Schizophrenia-related suicide mortality in the United States by sex, 1999–2023

Abstract

Purpose

People with schizophrenia experience markedly elevated risk of suicide, yet national-level data on schizophrenia-related suicide mortality in the United States (US) remain limited. This serial cross-sectional, population-based study quantifies temporal trends in schizophrenia-related suicide deaths from 1999 through 2023 by sex.

Methods

We used final mortality data from the CDC WONDER Multiple Cause of Death database among US residents aged 15–74 years whose underlying cause of death was intentional self-harm or sequelae of intentional self-harm and whose death certificate listed schizophrenia, schizotypal disorder, or other delusional disorders as a contributing cause. Counts and mortality rates per 100,000 population were extracted and stratified by sex. Temporal trends were examined using joinpoint regression, with annual percent change (APC) estimated for each identified trend segment.

Results

From 1999 to 2023, there were 3507 schizophrenia-related suicide deaths, of which 76.7% were male and 23.3% were female. The overall mortality rate increased from 0.04 per 100,000 in 1999 to a peak of 0.09 per 100,000 in 2021, a 122% increase. Overall, joinpoint models showed a 2.4% annual increase from 1999 to 2016, followed by a 5.8% annual increase from 2016 to 2023. Among female decedents, rates rose from 1999 to 2012 (APC = 3.6%) and then stabilized. Among male decedents, rates increased gradually from 1999 to 2015 (APC = 2.4%) and accelerated thereafter (APC = 5.8%).

Conclusions

Schizophrenia-related suicide mortality increased substantially in the US, particularly among male individuals in recent years. These findings underscore the need for enhanced suicide prevention and treatment efforts targeting individuals with schizophrenia.

Introduction

Premature mortality among individuals with schizophrenia is markedly high and is approximately 3.5 times greater than that of the general population [32]. Suicide is among the leading causes of death in this population, contributing to this disparity [24], and results in greater potential years of life lost than any other cause of premature mortality in this population [21]. Lifetime suicide risk among individuals with schizophrenia is estimated to be 5–10 times that of the general population, with as many as one in twenty dying by suicide over the life course [17, 34, 35]. Prior epidemiologic studies from the United States and other Western settings have consistently documented elevated suicide mortality among people with schizophrenia, yet fewer studies have characterized long-term national mortality trends using multiple-cause-of-death data. This is important to examine, given that the burden of schizophrenia is predicted to modestly increase globally by 2030 [46].

Clinically, suicide risk among individuals with schizophrenia arises from a combination of illness-related, psychological, and treatment factors. Prominent contributors include depressive symptoms, command auditory hallucinations, high insight into illness (the “insight paradox”), and pervasive feelings of hopelessness or demoralization [4, 27, 45]. Periods of acute psychotic exacerbation, the early years following diagnosis, or the weeks immediately after psychiatric hospitalization are particularly high-risk periods [10, 33]. Medication nonadherence and comorbid substance use further increase vulnerability by worsening symptom control and social functioning [15, 29, 41]. Conversely, effective antipsychotic treatment may attenuate severe clinical outcomes and psychological distress; clozapine has recognized anti-suicidal effects in schizophrenia, and other atypical antipsychotics have been investigated for possible suicide risk-reducing effects [36].

Despite well-documented disparities in suicide risk among people with schizophrenia, limited national evidence exists on how schizophrenia-related suicide mortality has changed over time. This gap is concerning given that schizophrenia carries one of the highest suicide risks among psychiatric disorders, with up to half of affected individuals attempting suicide during their lifetime [16, 34]. Over the past two decades, overall suicide rates in the United States have risen substantially [14, 39, 40], coinciding with major shifts in psychiatric care delivery, psychotropic medication use, and socioeconomic conditions [38]. The COVID-19 pandemic intensified psychosocial stressors, social disconnection, and mental health vulnerability among already vulnerable populations, including individuals with serious mental illness [1]. The economic burden associated with schizophrenia in the United States doubled between 2013 and 2019, reaching an estimated $343.2 billion in 2019, with premature mortality representing a substantial share of this cost [19]. These trends raise critical questions about whether individuals with schizophrenia have experienced similar increases in suicide mortality or whether suicide mortality among decedents with schizophrenia-spectrum conditions has changed in parallel with broader suicide patterns or whether these trends differ in magnitude or timing.

