Schizophrenia mortality data are from the National Death Index of Medicaid beneficiaries. General population mortality data are from the Centers for Disease Control and Prevention WONDER data.64Table 2 lists the 95% CIs associated with SMRs. COPD indicates chronic obstructive pulmonary disease; SMR, standardized mortality ratio (standardized for age, sex, race/ethnicity, and geographic region).
eTable 1. ICD-10-CM Codes Used to Define Primary Cause of Death Groups Used in Tables 1 through 3
eTable 2. ICD-10-CM Codes Used to Define Primary Cause of Death Groups in Table 4
eTable 3. Total Patients, Fee-for-Service Patients, Managed Care Patients, and Deaths in the Schizophrenia Cohort Stratified by State
eTable 4. Number of Observed Deaths of Adult Medicaid Beneficiaries Diagnosed With Schizophrenia by Disease Category, Sex, Age Group, and Race/Ethnic Group
eTable 5. Demographic Characteristics of a National Cohort of Medicaid Beneficiaries Diagnosed With Schizophrenia, Observed Deaths, Person Years, and All-Cause Mortality Rates
Olfson M, Gerhard T, Huang C, Crystal S, Stroup TS. Premature Mortality Among Adults With Schizophrenia in the United States. JAMA Psychiatry. 2015;72(12):1172-1181. doi:10.1001/jamapsychiatry.2015.1737
Although adults with schizophrenia have a significantly increased risk of premature mortality, sample size limitations of previous research have hindered the identification of the underlying causes.
To describe overall and cause-specific mortality rates and standardized mortality ratios (SMRs) for adults with schizophrenia compared with the US general population.
Design, Setting, and Participants
We identified a national retrospective longitudinal cohort of patients with schizophrenia 20 to 64 years old in the Medicaid program (January 1, 2001, to December 31, 2007). The cohort included 1 138 853 individuals, 4 807 121 years of follow-up, and 74 003 deaths, of which 65 553 had a known cause.
Main Outcomes and Measures
Mortality ratios for the schizophrenia cohort standardized to the general population with respect to age, sex, race/ethnicity, and geographic region were estimated for all-cause and cause-specific mortality. Mortality rates per 100 000 person-years and the mean years of potential life lost per death were also determined. Death record information was obtained from the National Death Index.
Adults with schizophrenia were more than 3.5 times (all-cause SMR, 3.7; 95% CI, 3.7-3.7) as likely to die in the follow-up period as were adults in the general population. Cardiovascular disease had the highest mortality rate (403.2 per 100 000 person-years) and an SMR of 3.6 (95% CI, 3.5-3.6). Among 6 selected cancers, lung cancer had the highest mortality rate (74.8 per 100 000 person-years) and an SMR of 2.4 (95% CI, 2.4-2.5). Particularly elevated SMRs were observed for chronic obstructive pulmonary disease (9.9; 95% CI, 9.6-10.2) and influenza and pneumonia (7.0; 95% CI, 6.7-7.4). Accidental deaths (119.7 per 100 000 person-years) accounted for more than twice as many deaths as suicide (52.0 per 100 000 person-years). Nonsuicidal substance-induced death, mostly from alcohol or other drugs, was also a leading cause of death (95.2 per 100 000 person-years).
Conclusions and Relevance
In a US national cohort of adults with schizophrenia, excess deaths from cardiovascular and respiratory diseases implicate modifiable cardiovascular risk factors, including especially tobacco use. Excess deaths directly attributable to alcohol or other drugs highlight threats posed by substance abuse. More aggressive identification and management of cardiovascular risk factors, as well as reducing tobacco use and substance abuse, should be leading priorities in the medical care of adults with schizophrenia.
Adults with schizophrenia are at markedly increased risk of premature death.1,2 Despite elevated rates of suicide and other unnatural causes of death, most of the excess mortality has been attributed to cardiovascular disease, respiratory disease, and other natural causes.3 One analysis from the United Kingdom found that suicide, homicide, and accidental deaths collectively accounted for 21 of 164 schizophrenia deaths.4 Accurate characterizations of premature mortality patterns are important to inform clinical and policy initiatives to improve services and reduce preventable deaths in this patient population.
Many factors, including economic disadvantage, negative health behaviors, and difficulties accessing and adhering to medical treatments, are thought to contribute to premature mortality in schizophrenia.5 Smoking,6 limited physical activity,7 obesity,8 elevated serum glucose level,9 hypertension,10 and dyslipidemia11 are all significantly more common in schizophrenia than in the general population. Adults with schizophrenia are also less likely than age-matched peers to receive adequate treatment for major medical conditions,12 which may compound risk of premature mortality.13
Research on early mortality in schizophrenia primarily derives from Western Europe.2 Because the United States differs from most Western European countries in its health and social welfare systems14 and in several relevant health indexes (including life expectancy,15 obesity,16 blood pressure,17 and tobacco use18), Western European mortality estimates for schizophrenia may not directly generalize to the United States.
