The Maristán stigma scale: a standardized international measure of the stigma of schizophrenia and other psychoses

The aim in this final phase of the study was to recruit up to 30 patients from each country, aged 18 years and over whose diagnoses of schizophrenia and related psychoses [F20-F29, International Classification of Disease version 10] was confirmed by the responsible psychiatrists. Participants were attending outpatient psychiatric services or day centres, had experienced a minimum of three years since the first known contact with services and had at least one year of continuous contact with the service. We chose three years to be sure that the diagnosis was established and stable, and for sufficient time to have passed for patients to have lived with the diagnosis and experienced stigma. Participants with moderate or profound intellectual disability and those with a history of illness of over 15 years since first contact with services were excluded. The latter was to restrict the sample to people with lesser degrees of institutionalisation, who receive care in model of community mental health. It was also our clinical observation that patients with longer term illness become accustomed or resigned to stigma and may be less likely to notice its effects on a day-to-day basis. Participants were identified from the clinical service registers in each participating centre. In the UK a key clinician invited patients to participate and if they agreed they were put in contact with a researcher who confirmed the diagnosis and inclusion criteria. In other countries researchers used a mixture of direct contact with patients and recruitment via clinicians. A written informed consent was obtained in keeping with each country’s guidance for good clinical practice in research. The study was approved by appropriate research ethical committees of the Faculties of Medicine and local health services in each country (Ethic Committees in National University of Lanus, Argentina, Maternal and Children’s Hospital Presidente Vargas, Porto Alegre, Brazil, Faculty of Medical Sciences in Nova University of Lisbon, Portugal, Faculty of Medicine in University of Concepcion, Chile, Ethic Committee in Human Research (CEIH) in Granada, Spain; and Camden and Islington Local Research Ethics Committee in UCL Medical School, UK).

The first draft of the Maristán Questionnaire that was developed for testing consisted of 38 statements to which the respondents indicate their degree of agreement. The statements concerned stigma in personal, family and social life, the attitudes of health professionals, and stigma in the public sphere and work. The format was a Likert scale with a score of 1 indicating complete disagreement and 7 indicating complete agreement with each statement. Where appropriate, the direction of scoring was reversed so that a high score indicates a high level of stigma. In all countries, the questionnaire was read out by an interviewer to make sure the meaning was clear.

As described, participants took part in individual face-to-face interviews in which the questionnaire was delivered by professionals who had clinical or other experience in the care of people with severe mental illness. All of the interviewees were invited to answer the questionnaire again after 30 days in order to estimate temporal stability. Each investigator kept field notes regarding the application of the questionnaire, including the degree of comprehension of each statement, the language terms used, views on the length of the interviews and any adverse comments about them.

The analysis was undertaken by means of SPSS, version 11.0. An exploratory factor analysis was used to evaluate the structure of the scale, using Principal Axis Factoring [PAF], followed by a Promax rotation. This approach is one of the most recommended strategies for analysis of the structure of a rating scale [36‐38]. In order to determine the number of factors to extract we used Horn’s parallel analysis combined with sample selection or bootstrapping [39‐41]. This analysis addresses weaknesses in more traditional factor analysis. For example, the Kaiser-Guttman rule dictates that all factors with an eigenvalue of more than 1.0 should be retained but ignores the possibility of random variation in eigenvalues obtained from any one sample population. Similarly, in the Catell analysis of the scree plot to determine the number of factors to be retained, it is often not possible to make an unequivocal decision about the cut-point, which leads to indecision on how many factors to retain [39, 42‐45]. Pearson’s product moment correlations were used to obtain correlations between factors. Cronbach’s alpha was used to estimate internal consistency and the intra-class correlation coefficient (with 95% confidence intervals) for test-retest reliability. A series of ANOVA were used to evaluate differences in factor scores between groups of patients.

