Abstract
Introduction
Metabolic comorbidities, including central adiposity, dyslipidemia, insulin resistance, and hypertension, are common in individuals with schizophrenia and contribute to an increased risk of cardiovascular disease and related mortality. There is limited evidence quantifying the economic burden of metabolic comorbidities in schizophrenia. This study aimed to assess the impact of cumulative metabolic comorbidity burden on medical costs in individuals with schizophrenia using claims data.
Methods
A retrospective cohort study was conducted using the STATinMED Real World Data Insights database covering the period 2018–2024 in the United States (US). Adults with schizophrenia were grouped by number of metabolic comorbidities (obesity, hyperlipidemia, hypercholesterolemia, diabetes, and/or hypertension). Propensity score matching was applied to balance selected baseline characteristics in individuals with and without metabolic comorbidities. The primary outcome was all-cause medical costs, during a 12-month follow-up, adjusted for insurance status and inflation.
Results
Overall, 122,248 individuals were eligible. After matching, the numbers for metabolic comorbidities and individuals in each group were: 0, n = 40,552; 1, n = 15,840; 2, n = 11,833; 3, n = 8119; 4, n = 4074; and 5, n = 686. Estimated total medical costs increased with the number of metabolic comorbidities; individuals with five metabolic comorbidities incurred mean costs of US$34,441 per person per year, over 4 times higher than the $8396 cost for individuals with none. A similar pattern was observed for estimated outpatient, inpatient, and stay-related costs, with 3.5-, 2.3- and 1.7-fold increases, respectively.
Conclusion
There is a substantial economic burden associated with cumulative metabolic comorbidities in individuals with schizophrenia. Prevention and management of metabolic comorbidities in this population includes early risk assessments, lifestyle modification, and the selection of antipsychotic medications with few metabolic adverse effects. Addressing metabolic comorbidities should be a key component of multidisciplinary care to reduce health-related and economic impacts in schizophrenia.
Supplementary Information
The online version contains supplementary material available at 10.1007/s40120-026-00973-5.
Keywords: Metabolic comorbidities, Schizophrenia, Medical costs, Claims data, United States
Plain Language Summary
Metabolic disorders include obesity, high fat levels in blood, high blood cholesterol, diabetes, and high blood pressure. When several occur together, this is known as metabolic syndrome. People with metabolic syndrome are at increased risk of heart disease and stroke. In the United States, about a third of people with schizophrenia (which is a serious mental health condition) have three or more metabolic disorders. Many available treatments for schizophrenia can increase the risk of developing metabolic disorders. This study compared medical costs in adults with schizophrenia who had at least one metabolic disorder with those who did not. The study used a large set of healthcare claims data from the United States, covering the years from 2018 to 2024. The results showed that the more metabolic disorders a person had, the more costly their healthcare. People with five metabolic disorders had medical costs that were about $34,000 a year in total, which is over four times more than people with schizophrenia without metabolic disorders. The costs for clinic visits and hospital stays increased as the number of metabolic disorders increased. Treating metabolic disorders and using medications for schizophrenia with fewer metabolic side effects, like weight gain, could help people with schizophrenia stay healthier and reduce the overall costs of medical care.
Supplementary Information
The online version contains supplementary material available at 10.1007/s40120-026-00973-5.
Key Summary Points
| Why carry out this study? |
| Metabolic comorbidities include central adiposity, atherogenic dyslipidemia, insulin resistance, and hypertension. |
| About one-third of individuals with schizophrenia in the United States (US) have three or more metabolic comorbidities, which contribute to increased risks of cardiovascular death in this population. |
| What was learned from the study? |
| Analysis of claims data in the US revealed that individuals with schizophrenia and five metabolic comorbidities (obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension) incurred total medical costs of US$34,441 per person per year, over four times higher than individuals with schizophrenia without such comorbidities. |
| Multidisciplinary care of patients with schizophrenia should include early risk assessment and management of metabolic comorbidities to reduce their health-related and economic impacts in schizophrenia. |
Introduction
Schizophrenia is a severely disabling mental health condition that typically manifests behaviorally in late adolescence or early adulthood and is estimated to affect approximately 23 million people worldwide [1]. Individuals with schizophrenia are at increased risk of adverse health outcomes, including cardiovascular disease (CVD) and stroke [2]. Life expectancy for people with schizophrenia was reported to be about 15 years less than for the general population in 2017 [3]. CVD is a major cause of death in people with schizophrenia, accounting for approximately one-quarter of deaths in men and one-third of deaths in women in high-income countries [4]. Although global mortality due to CVD has declined in the past 50 years because of risk factor reduction and improved medical care, people with schizophrenia have continued to experience higher mortality following diagnosis of CVD than those without schizophrenia [4].
Central adiposity, insulin resistance, elevated blood pressure, elevated triglycerides, and reduced high-density lipoprotein cholesterol (HDL-C) are risk factors for CVD [5]. The presence of three or more of these metabolic comorbidities is commonly defined as metabolic syndrome (MetS) [6, 7], which has an estimated prevalence of 32.5% in individuals with schizophrenia in the United States (US) [8]. However, the prevalence in individuals with schizophrenia varies depending on population demographics and the definition of MetS [9], and can be as high as 68% in patients treated in a rehabilitation setting [10].