Potential differences by sex represent an important but underexplored dimension of schizophrenia-related suicide mortality. In the general population, men are much more likely than women to die by suicide [14, 44], yet women with schizophrenia tend to exhibit higher rates of suicidal ideation and attempts [7, 43]. Biological, clinical, and social factors may contribute to these contrasting patterns. Men often experience earlier illness onset, more severe negative symptoms, and lower treatment adherence, which are factors associated with chronic functional impairment and heightened suicide risk [11, 12, 22]. In contrast, women with schizophrenia frequently have later onset, stronger social networks, and better treatment engagement, but may encounter elevated risk during hormonally sensitive periods such as postpartum and perimenopause [8, 11, 18, 22, 37]. These differences suggest that the mechanisms underlying suicide mortality in schizophrenia may operate differently by sex, with implications for both clinical management and suicide prevention strategies.

Despite clear evidence of sex differences in both schizophrenia and suicide [14, 44], few studies have examined long-term, sex-specific trends in schizophrenia-related suicide mortality at the population level. Most existing research has relied on cross-sectional or clinical samples, limiting generalizability and precluding robust temporal analyses. Given the persistent national rise in suicide rates [11, 14, 39, 40], there is an urgent need for population-based analyses that can clarify whether progress has been made in reducing suicide deaths over time among individuals with schizophrenia and whether these trends differ for male and female decedents. The objective of this descriptive epidemiologic study was to examine national trends in schizophrenia-related suicide mortality in the United States from 1999 through 2023, stratified by biological sex. We hypothesized that schizophrenia-related suicide mortality would increase over the study period, consistent with broader national suicide trends, and that increases would be more pronounced among male decedents than female decedents. By quantifying changes over time using national mortality data, this study aims to provide timely evidence to inform prevention and intervention strategies tailored to individuals living with schizophrenia.

Methods

We obtained final death certificate data from the CDC Wide-ranging ONline Data for Epidemiologic Research (WONDER) multiple causes of death database for male and female deaths occurring in 1999 through 2023. This database is maintained by the National Center for Health Statistics (NCHS) and is based on information abstracted from death certificates filed in all 50 states and the District of Columbia. Coverage is near complete, with more than 99% of deaths in the United States recorded in the NCHS vital statistics system each year. For each decedent, the certifier records the sequence of conditions leading to death [31] and any other significant conditions that contributed to death [30], NCHS mortality coders and nosologists then assign ICD-10 codes according to standardized World Health Organization and NCHS rules Thus, death certificates include the underlying cause of death (i.e., the disease or injury initiating the chain of events leading directly to death) as well as up to 20 multiple (contributing) causes of death, coded according to the International Classification of Diseases, Tenth Revision (ICD-10).

We obtained mortality data for U.S. residents aged 15–74 years whose deaths occurred between 1999 and 2023. We restricted analyses to this age range to minimize potential misclassification among pediatric and older adult deaths, for whom schizophrenia-related diagnostic coding is less reliable. Schizophrenia-spectrum diagnoses are rare and diagnostically complex in children and younger adolescents, where hallucinations and psychotic-like experiences may occur in other psychiatric or developmental conditions [9]. In older adults, new or documented psychosis may be difficult to distinguish from dementia, delirium, neurocognitive disorders, medical illness, or medication-related causes [42]. Decedents were included if the underlying cause of death was intentional self-harm (ICD-10: X60–X84) or sequelae of intentional self-harm (ICD-10: Y87.0), and any multiple (contributing) cause of death included schizophrenia, schizotypal disorder, or other delusional disorders (ICD-10: F20–F29). Deaths meeting these criteria were classified as schizophrenia-related suicide deaths. The F20–F29 grouping captures schizophrenia-spectrum and related psychotic disorders as recorded on the death certificate; it does not capture suicides among people with schizophrenia unless an F20–F29 code was listed as a contributing cause. We extracted both annual death counts and corresponding mortality rates per 100,000 population directly from CDC WONDER. Rates were based on annual population estimates from the U.S. Census Bureau and were stratified by biological sex (male, female) as well as combined totals across sexes. All rates were crude rather than age-adjusted, consistent with CDC WONDER default reporting, allowing direct comparability with other descriptive trend analyses using the same data source.

We first described the annual number and rate of schizophrenia-related suicide deaths from 1999 through 2023, overall and by sex. To quantify temporal trends, we employed Joinpoint regression analysis, a statistical method that identifies significant inflection points (“joinpoints”) in long-term trends. This approach is well-suited for descriptive surveillance because it identifies statistically significant changes in slope rather than imposing a single linear trend across the full study period [20]. Analyses were conducted using the Joinpoint Regression Program, version 5.4.0.0 (National Cancer Institute, Bethesda, MD). Models were fitted on a log-linear scale, beginning with zero joinpoints (a single straight line) and iteratively testing up to four joinpoints, a conventional upper bound for a 25 year annual time series that permits detection of meaningful inflection points while avoiding overfitting. The optimal model was determined using permutation tests and model fit statistics. For each segment defined by joinpoints, we estimated the annual percent change (APC) in mortality rates, along with 95% confidence intervals (CIs). APCs were considered statistically significant at a two-sided p-value < 0.05. We conducted all analyses separately for male and female decedents, as well as for combined totals, to evaluate potential sex-specific differences in temporal trends.

