Association between mortality-related health metrics and mental disorders




Systematic reviews have consistently demonstrated that those with mental disorders have an increased risk of premature mortality. We observed that people with mental disorders have 2-3 times higher mortality rates than those without mental disorders, and they experienced a life expectancy 7-10 shorter than persons from the general population. We have provided a comprehensive analysis of mortality associated with different types of mental disorders, presenting both Mortality Rate Ratios (MRRs) and premature mortality estimated using Life-Years-Lost (LYLs).




A study by O. Plana-Ripoll et al.

﹀

Mortality rates & mental disorders


Difference between Mortality Rates and Mortality Rate Ratios

Mortality rate is defined as the proportion of people dying in a given population and time period. It is estimated as a count of individuals who died per number of people in the population, per unit of time.

Our study examined the entire Danish population, and the unit of measurement is person-years. We show mortality rates for those with a mental disorder, and those without any mental disorder. As expected, for both groups, mortality rates increase constantly with age. However, mortality rates for those with a mental disorder are systematically higher compared with those without any mental disorder.

The mortality rates for the two groups across age show a ‘bulge’ during young adulthood (20-40 years). One way of comparing the two rates is through the ratio of them. The Mortality Rate Ratio is shown in the right-hand panel. It is not constant over age. Overall, Mortality Rate Ratio peaked around age 35 years old, when people with mental disorders have an 7-fold higher mortality rate, compared with those without mental disorders.

Sex-specific and Cause of death-specific Mortality Rates and Ratios

Sex-specific Mortality Rate Ratios for those with any mental disorder showed a similar pattern for males and females, although males had a higher Mortality Rate Ratio compared to females after age 20 years. Causes of death can also be categorised between natural and unnatural causes of death (see list of causes in each category). Mortality Rate Ratios linked with unnatural causes of death are much higher for people with mental disorders because of the impact of suicide. The next part of this webpage will explore the links between mental disorders and causes of death in more detail.


Mortality Rate Ratios by specific causes of death


Compared with people without mental disorders, those diagnosed with a mental disorder have between 2 and 3 times higher rates of dying during a specific period of observation (MRR=2.53 [95% CI 2.52-2.54]).

That is, people with mental disorders experience premature mortality. Our new study provides details on this risk by types of mental disorder (e.g. people with depression, substance use disorders, etc).

In addition, we can examine what was the underlying final cause of death (e.g. cancer, stroke, suicide, etc).

For historical reasons, Causes of Death are often divided into broad categories. Under the overall category “Any Cause”, there are two broad subcategories: (a) “Unnatural Causes” which includes suicide, accidents and being the victim of homicide; and (b) “Natural Causes” which includes all other causes, e.g. cancer (i.e. neoplasms), infectious diseases, and circulatory disorders.




Mortality Rate Ratios per specific cause of death for people with:

You can study sex together or separately:
Together
Separately
Select your kind of scale
Normal
Log

Sex-specific MRRs were systematically higher for males than females, except for suicide that was usually higher for females (MRR=17.5 for any mental disorder). It is also possible to split this analysis at the mental disorder level.

For instance, it appears that substance-use disorders are associated with very high MRRs, notably concerning unnatural causes of death and alcohol misuse.

Inversely, organic or behavioral disorders are linked with low MRRs; although always above 1 (i.e. higher mortality rates compared with those without a diagnosis).

The multi-panel plot below shows the associations between specific mental disorders and specific causes of death . Hover over specific causes of death to highlight these across all disorder panels.



You can study sex together or separately:
Together
Separately
Select your kind of scale
Normal
Log

Life Years Lost

Life Years Lost is another way to measure the impact of a disorder on premature mortality. This metric compares the expected remaining life expectancy after the diagnosis for each person with a diagnosis of a mental disorder, compared to a person of same age from the general reference population.

For example, men diagnosed with substance use disorders experience the largest Life Years Lost estimate: their remaining life expectancy after diagnosis is 14.8 years shorter than men of the same age from the general Danish population.

In addition, the total Life Years Lost are divided into specific causes of death: if you click to the mental disorder of interest, the Life Years Lost attributable to each cause are shown.

The total of 14.8 years for men diagnosed with substance use disorders can be decomposed into 9.4 years due to natural causes (including e.g. 5.7 years due to alcohol misuse and 0.9 years due to respiratory diseases) and 5.4 years due to unnatural causes (including e.g. 1.8 years due to suicide).

Being diagnosed with a mental disorder is always associated with a positive estimate of Life Years Lost (i.e. shorter life expectancies), both for natural and unnatural causes of death. Surprisingly, we found that people – especially men – with all types of mental disorders had negative estimates for Life Years Lost associated with neoplasm-related deaths compared to the general population of same age, which means that the general population loses more years of life due to neoplasm-related deaths than those with mental disorders.



Males
Females



About

This webpage aims to provide more information on our scientific publication in The Lancet. Pipeline of analysis is described here with reproducible code, dataviz section is here. You can see more of our work on Comorbidity in Mental disorder epidemiology on our webpage: nbepi.com.

NB Epi.com Paper Mail Data