Newspapers and other forms of mass media are influential in transmitting key information and ideas, including social commentary. ^{Reference Angermeyer and Schulze1,Reference Corrigan, Powell and Michaels2} Consequently, their role in reporting on mental illness is seen as one of the main conduits through which the public, especially those without personal exposure to mental illness, learn and come into contact with issues surrounding mental health. ^{Reference Dietrich, Heider, Matschinger and Angermeyer3,Reference Wahl4}

For many years it has been recognised that media reporting of mental illness is heavily biased in representing a misleading link between mental illness and violence, with preferential reporting for sensationalist stories depicting individuals with mental disorders as being aggressive, dangerous and unpredictable. ^{Reference Angermeyer and Schulze1,Reference Wahl4–Reference Nairn, Coverdale and Claasen7} Such widespread negative portrayals have reinforced the continued stigmatisation and lack of acceptance among those with psychiatric diagnoses, possibly increasing social distance and contributing to morbidity and distress. Of particular concern is the potential ‘internalisation’ of prejudices portrayed by the media, such that people with mental illnesses also begin to take on the view that they are dangerous and unstable. ^{Reference Biliæ and Georgaca8} Furthermore, there is a corresponding lack of voice for people of mental illness, ^{Reference Whitley and Berry6,Reference Biliæ and Georgaca8} with few personal accounts of their burden of disease and similarly little representation from qualified healthcare professionals. ^{Reference Whitley and Berry6,Reference Stuart9} All of this contributes to creating a stereotyped, isolated group without social status and acceptance. ^{Reference Thornton and Wahl10}

In a similar vein, there is a disappointing lack of coverage of stories on recovery and rehabilitation following mental illness. ^{Reference Corrigan, Watson, Gracia, Slopen, Rasinski and Hall5,Reference Whitley and Berry6} This potentially leads both the public and policy makers to erroneously believe that recovery is rare. We highlighted this in a newspaper survey 15 years ago in which we found that the reporting of mental illness and people with mental illness tended to be negative, whereas that of physical illness was much more likely to be positive in tone. ^{Reference Lawrie11} This may even influence health resource allocations to conditions that have a higher perceived chance of cure. ^{Reference Wahl4}

As far back as the 1960s, Nunally noted that mental health representations in media were few and distorted, limited to ‘neurotic behaviour’ in evening dramas where ‘the afflicted person often enters the scene staring glassy-eyed, with his mouth widely agape, mumbling incoherent phrases or laughing uncontrollably’. ^{Reference Nunally12} Steadman & Cocozza ^{Reference Steadman and Cocozza13} also showed that media depictions are often skewed towards extremes. Their findings on public perception of the ‘criminally insane’ (mentally disordered offenders) are salient, with 61% of the public surveyed responding that they feared former mentally disordered offenders ‘a lot’. ^{Reference Steadman and Cocozza13} However, as the authors noted, the mentally disordered offenders named by the public following intensive coverage tend not to have any legal claim to be ‘criminally insane’, and are merely assumed to be as such by the media.

Although those with mental illness are reportedly 2–6 times more likely to commit violence than matched controls, ^{Reference Corrigan, Watson, Gracia, Slopen, Rasinski and Hall5,Reference Taylor and Gunn14} the National Confidential Inquiry into Suicide and Homicide by People with Mental Illness ^{Reference Appleby, Kapur, Shaw, Hunt, While and Flynn15} found that only 5% of the homicides carried out in the general population between 2001 and 2011 were by those with an abnormal mental state (n = 52). However, an additional 14% had been in contact with mental health services in the preceding 12 months. Therefore, the vast majority of perpetrators of homicide do not have a mental illness at the time of the offence (n = 951). Contrast this with the statistics for suicide in the same report: 30% of all completed suicides in the 10-year period analysed were by patients already known to mental health services (n = 2678). The figures identified are 14 times higher than for homicide. From this, it is reasonable to infer that an individual with a mental health disorder is far more likely to self-harm than harm others. This important point is also often missing from media reports seeking to capitalise on the sensationalist angle and, by doing so, dehumanise the patient to the extent that they are defined by their illness. Similarly, it is clear that people with mental illness are much more likely to be ‘victims’ of crime than perpetrators. ^{Reference Hiroeh, Appleby, Mortensen and Dunn16,Reference Teplin, Mcclelland, Abram and Weiner17} This discrepancy in public perception between perpetrator and victim among those with mental illness has also been highlighted in other studies, ^{Reference Brekke, Prindle, Bae and Long18,Reference Walsh, Moran, Scott, Mckenzie, Burns and Tyrer19} acknowledging the need to shift attention away from the perceived threat to the public, to providing support and protection for the truly vulnerable. In this regard, the media has been counterproductive in reinforcing misleading stereotypes about public safety while simultaneously underreporting key statistics on mental health.

