Somatic health care professionals’ stigmatization of patients with mental disorder: a scoping review

In health research, it has become increasingly common to engage with stakeholders such as policy makers; clinicians; and patients, just as many research institutions offers research-based collaboration and advice to external parties. These collaborations can help improve study questions and provide more useful findings [29]. In the design of the literature study, we collaborated with The Danish Health Authority and the organization EN AF OS, who aims to destigmatize mental illness in Denmark. They were both involved in developing the research question, and the organization was additionally involved in qualifying keywords and search strings. As recommended by Arksey and O’Malley, we both considered relevant aspects of the research question (e.g. study population, phenomenon of interest and context) and were aware of developing a research question with a wide approach in order to generate breadth of coverage [28].

The first author developed the search strategy in collaboration with research librarians. The search strategy was based on the PICo model that specifies the population (P), phenomenon of interest (I), and context (Co). Following the PICo model the search consisted of two search strings. The first search string specified the population (P), which included patients with frequently occurring mental disorders. We used broad search terms such as ‘mental disorders’ and additionally performed searches on seven specific disorders, since a too broad search can result in missing relevant studies. The seven specific disorders included: anxiety, depression, bipolar disorder, borderline, schizophrenia, eating disorders, and self-harm. We chose these, because they are among the most common and disabling mental disorders [2]. The second search string specified the phenomenon of interest (I) and context (Co), which included stigmatizing behaviors, attitudes, and perceptions among health professionals, e.g. nurses and physicians, in the somatic health care system (see Additional file 1 for full search string). The search was based on keywords (subject headings/ MeSH terms) and free text searches (title, keywords, text). We tested the search string before we formed the final search string. To delimit the search, we applied proximity searching of two words. The search was carried out in May 2019 in three databases: Cinahl, MEDLINE and PsycINFO. Title, abstract and full paper screening were based on inclusion and exclusion criteria. To narrow down the search, we only included studies published 2008–2019 in peer-reviewed journals in English or Scandinavian languages. The temporal delineation was made as we wanted to find the most recent published literature in the field. Protocols, conference literature, book chapters, opinion papers and reviews were excluded. We excluded studies focusing on the mental health of health professionals as well as studies where somatic and psychiatric health professionals could not be separated. While we only included observational studies from the Western Hemisphere, we applied no geographical exclusion criteria for the intervention studies. This is because we wanted to gain insight into experiences from interventions conducted all over the world and because the wide range of interventions may contribute as inspiration for future prevention work. We made the choice of only including observational studies from the Western Hemisphere in collaboration with the organization EN AF OS, to narrow down the search.

We stored the studies in Endnote and removed duplicates. Next, we moved the studies to the review manager Covidence and performed another duplicate check, after which we began the screening process of the 11,798 identified studies. During the initial title screening, we excluded 10,622 studies for being out of scope. Additionally, we excluded 928 studies during abstract screening for not meeting the inclusion criteria. Of the remaining 248 studies, 111 studies were excluded for not meeting the inclusion criteria. Finally, we included a total of 137 studies in the scoping review. Figure 1 summarizes the literature search and study selection.

We used NVivo to organize and analyze the data. The data charting took the following information into consideration: author(s), year of publication, country of origin, study design, study population (health professionals in the somatic health system), sample size, and mental disorders. Additionally, we coded the scope, method, intervention type, instrument, and outcome in each study. Subsequently, we summarized the studies by study design: observational studies and intervention studies.

As briefly described above, we collaborated with The Danish Health Authority and the organization EN AF OS, in the development of the design of the literature review and of the research question. The organization was additionally involved in the process of further qualifying keywords and search strings to ensure that relevant keywords, including mental disorders and specific health professionals were not omitted. Thus, the organization played an important role in quality assuring the first two stages of the scoping review. Further, we discussed the findings of the scoping review with the organization and learned that our findings corresponded well with the organization’s knowledge of the phenomenon.

To create an overview of the wide range of different health professions included in the identified studies, we divided the health professionals into 13 categories (Fig. 2). Some studies focused on students rather than trained professionals, and we categorized these studies separately. Other studies included both students and professionals. Moreover, while some studies included well-defined groups of health professionals, such as nurses or general practitioners, others did not focus on specific health professionals. These studies were categorized under the heading ‘Health professionals’, e.g. various hospital employees. Similarly, the category ‘Health care students’ refers to various students within health education. However, we also identified studies examining both students and health professionals, which formed the broad category ‘Health professionals and health care students’. Finally, the category ‘Medical doctors’ covers all other types of medical doctors besides general practitioners, e.g. surgeons or different types of medical specialists.

We also categorized the identified studies according to the patients’ diagnoses and formed seven categories. Five of the categories cover specific diagnoses. In addition, we included a ‘Mixed mental disorders’ category. This category includes studies not confined to a specific mental disorder or studies that examined multiple diagnoses. Furthermore, the category ‘Dual diagnosis’ covers studies examining patients with a mental disorder and a substance use disorder.

