Self-stigma in Serious Mental Illness: A Systematic Review of Frequency, Correlates, and Consequences

Abstract

Self-stigma is associated with poor clinical and functional outcomes in Serious Mental Illness (SMI). There has been no review of self-stigma frequency and correlates in different cultural and geographic areas and SMI. The objectives of the present study were: (1) to review the frequency, correlates, and consequences of self-stigma in individuals with SMI; (2) to compare self-stigma in different geographical areas and to review its potential association with cultural factors; (3) to evaluate the strengths and limitations of the current body of evidence to guide future research. A systematic electronic database search (PubMed, Web of Science, PsycINFO, Scopus, and Ovid SP Cumulative Index to Nursing and Allied Health Literature [CINAHL]) following PRISMA guidelines, was conducted on the frequency, correlates, and consequences of self-stigma in SMI. Out of 272 articles, 80 (29.4%) reported on the frequency of self-stigma (n = 25 458), 241 (88.6%) on cross-sectional correlates of self-stigma and 41 (15.0%) on the longitudinal correlates and consequences of self-stigma. On average, 31.3% of SMI patients reported high self-stigma. The highest frequency was in South-East Asia (39.7%) and the Middle East (39%). Sociodemographic and illness-related predictors yielded mixed results. Perceived and experienced stigma—including from mental health providers—predicted self-stigma, which supports the need to develop anti-stigma campaigns and recovery-oriented practices. Increased transition to psychosis and poor clinical and functional outcomes are both associated with self-stigma. Psychiatric rehabilitation and recovery-oriented early interventions could reduce self-stigma and should be better integrated into public policy.

Introduction

For a long time, research on stigma in psychiatric disorders focused on public stigma, or the endorsement of negative beliefs and discriminating attitudes towards individuals by the general population. Over the last two decades, attention has shifted towards the effects of stigma on individuals with Serious Mental Illness (SMI). According to Link and Phelan,¹ stigma involves “elements of labeling, stereotyping, separation, status loss, and discrimination” that “co-occur in a power situation that allows the components of stigma to unfold.” The effects of stigma on individuals include perceived, experienced, anticipated, and self-stigma. Perceived stigma is defined as one’s beliefs about the attitudes of the general public towards people with SMI.² Experienced stigma refers to people with SMI’s experience of discrimination.³ Anticipated stigma—or the expectation that a person will be discriminated against because of their SMI—can occur even if the person has no previous experience of discrimination and contributes to social withdrawal and self-stigma.³ Self-stigma—or internalized stigma—describes the transformation process wherein a person’s previous social identity (defined by social roles such as son, brother, sister, friend, employee, or potential partner) is progressively replaced by a devalued and stigmatized view of oneself, known as “illness identity.” ⁴ Self-stigma occurs when a person moves beyond mere awareness of public stigma and actually agrees with it and applies it to themself.² Stigma resistance is defined as a person’s ability to deflect or challenge stigmatizing beliefs.^5,6

There is a growing body of evidence—including meta-analyses—on the effects of self-stigma on people with SMI.^7–12 In a meta-analysis of 127 articles, Livingston and Boyd⁷ reported that most studies were conducted in Europe or North America (77.5%), and that schizophrenia was the most common diagnosis (54.3%). Self-stigma is frequent in Europe (41.7% of 1229 participants with schizophrenia and 21.7% of 1182 participants with mood disorders^9,10) and North America (36.1% of 144 people with SMI¹³). Less is known about self-stigma in other geographic areas. The level of self-stigma might vary according to cultural factors (eg, causal attributions of mental illness¹⁴) and sociopolitical ideology.¹⁵ According to Yang,¹⁴ stigma affects “what matters most” in a local social world, by threatening one’s capacity to meet social expectations (eg, the ability to engage in key life activities such as work or marriage^14,16) and the whole family’s moral standing and socioeconomic status. Self-stigma might be more common in non-Western countries than in Western countries but this remains unproven.

Compared with those with nonpsychotic disorders, people at risk of psychosis face higher levels of public stigma that can lead to self-stigma.^17,18 People with BPD are prone to self-criticism and feelings of shame that can make them more vulnerable to self-stigma.¹⁹ They face high levels of stigma, not only from the general public but also from mental health professionals.²⁰ Gerlinger⁸ reported in a systematic review on stigma in schizophrenia the lack of studies comparing stigma at different stages of illness. It is still unclear whether self-stigma is greater in cases of prolonged psychosis compared with early psychosis, at-risk stages, or other SMI diagnoses.

To date, there have been no literature reviews comparing self-stigma frequency and correlates in different geographical and cultural areas (Europe, North America, South America, Middle East, South Asia, South-East Asia, and Oceania). Similarly, to our knowledge, no literature review has been conducted comparing internalized stigma in at-risk stages, schizophrenia, BD, MDD, BPD, and anxiety disorders. Self-stigma was negatively associated with self-esteem, self-efficacy, quality of life (QoL), and clinical and functional outcomes.⁷ However, it is difficult to disentangle the specific effects of self-stigma, as measures of perceived or experienced stigma are often used to assess the correlates of self-stigma (43.9% of the 127 studies⁷). Most studies were cross-sectional (86.7% of 127 studies⁷). The longitudinal effects of self-stigma remain largely unknown. Several interventions (combinations of psychoeducation and cognitive restructuring in most cases, making empowered decisions about disclosure in others) have been designed to reduce self-stigma and its impact on patients’ outcomes.²¹ Psychiatric rehabilitation brings together a wide range of recovery-oriented interventions.²² Psychiatric rehabilitation could indirectly reduce self-stigma (eg, through improved psychiatric symptoms, self-esteem, cognitive, social, and vocational functioning) although this remains to be investigated. However, the use of psychiatric rehabilitation services also carries the risk of increased labeling and self-stigma.^23,24

In summary, three reviews have been conducted on self-stigma since 2010, in schizophrenia⁸ and BD.^11,12 One meta-analysis investigated the correlates of stigma resistance in SMI.⁶ To our knowledge, there has been no review of self-stigma frequency and correlates in different cultural and geographic areas and for different SMI conditions; nor any specific reviews of self-stigma, excluding explicitly perceived or experienced stigma. The effects of nonspecific recovery-oriented practices on self-stigma remain unknown. Based on the literature, we would expect to find a higher frequency of self-stigma in Eastern countries compared with Western countries. We made the hypotheses that cultural factors would influence self-stigma and that self-stigma would be associated with poor recovery-related outcomes. The present review has three objectives: (1) to review the frequency, correlates, and consequences of self-stigma in individuals with SMI; (2) to compare self-stigma in different geographical areas and to review its potential association with cultural factors; (3) to evaluate the strengths and limitations of the current body of evidence to guide future research.