Although viral campaigns such as Rethink Mental Illness’ ‘Find Mike’ ^{Reference Gander20} have been highly successful in bringing mental illness to some public attention, it is difficult to assess their impact. Based on our previous survey, ^{Reference Lawrie11} we hypothesised that: mental health reporting would still be consistently more negative in tone compared with reporting of physical health issues, and patients would be consistently underrepresented in media depictions compared with physical health reports. Additionally, regarding the link with violence and criminality, we further hypothesised that the association between mental illness and violence would continue to pervade mental health reporting.

Method

The method followed that of Lawrie. ^{Reference Lawrie11} Nine weekday UK daily newspapers (Scottish editions) were surveyed over a period of 4 weeks between 4 March and 4 April 2014. The papers chosen represent a selection of popular broadsheet (The Times, The Telegraph, The Independent, The Guardian, The Scotsman, The Herald) and tabloid (Daily Mail, The Sun, Daily Record) publications in the UK, with a reported range in circulation of between 64 279 (The Independent) and 2 048 977 (The Sun) in February 2014 ²¹ Only national newspapers were included in this analysis, as they have the widest distribution.

Print copies of the newspapers were obtained from the Reference Library of Edinburgh on the day of publication and manually screened for headlines relating to mental health or psychiatry and general health issues. Two tabloids, The Sun and Daily Record, were purchased each day, as they were not subscribed to by library services, yet have the highest circulation figures in the UK.

Articles relating to mental health and physical health were identified initially through headline screening. The main body of text was then read and coded as positive, negative or neutral, if the article was mainly stating facts or had mixed elements. As such, it is possible to have a disparity between tone of headline and article. Each article was then analysed and coded to a schema with eight variables, which noted key themes (Table 1).

Table 1 Coding schema for newspaper articles

Variable	Description
Quote from patient, relative or charity	Direct or paraphrased quote from a patient with the illness in question, relative or advocacy organisation
Quote from professional	Quote from a healthcare professional in charge of the patient's care or spokesperson for government organisations, etc.
Association with violence	Any article where a person portrayed as having a mental illness or under the influence of drugs/alcohol engages in criminal or dangerous activity where others are put at risk or harmed. Suicide or self-harm is not included in this category
Mention of treatment/ rehabilitation	Either specific (e.g. Prozac), general (anti-depressants, cognitive–behavioural therapy), alternative (mindfulness, yoga) or relating to the Mental Health Act 1983 where treatment is noted under a compulsory treatment order
Substance misuse	Articles relating to addiction, overdose or psychiatric effects of substance misuse were classified as relating to mental health. Physical effects of drug usage, including side-effects, were classified under physical health
System failure/ negligence	Any article where substandard care is delivered by a single person or on an organisational level, resulting in adverse outcomes. Shortcomings range from IT failures and changes in social policy to deliberate mistreatment
Innovations in research	Articles where a new diagnostic method, treatment or better understanding of a named disease is reached
Medical advocacy/raising awareness	Where the primary purpose of the article is to transmit information about a particular disease profile or set of symptoms or charity efforts. For example Autism Awareness Week, United Nations World Down Syndrome Day, ‘Check 'Em Tuesday’ – a campaign run by The Sun to increase uptake of self-screening for breast cancer

Inclusion and exclusion criteria

The sports and business sections were surveyed, however property and automobile inserts were excluded due to likely lack of relevance. Primary newspaper articles, editorials and opinion pieces were included in the data collection. Advertisements, letters to the editor, film and entertainment reviews and obituaries were omitted. Articles with only a passing or chance reference to mental or physical health were also omitted.

All articles concerning mental disorders recognised by the DSM-5, ²² psychiatry or psychiatrists were included, as well as generic, undefined mental conditions with the exception of paedophilia and the paraphilias, as their classification as a mental illness in DSM-5 is controversial. For physical health, all articles relating to other general medical issues were collected. Articles concerning the National Health Service (NHS) or other organisational bodies were included only if there was a demonstrable mention of impact on patient health and safety.

Results

In total, 963 health-related articles were collected during the 4-week period. Of these, 763 (79.2%) concerned physical health, whereas 200 (20.8%) were about mental health. Although there was a slight preponderance towards neutral pieces (38.0%) in the physical health category, compared with positive (28.6%) and negative (32.9%) reporting, the majority of articles concerning mental health were negative in tone (50.5%), with only a small number of positive (22.0%) or neutral articles (27.5%) (Box 1). This striking disparity between mental and physical health reporting underscores a key finding that mental health reporting remains consistently more negative than physical health, which is in line with the results from Lawrie's ^{Reference Lawrie11} study over a decade previously (ratio of positive to negative articles for physical health: 218/251; and for mental health: 44/101; odds ratio (OR) 1.99, 95% confidence interval (CI) 1.34–2.97). In addition, we found a prominent and stigmatising association between mental illness and violence as 18.5% of mental health articles featured violence as a major theme, compared with 0.3% of physical health articles (OR = 86.37, 95% CI 20.60–362.1) (Table 2).

