Association of Specific Mental Disorders With Premature Mortality in the Danish Population Using Alternative Measurement Methods

Abstract

Importance: The association of mental disorders with premature mortality published in the Global Burden of Disease (GBD) studies has been underestimated because these analyses have recommended using only a small number of mental disorders as causes of death to estimate years of life lost (YLL). Alternative methods have been introduced, such as estimating life-years lost (LYL), to compare individuals with mental disorders with the general population.

Objectives: To generate register-based YLL and LYL estimates and to use these measurement methods to assess the association of specific mental disorders with premature mortality.

Design, setting, and participants: This population-based cohort study included all persons with and without mental disorders aged 0 to 94 years who were living in Denmark between January 1, 2000, and December 31, 2015. Data were analyzed from January to December 2019.

Main outcomes and measures: Danish health registers were used to identify mental disorder diagnoses, dates of death, and causes of death. The YLLs were estimated for the set of mental health-associated causes of death, and all-cause and cause-specific LYLs were estimated for 18 specific mental disorders and 3 broad categories of mental disorders that were recommended for use in the GBD studies. The association between the number of comorbid mental disorders (divided into categories of persons with ≥1 type of disorder, ≥2 types of disorders, ≥3 types of disorders, and ≥4 types of disorders) and LYL estimates was also examined.

Results: A total of 6 989 627 individuals (3 481 219 male persons [49.8%] and 3 508 408 female persons [50.2%]; mean [SD] age at study enrollment, 32.2 [24.4] years) were followed up for a total of 85 911 461 person-years. The YLL rates per 100 000 person-years were highest for alcohol use disorder (for male individuals, 568.7 [95% CI, 564.4-572.7]; for female individuals, 155.5 [95% CI, 153.5-157.9]) and suicide (for male individuals, 590.1 [95% CI, 583.8-596.5]; for female individuals, 202.3 [95% CI, 198.5-206.4]). Although only 3 of 18 mental and substance use disorders could be associated with YLL, all mental disorders were associated with shorter life expectancies when LYL was used for measurement. Male and female individuals diagnosed with any mental disorder had life expectancies that were shorter by 11.2 years (95% CI, 11.1-11.3 years) and 7.9 years (95% CI, 7.8-8.0 years), respectively, and remaining life expectancy decreased further among those with comorbid mental disorders. Drug use disorders were associated with the highest excess LYL estimates; however, common mental disorders, such as depressive and anxiety disorders, were also associated with substantial premature mortality.

Conclusions and relevance: Mental disorders were observed to be associated with reductions in life expectancy. This finding provides a foundation for future intervention programs designed to reduce the differential mortality gap associated with mental disorders. Register-based studies allow the calculation of precise individual-level YLLs and LYLs, and both measurement methods are informative for health care planning. Compared with YLL, the novel LYL measurement approach may more precisely capture the association of mental disorders with premature mortality and facilitates the exploration of comorbidity and specific causes of death in individuals with mental disorders.

Figure 1.. Excess Life-Years Lost Associated With Specific Causes of Death Among Persons With Specific Types of Mental Disorders vs the General Danish Population

A, Male. Persons with specific mental disorders were compared with persons of the same age and sex in the general Danish population. Excess life-years lost associated with specific causes of death could not be estimated for disorders with less than 3 cases (cocaine use disorder, amphetamine use disorder, eating disorders, attention-deficit/hyperactivity disorder, and conduct disorder). Estimates and 95% CIs are available in eTable 5 in the Supplement. B, Female. Persons with specific mental disorders were compared with persons of the same age and sex in the general Danish population. Excess life-years lost associated with specific causes of death could not be estimated for disorders with less than 3 cases (cocaine use disorder, amphetamine use disorder, eating disorders, attention-deficit/hyperactivity disorder, and conduct disorder). Estimates and 95% CIs are available in eTable 5 in the Supplement.

Figure 2.. Excess Life-Years Lost Associated With Specific Causes of Death Among Persons With Any Mental Disorder and Multiple Mental Disorders vs the General Danish Population

A, Male. Persons with any mental disorder and multiple mental disorders were compared with persons of the same age and sex in the general Danish population. Estimates based on less than 3 cases are not shown. Estimates and 95% CIs are available in eTable 5 in the Supplement. B, Female. Persons with any mental disorder and multiple mental disorders were compared with persons of the same age and sex in the general Danish population. Estimates based on less than 3 cases are not shown. Estimates and 95% CIs are available in eTable 5 in the Supplement.