Methods

A stepwise systematic literature review (PRISMA guidelines)²⁵ was conducted by searching PubMed, Medline, Web of Science, PsycINFO, the Scopus citation index, and Ovid SP Cumulative Index to Nursing and Allied Health Literature (CINAHL) for published, peer-reviewed articles using the following keywords: “schizophrenia” / “bipolar disorder” / “borderline personality disorder” / “major depression” / “depression” / “anxiety disorder” / “serious mental illness” AND “stigma” / “self-stigma” / “internalized stigma”/ “internalised stigma.” No time restriction was set. Only published articles in English or French were included in the review. The reference list of three meta-analyses^6,7,26 and three literature reviews^8,11,12 on stigma and schizophrenia or BD were screened for additional relevant articles. To be included in this review articles had to meet all of the following criteria: (1) report explicitly on self-stigma (ie, articles on public stigma or using measures of perceived or experienced stigma were excluded); (2) concern a diagnosis of schizophrenia, bipolar disorder, borderline personality disorder, major depression or anxiety disorders; (3) provide quantitative data on the prevalence, correlates, or consequences of internalized stigma or stigma resistance. The first author applied the eligibility criteria and screened the records to select the included studies. The last author reviewed each decision. Disputed items were solved through discussion and by reading the article in detail to reach a final decision. For each study, we extracted the following information: general information (author, year of publication, country, design, population considered, setting, total number of participants, mean age, or age range), outcome measure (scale/items used to measure self-stigma, reliability), the main findings, and variables relating to quality assessment (see supplementary table 2 for the detailed characteristics of the included studies). Quality assessment was performed using the Systematic Appraisal of Quality in Observational Research (SAQOR) tool. This tool comprises six domains (sample, control/comparison group, exposure/outcome measurements, follow-up, confounders, and reporting of data²⁷) and has been adapted for cross-cultural psychiatric epidemiology studies.²⁸ An overall quality score (high, moderate, or low) was determined based on adequacy in the six domains. Means and percentages were weighted for the number of cases per study to obtain prevalence data. Derived weighted means by geographical area and pooled standard deviations were calculated. One-way ANOVA was conducted from these summary data and post hoc pairwise test comparisons were computed using the Tukey–Kramer method. Weighted scores were calculated using proportions of rating scale scores. The frequency of self-stigma is often measured using the Internalized Stigma of Mental Illness Scale.²⁹ A score above the midpoint indicates a moderate to high level of self-stigma.^9,29 This choice was made for practical reasons (ie, facilitating comparisons between the studies) as there is no valid cutoff for measuring self-stigma. Stigma resistance is often measured using the ISMI stigma resistance subscale, which shows variable internal consistency (0.56 in Firmin et al meta-analysis⁶). Only the studies reporting internal consistency of above 0.50 were considered when extracting the correlates of stigma resistance. The study protocol was registered on the PROSPERO database on July 4, 2019 (ID 141282).

Results

Our search on July 5, 2019 found 3215 articles on PubMed and 11 472 on Web of Science. It was completed by searching PsycINFO, the Scopus citation index, and Ovid SP CINAHL on March 25, 2020 then updated on April 16, 2020. The search was completed on August 22, 2020 using the additional terms “psychosis,” “depression,” “internalized stigma,” and “internalised stigma.” This resulted in 5371 supplementary articles. After manually removing all duplicates, there were 4250 remaining references. Based on their titles and abstracts, 3811 articles were excluded for lack of relevance. Most of these articles focused on public stigma, perceived or experienced stigma, or self-stigma in other discriminated populations. Our search strategy yielded 429 full-text articles. After conducting a full-text analysis of all these articles and excluding those which did not meet the inclusion criteria, we ended up with 272 relevant articles (figure 1). See supplementary table 5 for the list of the excluded studies.

Fig. 1.

Review process (Prisma flow diagram)

The 272 articles included were characterized by the heterogeneity of the samples, methods, scales, and reported outcomes. Most were published after 2010 (244 studies; 89.7%) and used cross-sectional designs (231 studies; 84.9%) with only 41 studies (15.1%) reporting longitudinal outcomes. A total of 89 (95 studies (34.9%) were conducted in Europe, 76 (27.9%) in North America, 44 in South-East Asia (16.2%), 24 (8.8%) in the Middle East, 13 (4.8%) in Africa, 10 in South Asia (3.7%), 4 in South America (1.5%), 4 in Oceania (1.5%), and 2 studies (0.7%) compared internalized stigma in different geographical areas (see supplementary table 1 for the geographical distribution of the included studies).^30,31 Most studies included outpatients (211 studies, 77.6%). A total of 30 studies (11.3%) were conducted in a psychiatric rehabilitation context, 5 studies (1.8%) in consumer-operated service programs or advocacy groups and 6 studies (2.2%) in prison settings.

A total of 114 studies (41.9%) concerned schizophrenia, 14 (5.1%) BD, 13 (4.8%) MDD, and 13 (4.8%) at-risk stages or first episode psychosis, 2 (0.7%) anxiety disorders and 1 obsessive-compulsive disorder (0.4%). A total of 115 studies (42.3%) looked at SMI. There were large variations in the definition of SMI. Twenty-five studies (9.6%) defined SMI as schizophrenia, BD, or MDD. Fifty-four studies (20%) used a broad definition of SMI and included participants with anxiety disorders (AD, n = 25), obsessive-compulsive disorder (n = 11), and personality disorders (13 studies). Twelve studies (4.6%) investigated self-stigma in schizophrenia and BD (n = 6), schizophrenia and MDD (n = 1), mood disorders (n = 2), schizophrenia and AD (n = 1), AD and MDD (n = 1), and BPD and social phobia (n = 1). Twenty-four studies (8.1%) did not specify the different forms of SMI. Most studies included young (18–34 years old; 50 studies [18.4%]) or middle-aged participants (35–50 years old; 185 studies [68%]). See supplementary table 6 for the characteristics of studies that included people less than 18 years old. One hundred and ninety-two studies (70.6%) used the Internalized Stigma of Mental Illness²⁹ scale to measure internalized stigma and 33 studies (12.1%) the Self-Stigma of Mental Illness Scale³² (SSMI). Forty-seven studies (17.3%) measured self-stigma with other scales. Most of the instruments used to measure self-stigma in non-Western countries were adaptations of scales designed in Europe or North America (n = 74; 80%). The quality ratings of the included studies obtained using SAQOR²⁷ ranged from high to moderate (respectively, 57.7% and 42.3%). The results are shown in table 1.

Table 1.

Research Characteristics of the 272 Studies Included in the Review

Characteristic	All studies (n = 272) %
Publication date
Pre 2011	28 (10.3%)
2011 or later	244 (89.7%)
Study design
Cross-sectional	231 (84.9%)
Longitudinal	41 (15.1%)
Region of study
North America	76 (27.9%)
Europa	95 (34.9%)
South Asia	10 (3.7%)
South East Asia	44 (16.2%)
Middle East	24 (8.8%)
South America	4 (1.5%)
Africa	13 (4.8%)
Australia	4 (1.5%)
Others	2 (0.7%)
Study sites
Single site	270 (99.3%)
Multiple countries/sites	2 (0.7%)
Diagnosis
Severe mental illness	115 (42.3%)
Bipolar disorder	14 (5.1%)
Obsessive-compulsive disorder	1 (0.4%)
Major depressive disorder	13 (4.8%)
Anxiety disorder	2 (0.7%)
Schizophrenia	114 (41.9%)
At risk stages/first episode psychosis	13 (4.8%)
Mean age
<18 years old	3 (1.1%)
18–34 years old	50 (18.4%)
35–50 years old	185 (68%)
>50 years old	13 (4.8%)
Mixed	21 (7.7%)
% articles including people <18
Yes	24 (8.8%)
No	248 (91.2%)
Internalized stigma measures
Internalized Stigma of Mental Illness (ISMI)	192 (70.6%)
Self-Stigma of Mental Illness Scale (SSMIS)	33 (12.1%)
Others	47 (17.3%)
Patient status
Inpatient	28 (10.3%)
Outpatient	211 (77.6%)
Mixed	32 (11.8%)
Not reported	1 (0.3%)
Psychiatric rehabilitation
Yes	47 (17.3%)
No	225 (82.7%)
Quality rating
High	157 (57.7%)
Moderate	115 (42.3%)

Frequency of Self-stigma

Eighty articles (29.4%) reported data on self-stigma extent in SMI, or on the proportion of individuals with moderate to high self-stigma, or on both outcomes. Thirty-three studies were conducted in Europe, 20 studies in North America, 17 in South-East Asia, 9 in Africa, 12 in Middle East, 6 in South Asia, 1 in South America, 1 in Australia, and 1 in Austria and Japan. Nine studies compared the frequency of self-stigma in different countries,^{9,10,30,33,34} cities,³⁵ or settings.^13,36,37 The samples were mostly composed of individuals with schizophrenia (31 studies; 43.4%) or SMI (33 studies; 40.5%). The results are shown in table 2. Forty-seven articles reported on the proportion of moderate to high self-stigma (ISMI >2.5 or above the midpoint on other scales) in a total of 15 871 participants (7500 SMI, 5518 schizophrenia, 1582 BD, 1188 MDD, 64 BPD, 19 AD). Eighty-five articles documented self-stigma extent (ISMI mean total score and standard deviation) in 25 458 participants (11 028 with SMI, 9661 with schizophrenia, 2083 with BD, 2154 with MDD, 377 with AD, 155 with BPD).