In the United States, premature mortality has been well documented in diagnostically mixed samples of patients with severe mental illness.19,20 Investigations in the United States limited to schizophrenia have primarily involved small samples (<1000 patients) published more than a quarter century ago.21,22 A recent US study23 comparing mortality for schizophrenia research participants with a demographically matched general population reference group reported a standardized mortality ratio (SMR) of 2.80, with all 25 deaths in the schizophrenia group occurring from natural causes. In a large cohort of US military veterans, it was further found that veterans with schizophrenia were significantly more likely than those without mental disorders to die of heart disease.24
We conducted a national examination of premature mortality among adults with schizophrenia in the Medicaid program, the largest payer of health services for persons with schizophrenia in the United States.25 Mortality rates and mortality ratios standardized to the general population by age, sex, race/ethnicity, and geographic region were used to characterize the burden and excess mortality from several common medical diseases overall and stratified by demographic characteristics. By characterizing key sources of excess mortality in a large cohort with schizophrenia, the results provide a more comprehensive picture than was previously available of the gap in mortality, highlighting the need for more effective strategies to improve the medical care of this patient population.
The total resident population and death information were obtained from the January 1, 2001, to December 31, 2007, US Compressed Mortality File.26 Age, sex, race/ethnicity, and year-specific life expectancy data were obtained from the 2006 United States Life Tables.27 The schizophrenia cohort was extracted from the national Medicaid Analytic eXtract (MAX) data from the Centers for Medicare & Medicaid Services.63 It included data from 45 states, not including Arizona, Delaware, Nevada, Oregon, and Rhode Island. Dates and cause of death information for the schizophrenia cohort were derived from linkage to the National Death Index (NDI), which provides a complete accounting of state-recorded deaths in the United States and is the most complete resource for tracing mortality in national samples.28 The data, which are deidentified, were determined to be exempt from human participants review by the Rutgers University Institutional Review Board.
We identified a national retrospective longitudinal cohort of patients with schizophrenia 20 to 64 years old who received at least 2 outpatient claims or at least 1 inpatient claim for schizophrenia (ICD-10-CM code 295).29 The first observed day on which the inclusion criteria were met defined the start of follow-up. The cohort was followed forward until the loss of Medicaid eligibility, the date of death, or December 31, 2007 (end of NDI-linked MAX data), whichever came first.
All causes of death were first divided into natural and unnatural causes. Natural causes were partitioned into cardiovascular disease, cancer, diabetes mellitus, renal failure, influenza and pneumonia, sepsis, chronic obstructive pulmonary disease (COPD), liver disease, and other natural causes. Cardiovascular disease was subpartitioned into ischemic heart disease, nonischemic heart disease, stroke, and other circulatory diseases. Cancer was subpartitioned into lung, colon, breast, liver, pancreas, hematologic, and other cancer. Unnatural causes were partitioned into suicide, accidents, assault (homicide), and injuries with undetermined intent and other injuries (eTable 1 in the Supplement). In addition, an alternative overlapping Centers for Disease Control and Prevention classification was used to define substance-induced deaths, including drug-induced and alcohol-induced deaths, and firearm-related deaths.30 Drug-induced nonsuicidal deaths were also partitioned into those induced by drugs of abuse (opioids, cannabinoids, sedatives or hypnotics, cocaine, stimulants, and volatile solvents) and others. Deaths related to legal interventions (eg, encounters with law enforcement officials) were also examined (eTable 2 in the Supplement).
Based on Medicaid eligibility data, cohort members were classified by sex, age group (20-34, 35-44, 45-54, and 55-64 years), and race/ethnicity (Hispanic, white non-Hispanic [white], black non-Hispanic [black], and other non-Hispanic [other], including American Indian/Alaskan native, Asian, native Hawaiian/other Pacific Islander, and more than 1 race/ethnicity. Cohort members were also classified by geographic region (West, Midwest, South, and Northeast).
In the schizophrenia cohort, person-years of follow-up, number of deaths, and mortality rates per 100 000 person-years of follow-up were determined overall and stratified by demographic characteristics. To facilitate comparisons of the mean years of lost life per death across causes of death, the mean years of potential lost life per death were calculated overall and for each cause of death as the mean of the remaining life expectancy in years for each deceased schizophrenia cohort member at the age at death, as determined from the 2006 United States Life Tables based on the age at death, sex, and race/ethnicity.
