Mortality and Medical Comorbidity in the Severely Mentally Ill

Abstract

Background: Mentally ill patients die on average 10 years earlier than the general population, largely due to general medical disorders. This study is the first to explore in a large German sample the prevalence, mortality, and medical comorbidity in pa- tients with severe mental illness (SMI). The patients were affected by borderline personality disorder (BPD), psychotic disorders, bipolar disorder, or severe unipolar depression.

Methods: Our database consists of billing data from all adults with statutory health insurance in Germany. Twelve-month administrative SMI prevalence and medical comorbidity were estimated using cross-sectional data from 2016 (age ≥ 18; N = 59 561 310). Two-year mortality was established longitudinally in a randomly selected subset of the billing data (most recent mortality information available for 2012 to 2014; 2012: n = 15 590 107).

Results: Severe unipolar depression had the highest prevalence (2.01%), followed by psychotic disorders (1.25%), BPD (0.34%), and bipolar disorder (0.29%). While the prevalence of malignant neoplasms showed moderate deviations from reference values [severe unipolar depression: OR = 1.30 (95% CI = 1.29; 1.31), BPD: OR = 1.11 (1.09; 1.14), psychotic dis- orders: OR = 0.90 (0.89; 0.90), bipolar disorder: OR = 1.07 (1.06; 1.09)], other disease groups (infectious, endocrine/nutritional/ metabolic, circulatory, respiratory) were substantially elevated in all categories of SMI. Mortality rates for psychotic disorders, BPD, bipolar disorder, and severe unipolar depression were increased (OR = 2.38 [95% CI=2.32; 2.44], 2.30 [2.08; 2.54], 1.52 [1.42; 1.62], and 1.40 [1.37; 1.44], respectively), with a loss of 2.6 to 12.3 years, depending on age, sex, and SMI.

Conclusion: Mortality is substantially elevated in all SMI patients. The results underline the need to remove barriers to adequate general medical care, both on the patient and the provider side, to reduce excess mortality.

Figure 1

Mortality rates in patients with severe mental illness (SMI) compared with patients without the respective SMI (relative risk) over the whole adult life span. The elevated risk of death of SMI patients is particularly high at young age. Data from statutory health insurance funds; subsample (n = 15 590 107) with information on mortality (deaths in this subsample, 2013 to 2014: n = 288 503); results based on a simple descriptive analysis relating age- and sex-specific mortality rates of patients without a specific SMI to those of patients with the respective SMI; ICD-10 diagnoses included: F60.3x, F2x, F30.x/F31.x, F32.2/F33.2/F32.3/F33.3

Figure 2

Estimated loss of life years by age for patients with severe mental illness (SMI) compared with total population. Data from statutory health insurance funds; subsample (n = 15 590 107) with information on mortality (deaths in this subsample, 2013 to 2014: n = 288 503); ICD-10 diagnoses included: F60.3x, F2x, F30.x/F31.x, F32.2/F33.2/F32.3/F33.3

eFigure

Age-dependent mortality in patients with SMI and in patients without the specific SMI, together with the relative risk (SMI vs. no SMI) over the whole adult life span. Data from statutory health insurance funds; subsample (n = 15 590 107) with information on mortality (deaths in this subsample, 2013 to 2014: n = 288 503). The results are based on a descriptive analysis: age- and gender-specific mortality rates of patients with SMI (ICD-10 diagnoses: F60.3x, F2x, F30.x/F31.x, F32.2/F33.2/F32.3/F33.3) to patients without the respective SMI SMI = severe mental illness