Table 2.

Frequency of Internalized Stigma

Area		Study	Country	N	Mean IS total score		SD	n	High IS (%)
Europe	SMI	Evans-Lacko51	Multi-site	1835	2.2		0.5	-	-
		Grambal169	Czech Republic	184	2.19		0.5	-	-
		Rusch58	Switzerland	186	1.9		0.6	-	-
		Xu57	Switzerland	141	1.91		0.61	-	-
		Kamaradova170	Czech Republic	332	2.10		0.5	-	-
		Bradstreet45	UK	272	2.2		0.49	-	-
		Krajewski33	Multi-site	786	2.29 (1.96–2.71)		0.5	786	33% (15.2%–57.4%)
		Dubreucq93	France	738	2.20		0.51	738	31.2%
		Oexle96	Switzerland	222	2.10		0.55	-	-
		Rusch107	Switzerland	116	2.15		0.54	-	-
		Szczesniak171	Poland	114	2.23		0.5	-	-
	Weighted total			4926	2.18		0.26	1524	32%
	Schizophrenia	Brohan9	Multi-site	1129	2.40 (2.0–2.97)		0.56	1129	41.7%
		Vidovic126	Croatia	149	2.13		0.93	149	22.8%
		Sibitz172	Austria	157	2.09		0.67	-	-
		Galderisi173	Italia	921	2.1		0.5	-	-
		Rossi174	Italia	910	2.2		0.44	-	-
		Aukst-Margetic81	Croatia	117	2.13		0.44	-	-
		Hofer30	Austria	52	2.01		0.51	-	-
		Bouvet175	France	62	2.23		0.46	-	-
		Vrbova176	Czech Republic	197	2.18		0.46	-	-
		Surmann177	Germany	80	2.12		0.47	-	-
		Dubreucq93	France	466	2.18		0.51	466	29.8%
		Holubova178	Czech Republic	103	2.20		0.47	-	-
		Szczesniak171	Poland	51	2.19		0.5	-	-
		Uhlmann179	Germany	23	2.0		0.5	-	-
		Vrbova176	Czech Republic	197	2.18		0.47	-	-
		Switaj83	Poland	110	2.39		0.53	-	-
	Weighted total			4724	2.21		0.27	1744	36.9%
	Bipolar disorders/MDD	Brohan10	Multi-site	577	1.94 BD	1.97	0.87	1182	21.7%
				603	2.11 MDD
	Bipolar disorders	Post180	Austria	60	1.9 BD		0.57	-	-
		Dubreucq93	France	117	2.2		0.53	117	29.9%
		Quenneville181	Switzerland	69	2.27		0.5	-	-
		Szczesniak171	Poland	19	2.39		0.5	-	-
	MDD	Lanfredi182	Multi-site	516	2.2		0.5	-	-
		Dubreucq93	France	27	2.29		0.56	27	40.7%
		Holubova178	Czech Republic	80	1.96		0.42	-	-
		Prasko95	Czech Republic	72	2.36		0.49	-	-
		Szczesniak171	Poland	42	2.46		0.6	-	-
	Weighted total	BD MDD		842	2.01		0.60	1326	22.7%
				1340	2.16		0.47
	Anxiety disorders	Ociskova183	Czech Republic	109	2.24		0.49	-	-
		Grambal169	Czech Republic	37	1.98		0.54
		Dubreucq93	France	19	2.35		0.56	19	42.1%
	OCD	Moritz184	Germany	50	1.99		0.48	-	-
	Weighted total	-		215	2.14		-	19	42.1%
	BPD	Grambal169	Czech Republic	35	2.45		0.50	-	-
		Dubreucq93	France	64	2.36		0.47	64	43.8%
		Quenneville181	Switzerland	39	2.56		0.66	-	-
	Personality disorder	Kamaradova170	Czech Republic	17	2.37		0.58	-	-
	Weighted total	-		155	2.43		-	64	43.8%
North America	SMI	Livingston36	Canada	91	2.13		0.38	91	10%
		Livingston37	Canada	71	2.10		0.35	71	9%
		Livingston43	Canada	94	2.22		0.49	94	23.4%
		West13	USA	144	1.72		0.57	144	36.1% (31%–41%)
		Ritscher-Boyd112	USA	-	-		-	82	28%
		Ritscher-Boyd185	USA	149	2.19		0.52	149	24.8%
		Drapalski186	USA	100	2.3		0.4	100	35%
		Harris38	USA	235	2.31		0.47	235	37%
		Chronister187	USA	101	2.2		0.45	101	25%
		Kira188	USA	330	2.16		0.72	330	40.1%
		Tomar189	USA	108	2.31		0.5	108	40.7%
		Pearl105	USA	319	2.11		0.53	-	-
		Jahn190	USA	516	2.25		0.49	-	-
		Villotti191	USA	170	1.95		0.47	-	-
	Weighted total	-		2428	2.16			1505	31.4%
	Schizophrenia	Firmin192	USA	-	-		-	111	36.5% (25%–45%)
		O’Connor133	USA	-	-		-	353	50%
		Link193	USA	-	-		-	65	26.2%
		Lysaker101	USA	70	2.36		0.54
	Weighted total	-		70	2.36		0.29	529	44.2%
	Bipolar disorders	Howland42	USA	115	2.22		0.48	115	26%
		Bassirnia194	USA	112	1.94		0.47	-	-
	Weighted total			227	2.08		0.22
South America	Schizophrenia	Caqueo-Uritzar34	Multi-site	253	2.3		0.54	253	38.6% (28.6%–48.7%)
	Weighted total			253	2.3		0.29
Australia	Schizophrenia	Hill195	Australia	60	2.56		0.49	-	-
	Weighted total			60	2.56		0.24
South Asia	SMI	Grover196	India	1,403	2.29		0.47	1403	29.4%
		Maharjan197	Nepal	-	-		-	180	52%
	Weighted total	-		1403	2.29		0.22	1583	28.6%
	Schizophrenia	Grover196	India	707	2.37		0.51	707	37.9%
		Singh198	India	100	2.3		0.40	100	29%
		Pal199	India	32	2.74		0.40	-	-
	Weighted total	-		839	2.37		0.24	807	36.8%
	Bipolar disorders	Grover196	India	344	2.23		0.38	344	20.6%
		Pal199	India	59	2.25		0.38	-	-
		Grover200	India	185	2.33		0.43	185	28%
	Weighted total	-		588	2.26		0.15	529	23.2%
	MDD	Grover196	India	352	2.19		0.45	352	21%
		Sahoo201	India	107	2.39		0.57	107	40.1%
	Weighted total	-		459	2.23		0.23	459	25.5%
	Anxiety disorders	Pal199	India	30	1.97		0.37	-	-
South-East Asia	SMI	Picco202	Singapore	280	2.37		0.54	280	43.6%
		Ho97	China	136	2.3		0.6	136	36.8%
		Young35	China	474	2.42 (2.34–2.50)		0.52	474	43.5% (38.3%–49.5%)
		Kim39	South Korea	160	2.06		0.36	160	8.1%
		Kao40	Taiwan	251	2.26		0.51	251	50%
	Weighted total	-		1301	2.32		0.26	1301	39.7%
	Schizophrenia	Lv138	China	95	2.17		0.38	95	20%
		Hsiao61	Taiwan	111	2.33		0.53	111	27%
		Lien76	Taiwan	170	2.36		0.52	170	39.4%
		Kao40	Taiwan	151	2.42		0.44	151	51%
		Li139	China	384	2.30		0.39	-	-
		Hofer30	Japan	60	2.16		0.42	-	-
		Ran127	China	232	2.46		0.28	-	-
		Mak67	China	162	2.34		0.61	-	-
		Shin88	South Korea	70	2.11		0.53	-	-
		Pribadi203	Indonesia	300	2.5		0.94	300	34%
		Picco202	Singapore	74	2.41		0.52
		Kim204	South Korea	123	2.15		0.50	-	-
		Lu205	China	92	2.11		0.45	-	-
		Park206	South Korea	321	2.51		0.65	-	-
	Weighted total	-		2345	2.36		0.32	827	32.4%
	Bipolar disorder	Ran127	China	39	2.36		0.30	-	-
	Weighted total				2.36		0.09
	MDD	Picco202	Singapore	74	2.44		0.55	-	-
		Woon207	Malaysia	99	2.0		0.6	99	23.2%
		Ran127	China	182	2.43		0.28	-	-
	Weighed total			355	2.31		0.20	99	23.2%
	Anxiety Disorders OCD	Picco202	Singapore	71	2.23		0.56	-	-
		Picco202	Singapore	61	2.41		0.49	-	-
	Weighted total	-		132	2.31
Middle East	SMI	Ghanean208	Iran	-	-		-	138	39%
		Tanriverdi209	Turkey	217	2.59		0.33	-	-
		Hasson-Ohayon210	Israel	107	2.28		0.49	-	-
		Korkmaz211	Turkey	224	2.53		0.48	-	-
	Weighted total			548	2.50		0.18
	Schizophrenia	Sarisoy41	Turkey	109	2.17		0.51	109	29.4%
		Çapar212	Turkey	250	2.76		0.37	-	-
		Olçun213	Turkey	76	2.51		0.55	-	-
		Yilmaz214	Turkey	63	2.63		0.49	-	-
		Yildirim94	Turkey	200	2.74		0.46	-	-
		Tanriverdi209	Turkey	46	2.54		0.25
	Weighted total	-		744	2.61		0.32	109	29.4%
	Bipolar disorders	Sarisoy41	Turkey	118	2.10		0.46	118	18.5%
		Cerit215	Turkey	80	2.12		0.39	-	-
		Sadighi216	Iran	126	1.90		0.87	126	27.6%
		Tanriverdi209	Turkey	63	2.52		0.32	-	-
	Weighted total	-		387	2.10		0.35	244	19.2%
Africa	SMI	Adewuya217	Nigeria	-	-		-	340	21.6%
		Girma53	Ethiopia	422	2.32		0.30	422	25.1%
		Ibrahim55	Nigeria	-	-		-	370	22.5%
		Asrat218	Ethiopia	-	-		-	317	32.1%
	Weighted total			422	2.32		0.09	1449	25.1%
	Schizophrenia	Mosanya125	Nigeria	256	1.94		0.68	256	18.8%
		Fadipe219	Nigeria	370	2.09		0.43	370	16.5%
		Assefa220	Ethiopia	-	-		-	212	46.7%
		Bifftu124	Ethiopia	-	-		-	411	48.6%
	Weighted total	-		626	2.03		0.29	1249	32.6%