Patients’ social networks can be both a source of support and of stigma. Stigma complicates recovery and reduces self- esteem. When it arises in a person’s close relationships, it can have a severe knock-on effect on the take-up of opportunities for work, use of services, symptoms, hospitalization, trajectory of the illness, and quality of life. Although the family may be victim of stigma because one of its members has a mental illness [20], family members themselves can hold stigmatizing attitudes. The relationship between both processes is not yet clear; however recent research indicates that as people with schizophrenia become increasingly isolated, their family becomes the closest and most important support network. This may mean that patients are focused on this intimate relationship, even when it risks becoming a source of difficulties [33]. Few scales cover this issue in detail and it might be most relevant in societies with a lesser developed community mental health system.

Unfortunately, formal care givers appear to be an important source of stigma. Two of the factors [socio-institutional and stigma from health professionals] touched on this source of stigma. Thus it may be important to include both the system and professionals when developing strategies to reduce stigma. A frequent complaint against health professionals concerns their attitudes, while perceptions of socio-institutional stigma concern processes of care and scarcity of resources. The latter requires political and public health solutions in each country. Stigma arising from professionals and institutions may be less likely to be experienced by patients with less than three years of illness or those who have never been institutionalised. One modification when applying the Maristán scale to this group is to use only the subscales ‘informal social networks’ and ‘self-stigma’ , both of which were strongly correlated with each other and both of which tapped important, albeit incomplete elements, of perceived stigma in this group of people.

The relation between stigma and the demand for mental health care does not appear to be direct. A number of authors have suggested that perceived stigma may not affect demand for mental health care in particular populations, but this may vary with the characteristics of the population affected [46]. Furthermore, people with schizophrenia are in long-term contact with specialized services and if they perceive stigma arising from such institutions and from health professionals, specific measures need to be taken to address it.

Both families and professionals were sources of stigma. Previous research is contradictory on this issue. There have been suggestions that close interaction with patients will reduce stigmatising or negative attitudes [47], however more recent research indicate that those with the closest contact may be the most discriminating of all [33]. Thus, we need closer study of just how the degree of intimacy leads to an increase or decrease in stigma.

Of the four factors we derived, three identify sources of stigma, while self-stigma, one of the most commonly found in other instruments to measure stigma [12, 21], is the internalization of the public stigma; it is the experience of the self [17]. Our qualitative work however, showed that the other three areas of stigma were of much greater concern to patients and carers and thus this section of the scale is brief. The items in the self-stigma subscale capture internalized or subjective stigma and they reflect the prejudice which people with mental illness may turn against themselves [21], rather than explicit behaviours which give rise to discrimination. Although it includes a general item “people don’t respect me”, stigma perceived as arising from informal social networks included mainly family and other close persons and reflects the limited social networks which patients with psychotic mental illness can depend on. The fact that both factors [self-stigma and informal networks] are the most stable over time and have the greatest intra-factor correlation suggests they are closely related.

As opposed to other scales [14] there were no positive elements associated with suffering a severe mental illness. This may reflect the way in which the qualitative data were collected. In focus group discussions about stigma; patients and carers may be most concerned about discussing what is wrong in their lives rather than what is going well.

Users of services in Argentina, a country where psychiatric care is still largely concentrated in hospitals, had higher (worse) mean scores on the factors ‘informal networks’ , ‘socio-institutional’ and ‘health professionals’ than users in the United Kingdom, Spain, Portugal and Brazil, where community care appears to be better developed. This suggests that community based service provision is less stigmatising and professionals’ attitudes are less institutionalised and hierarchical, than in hospital based services.

There are a number of limitations to our study. First, the instrument has been tested in patients in contact with mental health services, a population that may have different perceptions of stigma than those who have no contact with services. Second, the questionnaire was validated in a relatively middle aged population that had been in contact with community services for a mean of over 12 years but had not had experience of long stay hospital care. We explain our reasons for that above. Third, the interviewers noted that a 5 point response scale might be simpler and thus the pattern of scoring may be revised in later testing of the instrument. Finally although we did not include any measure of criterion validity, we omitted this for two reasons: 1. No stigma scale exists that has universal application across these six cultures and 2. We believe that its multicultural, bottom-up development provides the validity needed.

We have presented the development and standardisation of the MARISTAN Stigma Scale. Further evidence on its structure, in particular confirmatory factor analyses, is needed, as well as its application in treatment services for schizophrenia and related psychoses around the world.