The main purpose of the observational studies was to investigate the magnitude of stigmatizing attitudes and behaviors among health professionals in the somatic health care system toward patients with mental disorders. In total, we identified 73 observational studies, all from the Western Hemisphere. About half of the studies were from Europe (n = 41), including a large proportion of studies from England (n = 16), followed by North America (n = 15), Oceania (n = 11), and Asia (n = 3). Furthermore, we identified three studies comparing populations across countries. Most studies used quantitative methods (n = 58); however, we also identified qualitative studies (n = 13) and mixed methods studies (n = 2).

As illustrated in Fig. 2, we found that most studies (n = 13) focused on emergency personnel, followed by medical doctors (n = 10) and health professionals (n = 8). Most studies (n = 40) focused on several, different diagnoses (Table 1). In studies focusing on single, specific diagnoses, the most frequent mental disorders were self-harm (n = 10) and schizophrenia (n = 9).

We combined type of health profession with diagnosis as shown in Table 1. The table shows that a relatively large proportion of studies examined attitudes and behaviors among emergency staff toward patients who self-harm (n = 6). In contrast, studies examining attitudes and behaviors among medical doctors (n = 10) were divided into a wide range of diagnoses, such as depression (n = 1), schizophrenia (n = 1), eating disorders (n = 2), self-harm (n = 2), mixed mental disorders (n = 3), and dual diagnosis (n = 1).

For a detailed description of study design, target group, sample size and diagnosis on observational studies, see Table 2.

The main purpose of the intervention studies was to evaluate interventions to reduce health professionals’ stigmatizing attitudes and behaviors toward people with mental disorders. We identified 64 intervention studies, most of which were from North America (n = 19), followed by studies from Europe (n = 16), Oceania (n = 13), Asia (n = 13), and Africa (n = 1). In addition, two interventions studies were comparative studies. Most studies were based on quantitative methods (n = 51); seven were mixed methods studies, and six were qualitative studies.

Most intervention studies focused on changing students’ attitudes and behaviors toward patients with mental disorders; 20 of these studies focused on medical students, 20 focused on nursing students and six focused on pharmacy students, as illustrated in Fig. 3. Furthermore, eight studies focused on general practitioners, constituting the third-largest category when distributing the studies by health profession.

Most intervention studies (n = 45) did not focus on patients with a specific mental disorder, but typically on attitudes and behaviors toward multiple mental disorders or mental disorder in general as shown in Table 2. Depression (n = 5), eating disorder (n = 4), and schizophrenia (n = 4) were among the most common diagnoses.

By combining health profession with diagnosis (Table 3), we found that most of the studies focusing on attitudes and behaviors among nursing students did not focus on a specific mental disorder, as 17 of the 20 identified studies looked at multiple diagnoses or mental disorder in general. Similarly, 16 of 20 identified studies examining attitudes and behaviors among medical students focused on multiple diagnoses or mental disorder in general. We found a similar pattern for studies examining attitudes and behaviors among pharmacy students (n = 6) and health care students (n = 1).

For a detailed description of study design, target group, sample size, and diagnosis in the intervention studies, see Table 4.

Of the 64 included intervention studies, 47 targeted health care students while 17 targeted health professionals. To provide a more detailed characterization of the type and content of the identified intervention studies, we categorized the interventions into eight main types: four targeting students and four targeting health professionals. We categorized the interventions targeting students based on the content of the intervention, whereas we categorized interventions targeting health professionals according to the content and specific health profession (e.g. nurses or medical doctors) because these were often closely related. We present one example of each intervention type, focusing on examples that are illustrative of the intervention types, well-described in the articles, and show a geographical breadth and variation between health care students and professionals. For a detailed description of the intervention studies, see Table 4.

We identified 17 intervention studies targeting health professionals, including general practitioners and other medical doctors (n = 9), emergency personnel (n = 3), nurses (n = 1), pharmacists (n = 2), and non-specific groups of health professionals (n = 2). We categorized these interventions into four intervention types: (1) Interventions targeting general practitioners and medical doctors, (2) Interventions targeting pharmacists, (3) Interventions targeting emergency personnel, and (4) Interventions targeting non-specific groups of health professionals.

This review has several limitations. First, as the search was limited to studies published after 2008, as well as studies in English or Nordic languages, we may have overlooked relevant studies and anti-stigmatizing interventions. Second, as mentioned above and as applies to scoping reviews in general, we did not assess the quality of the included studies.

Despite these limitations, we note several strengths of this study. First, a methodological strength is the fundamental systematic approach to charting the studies. We have systematically selected keywords and searched in selected databases based on several inclusion and exclusion criteria. Additionally, the screening process was undertaken on different levels: by title screening, abstract screening, and full-text screening in the review manager Covidence, after which all included studies were systematically organized and analyzed in NVivo. The use of NVivo contributed to a systematic organization of the included studies, which provided a useful tool for creating overview and high-level systematization of all included studies. Additionally, we consider it a strength that we have included intervention studies from the entire world. Although organization and practice within the somatic health care system are diverse worldwide, which reduces the transmissibility of the interventions, this knowledge contributes to the research field and can further inspire and mobilize new interventions. Finally, we consider it a strength that we have included a consultation element (e.g. stage 6) in our review process, as recommended by Arksey and O’Malley [28].