Note: MDD, Major Depressive Disorder; OCD, Obsessive Compulsive Disorder; BPD, Borderline Personality Disorder; SMI, Severe Mental Illness. Mean Internalized Stigma refers to ISMI mean total score. High Internalized Stigma refers to the proportion of patients with ISMI > 2.5 or above the midpoint on other scales. Bold faces represents the total sample and the weighted means and proportions.

About one-third of people with SMI (31.3%) reported elevated self-stigma. The highest frequency was found in schizophrenia (35.8%). Higher frequency was found in South-East Asia and the Middle East for SMI, and in North America, Europe, and Africa for schizophrenia. Significant between-group differences in mean self-stigma scores were found between Europe and South Asia or South-East Asia for participants with SMI (P < 0.001; weighted mean difference = 0.150 and 0.140), schizophrenia (P < 0.001; 0.159 and 0.143), bipolar disorder (P < 0.001; 0.253 and 0.35), and MDD (P < 0.001; 0.148 and 0.07). Compared with Europe, self-stigma was higher in the Middle East and Africa for SMI (P < 0.001; 0.322 and 0.140) and in the Middle East and South America for schizophrenia (P < 0.001; 0.401 and 0.08). Box plots on the differences by geographical area are provided on table 3. Self-stigma did not differ between South-East Asia and Africa for SMI, South Asia and South-East Asia for schizophrenia, and North America and the Middle East for bipolar disorder. The results are shown in supplementary table S4. There were significant country-related differences in the proportion of people with SMI who reported high self-stigma in Europe (from 15.2% in Sweden⁹ to 57% in Croatia³³), North America (from 9% in Canada³⁷ to 37% in USA³⁸) and in South-East Asia (from 8.1% in South Korea³⁹ to 50% in Taiwan⁴⁰). Setting-related differences were also found.^13,35 Similar variations were found for schizophrenia, BD (18.5% Turkey,⁴¹ in remitted BD; 26% USA,⁴² in non-adherent BD), MDD, obsessive-compulsive disorder and anxiety disorders.

Table 3.

Between-groups Differences by Geographical Area

Means and percentages were weighted for the number of cases per study to obtain prevalence data. Derived weighted means by geographical area and pooled standard deviations were calculated. One-way ANOVA was conducted from these summary data and Post hoc pairwise test comparisons were computed with Tukey–Kramer method. Plots shows the ISMI mean score for each studies and each area with 95% confidence interval.

Cross-sectional Correlates of Self-stigma

Two hundred and forty-one studies (88.6%) reported on cross-sectional internalized stigma correlates. The most common diagnosis was schizophrenia (n = 106; 44%), followed by BD (n = 11; 4.6%) and at-risk states or early psychosis (n = 9; 3.7%). Ninety studies (37.3%) concerned SMI. The results are shown in table 4. See supplementary table 3 for the detailed list of correlates/consequences.

Table 4.

