At this stage, it is pretty well known that negative attitudes held towards people with mental health problems (stigma) are responsible for harm against an already vulnerable group. However, such stigma is still very common, so like Winnie the Pooh wallpaper, it bears repeating. This is not about offence; although some views held can be pretty offensive, but actual harm.

Stigma can lead to problems accessing any available help, inability to gain employment, reduction in financial options and social isolation, as well as personal distress. In summary, stigma exists and stigma is a bad thing.

As with any bad thing, we should try to stop or reduce it. Previously we have seen how anti-stigma interventions are (broadly) successful. For example, interpersonal contact has been associated with more positive attitudes towards people with mental health problems. Such social contact interventions, where people affected by mental illness share their individual stories, have been developed and are common components of mental health anti-stigma interventions. However, it is not clear how these interventions might work. It’s not as if people don’t get social contact with people with mental health problems in their general lives. It’s very common. And stigma still clearly exists.

In addition, investigations into these interventions only tend to look at the short-term impact on stigma and discrimination and are predominantly conducted in the USA or Australia, i.e. high-income settings. They’re also settings with a reasonably high number of soap operas, but it’s probably unfair to blame that. Probably. Unfortunately, it isn’t the case that negative beliefs about people with mental health problems only have an impact in rich countries over the course of a week. We need to know if interventions work in the long-term and whether they can be applied in different country settings. As such, a review was conducted to bring together what is known globally about interventions to reduce mental illness-based stigma, in relation to their effectiveness in the medium and long term (minimum 4 weeks), and in low- and middle-income countries (LMICs).

Method

The usual electronic databases were searched for studies to include in the review; Medline, PsycINFO, the Cochrane Library and the Bumper Book of Lists of Psychological Studies. The search was restricted to results between 1980 and 2013 and studies on human beings. Apparently there weren’t many studies regarding stigma against people with mental illness from before 1980. There also aren’t many studies regarding stigma against non-humans with mental illness. Surprisingly.

The search was supposedly not limited by language, and where relevant, studies were translated by fluent speakers in French and Spanish. The search may have been limited by language.

Studies were included in the review if they described an intervention aiming to change mental health-related stigma or at least one of the following outcomes:

• Stigma (any)
• Prejudice (attitudes and related outcomes)
• Discrimination
• Internalised/self-stigma
• Public mental health awareness/literacy

To be included, studies also had to address one of the two research questions:

1. To have at least one follow-up point at least 4 weeks after the intervention was completed
2. The intervention was carried out in an LMIC setting

All reported stigma outcomes were classified into the categories of:

• Knowledge (changed what people know about mental health)
• Attitudes (changed what people think about mental health)
• Behaviour (changed what people do about mental health)

If possible, effect sizes for the interventions were calculated.

Two subgroup analyses on type of intervention were conducted:

1. A comparison of direct, indirect or no social contact interventions
2. A comparison of target groups

Results

Studies

A total of 80 studies were identified for inclusion in the review:

• 72 studies addressed the long-term effectiveness of anti-stigma interventions
• 11 studies addressed the setting of anti-stigma interventions

Medium-or long-term outcomes

Final follow-up assessment was 1–6 months after the intervention had ended in 69% of studies, with longer follow-up (1–10 years post-intervention) in 21% of studies.

• Knowledge outcomes: The median effect size was 0.54, indicating a medium effect in increasing knowledge about mental health
• Attitude outcomes: The median effect size was -0.26, indicating a small reduction in stigmatising attitudes against people with mental health problems
• Behavioural outcomes: The standardised mean difference could only be calculated in one intervention, which showed a small effect (0.22) in reducing stigmatising behaviour against people with mental health problems

Setting

Out of the 11 studies from LMIC settings, 8 were from upper middle income countries and 3 were from lower middle-income countries. No studies met the review inclusion criteria from a low income country.

Across the 11 studies included, there were 16 interventions (five measuring knowledge outcomes and 14 measuring attitude outcomes). None of the studies included behavioural outcomes.

There were sufficient data to calculate an effect size in only one of the studies; an education programme for the caregivers of patients with schizophrenia in Chile. In this study, the standardised mean difference for stigmatising attitudes was -2.11, indicating a large effect for the intervention in improving attitudes regarding people with mental health problems. However, this is just one study and there’s no space to analyse it here, so in terms of this review, it can pretty much be ignored.