Cause-specific mortality rates and SMRs with 95% CIs were calculated for the entire schizophrenia cohort and stratified by age, sex, and race/ethnicity. Standardized mortality ratios are the ratio of the observed number of deaths in the schizophrenia cohort to the number of deaths expected in the same cohort based on data from the general US population (January 1, 2001, to December 31, 2007, US Compressed Mortality File). A software program (SAS PROC STDRATE; SAS Institute Inc) was used to derive SMRs indirectly standardized by age, sex, race/ethnicity, and geographic region.
The schizophrenia cohort included 1 138 853 individuals, 4 807 121 years of follow-up, and 74 003 deaths, of which 65 553 had a known cause. The cohort sample sizes are stratified by state and payer type in eTable 3 in the Supplement, and the death counts are stratified by demographic group and specific mortality in eTable 4 in the Supplement.
The crude all-cause mortality rate for adults with schizophrenia was higher for men than for women and increased with age. The rate was higher for persons of white race/ethnicity than for other racial/ethnic groups. These results are detailed in eTable 5 in the Supplement.
Compared with the general population, the all-cause SMR for the schizophrenia cohort was significantly increased in the total sample and in each demographic group. The all-cause SMRs were higher for women than for men, for older adults than for younger or middle-aged adults, and for persons of white race/ethnicity than for the “other” racial/ethnic group, Hispanics, and blacks (Tables 1, 2, and 3). Standardized mortality ratios varied across the leading causes of death and age groups (Figure).
In the schizophrenia cohort, natural causes accounted for most of the known-cause deaths. Standardized mortality ratios from all natural causes of death were significantly elevated in the total schizophrenia cohort and in each demographic subgroup (Tables 1, 2, and 3).
Cardiovascular disease had the highest mortality rate of all disease groups examined, accounting for approximately one-third of all natural deaths. Approximately one-half of cardiovascular deaths were due to ischemic heart disease. The cardiovascular disease mortality rate was higher for men than for women, increased with age, and was highest for persons of white race/ethnicity and lowest for the “other” ethnic/racial group. Standardized mortality ratios for cardiovascular disease were significantly elevated in each demographic group, particularly among women (4.6; 95% CI, 4.5-4.7), young adults (4.5; 95% CI, 4.1-4.8), and individuals of white race/ethnicity (4.9; 95% CI, 4.8-5.0) (Tables 1, 2, and 3).
Cancer accounted for approximately 1 in 6 natural deaths. Lung cancer had the highest mortality rate of the 6 selected specific cancers. The lung cancer mortality rate was higher for men than for women, for persons of white race/ethnicity than for other racial/ethnic groups, and for older adults than for middle-aged or younger adults.
The SMR for cancer was significantly elevated in the total cohort but was only approximately half as large as the SMR for cardiovascular disease. The cancer SMR was significantly elevated among all demographic groups except young adults (Tables 1, 2, and 3). The SMR for lung cancer was considerably larger than SMRs for the other specific cancers (Table 1).
Among the other specific natural causes of death, COPD, diabetes mellitus, and influenza and pneumonia had the highest mortality rates (Table 1). For each of these diseases, SMRs were significantly increased overall and in each demographic subgroup. Particularly high SMRs were evident for COPD and influenza and pneumonia except among black adults with schizophrenia. The diabetes mellitus SMR was significantly higher among young adults than among middle-aged or older adults (Table 2).
Unnatural causes of death accounted for approximately 1 in 7 known-cause deaths. Mortality due to unnatural causes was higher for men than for women and for middle-aged adults than for younger or older adults. Accidents followed a similar pattern. Among accidental deaths, poisoning and nonpoisoning accounted for similar numbers of deaths, although SMRs for poisoning were significantly larger than those for nonpoisoning accidental deaths except among older adults.
Suicide accounted for approximately one-quarter of unnatural deaths. Among all causes of death, suicide was associated with the highest mean years of potential life lost per death. Suicide mortality was higher in men than in women, decreased with age, and was highest for persons of white race/ethnicity. Suicide SMRs were significantly elevated in all demographic groups. The homicide SMR was not significantly increased in the total schizophrenia cohort, although it was significantly increased among women, middle-aged and older adults, and persons of white race/ethnicity (Tables 1, 2, and 3).
Under the alternative Centers for Disease Control and Prevention classification, substance-induced deaths accounted for 8.2% of known-cause deaths and were most commonly nonsuicide deaths. Alcohol-induced and drug-induced deaths from drugs of abuse collectively accounted for most substance-induced deaths that were not classified as suicides. Standardized mortality ratios were significantly elevated for substance-induced suicide and nonsuicide deaths, firearm deaths, and deaths due to legal interventions (Table 4).