Cross-sectional Correlates and Consequences of Self-stigma

	A		B		C		D		E
	Number of studies (n = 272)		Non-significant relationship (P > 0.05)		Significant relationship (P < 0.05)		Positive relationship (P < 0.05)		Negative relationship (P < 0.05)
Variables	n	%	n	%	n	%	n	%	n	%
Sociodemographic
Gender (1)	62	22.8	50	80.7	12	19.3	3	25.0	9	75.0
Age	68	25.0	49	72.1	19	27.9	12	63.2	7	36.8
Education	60	22.1	37	61.7	23	38.3	10	43	13	57
Employment(2)	37	13.6	16	43.2	21	56.8	2	9.5	19	90.5
Marital status(3)	40	14.7	28	70	12	30	3	25.0	9	75.0
Income	16	5.9	5	31.2	11	68.8	0	0	11	100
Source of income (4)	11	4.0	6	54.5	5	45.5	3	60	2	40
Immigrant status	1	0.4	0	0	1	100	1	100	0	0
History of incarceration/homelessness	2	0.7	0	0	2	100	2	100	0	0
Experience of victimization	2	0.7	0	0	2	100	2	100	0	0
Parenting status (5)	2	0.7	1	50	1	50	1	100	0	0
Internalizing personality traits	6	2.2	0	0	6	100	6	100	0	0
Self-compassion/mindfulness	2	0.7	1	50	1	50	0	0	1	100
Locality (urban/rural) (6)	7	2.6	4	57.1	3	42.9	1	33.3	2	66.7
Illness-related
Severity of psychiatric symptoms	65	23	10	15.4	55	84.6	55	100	0	0
Higher distress from sub-threshold psychotic symptoms/psychiatric symptoms	2	0.7	0	0	2	100	2	100	0	0
Age of onset	26	9.6	19	73.0	7	27.0	0	0	7	100
Illness duration	32	11.8	22	68.7	10	31.3	7	70.0	3	30.0
Stage of illness	1	0.4	1	100	0	0	-	-	-	-
Diagnosis(7)	38	14.0	20	52.6	18	47.4	NA	-	NA	-
Insight	32	11.8	0	0	32	100	32	100	0	0
Parental insight	1	0.4	0	0	1	100	1	100	0	0
Subjective clinical severity	3	1.1	0	0	3	100	3	100	0	0
Perceived cognitive dysfunction	4	1.5	0	0	4	100	4	100	0	0
Hospitalizations	38	14.0	23	60.5	15	39.5	14	93.3	1	6.7
Treatment setting (8)	6	2.2	3	50	3	50	1	33.3	2	66.7
Involuntary hospitalizations (IH)	3	1.1	1	33.3	2	66.7	2	100	0	0
Negative emotional reactions to IH	2	0.7	0	0	2	100	2	100	0	0
Forensic patient status	3	1.1	2	66.7	1	33.3	1	100	0	0
Compulsory community treatment	1	0.4	1	100	0	0	-	-	-	-
History of suicide attempt	6	2.2	1	16.7	5	83.3	5	100	0	100
Social anxiety	3	1.1	0	0	3	100	3	100	0	0
Comorbid post-traumatic stress disorder	2	0.7	0	0	2	100	2	100	0	0
Comorbid personality disorder	2	0.7	0	0	2	100	2	100	0	0
Cormorbid substance use disorder	3	1.1	1	33.3	2	66.7	2	100	0	0
Drug extra-pyramidal side-effects	4	1.5	0	0	4	100	4	100	0	0
Cognitive functioning	6	2.2	1	16.7	5	83.3	0	0	5	100
Social cognition	4	1.5	1	25	3	75	0	0	3	100
Metacognitive abilities	7	2.6	0	0	7	100	0	0	7	100
Dysfunctional attitudes	1	0.4	0	0	1	100	1	100	0	0
Use of negative coping strategies	8	2.9	0	0	8	100	8	100	0	0
Environment-related
Country level of public stigma	6	2.2	0	0	6	100	6	100	0	0
Perceived stigma	18	6.6	0	0	18	100	18	100	0	0
Self-labelling	3	1.1	0	0	3	100	3	100	0	0
Stigma stress	4	1.5	0	0	4	100	4	100	0	0
Group value	3	1.1	0	0	3	100	0	0	3	100
Perceived legitimacy	2	0.7	0	0	2	100	2	100	0	0
Social network	11	4.0	0	0	11	100	0	0	11	100
Sense of belonging	2	0.7	0	0	2	100	0	0	2	100
Perceived social support	21	8.4	0	0	21	100	0	0	21	100
Loneliness	2	0.7	0	0	2	100	2	100	0	0
Family sense of coherence	1	0.4	0	0	1	100	0	0	1	100
Experienced and anticipated stigma	13	4.8	0	0	13	100	13	100	0	0
Expressed emotion in families	1	0.4	0	0	1	100	1	100	0	0
Associative stigma in MHP	1	0.4	0	0	1	100	1	100	0	0
Cultural factors
Attributing mental illness to supernatural causes	3	1.1	0	0	3	100	3	100	0	0
History of traditional treatment	2	0.7	0	0	2	100	2	100	0	0
Loss of face in Eastern countries	2	0.7	0	0	2	100	2	100	0	0
Psychosocial
Self-efficacy	13	4.8	0	0	13	100	0	0	13	100
Empowerment	11	4.0	0	0	11	100	0	0	11	100
Self-esteem	44	16.2	0	0	44	100	0	0	44	100
Hope	19	7.0	0	0	19	100	0	0	19	100
Depression	41	15.1	0	0	41	100	41	100	0	0
Suicide risk	15	5.5	1	6.7	14	93.3	14	100	0	0
Help-seeking/Therapeutic alliance	5	1.8	0	0	5	100	0	0	5	100
Treatment adherence	15	5.5	0	0	15	100	0	0	15	100
Subjective social status	1	0.4	0	0	1	100	0	0	1	100
Psychosocial function	27	9.9	0	0	27	100	0	0	27	100
Activity	1	0.4	0	0	1	100	0	0	1	100
Self-reported physical health	2	0.7	0	0	2	100	0	0	2	100
Capacity for intimacy/satisfaction in intimate relationships	4	1.5	0	0	4	100	0	0	4	100
Self-reported parenting experiences	1	0.4	0	0	1	0	0	0	1	100
Quality of life	41	15.1	0	0	41	100	0	0	41	100
Wellbeing	9	3.3	0	0	9	100	0	0	9	100
Satisfaction with life	8	2.9	0	0	8	100	0	0	8	100
“why try effect”	2	0.7	0	0	2	100	2	100	0	0
Coming out (CO)/CO assertiveness	1	0.4	1	100	0	0	-	-	-	-
Benefits from being out	1	0.4	0	0	1	100	0	0	1	100
Stigma resistance	11	4.0	0	0	11	100	0	0	11	100
Resilience	7	2.6	0	0	7	100	0	0	7	100
Personal recovery	22	8.1	0	0	22	100	0	0	22	100

1)Columns D + E; 1 = Men.

2) Columns D + E; 1 = Employed.

3) Columns D + E; 1 = Married.

4) Columns D + E; 1 = Income earner.

5) Columns D + E; 1 = Mothers.

6) Columns D + E; 1 = Urban.

7) Columns D+ E; not applicable (heterogeneity of the samples).

8) Columns D + E; 1= Inpatients.

Table 4 presents the relationships between sociodemographic, illness-related, environment-related, cultural and psychosocial variables with self-stigma (includes 272 studies).

Few sociodemographic characteristics correlated significantly with self-stigma. Immigrant status, history of incarceration or homelessness,^36,43 parenting status (mothers > fathers⁴⁴), shame proneness, and avoidant or self-defeating personality traits (n = 6 studies) were associated with higher self-stigma. The results were contrasted for all other sociodemographic variables (age, gender, education level, employment, marital status, income, and source of income). Other personal characteristics (attachment style, self-compassion) were not associated with self-stigma.^45,46 Residual psychiatric symptomatology, positive and negative symptoms for schizophrenia, and depressive symptoms for BD were associated with higher self-stigma in most studies (84.6% significance; n = 65). Social anxiety (n = 3) and distress from sub-threshold psychotic symptoms⁴⁷ were positively correlated with self-stigma. Self-stigma was equally severe in participants with ultra-high risk and established psychosis.⁴⁸ Internalized shame about mental illness and fear accuracy in an emotion recognition task were negatively associated in people at risk of psychosis.⁴⁹ Stigma stress, identified as a predictor of self-stigma in several studies (n = 4), was positively associated with transition to psychosis.⁵⁰ Comorbid post-traumatic stress disorder (n = 2) and an increased number of drug side effects (n = 4) were positively associated with self-stigma.