In individuals with schizophrenia, there is evidence that treatment with antipsychotic medications increases the risks of MetS [9, 11]. In individuals with schizophrenia, there is evidence that treatment with antipsychotic medications increases the risks of MetS [9, 11]. Although it is unclear whether duration of treatment with antipsychotic medications is associated with metabolic comorbidities, published data suggest that MetS is closely associated with the use of antipsychotic medications that are currently available for the treatment of schizophrenia, particularly second-generation antipsychotics [9, 11]. There is also evidence suggesting a link between psychosis and metabolic dysregulation independent of antipsychotic use [11]. Abnormal glucose metabolism and insulin resistance are often already present in individuals who are experiencing their first episode of psychosis, even before the introduction of antipsychotics [12]. In addition, several behavioral risk factors, including physical inactivity, smoking, and high alcohol use, are associated with an increased risk for both metabolic comorbidities and mental health disorders [11, 13].
In the general US population, total healthcare costs have been estimated to be 1.6 times greater in individuals with MetS than in those without, with costs increasing with each additional risk factor for MetS [14, 15]. Data on healthcare costs associated with metabolic comorbidities in individuals with schizophrenia are lacking; however, evidence from a systematic literature review in 2022 suggests that comorbidities, including metabolic disorders, are key drivers of increased healthcare costs in individuals with mental health conditions, including schizophrenia [16].
Given the high prevalence and risks of metabolic comorbidities in people with schizophrenia, understanding the economic impact of cumulative metabolic comorbidity burden in this population would help to address a key evidence gap for the burden of illness. [6, 7, 9]
In this study using claims data, we aimed to investigate the impact of cumulative burden of metabolic comorbidities—comprising obesity, hypertension, hyperlipidemia, hypercholesterolemia, and diabetes (glucose intolerance)—on medical costs for individuals with schizophrenia.
Methods
Study Design
This was a retrospective cohort study using secondary data from individuals with schizophrenia included in the US STATinMED RWD (real-world data) Insights database. The full study period was from January 2018 to December 2024. Individuals with schizophrenia were identified from January 2019 to December 2022 (Fig. 1). This study is an analysis of deidentified claims data, and therefore no ethical approval was required.
Fig. 1.
Study design, periods, and timelines. aSchizophrenia was defined based on ≥ 2 outpatient claims on separate days or ≥ 1 inpatient claim related to schizophrenia; at least one of these claims was required to occur in the 12 months before the index date. bMetabolic comorbidities comprise obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension. cAll individuals (in case and control groups) must have ≥ 12 months of continuous enrollment leading up to and following the index date
The definitions of metabolic comorbidities in recent clinical trials investigating the effects of antipsychotics in patients with schizophrenia and concurrent MetS have been based on the National Cholesterol Education Program (NCEP) Adult Treatment Panel 3 (ATP3) criteria (NCEP-ATP3) [6, 17–19], as follows: a waist circumference of ≥ 40 inches (101.6 cm) in men and ≥ 35 inches (88.9 cm) in women; serum triglycerides level of ≥ 150 mg/dL or pharmacological treatment for elevated triglycerides; reduced HDL-C of < 40 mg/dL in men or < 50 mg/dL in women; elevated fasting glucose of ≥ 100 mg/dL or pharmacological treatment for elevated glucose; and elevated blood pressure (BP) of systolic ≥ 130 mmHg or diastolic ≥ 85 mmHg, or pharmacological treatment for hypertension [6]. For a MetS diagnosis, NCEP-ATP3 requires the presence of at least 3 out of the 5 metabolic comorbidities.
In this claims study, increased waist circumference, triglyceridemia, elevated fasting glucose, and elevated BP were mapped to diagnoses of obesity, hyperlipidemia, diabetes, and hypertension, respectively. The diagnosis of hypercholesterolemia was used as the proxy for low HDL because both are known to be associated with increased CVD mortality [20].
Individuals with schizophrenia and one or more metabolic comorbidity (obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension) comprised the case group (Fig. 1). The index date was the earliest date of claim for a metabolic comorbidity in the identification period (January 2019–December 2022). In individuals with more than one metabolic comorbidity, the index date was the earliest date of the metabolic comorbidity that occurred most recently. The control group comprised individuals with schizophrenia and without any metabolic comorbidity (Fig. 1). The index date for controls in the identification period was randomly assigned, based on differences in the distributions of index date of cases and the eligibility date of controls in the matched cohorts, using an automated method described in Harvey et al. 2013 [21].
For cases and controls, there was a pre-index baseline period of 6 months, and a 12-month follow-up period (Fig. 1).