All analyses were based on publicly available, deidentified, and aggregated data and therefore were determined to be exempt from review by the University of Illinois Urbana-Champaign Institutional Review Board. Because data were aggregated at the population level, informed consent was not required.

Results

From 1999 to 2023, there were a total of 3507 schizophrenia-related suicide deaths among U.S. residents aged 15–74 years, of which 76.7% occurred among male decedents and 23.3% among female decedents (Table 1). Across the 25 year study period, the number of schizophrenia-related suicide deaths generally increased over time, peaking in 2021 (n = 227). The greatest number of deaths among male decedents also occurred in 2021 (n = 191), whereas among female decedents, the number peaked slightly later, in 2023 (n = 51). Throughout the entire period, annual counts of schizophrenia-related suicide deaths were consistently higher among male decedents than among female decedents.

Table 1.

Number and rate of schizophrenia-related suicide deaths by sex and year, United States, 1999–2023

Year	Male		Female		Combined
	Number of deaths	Mortality rate per 100,000	Number of deaths	Mortality rate per 100,000	Number of deaths	Mortality rate per 100,000
1999	62	0.062	21	0.020	83	0.041
2000	74	0.073	25	0.024	99	0.048
2001	61	0.059	19	0.018	80	0.039
2002	72	0.069	19	0.018	91	0.043
2003	86	0.082	19	0.018	105	0.050
2004	90	0.085	27	0.025	117	0.055
2005	71	0.066	33	0.030	104	0.048
2006	106	0.097	32	0.029	138	0.063
2007	87	0.079	29	0.026	116	0.052
2008	90	0.081	31	0.027	121	0.054
2009	98	0.087	30	0.026	128	0.056
2010	91	0.080	29	0.025	120	0.052
2011	98	0.086	35	0.030	133	0.057
2012	99	0.086	38	0.032	137	0.059
2013	88	0.075	45	0.038	133	0.056
2014	119	0.101	32	0.027	151	0.063
2015	100	0.084	45	0.037	145	0.060
2016	117	0.098	36	0.030	153	0.063
2017	137	0.114	30	0.024	167	0.069
2018	139	0.115	36	0.029	175	0.072
2019	133	0.110	45	0.036	178	0.073
2020	143	0.117	40	0.032	183	0.074
2021	191	0.154	36	0.029	227	0.091
2022	170	0.136	35	0.028	205	0.082
2023	167	0.134	51	0.040	218	0.087

The overall schizophrenia-related suicide mortality rate was highest in 2021, at 0.09 deaths per 100,000 population, representing a 122% relative increase from 1999 (0.04 deaths per 100,000 population). This pattern closely paralleled the temporal pattern in death counts. In line with these counts, the mortality rate was consistently higher among male decedents than among female decedents across all years of observation (Fig. 1).

Fig. 1.

Rate of schizophrenia-related suicide deaths by sex and year, United States, 1999–2023

Trends in schizophrenia-related suicide mortality also differed meaningfully by sex. Results from the joinpoint regression models showed that, among male and female decedents combined, the overall rate of schizophrenia-related suicide mortality increased by an average of 2.4% per year from 1999 to 2016 and accelerated to a 5.8% annual increase from 2016 to 2023 (Table 2). When examined separately, divergent temporal patterns emerged between male and female decedents. Among female decedents, the schizophrenia-related suicide mortality rate increased modestly by 3.6% per year from 1999 to 2012, after which the trend leveled off and remained relatively stable through 2023. In contrast, among male decedents, the mortality rate rose by 2.4% per year from 1999 to 2015 and then accelerated markedly, increasing by 5.8% annually from 2015 through 2023. Overall, these findings indicate that the rise in schizophrenia-related suicide mortality observed in recent years was driven primarily by increases among male decedents, whereas rates among female decedents plateaued after the early 2010s.

Table 2.