Public stigma⁵¹ and other dimensions of stigma were associated with self-stigma (perceived stigma, n = 18 studies; perceived stigma from mental health providers,⁵²n = 1; experienced stigma, n = 9; anticipated stigma, n = 3). Cultural factors such as attributing mental illness to supernatural causes,^53–55 a history of traditional treatment,^53,55 and loss of face in Eastern countries^35,56 were associated with self-stigma. Concerns about losing face (or the fear of losing face because of being diagnosed with SMI) mediated the relationship between perceived stigma and self-stigma.⁵⁶ Stigma stress (n = 4) and negative emotional reactions to involuntary psychiatric admission^57,58 were significant correlates of self-stigma, in contrast with compulsory community treatment,³⁷ forensic status,³⁶ and the number of involuntary admissions.^57–59 In-group value (ie, how people with SMI see their own group,⁶⁰ social networks and support (n = 11; n = 21), membership of an advocacy group, and family support protected against self-stigma.^61,62 Family expressed emotion and associative stigma in mental health professionals were associated with higher self-stigma.^63,64 Self-stigma mediated the effects of family expressed emotion and in-group value on psychosocial function and personal recovery.^64,65

Insight into illness (n = 32 studies), parental insight,⁶⁶ self-perception of clinical severity (n = 3), perceived cognitive dysfunction (n = 4), and attributions of personal responsibility⁶⁷ were associated with higher self-stigma. Impairments in cognitive and metacognitive function (n = 5; n = 7), dysfunctional attitudes⁶⁸ and avoidant coping strategies (n = 8) were associated with higher self-stigma. Conversely preserved cognitive abilities,^69–71 empowerment, self-efficacy, self-agency, and stigma resistance protected against self-stigma.^9,10,62,72 Mixed results were found for other illness-related correlates (age of onset, psychiatric diagnosis, illness duration, history of suicide attempts, inpatient status, past psychiatric admission, and number of hospitalizations).

In general, self-stigma was positively associated with depressive symptoms (n = 41) and suicidal ideation (n = 14) and negatively correlated with hope (n = 19), help seeking (n = 5), and treatment adherence (n = 15). Single studies found negative associations with therapeutic alliance⁷³ and shared decision making.⁷⁴ Insight into illness,^70,75–80 avoidant personality traits,⁸¹ and coping strategies,⁸² loneliness,⁸³ and resilience⁸⁴ moderated the relationship with depression and self-esteem mediated the effects of self-stigma on hope.⁸⁵ Self-stigma mediated the effects of perceived cognitive dysfunction and experienced stigma on suicidality^86,87 and QoL.⁸⁸ Self-stigma was negatively associated with QoL (n = 41), self-esteem (n = 44), self-efficacy (n = 13), well-being (n = 9), life satisfaction (n = 8), empowerment (n = 11), resilience (n = 7), stigma resistance (n = 11), and personal recovery (n = 22). Self-stigma positively correlated with the “why try effect” ⁸⁹ and later stages of self-stigma with a higher impact on hope, self-esteem, psychosocial function, and personal recovery.^89,90 Participants in the late stages of self-stigma reported more reasons for not disclosing their psychiatric diagnosis, in contrast with those in the early stages who reported greater benefits from being “out”.⁹¹

Self-stigma was negatively associated with global functioning (n = 27). Higher demoralization and decreased resilience mediated the relationship between self-stigma and psychosocial function.^75,92 Self-stigma was negatively associated with relational satisfaction.^41,93 Self-stigma positively correlated with sense of loneliness,^83,94 fear of intimate relationships,⁴¹ and self-stigma on parenting abilities for mothers living without their children.⁴⁴

Self-stigma and Longitudinal Outcomes

Forty-one studies reported longitudinal outcomes associated with internalized stigma. These studies mainly included participants with SMI (n = 15; 36.6%), schizophrenia (n = 16; 39%) or at-risk stages/early psychosis (n = 5; 12.2%). Twelve studies included young participants (<18, n = 1; 18–35, n = 11) and three individuals over 50 years old (10.7%). Fifteen were conducted in psychiatric rehabilitation settings (36.5%). The duration of follow-up ranged from 6 weeks⁹⁵ to 2 years.⁹⁶ The results are shown in table 5.

Table 5.

Longitudinal Correlates and Consequences of Self-stigma

	A.		B.		C.		D.		E.
	Number of studies (n = 41)		Non-significant relationship (P > 0.05)		Significant relationship (P < 0.05)		Positive relationship (P < 0.05)		Negative relationship (P < 0.05)
Variables	n	%	n	%	n	%	n	%	n	%
Correlates of increased self-stigma at follow-up
Female sex	5	12.2	4	80	1	20	1	100	0	0
Hospitalizations	1	2.4	0	0	1	100	1	100	0	0
Duration of untreated psychosis	1	2.4	0	0	1	100	1	100	0	0
Baseline psychiatric symptoms	5	12.2	0	0	5	100	5	100	0	0
Shame, self-contempt about IH and stigma stress	2	4.9	0	0	2	100	2	100	0	0
Negative coping strategies	1	2.4	0	0	1	100	1	100	0	0
Correlates of reduced self-stigma at follow-up
Attending to recovery-oriented daycare/vocational rehabilitation/COSP	5	12.2	0	0	5	100	5	100	0	0
Work without experienced discrimination	1	2.4	0	0	1	100	1	100	0	0
Receiving no disability benefits during PR	1	2.4	0	0	1	100	1	100	0	0
Longitudinal consequences of self-stigma
Psychiatric symptoms	5	12.2	0	0	5	100	5	100	0	0
Social anxiety	1	2.4	0	0	1	100	1	100	0	0
Depression	6	14.6	0	0	6	100	6	100	0	0
Suicide risk	4	9.8	0	0	4	100	4	100	0	0
Risk of hospitalizations	1	2.4	0	0	1	100	1	100	0	0
Self-esteem	3	7.3	0	0	3	100	0	0	3	100
Treatment adherence	3	7.3	0	0	3	100	0	0	3	100
Psychosocial function	9	21.9	0	0	9	100	0	0	9	100
Quality of Life	3	7.3	0	0	3	100	0	0	3	100
Life satisfaction	1	2.4	0	0	1	100	0	0	1	100
Personal recovery	2	4.9	0	0	2	100	0	0	2	100

Note: COSP, consumer-operated service program; IH, involuntary hospitalization; PR, Psychiatric rehabilitation.

Table 5 presents the longitudinal relationships between sociodemographic, illness-related, environment-related, and psychosocial variables with self-stigma (includes 41 studies).

Fifteen studies reported on the baseline factors influencing the level of self-stigma at follow-up. Residual psychiatric symptoms,^97–100 negative emotional reactions to involuntary hospitalization,^57,99 and emotional distress^98,101 were the most significant baseline factors associated with self-stigma at follow-up. Duration of untreated psychosis⁹⁷ and baseline coping strategies¹⁰² were associated with higher self-stigma in single studies. Mixed results were found for self-stigma stability over time with no specific intervention.^86,102–104 Attending psychiatric rehabilitation was associated with significant reductions in self-stigma extent (from a mean total ISMI score of 2.31 on admission to 1.96 at discharge; ³⁸ 2.36–2.20; ¹⁰¹ 2.11–1.96; ¹⁰⁵ 2.1–2.04; ¹⁰⁶ 1.80–1.50 in participants having worked in the past year without being discriminated against¹⁰⁷) and in the proportion of participants with high self-stigma (ISMI > 2.5 from 37% to 13.7%; ³⁸ reduction in ISMI levels at follow up > 25% in 38% of the participants with mean self-stigma > 2¹⁰³). Not receiving disability benefits during psychiatric rehabilitation was associated with a greater reduction in self-stigma.³⁸ Attending consumer-operated service programs was associated with self-stigma reduction.¹⁰⁸

Twenty-eight studies reported on the longitudinal consequences of self-stigma. Self-stigma at baseline was associated with increased positive symptoms,¹⁰⁹ emotional discomfort,¹⁰⁰ social anxiety,¹¹⁰ depression,^111,112 suicidal ideation,^96,113 and an increased risk of psychiatric hospitalization¹¹⁴ at follow-up. Participants with high self-stigma reported reduced self-esteem,¹¹² decreased life satisfaction,¹¹⁵ lower personal recovery,¹¹⁶ less use of adaptive coping strategies,¹⁰² and lower treatment adherence¹¹⁷ at follow-up. Baseline self-stigma was associated with poorer social and vocational functioning^109,118 at follow-up, and less benefits from vocational rehabilitation.¹¹⁹ A change in self-stigma during follow-up predicted depression. Increases in self-stigma were associated with more depressive symptoms¹⁰⁹ and higher suicidality.^99,120 Decreases in self-stigma were associated with less depression.^104,105 Increased self-stigma was associated with more negative attitudes towards psychiatric medication,^120,121 poorer social function,¹²² reduced self-esteem,⁹⁹ and lower personal recovery.¹¹⁶ Reduced self-stigma was associated with decreased subjective clinical severity,¹⁰⁵ higher self-esteem,¹⁰¹ and improved global functioning.¹⁰⁵ Baseline self-stigma was not associated with QoL at follow-up.³⁷ Decreased self-stigma during follow-up was, however, associated with improvements in QoL¹⁰⁵ during psychiatric rehabilitation.