Data Source
The STATinMED RWD Insights (https://statinmed.com/data/) is a large-scale statewide database that systematically aggregates medical and pharmacy claims from a variety of payer sources, including Medicaid. This database covers approximately 80% of the US healthcare system, encompassing numerous payers sourced directly from claims clearinghouses, which are responsible for managing claims transactions between payers and providers across the US. At the time of the study, STATinMED contained data for > 1.8 million individuals with schizophrenia. Data were obtained from the STATinMED RWD Insights database under a valid and active license agreement, and analyses and publication of findings were conducted in accordance with the terms of such agreement.
Study Population
Individuals were included if they were aged ≥ 18 years, had ≥ 12 months of continuous enrollment immediately before and after the index date, and ≥ 2 outpatient claims on separate days or ≥ 1 inpatient claim related to schizophrenia, based on the International Classification of Disease Tenth Revision-Clinical Modification (ICD-10-CM) codes during the identification period. At least one of these claims was required to have occurred in the 12 months before the index date. Individuals with any diagnosis of bipolar disorder, major depressive disorder, or schizoaffective disorder during the entire study period (January 2018–December 2024) were excluded from the analysis.
Included individuals were either part of the case or the control group. The case group comprised individuals with schizophrenia and one or more of the following metabolic comorbidities: obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension. ICD-10-CM codes were employed to identify diagnoses of schizophrenia (Supplemental Table S1) and metabolic comorbidities (Supplemental Table S2). The control group comprised individuals with schizophrenia and no ICD-10-CM codes for any of the five metabolic comorbidities.
The cumulative metabolic comorbidity burden (proxy for MetS) was the primary independent variable (exposure) of interest. In order to ensure baseline covariate balance between the case and control groups, the models adjusted for baseline characteristics of demographics (age and sex), other mental health disorders [anxiety disorder, substance use disorder, post-traumatic stress disorder, and personality disorder (see Supplemental Table S1 for a list of ICD-10-CM codes)], and geographic region in the US, using 1:1 propensity score matching. Other baseline characteristics were the use of classes of psychotropic (anti-anxiety, antidepressant, antipsychotics, hypnotics, and mood stabilizers) and nonpsychotropic medications (antihypertensive, antidiabetics, glucagon-like peptide-1 receptor agonists, and lipid-lowering agents). Supplemental Table S3 lists the psychotropic and nonpsychotropic medications that were grouped by class for the purpose of the study; however, individual medications may have been prescribed for reasons other than the primary indication(s) for the drug class.
Outcomes
The primary outcome was total all-cause medical costs, which comprised outpatient, inpatient (hospital), emergency room, stay-related (nursing home, group home, and hospice), and other unspecified costs in the 12-month follow-up period (this analysis did not include outpatient pharmacy costs), stratified by the number of metabolic comorbidities (0 in controls; 1–5 in cases). These costs were adjusted for insurance status in the analysis and according to the consumer price index for 2024 and are reported rounded to the nearest US dollar.
Statistical Analysis
The nearest-neighbor matching approach implemented in the MatchIt Package [22] from R (The R Foundation, Vienna, Austria) was employed for propensity score matching of case and control groups. Balance between groups was assessed using standardized mean differences (SMD); values < 0.20 were considered acceptable [23]. Descriptive statistics including mean [standard deviation (SD)] or median [interquartile range (IQR)] were reported for continuous variables, and counts and proportions were reported for categorical variables.
A generalized linear model (GLM) was employed to assess annualized healthcare cost per individual per year stratified by the number of metabolic comorbidities with adjustment for insurance status. Estimated marginal mean costs were reported from the GLMs for each type of medical cost (outpatient, inpatient, emergency, and stay-related) and total medical cost. The inpatient total cost analysis was restricted to a complete-case cohort of verified, adjudicated claims [24]. This approach avoids the downward bias inherent in including capitated or unpriced encounters, which represent a different payer-contracting mechanism rather than a reduction in clinical intensity. Additionally, a sensitivity analysis was conducted using a two-part hurdle model that separately evaluated the probability of incurring any medical cost via logistic regression, and the cost intensity among utilizers via a Gamma GLM, thereby calculating the unconditional expected cost across the entire study population, including those with zero or missing cost records.
In each GLM cost model, a log link with gamma distribution was used to account for skewness and heteroscedasticity of the data. The model only included nonzero cost observations. Therefore, the model-based estimates reflected the mean costs in individuals with schizophrenia, stratified by the number of metabolic comorbidities, who incurred any medical cost rather than the overall mean cost, which included zero or missing values. The GLM analysis was conducted using SAS 9.4 (SAS Institute) and R statistical software version 4.4.0 [25]. Differences in cost variables stratified by number of metabolic comorbidities were assessed using unpaired t test (between-group mean) or Wilcoxon test (between-group median) depending on the Shapiro–Wilk normality test for continuous variables. A P value ≤ 0.008 is considered significant after Bonferroni correction for multiple pairwise comparisons.
Results
Study Population and Baseline Characteristics
Overall, 122,248 individuals were eligible before being matched, of whom 71,570 had at least one metabolic comorbidity (case group) and 50,678 had no metabolic comorbidity (control group) (Table 1). Demographic and clinical characteristics of all participants in the case and control groups, focusing on the characteristics used for matching (age, sex, other mental health disorders, and geographic region), are shown in Tables 2, 3 shows these characteristics after matching. Prior to matching, the individuals in the case group were older than those in the control group (mean age: 57 years vs. 39 years); other differences were less marked. Detailed characteristics of the unmatched case and control groups are listed in Table S4.