Nonelderly adults with schizophrenia in the Medicaid program die at approximately 3.5 times the rate of the general population. Their increased risk of mortality was distributed across several diseases but was particularly elevated for COPD, influenza and pneumonia, diabetes mellitus, cardiovascular disease, and suicide. In absolute terms, the leading identified causes of death were cardiovascular disease, cancer, and accidents. These patterns have implications for the medical care of patients with schizophrenia.
The 3.7 SMR for all-cause mortality was higher than the corresponding 2.98 SMR from a meta-analysis of 38 studies that collectively included 22 296 deaths.3 Our higher figure is consistent with a trend in the meta-analysis toward an increasing all-cause SMR in recent decades (the statistical test in the meta-analysis was significant at P = .03).3 A French study31 of 3470 patients with schizophrenia aged 18 to 64 years old, conducted between 1993 and 2005, reported all-cause SMRs of 3.6 for men and 4.3 for women, resembling the present findings.
Increased relative risk of cardiovascular mortality was observed for 3 age groups (20-34, 35-54, and 55-64), both sexes, and all 4 racial/ethnic groups. The number of age groups was reduced in the cause of death analysis to simplify the data presentation. Previous schizophrenia studies have reported significant, although smaller, increases in the relative risk for cardiovascular mortality for men,32 women,33 and younger adults.34 Incomplete follow-up, sampling from hospital discharges, and short follow-up periods may have depressed prior estimates. The relative risk of cardiovascular mortality was lower among black adults than among white or Hispanic adults in part because of the higher background cardiovascular mortality in the general black population.35 Excess cardiovascular mortality was evident even in young adults. These patterns highlight the importance of an early clinical focus on cardiovascular health in the management of schizophrenia. Addressing the disparity in cardiovascular death will likely require increased focus on primary prevention and on the identification and management of conditions contributing to cardiovascular mortality risk, including diabetes mellitus, hypertension, hyperlipidemia, and coronary artery disease. Because some antipsychotic medications are known to increase risk for cardiovascular disease and have been associated with increased risk of sudden cardiac death, myocarditis, and cardiomyopathy,36 long-term studies are needed to determine whether and to what extent antipsychotic treatment contributes to cardiovascular mortality in schizophrenia.
High mortality rates and SMRs were also observed for COPD, lung cancer, and influenza and pneumonia. Together with elevated cardiovascular death risk, these patterns strongly implicate smoking as a major risk factor for premature mortality in schizophrenia in the United States. Approximately two-thirds of adults with schizophrenia smoke,37 and they tend to smoke more heavily than smokers in the general population.38 The recent general decline in smoking in the United States did not extend to adults with serious psychiatric disorders.39 Despite professional guidelines calling for screening all psychiatric patients for tobacco use and providing evidence-based treatments to those interested in quitting,40 rates of screening and counseling for tobacco use and nicotine therapy remain low.41
Schizophrenia was also associated with increased risk of mortality from cancer of the lung, colon, breast, liver, pancreas, blood, and other organs. Greater excess mortality due to lung cancer than owing to other cancers has not previously been reported because of insufficient statistical power.31 Research on the underlying association between schizophrenia and cancer incidence has yielded mixed results, with reports of increased,42 equivalent,43 and decreased44 risks that may reflect differential mortality across age groups and competing risks of death from other causes. Because schizophrenia could influence not only cancer risk but also cancer detection and treatment, mortality data cannot confirm a causal association between schizophrenia and incident cancer. However, because patients with schizophrenia are at increased risk of mortality from a wide range of cancers, including especially lung cancer, physicians should aggressively seek to reduce modifiable risk factors and pursue early cancer detection in their patients with schizophrenia.
Consistent with most prior research,5,45 the relative risk of suicide was increased and the absolute risk greater among men than among women,46,47 among persons of white race/ethnicity than among those of nonwhite race/ethnicity,45,47 and among younger adults than older adults.46,47 Previous studies with younger samples47,48 and cohorts assembled following psychiatric hospital admission49 have reported suicide mortality rates considerably higher than those herein. Differences in suicide risk estimates are likely related to variations in the age and disease stage distributions of the various study populations.