Self-stigma and Stigma Resistance

Thirty-one studies (11.9%) reported data on stigma resistance frequency in SMI. The results are shown in table 6. Stigma resistance was higher in mood disorders (59.7%¹⁰) than in schizophrenia (53.1%; n = 5). Stigma resistance in schizophrenia varied within Europe (from 49.2%⁹ to 63%¹²³ and Africa (from 49.4% in Ethiopia¹²⁴ to 72.7% in Nigeria¹²⁵). Stigma resistance was negatively correlated with self-stigma in Austria,^30,123 Croatia,¹²⁶ Nigeria,¹²⁵ and South Africa.⁶² In some countries, self-stigma and stigma resistance were both high (USA,⁴² Turkey, Ethiopia,¹²⁴ India, Bolivia,³⁴ Peru,³⁴ Chile,³⁴ and China¹²⁷) and in Canada these were both low.^36,37 Metacognitive abilities,^128–130 social network,^123,131 social power,¹³² self-compassion,¹³¹ psychological flexibility,¹³¹ fear of negative evaluation,¹²⁸ perceived stigma,^60,114 negative symptoms,¹³⁰ and coping strategies^129,133 were associated with high stigma resistance.

Table 6.

Prevalence of Stigma Resistance

Area		Study	Country	n	Mean SR	SD	n	High SR (%)
Europe	SMI	Grambal169	Czech Republic	184	2.52	0.57	-	-
		Dubreucq93	France	738	2.54	0.51	693	54.1
	Weighted total			922	2.53
	Schizophrenia	Brohan9	Multi-site	1129	2.47 (2.29–2.7)	0.51	1129	49.2%
		Vidovic126	Croatia	149	2.37	0.99	149	54%
		Sibitz123	Austria	157	2.73	0.76	157	63.3%
		Hofer30	Austria	52	2.94	0.53	-	-
		Bouvet175	France	62	2.46	0.55	-	-
		Vrbova176	Czech Republic	197	2.56	0.48	-	-
		Surmann177	Germany	80	2.35	0.58	-	-
		Dubreucq93	France	466	2.56	0.51	466	55.4%
	Weighted total	-		2292	2.51		1435	51.2%
	Bipolar Disorders/MDD	Brohan10	Multi-site	1182	2.81	0.98	1182	59.7%
		Dubreucq93	France	117 BD	2.52	0.53	117	52.1%
				27 MDD	2.58	0.53	27	59.3%
	Weighted total			1326	2.77		1326	59.02%
	Anxiety disorders	Grambal169	Czech Republic	37	2.30	0.56	-	-
		Dubreucq93	France	19	2.44	0.68	19	36.8%
	Borderline Personality Disorder	Grambal169	Czech Republic	35	2.27	0.57	-	-
		Dubreucq93	France	64	2.52	0.48	64	51.6%
North America	SMI	Livingston37	Canada	71	2.07	0.38	-	-
		Livingston43	Canada	94	2.10	0.50	94	15.9%
		Ritsher-Boyd112	USA	-	-	-	82	29%
	Weighted total	-		165	2.08		176	22%
	Bipolar Disorders	Howland42	USA	115	3.06	0.46	-	-
South America	Schizophrenia	Caqueo-Uritzar34	Multi-site	253	2.4	0.61	-	–
South Asia	Schizophrenia	Singh198	India	100	2.3	0.70	100	45%
		Pal199	India	32	2.84	0.50	-
	Weighted total	-		132	2.43		100	45%
	Bipolar Disorders	Pal199	India	59	2.43	0.42	-	-
		Grover200	India	185	2.21	0.51	-	-
	Weighted total	-		244	2.26		-	-
South-East Asia	SMI	Ran127	China	453	2.49	0.42	-	-
		Lau221	Singapore	280	2.87	0.47	280	82.9%
	Weighted total			733	2.63
	Schizophrenia	Li139	China	384	2.28	0.45	-	-
		Hofer30	Japan	60	2.35	0.46	-	-
		Ran127	China	232	2.50	0.36	-	-
	Weighted total	-		676	2.36		-	-
Middle East	SMI	Ghanean208	Iran	138	2.46	0.39	-	-
	Schizophrenia	Sarisoy41	Turkey	109	2.56	0.62	-	-
		Çapar212	Turkey	250	2.41	0.35	-	-
		Olçun213	Turkey	76	2.60	0.60	-	-
		Karakas222	Turkey	60	2.62	0.63	-	-
		Yilmaz214	Turkey	63	2.66	0.43	-	-
	Weighted total score	-		558	2.51		-	-
	Bipolar Disorders	Sarisoy41	Turkey	118	2.38	0.54	-	-
		Sadighi216	Iran	126	2.76	0.98	126	57.49%
	Weighted total score			246	2.55
Africa	SMI	Girma53	Ethiopia	422	2.41	0.40	-	-
		Sorsdahl62	South Africa	142	2.90	-	-	-
	Weighted total	-		564	2.53		-	-
	Schizophrenia	Mosanya125	Nigeria	256	2.79	0.53	256	72.7%
		Bifftu124	Ethiopia	411	2.52	-	411	49.4%
	Weighted total			667	2.62		667	58.34%

Note: MDD, Major Depressive Disorder; SMI, Severe Mental Illness. Mean stigma resistance refers to ISMI stigma resistance subscale mean score. High stigma resistance refers to the proportion of patients with ISMI stigma resistance subscale > 2.5. Bold faces represents the total sample and the weighted means and proportions.

Discussion

Research on self-stigma in SMI has considerably progressed over the past decade (244 articles since 2010 and 28 studies published prior to this).⁷ Self-stigma was present in all geographical areas with higher frequency in South-East Asia and the Middle East for SMI (>39%). Self-stigma in SMI was higher in the Middle East, South Asia, South-East Asia, and Africa than in Europe or North America. Schizophrenia was associated with high self-stigma in all geographical areas (32.6%–44.2%). Self-stigma in schizophrenia was higher in the Middle East, South-East Asia, South Asia, and South America than in Europe. Self-stigma differed according to the geographical area in mood disorders (higher self-stigma in South Asia and South-East Asia in comparison with Europe, North America, or the Middle East). Variations in patterns of stigma according to the cultural area might explain these differences. The greater public stigma relating to SMI in Eastern countries compared with Western countries could explain the higher levels of self-stigma in South Asia and South-East Asia.¹³⁴ Other potential factors contributing to higher self-stigma include feelings of shame about not meeting the social expectations of “what matters most” in one’s social group,^14,16 moral or supernatural causal attributions of mental illness, and social concerns about the spillover effects disclosing mental illness might have on the social and economic status of family members.^53,55,134 The frequency of self-stigma is consistently high across the world. This concurs with several cross-national studies on perceived, experienced, or anticipated stigma.^3,135–137 Lower rates of experienced stigma, but similar levels of self-stigma for people with schizophrenia were reported in China and India compared with Western countries.^16,138,139 Cultural factors (eg, concerns about disclosure spillover on family members) leading to higher self-stigma and social withdrawal might explain these variations.^{14,16,138,139} Most of the instruments used for assessing self-stigma in non-Western countries were adapted from scales developed in Europe or North America and did not include culture-specific items. These scales might not reflect all the culture-specific forms of stigma.^55,140 Anti-stigma campaigns and self-stigma reduction interventions should take into consideration cultural factors (eg, cost/benefits of strategic disclosure in a given cultural context^35,54,141).