Table 1.
Number of individuals by eligibility in the case and control groups
| Case group | N |
|---|---|
| Presence of ≥ 1 metabolic comorbiditya (obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension) | 24,191,864 |
| ≥ 12 months of continuous enrollment immediately before and after the index date | 21,793,372 |
| ≥ 2 outpatient claims on separate days or ≥ 1 inpatient claim for schizophrenia, with at least one claims in the year prior to the index date | 372,919 |
| ≥ 18 years at index date | 371,354 |
| No evidence of bipolar disorder, major depressive disorder, and schizoaffective disorder based on ICD-10-CM codesb | 71,570 |
| Control group | N |
| ≥ 2 outpatient claims on separate days or ≥ 1 inpatient claim for schizophrenia, with at least one of the claims in the year prior to the index date | 821,454 |
| ≥ 12 months of continuous enrollment immediately before and after the index date | 750,346 |
| ≥ 18 years at index date | 741,543 |
| Without any metabolic comorbiditya (obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension) | 169,439 |
| No evidence of bipolar disorder, major depressive disorder, and schizoaffective disorder based on ICD-10-CM codesb | 50,678 |
ICD-10-CM International Classification of Disease, Tenth Revision-Clinical Modification
aSee Supplemental Table S1
bBipolar disorder, F31X; depressive disorder, F32X/F33X; schizoaffective disorder, F25X
Table 2.
Demographic and clinical characteristics employed for propensity score matching in unmatched cases and controls
| Characteristics (unmatched groups) | Case group (N = 71,570) | Control group (N = 50,678) |
|---|---|---|
| Age at index date, median (IQR), years | 57 (44–65) | 39 (30–54) |
| Sex, n (%) | ||
| Female | 24,075 (34) | 13,153 (26) |
| Male | 47,495 (66) | 37,525 (74) |
| US region, n (%) | ||
| Midwest | 12,758 (18) | 10,507 (21) |
| Northeast | 15,549 (22) | 9331 (18) |
| South | 23,795 (33) | 16,198 (32) |
| West | 19,468 (27) | 14,642 (29) |
| Mental health disordersa n (%) | ||
| Anxiety disorder | 8900 (12) | 5319 (10) |
| Substance use disorder | 7962 (11) | 7435 (15) |
| PTSD | 1026 (1) | 1024 (2) |
| Personality disorder | 1004 (1) | 781 (2) |
ICD-10-CM International Classification of Disease Tenth Revision-Clinical Modification, IQR interquartile range, PTSD post-traumatic stress disorder, US United States
aBased on ICD-10-CM codes (bipolar disorder: F31X; depressive disorder; F32X/F33X; schizoaffective disorder: F25X)
Table 3.
Demographic and clinical characteristics for the matched case and control groups, with SMDs of matched covariates after propensity score matching
| Characteristics (matched groups) | Cases (N = 40,552) | Controls (N = 40,552) | SMD for matched covariates |
|---|---|---|---|
| Age at index date, median (IQR), years | 47 (37–60) | 44 (35–57) | 0.15a |
| Sex, n (%) | |||
| Female | 12,727 (31) | 11,335 (28) | – |
| Male | 27,825 (69) | 29,217 (72) | 0.07 |
| US region, n (%) | |||
| Midwest | 7540 (19) | 7735 (19) | 0.01 |
| Northeast | 8951 (22) | 7941 (20) | 0.06 |
| South | 14,087 (3) | 12,932 (32) | 0.06 |
| West | 9974 (25) | 11,944 (29) | 0.11 |
| Insurance status, n (%) | |||
| Commercial only | 5874 (14) | 7417 (18) | – |
| Medicaid only | 13,604 (34) | 15,967 (39) | – |
| Medicare only | 8842 (22) | 8888 (22) | – |
| Dual type | 11,176 (28) | 7115 (18) | – |
| Other | 1056 (3) | 1165 (3) | – |
| Metabolic comorbiditiesb n (%) | |||
| Hyperlipidemia | 22,034 (54) | – | – |
| Diabetes | 15,140 (37) | – | – |
| Hypercholesterolemia | 4893 (12) | – | – |
| Obesity | 16,360 (40) | – | – |
| Hypertension | 25,162 (62) | – | – |
| Mental health disordersc n (%) | |||
| Anxiety disorder | 6377 (16) | 4518 (11) | 0.13 |
| Substance use disorder | 6355 (16) | 5550 (14) | 0.06 |
| PTSD | 802 (2) | 713 (2) | 0.02 |
| Personality disorder | 868 (2) | 598 (2) | 0.05 |
| Psychotropic medicationd n (%) | |||
| Antipsychotics | 12,779 (32) | 12,183 (30) | – |
| Antidepressants | 4845 (12) | 4478 (11) | – |
| Antianxiety | 2649 (7) | 2520 (6) | – |
| Hypnotics | 1028 (3) | 780 (2) | – |
| Mood stabilizers | 1994 (5) | 1678 (4) | – |
| Nonpsychotropic medicationd n (%) | |||
| Antihypertensive | 7362 (18) | 1961 (5) | – |
| Lipid-lowering agents | 5313 (13) | 1261 (3) | – |
| Antidiabetics | 3849 (10) | 536 (1) | – |
| GLP-1RAs | 354 (1) | 39 (< 1) | – |
GLP-1RA glucagon-like peptide-1 receptor agonist, ICD-10-CM International Classification of Disease Tenth Revision, Clinical Modification, IQR interquartile range, PTSD post-traumatic stress disorder, SD standard deviation, SMD standardized mean differences, US United States
aMean (SD) ages in cases and controls were 48.1 (14.7) and 46.0 (13.8) years, respectively
bBased on ICD-10-CM codes (see Supplemental Table S2)
cBased on ICD-10-CM codes (see Supplemental Table S1)
dA complete list of medications is provided in Supplemental Table S3
