meaning that over time Agreement scores decreased in both conditions. No significant effect of condition was found for Agreement, F(1,39) = .00, p = .98, and no significant interaction effect was found, F(1,39) = 1.31, p =.26. suggesting that the ACT-training had no effect on the ‘Agreement ‘ subscale, compared to the control condition.

For the subscale ‘Application’, the analysis demonstrated no significant effect of time, F (1,39) = .11, p = .74 or condition, F(1,39) = .19, p = .67. However, a significant interaction effect of time and condition was found, F (1,39) = 9.33, p < 0.01, η_p ² = .85 ( Figure 4 ). This finding suggests that the ACT-training had a significant favorable effect on the ‘Application’ subscale, compared to the control condition with an effect size of .193 (50, 51). The increase in ‘Application’ scores in the control condition was approximately as large as the decrease in the experimental condition ( Figure 5 ).

Hurt-self: On the subscale no effect of time F(1,39) = 1.90, p = .18. or condition F(1,39) = 1.66, p = .21. was found. Furthermore, a non-significant trend interaction effect was found F(1,39) = 3.78, p = .056. suggesting a potential favorable effect in the intervention condition.

The focus of this study was to target self-stigmatization as it is known to be prevalent and detrimental for many patients in mental healthcare. To the best of our knowledge, the present study has been the first study to examine the impact of training MHCPs in ACT on reducing self-stigmatization among their patients. A significant effect of the intervention on the subscale Application was found; indicating a positive effect on reducing the internalization of mental health stereotypes among patients. Interestingly, scores on the Application subscale increased over time in the control condition. The hypothesis was not supported for the other subscales: Awareness, Agreement and Hurt-self (although favorable effect concerning Hurt-self were observed at a trend level). These findings suggest that the intervention did not result in significant changes in awareness of, agreement with, commonly held stereotypes, or self-esteem.

Self-stigmatization is commonly conceptualized as a process, encompassing identification with a stigmatized group and the internalization of negative attributes and stereotypes associated with that group (17). The subscales utilized in this study assess different stages of the self-stigmatization process and the subscales can only be interpreted separately. Awareness: this stage involves individuals becoming conscious of the existing stereotypes about mental illnesses, such as beliefs held by the public or mental healthcare professionals that attribute blame to individuals with mental illnesses for their problems (e.g., “The public/MHCP believes most people with mental illnesses are to blame for their problems.”). Agreement: After becoming aware of these stereotypes, individuals may then internalize and agree with them, reflecting their personal endorsement of such beliefs (e.g., “I think most people with mental illness are to blame for their problems.”). The first two stages of the process are not ‘personal’ yet; one can be aware of or agree with the statements without experiencing self-stigmatization. Application: At this stage, individuals apply the stereotypes onto themselves, attributing the negative stereotypes to their own experience with mental illness (e.g., “Because I have a mental illness, I am to blame for my problems.”). Hurt-self: As a consequence of the preceding stages, individuals may experience negative impacts on their self-esteem and self-worth, perceiving a loss of respect for themselves due to the belief that they are responsible for their problems (e.g., “I currently respect myself less because I am to blame for my problems.”). One implication of this model is that the ‘harmful’ effects of self-stigmatization do not occur until the later stages (Application and Hurt-self) when the person has internalized the stigma (32).

The subscale “Application” measures the internalization of negative stereotypes about mental health problems: from all the subscales this definition is most in line with the most commonly used definition of self-stigma (52) and was the only subscale that changed significantly. We propose that there was no significant effect found on the recognition of or the agreement with stereotypes alone (as measured by the Awareness and Agreement subscales) because these subscales measure beliefs that have been formed over a lifetime of socialization and were not affected by the ACT intervention. ACT does not attempt to decrease awareness or to change beliefs but to have more awareness of how these beliefs work and of how to create cognitive flexibility around them (for example to make people aware of the possibility not to apply the stereotypical beliefs they have onto themselves). So perhaps ACT doesn’t quickly modify the presence of self-stigmatizing psychological content, but alters its impact or function (e.g. believability). Varra et al. showed that ACT reduced the believability of barriers and greater psychological flexibility mediated the impact of the intervention in clinicians (53). Their results support the idea that acceptance-based interventions may be helpful in addressing the psychological factors related to poor adoption of evidence-based treatments. Unlike our study Hayes et al. observed that ACT also significantly changed the believability of stigmatizing attitudes (54).

Interestingly, the increase in Application scores in the control condition was approximately as large as the decrease in the experimental condition. A theory that could explain the observed increase of Application scores in the control condition can be found in the Ironic Process Theory. This theory states that negative (self)beliefs are partly suppressed into the subconscious (55). Addressing beliefs for instance through measurement with stereotypical statements has the potential to increase individuals’ awareness of them and hence can result in higher scores over time (56). An explanation for why this effect did not appear in the other subscales (Awareness and Agreement) is that the latter two scales measure ‘general’ beliefs, the subscale “Agreement” aims to measure the agreement with commonly held stereotypes about individuals with mental health problems and the subscale of Awareness measures the knowledge of common mental health stereotypes in society. Possibly these scores are not affected by the Ironic Process Theory because general beliefs are less likely to be suppressed (52).

On the subscale ‘Agreement’ patients in both conditions scored lower over time meaning that, because treatment as usual which both conditions received or because of other factors, they agreed less with the stereotypical beliefs as stated in the ISMI-SF. Interestingly, despite a decrease in ‘Agreement’ with the commonly held stereotypes observed in the control condition, there was still an increase in the application of these stereotypes onto oneself. This indicates that even though patients agreed less with stereotypes over time they still continued to apply them to themselves and internalized them. It could be that somehow the explicit beliefs (conscious beliefs) changed but the implicit (subconscious) beliefs did not which caused the application of the beliefs onto oneself to remain.

Furthermore, it is noteworthy that the intervention did not result in significant changes in the Hurt-self subscale. This subscale holds significance within the stigmatization process, as it has been found to be correlated with self-esteem, self-efficacy, empowerment, and hope (29). It is possible that these deeply rooted self-esteem beliefs are less amenable to change within the scope of a single workshop and session with mental healthcare professionals. Additional interventions or prolonged engagement may be required to stimulate change in this particular area.