Accidents accounted for more than twice as many deaths as suicide. Although some previous studies4,50 (but not another prior study51) have noted increased relative risk of accidental deaths in schizophrenia, these estimates have been based on small numbers. Little clinical or academic attention has been devoted to defining risks and prevention of medically serious accidents in schizophrenia. We found that increased accidental deaths were distributed across demographic groups. The relative risk of death from poisoning accidents tended to be larger than that from nonpoisoning accidental deaths, perhaps due to lower rates of driving and related motor vehicle crashes among adults with schizophrenia.52 Drug-induced deaths, whether accidental or intentional, were also a common source of mortality. In evaluating these patterns, it is important to recognize that some unnatural deaths of unintentional or indeterminate intent may be misclassified suicide deaths.53
A series of recent mass shootings in the United States has focused attention on mental illness and gun policy.54 In these discussions, little attention has been devoted to the extent to which adults with schizophrenia or other severe mental illnesses are casualties of firearm-related deaths. Suicide accounted for more than twice as many firearm deaths as homicides. Although adults with schizophrenia were not at increased risk of firearm-related homicide, an overall increased risk of homicide was apparent for women, for older adults, and for persons of white race/ethnicity and Hispanics with schizophrenia. Although adults with schizophrenia had a decreased overall risk of firearm-related homicide, they had more than a 3-fold increased risk for death related to legal interventions. These findings support efforts to improve mental health training of law enforcement professionals.55
This study has several limitations. First, we have no means of validating the accuracy of schizophrenia diagnoses in the Medicaid claims data, although we used a validated case identification algorithm.27 Second, different results may have been obtained if privately insured and uninsured individuals with schizophrenia were included in the analysis; however, approximately two-thirds of noninstitutionalized adults with schizophrenia in the United States are covered by Medicaid,25 and rates of all-cause hospitalization are similar in commercially insured and Medicaid-insured patients with schizophrenia.56 Third, information was not available concerning key health risk factors such as smoking status, body mass index, and substance abuse. Fourth, stigma may have contributed to underreporting of suicide and spuriously inflated accidental deaths,57 and the cause of approximately 11.4% of the deaths was unknown, undetermined, or unspecified. Fifth, mortality risk estimates are sensitive to the age composition of the study population and the length of follow-up. Because the SMR of adults with schizophrenia typically declines with older age as mortality rises in the general population,1 restricting the cohort to 64 years or younger likely increased the reported SMRs compared with what they would have been if elderly patients had been included. This conjecture may be especially true for the leading causes of death among older adults such as heart disease and cancer.58 However, because the SMR for suicide in schizophrenia may be highest among younger patients,59 restricting the cohort to adults with schizophrenia who were at least 20 years old may have lowered the SMR for suicide. Sixth, the data reflect mortality patterns from January 1, 2001, to December 31, 2007, and the patterns may have changed since that time. Seventh, Medicaid encounter data provide incomplete coverage of patients enrolled in some managed care plans.60
The results from this study confirm a marked excess of deaths in schizophrenia, particularly from cardiovascular and respiratory disease, that is evident in early adulthood and persists into later life. Especially high risks of mortality were observed from diseases for which tobacco use is a key risk factor. These findings support efforts to train mental health care professionals in tobacco use prevention and treatment and in implementation of policies that incentivize smoking control interventions in settings treating patients with schizophrenia.61 The excess of cardiovascular deaths further supports initiatives aimed at improving monitoring of blood pressure, serum glucose level, and serum lipid levels,62 as well as assertive medical and behavioral interventions, including more consistent management of diabetes mellitus, hypertension, obesity, dyslipidemia, and other cardiovascular risk factors. Finally, the large excess of deaths due to drugs of abuse highlights the critical importance of harm reduction interventions for adults with schizophrenia and comorbid substance abuse.
Submitted for Publication: May 26, 2015; final revision received July 28, 2015; accepted July 30, 2015.
Corresponding Author: Mark Olfson, MD, MPH, Department of Psychiatry and New York State Psychiatric Institute, College of Physicians and Surgeons, Columbia University, 1051 Riverside Dr, New York, NY 10032 (email@example.com).
Published Online: October 28, 2015. doi:10.1001/jamapsychiatry.2015.1737.
Author Contributions: Dr Huang had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Olfson, Gerhard, Crystal, Stroup.
Acquisition, analysis, or interpretation of data: Gerhard, Huang, Crystal, Stroup.
Drafting of the manuscript: Olfson, Gerhard, Crystal.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Gerhard, Huang, Crystal.
Obtained funding: Gerhard, Crystal, Stroup.
Conflict of Interest Disclosures: Dr Olfson reported being the principal investigator of a grant to Columbia University from Sunovion Pharmaceuticals. No other disclosures were reported.
Funding/Support: This research was supported by grant U19 HS021112 from the Agency for Healthcare Research and Quality and by the New York State Psychiatric Institute.
Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation or approval of the manuscript.