In total, 40,552 individuals in the case group were matched to the same number of individuals in the control group. After matching, SMD for age, sex (male), psychiatric diagnoses, and geographic regions in the US ranged from 0.012 to 0.148 (Table 3). Median age was 47 years in the case group and 44 years in the control group. The majority of individuals in both groups were male (cases: 69%; controls: 72%). Insurance status was similar between the case and control groups. Medicaid-only insurance claims were most common in both the case (34%) and control (39%) groups.
The percentages of mental health conditions (other than schizophrenia) and psychotropic medication dispensations were broadly similar between the case and control groups after matching (Table 3). Although the diagnoses of bipolar disorder and major depression were excluded in case and control groups, small and similar percentages of individuals (4–12%) in these groups were on mood stabilizers and/or antidepressants, suggesting that the medications in these classes may have been dispensed for other indications. As expected, a greater percentage of individuals in the case group than in the control group were prescribed medications for treatment of metabolic comorbidities: antihypertensives (18% vs. 5%), antidiabetics (10% vs. 1%), lipid-lowering agents (13% vs. 3%), and glucagon-like peptide-1 receptor agonists (1% vs. < 1%).
All-Cause Medical Costs
After matching, the numbers for metabolic comorbidities and individuals in each group were: 0, n = 40,552; 1, n = 15,840; 2, n = 11,833; 3, n = 8119; 4, n = 4074; and 5, n = 686. The distribution of observed medical costs per person per year (PPPY) by number of metabolic comorbidities (0 to 5) is shown in Fig. 2. In individuals with schizophrenia, the mean all-cause total medical costs estimated by the GLM increased with the number of metabolic comorbidities (0, US$8396; 1, $14,069; 2, $16,999; 3, $18,637; 4, $22,451; and 5, $34,441); groups of individuals with one or more metabolic comorbidity had significantly greater medical costs than the group with none (P < 0.0001; Table 4). Individuals with all five metabolic comorbidities incurred estimated total medical costs of $34,441 PPPY, over 4 times higher than costs of $8396 for individuals without any metabolic comorbidity (Table 4).
Fig. 2.
Probability densitya Plot showing the distribution of observed PPPY all-cause medical costb by the number of metabolic comorbiditiesc a Probability density is scaled such that the total area under each curve in the plot equals to 1. b All-cause medical cost comprised outpatient, inpatient, emergency room, and stay-related costs. c Metabolic comorbidities comprised obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension. PPPY per person per year
Table 4.
Adjusted estimated medical costsa in individuals with schizophrenia by the number of metabolic comorbiditiesb
| All-cause medical costs | Number of metabolic comorbiditiesb | Estimated meana (95% CI) | P valued |
|---|---|---|---|
| Outpatient | 0 | 3912 (3838–3987) | Reference |
| 1 | 5077 (4951–5206) | < 0.001 | |
| 2 | 6050 (5883–6220) | < 0.001 | |
| 3 | 6943 (6721–7172) | < 0.001 | |
| 4 | 8242 (7888–8613) | < 0.001 | |
| 5 | 13,825 (12,469–15,327) | < 0.001 | |
| Inpatient (hospital) | 0 | 21,939 (20,940–22,983) | Reference |
| 1 | 29,437 (27,993–30,955) | < 0.001 | |
| 2 | 38,105 (35,900–40,441) | < 0.001 | |
| 3 | 38,052 (35,515–40,774) | < 0.001 | |
| 4 | 42,907 (39,148–47,033) | < 0.001 | |
| 5 | 51,062 (41,835–62,324) | < 0.001 | |
| Emergency room | 0 | 1093 (1050–1137) | Reference |
| 1 | 1171 (1109–1236) | 0.02 | |
| 2 | 1303 (1220–1392) | < 0.001 | |
| 3 | 1070 (989–1158) | 0.6 | |
| 4 | 1026 (923–1139) | 0.2 | |
| 5 | 911 (715–1161) | 0.1 | |
| Stay-related (nursing home, group home, and/or hospice) | 0 | 5757 (5205–6369) | Reference |
| 1 | 5664 (5100–6291) | 0.8 | |
| 2 | 5694 (5150–6296) | 0.8 | |
| 3 | 5604 (5022–6254) | 0.6 | |
| 4 | 7150 (6128–8342) | 0.01 | |
| 5 | 9685 (7007–13,386) | 0.002 | |
| Total medical costc | 0 | 8396 (8228–8568) | Reference |
| 1 | 14,069 (13,701–14,445) | < 0.001 | |
| 2 | 16,999 (16,495–17,517) | < 0.001 | |
| 3 | 18,637 (17,987–19,310) | < 0.001 | |
| 4 | 22,451 (21,388–23,565) | < 0.001 | |
| 5 | 34,441 (30,709–38,631) | < 0.001 |
CI confidence interval, GLM generalized linear model, ICD-10-CM International Classification of Diseases Tenth Revision-Clinical Modification, MetS metabolic syndrome
aMean costs were estimated from GLMs for matched cohort adjusted for individual insurance status. In the GLMs, for each medical cost type and total medical cost, the model-based estimates reflect mean cost among patients who incurred any medical cost (nonzero cost observations) stratified by number of metabolic comorbidities, rather than the overall mean observed costs