Cultural and socio-ideological factors might account for the large country-related variations that were found within geographical areas. Higher public stigma was found in Eastern/Southern Europe compared with Western Europe.^142–144 Compared with Western Europe or Canada, higher levels of self-reported sociopolitical conservatism were found in Eastern Europe and the USA.^145,146 Gonzales¹⁴⁷ and De Luca¹⁴⁸ found that self-reported political conservatism and right-wing authoritarianism were associated with increased public stigma. Cultural factors (eg, the endorsement of traditional cultural values, higher in China than in Taiwan or South Korea¹⁴⁹) might contribute to the differences observed between Guangzhou and Hong Kong³⁵ or between Taiwan and South Korea.^39,40 These variations could be related to sample characteristics (eg, higher reliance on family support in Guangzhou compared with Hong Kong; ³⁵ higher proportion of patients with BD in Korea than in Taiwan^39,40).

Sociodemographic and illness-related correlates yielded mixed results in line with previous reviews.^7,8 Self-stigma was high in BPD but this is based on a small number of studies with small sample sizes. Self-stigma was equally severe in the at-risk stages as in psychosis.¹¹¹ Further research is needed to confirm this result. Self-stigma was closely associated with perceived and experienced stigma. These concepts are distinct and should be better differentiated between, as stereotype awareness and self-labeling do not necessarily imply stereotype agreement, self-application, and increased self-stigma.^91,150 Self-support groups and recovery-oriented services promoting positive group identification^60,106 should be further developed to prevent or reduce self-stigma. Reducing self-stigma implies targeting the explicitly negative views about the self that relate to being diagnosed with SMI. Making an empowered decision about disclosing an SMI diagnosis might be effective for adolescents or people in the early stages of self-stigma.^91,151 People in the late stages of self-stigma may need to take part in group interventions combining psychoeducation and cognitive restructuring.^21,57 Interventions should be proposed to each individual according to his/her personal needs and level of self-stigma.

The association between self-stigma with treatment setting varies (50% significance). Two studies reported higher self-stigma in outpatients compared with inpatients^152,153 and one the opposite.⁸³ Loneliness, low social support, perceived stigma, experienced stigma, and anticipated stigma might contribute to higher self-stigma in outpatients.^152,153 Participating in community activities, good social support, and attending psychiatric rehabilitation services or consumer-operated service programs protect against self-stigma.^9,62,105,108 Stigma stress, negative emotional reactions to involuntary hospitalization, and the use of avoidant coping strategies after discharge contribute to higher self-stigma.^57,58,102 Improved inpatient care (ie, the implementation of recovery-oriented practices and interventions targeting stigma stress, therapeutic alliance and coping strategies) might result in better patient outcomes after discharge, although this remains to be investigated.

The development of recovery-oriented practices in mental health facilities should be encouraged as it could reduce perceived stigma, stigma stress,^57–59 and negative emotional reactions to involuntary admissions.^57–59 Peer-supported self-management interventions, Joint Crisis Plans, “No Force First” policies, and selective disclosure programs could improve self-stigma through reduced stigma stress and perceived coercion.^{151,154–156} Recovery-oriented training programs for mental health professionals improve personal recovery in people with SMI.¹⁵⁷ They may also improve mental health professionals’ job satisfaction, burnout, and associative stigma of mental illness.^15,63 Their effectiveness in reducing self-stigma in patients should be investigated.

Given the potential relationships with stigma stress,⁵⁰ duration of untreated psychosis,⁹⁷ distress from sub-threshold psychotic symptoms,⁴⁷ and transition to psychosis,⁵⁰ the effects of recovery-oriented early interventions on self-stigma and its consequences should be further investigated. Strategic disclosure programs result in people making empowered decisions about whether to disclose a diagnosis of SMI or not. They result in improved stigma stress and self-stigma in adolescents with SMI¹⁵¹ and should be integrated into recovery-oriented early intervention services.

As expected,^7,8 self-stigma was negatively associated with recovery-related outcomes and positively associated with depression and suicidal ideation. Cognitive impairments, dysfunctional attitudes, and avoidant coping strategies were positively associated with self-stigma. Insight into illness was the most significant moderator of internalized stigma. Perceived cognitive dysfunction, perceived and experienced stigma all had indirect effects on clinical and functional outcomes via self-stigma. Baseline self-stigma was associated with poorer recovery-related outcomes and less benefit from vocational rehabilitation at follow-up.^8,119 Reduction in self-stigma was associated with improved depression, suicidality, attitudes towards medication, self-esteem, QoL, and social function at follow-up.

Improved treatment (ie, recovery-oriented practices and nonspecific interventions targeting therapeutic alliance, dysfunctional attitudes, self-esteem, or coping strategies) could indirectly reduce self-stigma.^158–160 Recovery-oriented psychoeducation improves treatment adherence and reduces the risk of hospitalization.¹⁶¹ Improved therapeutic alliance is associated with better recovery-related outcomes after attending to early interventions services.¹⁵⁸ Other interventions such as cognitive behavioral therapy, cognitive remediation, or social skills training might reduce self-stigma through improved symptoms, dysfunctional attitudes, and functioning.^{22,159,160,162} Given the potential relationship between expressed emotion in the families of people with SMI and self-stigma and recovery-related outcomes,^64,65 family psychoeducation could be effective for self-stigma.¹⁶³ Family psychoeducation should be recovery-oriented and address both public stigma and self-stigma.^164–167 The relationship between self-stigma in people with SMI and in their relatives is still unclear and should be further investigated.¹⁶⁸

Stigma resistance and self-stigma were negatively associated with each other⁶ but with different patterns. Self-stigma and stigma resistance are distinct constructs and should be measured using more specific scales.⁶

Limitations

There are some limitations to this review due to the heterogeneity in the definition of SMI and in the samples, settings, methods, scales, and reported outcomes. Few articles reported longitudinal outcomes with a limited number of studies conducted in psychiatric rehabilitation settings. This review excluded studies where self-stigma was not the main focus, which means that stigma in all its forms (ie, perceived, experienced or anticipated stigma, and self-stigma) could actually have more wide-ranging effects on people with SMI. However, by focusing on self-stigma, this review provides a more accurate understanding of its effects on people with SMI. The heterogeneity of the samples, methods, scales, and reported outcomes in the included articles limited the possibilities for extracting comparable data. The large number of studies included in this review and the range of countries represented is however a considerable strength. The under-reporting of negative or nonsignificant results due to publication bias and the exclusion of unpublished studies from this review might have limited the accuracy of the synthesis. The present systematic review does not include a meta-analysis. This decision was made due to the large number of studies and the heterogeneity of the samples, methods, scales, and reported outcomes. Statistical analyses were used to compare self-stigma frequency in different geographical areas (the second objective of the present study). Future meta-analyses with a more limited focus (eg, on the impact of self-stigma on recovery-related outcomes) could be conducted to explore the present findings in more detail.

Conclusions

In short, self-stigma is a severe problem in all SMI conditions (including the at-risk stages) and all geographical areas and is associated with poor clinical and functional outcomes. Levels of public, perceived, and experienced stigma (including from mental health providers) are significant predictors of self-stigma, pleading for the reinforcement of anti-stigma campaigns and the development of recovery-oriented practices in mental health settings. The respective associations between the duration of untreated psychosis, self-stigma, and transition to psychosis support the development of recovery-oriented early intervention programs. Psychiatric rehabilitation could be an effective means of reducing self-stigma and should therefore be further developed in public policies.

Funding

This research did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.