Joinpoint regression results for trends in the rate of schizophrenia-related suicide deaths by sex and year, United States, 1999–2023

Years	APC (95% CI)
Male
1999–2015	2.0** (0.8, 3.3)
2015–2023	5.8** (2.1, 9.5)
Female
1999–2012	3.6** (1.1, 6.2)
2012–2023	0.4 (− 2.8, 3.6)
Combined
1999–2016	2.4*** (1.5, 3.2)
2016–2023	4.7** (1.4, 8.1)

APC, Annual percent change; CI, Confidence interval; ***p < 0.001; **p < 0.01; bold values are statistically significant

Discussion

The main finding of this serial cross-sectional, population-based trend analysis is that schizophrenia-related suicide mortality increased in the United States from 1999 through 2023, with the steepest recent increases observed among male decedents. During this 25 year period, a total of 3507 schizophrenia-related suicide deaths occurred among individuals aged 15–74 years, with male decedents accounting for nearly three-quarters of all deaths. The overall male-to-female ratio of approximately 3.3–1 underscores a persistent sex disparity in suicide mortality among individuals with schizophrenia. Our findings reveal a concerning upward trajectory in schizophrenia-related suicide mortality over time, particularly among male decedents.

Although rates increased for both male and female decedents, the acceleration in male mortality beginning around 2015 suggests that suicide risk in this population has intensified in recent years. This pattern occurred during a period in which overall U.S. suicide mortality also increased substantially, particularly among male individuals. [14]. These findings should be interpreted as describing schizophrenia-related suicide deaths identified through death certificate coding, rather than all suicides among people living with schizophrenia. The temporal pattern is broadly consistent with national increases in overall U.S. suicide mortality during much of the same period [14], but the present study was not designed to determine whether the observed pattern is unique to schizophrenia-spectrum conditions. Similar temporal patterns may be present among people with other psychiatric disorders, including mood, substance use, or other severe mental illnesses. Therefore, the results are best interpreted as evidence that schizophrenia-related suicide mortality has not declined in parallel with prevention efforts, and that comparator analyses across psychiatric diagnostic groups are needed to clarify whether schizophrenia-related trends differ in timing, magnitude, or underlying drivers.

In contrast, mortality among female decedents increased modestly through 2012 and then plateaued, suggesting that suicide risk in this subgroup may have stabilized. These diverging trends by sex may reflect underlying differences in the clinical course of schizophrenia, patterns of care engagement, and social determinants of health. For example, male individuals with schizophrenia typically experience earlier onset, greater functional impairment, and lower treatment adherence [11, 12], all of which can compound suicide risk [5]. Moreover, they are also more likely to experience unemployment and homelessness [13, 23, 25], circumstances that exacerbate both psychiatric symptoms and suicide risk [2, 26]. In addition, men tend to use more lethal methods and are less likely to seek help for suicidal ideation, which may partially explain their higher suicide mortality despite comparable or even lower rates of suicidal thoughts relative to women [6, 14, 28]. Female individuals, conversely, often have later onset, stronger social support networks, and greater treatment engagement, though their risk may fluctuate during hormonally sensitive periods such as the postpartum or perimenopausal years [8, 11, 12]. These distinct trajectories underscore the need for sex-specific prevention and treatment strategies that address both biological and contextual risk factors.

The peak in schizophrenia-related suicide mortality in 2021 coincided with the COVID-19 pandemic, which profoundly disrupted mental health care and social support systems worldwide. People with schizophrenia were disproportionately affected by the pandemic, owing to increased social isolation, reduced access to in-person psychiatric care, and higher rates of medical comorbidity [3]. These pandemic-related disruptions may have intensified depressive symptoms, loneliness, psychological distress, and barriers to treatment continuity among individuals already vulnerable to suicide. These disruptions may have exacerbated preexisting vulnerabilities, particularly among male decedents, who already face greater barriers to consistent care and social connection. The continued increase in male mortality rates through 2023 suggests that the downstream mental health effects of the pandemic may persist well beyond its acute phase.

In contrast, schizophrenia-related suicide mortality among female decedents did not show the same pronounced pandemic-era increase. Several explanations are possible. Women with schizophrenia may have had comparatively stronger social support networks, greater treatment engagement, or more frequent help-seeking, which could have buffered some effects of pandemic-related disruption. Involvement in caregiving or household roles may also have preserved interpersonal contact for some women during periods of social restriction. However, these interpretations should be considered cautiously because annual counts among female decedents were small, and death certificate data do not include information on social support, caregiving roles, treatment continuity, or pandemic-related stressors. Future research should examine whether sex differences in social connection, care engagement, and telehealth access contributed to these divergent patterns.