Subgroup analyses

• Type of intervention: Interventions containing direct social contact had a smaller median effect size for stigmatising attitudes (-0.17) than those with indirect social contact (-0.32) or no social contact (-0.33)
• Target group: Interventions targeted at health professionals had a higher median effect size (-0.41) than those targeting school pupils (-0.21) or university students (-0.13)

Conclusions

The authors concluded that interventions aiming to reduce mental health-related stigma typically had a medium-sized effect on knowledge outcomes and a small effect on attitude outcomes in the long-term. There were insufficient data on the effect of interventions on behaviour in relation to people with mental health problems to come to any conclusions regarding the effectiveness, in the long-term or otherwise, of interventions aiming to reduce discrimination.

The authors also stated that there is a clear need for more stigma reduction studies, particularly from low-income countries.

Limitations

With any review, there are five potential areas of limitation (a good alternative name for the original Take That):

1. The studies available
2. The search conducted
3. The criteria for selecting the studies to include
4. The statistical analysis (if any)
5. The conclusions drawn

The studies available appeared to be the biggest problem for the authors for this review. They noted that the studies had a large variation in methodology and study quality. Basically, anti-stigma research is the scientific equivalent of a sock drawer with no matching socks, i.e. a sock drawer. For example, in the studies included in this review, there were 136 outcomes and 55 scales for measuring them. Stigma is undoubtedly a complex issue, but at least for comparison of research and for the design of effective interventions, there needs to be an agreed definition and way to measure it. Without this people will just carry on inventing stigma measures and interventions that address that one measure. It might be a good intervention. It might be a good measure. But without widespread, validated application we’ll never know. As such, without this we’ll delay advances towards reducing stigma against people with mental health problems.

The authors state that their search for studies was not limited by language and that, where possible, French and Spanish studies were translated by native speakers. This really means that the study wasn’t limited by these languages. There are probably more than three languages. It’s possible that there will be at least some studies that may otherwise have been eligible for inclusion, but were missed due to the language they were in. This is especially important given the relatively low number of studies that can be included, particularly from LMICs.

The authors acknowledge the possibility of publication bias. Intervention studies showing no impact on stigma may be less likely to be published than those that identify intervention benefits. Given that the effect sizes calculated in this review tended to be small to modest, this is troubling if we’re considering the overall effectiveness of stigma reduction and could lead us to suspect that anti-stigma interventions don’t work. However, we can’t know this from this review.

This review didn’t particularly look at specific interventions other than by grouping them by the extent of social contact. It’s all very well knowing if interventions in general are effective in the long-term, but arguably it would be more important to know exactly which interventions work in the long-term. It’s easier to do something if you know what that “something” is.

In terms of the criteria for selecting the studies, the statistical analysis and the conclusions drawn there appear to be no major problems. The largest issue here is the lack of any conclusion that can be drawn given the studies available. The authors use this to highlight the many gaps in the evidence for anti-stigma research, especially in terms of their efficacy in impacting behaviour and in LMICs.

Summary

This review has highlighted gaps in the evidence for anti-stigma interventions and a need for outcomes in this field of research to be made more comparable.

Unfortunately, not much can be said about whether interventions work to reduce stigma against people with mental health problems in the long term or interventions used in LCIMs. Almost nothing can be said about any one particular intervention. As always, the main message to take away is “more research is needed.” However, this is an important message, that like a tenuous pun about wallpaper of Winnie the Pooh making his own bread, bears and needs repeating. More accurately, more good quality, comparable research is required. Now get to it.

Links

Primary paper

Mehta N, et al. Evidence for effective interventions to reduce mental health-related stigma and discrimination in the medium and long term: systematic review. Br J Psychiatry 2015;207:377–384. [Abstract]

Other references

Oexle N, et al. Mental illness stigma, secrecy and suicidal ideation. Epidemiol Psychiatr Sci 2015;26:1–8. [PubMed abstract]

Griffiths KM, et al. Effectiveness of programs for reducing the stigma associated with mental disorders. A meta-analysis of randomized controlled trials. World Psychiatry 2014;13(2):161–175.

Photo credits

• Neal Fowler CC BY 2.0
• Steven Depolo CC BY 2.0
• Jake Stimpson CC BY 2.0
• Marco Castellani CC BY 2.0
• US Department of Agriculture CC BY 2.0
• Anne-Lise Heinrichs CC BY 2.0