bMetabolic comorbidities comprised diagnoses of obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension, based on ICD-10-CM codes (Supplemental Table S1)
cTotal medical costs comprised outpatients, inpatients, emergency visits, stay-related, and other unspecified costs. The estimated total medical costs were lower than estimated inpatient costs because data on inpatient costs were largely missing and the total medical cost reflected the costs in the other categories, which were typically smaller
dP values were estimated using patients with 0 metabolic disorders (control) as the reference group and were considered statistically significant at P < 0.008 based on Bonferroni correction
Outpatient and inpatient costs contributed to the increase in total medical costs associated with metabolic comorbidities (Table 4). The mean estimated outpatient costs were 3.5 times greater in individuals with five metabolic comorbidities than in those with none ($13,825 PPPY vs. $3912 PPPY; P < 0.0001; Table 4). Compared with individuals without metabolic comorbidities, mean estimated inpatient costs in those with five comorbidities were 2.3 times greater ($51,062 PPPY vs. $21,939 PPPY; P < 0.0001; Table 4). However, inpatient cost variables were incomplete for a subset of records, leading to per-person cost estimates that were greater than those modeled for total medical costs because the costs in the remaining categories that comprise total medical costs were typically smaller. The findings on trends in the sensitivity analysis, which included patients with zero or missing cost records, were highly consistent with the main analysis.
Differences in stay-related and emergency room costs among groups stratified by number of metabolic comorbidities were less evident. The mean stay-related estimated costs were about 1.7 times greater in individuals with five metabolic comorbidities than in those without any metabolic comorbidity, but the increase was not considered statistically significant after Bonferroni correction for multiple comparisons ($9685 PPPY vs. $5757 PPPY). Mean emergency room estimated costs did not increase consistently with the number of metabolic comorbidities (Table 4).
Discussion
In this retrospective case–control study of claims data from more than 80,000 individuals with schizophrenia in the US from 2018 to 2024, we found that all-cause total medical costs increased incrementally with the number of metabolic comorbidities. These comorbidities, comprising obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension, were identified based on ICD-10-CM codes and employed as proxy components of cumulative metabolic comorbidity burden. Estimated mean total medical costs in individuals with five metabolic comorbidities were more than four times those individuals without any metabolic comorbidity. This pattern was also observed for outpatient and inpatient costs, and to a lesser extent for stay-related costs but not emergency room costs. As expected, a greater percentage of individuals with metabolic comorbidities than those without had claims for medications indicated for these disorders, likely accounting in part for the increased medical costs in this group.
The association between cumulative metabolic comorbidity burden and increased healthcare costs has been reported in other populations. In a study using data from the US general population, total PPPY healthcare costs were 1.6 times greater in individuals with MetS than those without ($5732 vs. $3581); these costs increased by an average of 24% per additional metabolic comorbidity [14]. The same study reported increased inpatient care, primary and other outpatient care, and pharmacy costs; however, it did not provide data on emergency room costs [14]. In a separate study of 3789 individuals aged ≥ 65 years in the general US population, cumulative 10-year total Medicare costs were 20% greater in those with MetS than in those without [26].
The economic burden of metabolic comorbidities may be compounded by indirect effects. For example, the pathophysiology of schizophrenia may be affected by MetS-induced peripheral sub-inflammatory states and kynurenine pathway interactions, leading to an increase in neurotoxic and neuroactive metabolites that exacerbate symptom severity and cognitive deficits [27]. Resultant treatment with higher doses of antipsychotics is associated with MetS risk, perpetuating a vicious clinical cycle. The trajectory of metabolic risk and the presentation of schizophrenia are also both influenced by sex differences [28–30]. A large epidemiological cohort study showed an association between schizophrenia and subsequent cardiovascular disease events that was more pronounced in women than in men [31]. The current study matched cases and controls for potential confounders, including sex, and the potential effects of sex on cost outcomes were thus not independently assessed.
