The Global Burden of Disease (GBD) study recently reported that smoking and poor nutrition were the most significant causes of premature mortality and disability among people living in England, followed closely by obesity, hypertension, alcohol and drug misuse, high cholesterol and renal disease. ^{Reference Newton, Briggs, Murray, Dicker, Foreman and Wang1} Life expectancy in the UK is higher in less-deprived areas, with poverty and socioeconomic variations in the UK showing substantial immovable impact; 40% of the NHS budget is spent in preventable illness. The GBD project offers powerful analyses to guide policy and practice when trying to tackle global and national inequalities. Socioeconomic status explains 50% of the country variation in disabilities associated with infectious, maternal, neonatal, nutritional, musculoskeletal and other non-communicable diseases; socioeconomic status accounts for less than 10% of country variation in cirrhosis, unintentional injuries, self-harm and interpersonal violence, as well as cardiovascular disease, chronic respiratory diabetes, urogenital and endocrine disorders. ^{Reference Murray, Barber, Foreman, Ozgoren, Abd-Allah and Abera2}

The GBD findings from high- and low-income countries do not show premature mortality in people with mental illness but significant disability over the life course. The limitation of identifying a single cause of death in the GBD study means that the majority of people facing mental disorders who suffer other illnesses are not captured as dying from mental illness. ^{Reference Bhui3} Thus, the years of life lost due to mental illnesses (schizophrenia, depression, intellectual disabilities, autism, bipolar disorder, dementia, drug and alcohol use) are significantly underestimated by GBD methods. ^{Reference Charlson, Baxter, Dua, Degenhardt, Whiteford and Vos4} Men and women in England now live longer, and similar trends are seen in all high-income economies. However, the rates of premature death among people with mental illnesses have not shifted for two decades, attracting the epithet of lethal discrimination. ^{Reference Thornicroft5,6} Taggart & Bailey (pp. ) and Shiers et al (pp. ) remind us of the disastrous life expectancy statistics facing people with severe mental illness, revealing a failure of social policy and health service provision. Specifically, prescribing of antipsychotics to patients with psychosis without due attention to weight management and more active management of cardiometabolic risks are thought to account for the excess mortality, although smoking and cancer are also significant targets for prevention of premature deaths, as they are in the entire population. Surprisingly, stigma and discrimination are highest in countries with favourable scores on the Human Development Index, reflecting longer life, and better access to knowledge and standards of living (see Lasalvia et al, pp. ) and so premature deaths cannot be uniquely attributed to a lack of education or poverty.

A substantial number of people with mental illnesses still do not access the right treatment, leading to a very poor prognosis in the long term, as illustrated by Ran et al's 14-year follow-up in Chengdu, China (pp. ). The untreated minority more often were from rural regions, were older, homeless, socially isolated without a carer, experienced poor family attitudes and they were more likely to die from other causes. Partial or complete remission was higher in the patients receiving treatment (57% v. 29%), calling into question earlier studies suggesting that schizophrenia has a better prognosis in low-income countries. ^{Reference Cohen, Patel, Thara and Gureje7} There is a significant ‘treatment gap’ even in high-income countries, so lower-income countries struggle with screening and recognition of mental illnesses. Jordans et al's new Community Informant Detection Tool (see pp. ), and related task-shifting methods, offer a new approach to improve public mental health for improved outcomes.

Other global causes of preventive concern are violence and suicide; 35% of women worldwide suffer intimate partner violence or non-partner sexual violence in their lifetime, and 30% of women in a relationship report physical or sexual violence by their partner. ⁸ De Mooij et al (pp. ) show that violent victimisation is more common among those with severe mental illness, especially patients admitted to hospital where property crimes were more common; and intimate partner violence is a significant cause of concern, ^{Reference Khalifeh, Oram, Trevillion, Johnson and Howard9} especially for people with severe and chronic mental illnesses and multiple disadvantage. ^{Reference Dean, Walsh, Moran, Tyrer, Creed and Byford10} Preventive interventions for suicide and self-harm continue to evolve, ^{Reference Scott and Guo11} yet new risks are emerging with social media platforms. Westerlund et al's remarkable study (pp. ) shows posts before and after a suicide, with worrisome findings; and yet there are also powerful preventive opportunities.

Preventive interventions may be offered to entire populations, and to cohorts at the first episode of psychosis. Hodgekins et al (pp. ) finds a typology of social recovery in a national cohort of individuals with first-episode psychosis; the majority of patients face poor social recovery (Low Stable, 66%; Moderate-Increasing, 27%) and even those with little social disability show some deterioration over time (High-Decreasing, 7%). Relatively worse social recovery was found in men, ethnic minorities, those who were younger at onset of psychosis, with increased negative symptoms and poor premorbid adjustment. The concept of recovery needs better methods of measurement and intervention (see Slade et al, pp. , and Williams et al, pp. ) that target recovery-oriented relationships and working practices of staff. Adverse pathways into care in a first episode of psychosis are associated with a longer duration of untreated psychosis and should be avoided; criminal justice pathways (CJPs) at a first onset of psychosis are more common if there is violence, substance misuse or psychopathy, but these do not explain or mediate ethnic variations of CJPs (see Bhui et al, pp. ). So the long-standing finding of Black Caribbean and Black African patients, even at first onset, entering care through CJPs remains a controversial and unresolved cause for concern. ^{Reference Morgan, Mallett, Hutchinson and Leff12} Ethnic variations in access have previously been reported for severe mental illness; less is known about neuroses, so Fernandez de la Cruz et al's findings (pp. ) of underrepresentation for obsessive–compulsive disorders in an inner-London service, and more severe underrepresentation than for depression, raises further questions about how people with mental illness and multiple disadvantage seek help. Irrespective of the socioeconomic and cultural environment, probing and inquisitive investigations of genetic and phenotypic biomarkers (see Li et al, pp. , and Cederlöf et al, pp. ) and paternal depression (Nath et al, pp. ) offer optimism for better understanding of aetiological mechanisms, so as to make life-course epidemiology and prevention a reality. ^{Reference Ben-Shlomo and Kuh13} Margaret Mead's biography, Blackberry Winter, alludes to her experiences as an outsider, a woman and culturally isolated, writing that ‘those who reject and are rejected, and usually both, suffer irreversible character damage’. ^{Reference Mead14} Lethal discrimination needs more active policy and practice priority for a progressive society in which all babies, children and adults flourish and have equal chances of healthy and long lives.