Table 2 Comparison of variables between physical and mental health reporting

Variable	Physical health, n (%)	Mental health, n (%)
Total number of articles	763	200
Positive	218 (28.6)	44 (22.0)
Negative	251 (32.9)	101 (50.5)
Neutral	290 (38.0)	55 (27.5)
Quote from patient/charity	150 (19.7)	45 (22.5)
Quote from professional	285 (37.4)	58 (29.0)
Association with violence	2 (0.3)	37 (18.5)
Mention of treatment/rehabilitation	217 (28.4)	48 (24.0)
Substance misuse	31 (4.1)	22 (11.0)
Medical advocacy/raising awareness	131 (17.2)	26 (13.0)
System failure/negligence	118 (15.5)	25 (12.5)
Innovations in research	184 (24.1)	22 (11.0)

Although previous studies have shown that patients with mental illness are underrepresented in newspaper articles, ^{Reference Whitley and Berry6,Reference Biliæ and Georgaca8} we found this not to be the case, with a higher proportion of quotes from patients with mental illness, their relatives or mental health charities (22.5%) compared with physical health (19.7%) (ratio of quotes in physical health: 150/613; mental health: 45/155; OR=0.84, 95% CI 0.58–1.23). There appears to be comparatively fewer quotes from professionals in relation to mental health however, highlighting the need of the psychiatric profession to make a conscious effort to engage with the media, and vice versa. ^{Reference Salter and Byrne23} Interestingly, whereas previous studies have highlighted a lack of reporting on treatment and rehabilitation for mental health, ^{Reference Wahl4–Reference Whitley and Berry6} here we found the rates to be comparable to that of physical health articles (24.0% v. 28.4% respectively) and previous reports. ^{Reference Whitley and Berry6}

Box 1 Examples of positive, negative and neutral headlines for mental and physical health

Physical health

Positive

‘US Aids baby may be cured’

‘Cholesterol drugs may slow MS, study shows’

Negative

‘Three-year old died after hospital delay’

‘ “Sickness culture” claim as 1 in 3 has long-term illness’

Neutral

‘Statins have no side-effects, study finds’

‘Smoking link to breast cancer in older women’

Mental health

Positive

‘Depression and self-harming are taboo … that's why I'm determined to speak out’

‘Ketamine trial provides fresh hope for long-term depressed’

Negative

‘Mentally ill man in bus knife attack’

‘School accused of excluding girl with anorexia to avoid copycats’

Neutral

‘Give dementia patients trial drugs en masse’

‘Depression must be seen as a societal disorder’

The disorders most commonly associated with violence were personality disorders as well as undefined ‘mental illnesses’ and post-traumatic stress disorder, which make up 13.5% of all articles pertaining to mental health. Surprisingly, sensationalist stories about schizophrenia were minimal and comparable to reporting on other mental disorders, whereas dementia and eating disorders were least likely to be reported as associated with violence (Table 3).

Table 3 A breakdown of newspaper articles according to diagnosis and association with violence

Diagnosis	Articles n (%)	Association with violence, n (%)
Suicide	31 (15.5)	0
Mood disorder	26 (13)	2 (1.0)
Dementia	32 (16)	0
Eating disorder	12 (6)	0
Developmental disorder	17 (8.5)	3 (1.5)
Substance misuse	13 (6.5)	2 (1.0)
Personality disorder	11 (5.5)	14 (7.0)
Schizophrenia/psychosis	10 (5)	3 (1.5)
Other a	49 (24.5)	13 (6.5)

a. Includes multiple disorders, undefined and post-traumatic stress disorder.

Based on the previous data that show people with mental illnesses are more likely to be victims of crime rather than perpetrators, we were interested to see whether this correlated with media reports. We found that people with mental illness are twice as likely to be portrayed as a perpetrator (34/200) compared with a victim (17/200), in direct contrast to the available data. ^{Reference Hiroeh, Appleby, Mortensen and Dunn16,Reference Brekke, Prindle, Bae and Long18,Reference Walsh, Moran, Scott, Mckenzie, Burns and Tyrer19,Reference Choe, Teplin and Abram24}

Given that 18.5% of articles on mental health concern violent acts and are seen as a threat to the public, we asked what made up the rest of the negative reporting. On further analysis, it appears that a significant proportion of the negative articles pertain to the stigmatisation and abuse of people with mental health within society, as well as lack of funding and adequate care provision. Examples include headlines such as: ‘The great mental health betrayal’ (The Independent, 13 March 2014), an article about the unlawful detaining of people with mental health illnesses in care homes and hospitals, and ‘Mental health and race – the blight of dual discrimination’ (The Guardian, 26 March 2014), which highlights the additional burden of being an ethnic minority with mental illness. The remainder of the negative articles are reports on suicide (15.5%).

Overall, there was no significant difference between mental health reporting in broadsheets compared with tabloids.