Long-term management of schizophrenia includes treatment with antipsychotic medications [32]. Adverse effects of most available antipsychotic medications include conditions or metabolic disorders related to MetS, for example, weight gain, increased waist circumference, dyslipidemia, insulin resistance, type 2 diabetes, and hypertension [33]. These side effects are associated with first-generation antipsychotics, which are dopamine receptor antagonists [9], but have been reported particularly for clozapine and olanzapine, which are second-generation antipsychotics that have been associated with a higher risk of metabolic dysregulation rates than other antipsychotic agents [34].
Given the high prevalence of MetS in schizophrenia, antipsychotic medications associated with fewer adverse metabolic effects are needed and have been developed or are currently being developed for the treatment of schizophrenia [17–19, 35–40]. New antipsychotic classes with highly favorable metabolic profiles have been shown to be efficacious and associated with a lower risk of the common adverse effects of currently available antipsychotics, including decreased risks of weight gain and metabolic effects. [17]
Prevention and management of metabolic comorbidities in individuals with schizophrenia should encompass selection of antipsychotic medications with few adverse metabolic effects [11]. While pharmacological selection is one pillar of management, non-pharmacological interventions are equally vital. Holistic management entails a multidisciplinary team with expertise in different areas, including movement therapy. Aerobic exercise and improved functional mobility have proven benefits on cognition and overall quality of life in individuals with schizophrenia [41].
This claims-based study relied on the large number of individuals with schizophrenia in the STATinMED database, which facilitated matching of cases and controls. A primary limitation of this study is the inherent nature of administrative claims data, which lack the granular clinical measurements—such as waist circumference and specific laboratory values for HDL and triglycerides—required to formally diagnose MetS. Consequently, our analysis relies on ICD-10-CM diagnosis codes as proxies to measure cumulative metabolic comorbidity burden rather than clinically confirmed MetS. Although the use of diagnostic codes to proxy metabolic burden is a standard and validated approach in retrospective health economic analyses, it may result in the under-identification of patients who meet the biometric criteria for MetS but lack the corresponding billing codes during the study period. Nonetheless, the likelihood of an individual having MetS should increase with the number of metabolic comorbidities. Diagnostic codes were used as proxies for MetS components and as such did not include other conditions of metabolic dysfunction such as metabolic liver disease. The study may have underestimated the true economic burden of inpatient costs due to incomplete cost variables for a subset of records in the main analysis. However, trends in the sensitivity analysis, which included patients with zero or missing cost records, were highly consistent with those in the main analysis. The findings in this study can be generalized to individuals who have healthcare insurance in the US. However, the database lacks data on healthcare resource utilization that could be analyzed to provide further information on the economic impact of cumulative metabolic comorbidity burden in individuals with schizophrenia.
Conclusion
This claims-based study showed that, in individuals with schizophrenia, total all-cause medical costs increased with the number of metabolic comorbidities. From 2018 to 2024, the estimated total medical costs in US individuals with obesity, hyperlipidemia, hypercholesterolemia, diabetes, and hypertension were $34,441 PPPY, nearly four times greater than in individuals without these comorbidities. Outpatient, inpatient, and stay-related costs also increased with the number of metabolic comorbidities.
Given the high prevalence of MetS and the medical costs of associated metabolic comorbidities in individuals with schizophrenia, the metabolic and cardiovascular side effects of available treatments for schizophrenia should be considered as part of a multidisciplinary assessment for the treatment of schizophrenia. Prevention and management of metabolic comorbidities in individuals with schizophrenia should encompass early risk assessment, lifestyle modifications (including increasing aerobic exercise and functional mobility), and selection of antipsychotic medications with few adverse metabolic effects.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
This work was informed by exploratory analyses previously performed by Sumitomo Pharma, for which we thank Sasagu Tomioka for early conceptual input. Shivanshu Awasthi is no longer affiliated with Otsuka Pharmaceutical Development & Commercialization, Inc. Shivanshu Awasthi is now affiliated with Merck.
Medical Writing/Editorial Sssistance
Medical writing support was provided by Mike Lee, PhD, of Oxford PharmaGenesis, Oxford, UK, and The Medicine Group, New Hope, PA, USA, in accordance with Good Publication Practice (GPP 2022) guidelines (www.ismpp.org/gpp-2022), and funded by Otsuka Pharmaceutical Development & Commercialization, Inc., Princeton, NJ, USA.
Author Contributions
Xue Han contributed to all aspects of the study. Seth C. Hopkins, Zhen Zhang, and Shivanshu Awasthi were involved in the study design, data acquisition and analysis, and interpretation of results. All authors contributed to drafting and revising the manuscript. All authors read and approved the final version of the manuscript for submission.
Funding
Sponsorship for this study and the journal’s Rapid Service Fee were funded by Otsuka Pharmaceutical Development & Commercialization, Inc.
Data Availability
The datasets generated during and/or analyzed during the current study are not publicly available due to the commercially owned, proprietary nature of the datasets but are available from the corresponding author on reasonable request.