Taken together, these findings indicate that schizophrenia-related suicide mortality in the U.S. has not declined despite national suicide prevention initiatives, highlighting a continued gap in effective risk reduction for this high-need population. Prevention strategies may require more targeted outreach to individuals with severe mental illness, enhanced continuity of care, and stronger integration between psychiatric and primary care services. Sex-specific interventions may be particularly critical, addressing male underutilization of care and tailoring support to female patients’ unique psychosocial and hormonal risk factors.

Limitations

Several limitations should be considered when interpreting these findings. This analysis relied on death certificate data, which may underestimate the true burden of schizophrenia-related suicide due to potential misclassification of mental health diagnoses or manner of death. Although U.S. vital statistics data provide near-complete death registration, this coverage should not be interpreted as evidence of complete or perfectly accurate ascertainment of psychiatric diagnoses listed as contributing causes. Diagnostic accuracy may vary depending on available medical records, clinical documentation, certifier knowledge of psychiatric history, and local death investigation practices. Psychiatric conditions may be less consistently recorded on death certificates than acute medical causes or injuries, particularly when they are not viewed as directly contributing to the fatal event. Under-ascertainment may also reflect administrative burden, incomplete clinical documentation, limited access to psychiatric history at the time of death certification, and stigma surrounding severe mental illness.

Additionally, the use of ICD-10 codes does not allow for differentiation among schizophrenia subtypes or comorbid psychiatric conditions, which could influence risk. The F20–F29 category also includes a heterogeneous group of schizophrenia-spectrum and related psychotic disorders that may differ in chronicity, etiology, clinical course, and suicide prevention needs. For example, some psychotic states may be more transient or substance-related, whereas others reflect chronic schizophrenia-spectrum illness. Because annual counts were small, especially among female decedents, subgroup analyses by individual F20–F29 diagnoses were not conducted. Moreover, suicide deaths in which schizophrenia, schizotypal disorder, or other delusional disorder was not listed as a contributing factor on the death certificate were not included. Therefore, the 3507 deaths identified in this analysis represent a subset of suicide deaths among people with schizophrenia-spectrum disorders, rather than the full burden of suicide mortality in this population. Given that schizophrenia is highly stigmatized, the true public health burden of schizophrenia-related suicide mortality is likely to be underestimated. Finally, because schizophrenia-related suicide deaths are relatively rare events, annual counts, particularly among female decedents, were small, limiting precision in trend estimates.

Conclusions and future directions

Overall, the findings from the current study demonstrate that schizophrenia-related suicide mortality in the United States has increased over the past 25 years, with recent increases driven primarily by male decedents. These findings underscore the need to strengthen suicide prevention for individuals with schizophrenia-spectrum and related psychotic disorders, particularly through sustained risk assessment, improved continuity of psychiatric care after hospitalization, support for medication adherence, integrated treatment for comorbid substance use, and outreach strategies that address male underutilization of mental health services. Expanding access to long-acting injectable antipsychotics may also reduce relapse and associated suicide risk. Given the elevated mortality observed among men, gender-responsive approaches, such as outreach embedded in community settings, peer support models, and interventions that address stigma and barriers to help-seeking, are particularly warranted. Although these findings are generalizable to U.S. death certificate-coded schizophrenia-related suicide mortality among individuals aged 15–74 years, they should not be interpreted as capturing all suicides among people living with schizophrenia.

Future studies should compare mortality trends across psychiatric diagnostic groups, including schizophrenia-spectrum disorders, mood disorders, and substance use disorders, to determine whether the observed trends are specific to schizophrenia-related suicide mortality or reflect broader patterns among people with severe mental illness. Analyses that disaggregate F20–F29 diagnoses may also help determine whether temporal trends differ across chronic schizophrenia, delusional disorders, and other psychotic conditions. Longitudinal studies linking mortality data with clinical, medication, and service utilization records could help clarify the role of treatment adherence, comorbid substance use, and psychiatric care engagement in suicide risk. Additionally, qualitative and mixed-methods research is needed to understand how gendered experiences, such as stigma, social isolation, and help-seeking behavior, shape suicide trajectories among people with schizophrenia. Reducing schizophrenia-related suicide mortality will require rigorous evaluation of emerging interventions and sustained investment in accessible, coordinated, and equitable mental health care.

Author contributions

R.H. wrote the main manuscript text with sections contributed by T.H., S.O., M.A.I.A., A.N., T.L., and R.Y. RH. conducted the analysis and created the tables/figure. All authors reviewed the manuscript.

Funding

This research was supported by the Scott Edward Erickson Endowed Research Award of the University of Illinois Urbana-Champaign.

Data availability

All data are publicly available via CDC WONDER at https://wonder.cdc.gov/

Declarations

Ethics approval

Not applicable.

Consent to participate

Not applicable (not human subjects research).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.