Declarations
Conflict of Interest
Xue Han, Seth C. Hopkins, and Zhen Zhang are employees of Otsuka Pharmaceutical Development & Commercialization, Inc. Shivanshu Awasthi was an employee of Otsuka Pharmaceutical Development & Commercialization, Inc., at the time the study was conducted. Data were obtained from the STATinMED RWD Insights database under a valid and active license agreement, and analyses and publication of findings were conducted in accordance with the terms of such agreement.
Ethical Approval
This study is an analysis of deidentified claims data, and therefore no ethical approval was required.
Footnotes
Prior presentation: A portion of this research was previously presented at the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) 2026 Annual Meeting: May 17–20, 2026, Philadelphia, PA, USA as poster number EE237.
References
- 1.World Health Organization. Schizophrenia [Fact sheet]. https://www.who.int/news-room/fact-sheets/detail/schizophrenia. (Accessed Oct 27, 2025).
- 2.Lee H, Lee JH, Lee S, Lim JS, Kim HJ, Park J, et al. Comorbid health outcomes in patients with schizophrenia: an umbrella review of systematic reviews and meta-analyses. Mol Psychiatry. 2025;30(3):1127–37. 10.1038/s41380-024-02792-2. [DOI] [PubMed] [Google Scholar]
- 3.Hjorthoj C, Sturup AE, McGrath JJ, Nordentoft M. Years of potential life lost and life expectancy in schizophrenia: a systematic review and meta-analysis. Lancet Psychiatry. 2017;4(4):295–301. 10.1016/S2215-0366(17)30078-0. [DOI] [PubMed] [Google Scholar]
- 4.Ali S, Santomauro D, Ferrari AJ, Charlson F. Schizophrenia as a risk factor for cardiovascular and metabolic health outcomes: A comparative risk assessment. Epidemiol Psychiatr Sci. 2023;32:e8. 10.1017/S2045796023000045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Romeo S, Vidal-Puig A, Husain M, Ahima R, Arca M, Bhatt DL, et al. Clinical staging to guide management of metabolic disorders and their sequelae: A European Atherosclerosis Society consensus statement. Eur Heart J. 2025;46(38):3685–713. 10.1093/eurheartj/ehaf314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA, et al. Diagnosis and management of the metabolic syndrome: An American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation. 2005;112(17):2735–52. 10.1161/CIRCULATIONAHA.105.169404. [DOI] [PubMed] [Google Scholar]
- 7.Saklayen MG. The global epidemic of the metabolic syndrome. Curr Hypertens Rep. 2018;20(2):12. 10.1007/s11906-018-0812-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mitchell AJ, Vancampfort D, Sweers K, van Winkel R, Yu W, De Hert M. Prevalence of metabolic syndrome and metabolic abnormalities in schizophrenia and related disorders--A systematic review and meta-analysis. Schizophr Bull. 2013;39(2):306–18. 10.1093/schbul/sbr148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Papanastasiou E. The prevalence and mechanisms of metabolic syndrome in schizophrenia: A review. Ther Adv Psychopharmacol. 2013;3(1):33–51. 10.1177/2045125312464385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tirupati S, Chua LE. Obesity and metabolic syndrome in a psychiatric rehabilitation service. Aust N Z J Psychiatry. 2007;41(7):606–10. 10.1080/00048670701392841. [DOI] [PubMed] [Google Scholar]
- 11.Penninx B, Lange SMM. Metabolic syndrome in psychiatric patients: overview, mechanisms, and implications. Dialogues Clin Neurosci. 2018;20(1):63–73. 10.31887/DCNS.2018.20.1/bpenninx. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lee J, Xue X, Au E, McIntyre WB, Asgariroozbehani R, Panganiban K, et al. Glucose dysregulation in antipsychotic-naive first-episode psychosis: In silico exploration of gene expression signatures. Transl Psychiatry. 2024;14(1):19. 10.1038/s41398-023-02716-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Vaccarino V, Prescott E, Shah AJ, Bremner JD, Raggi P, Dobiliene O, et al. Mental health disorders and their impact on cardiovascular health disparities. Lancet Reg Health Eur. 2025;56:101373. 10.1016/j.lanepe.2025.101373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Boudreau DM, Malone DC, Raebel MA, Fishman PA, Nichols GA, Feldstein AC, et al. Health care utilization and costs by metabolic syndrome risk factors. Metab Syndr Relat Disord. 2009;7(4):305–14. 10.1089/met.2008.0070. [DOI] [PubMed] [Google Scholar]
- 15.Nichols GA, Moler EJ. Metabolic syndrome components are associated with future medical costs independent of cardiovascular hospitalization and incident diabetes. Metab Syndr Relat Disord. 2011;9(2):127–33. 10.1089/met.2010.0105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Simon J, Wienand D, Park AL, Wippel C, Mayer S, Heilig D, et al. Excess resource use and costs of physical comorbidities in individuals with mental health disorders: a systematic literature review and meta-analysis. Eur Neuropsychopharmacol. 2023;66:14–27. 10.1016/j.euroneuro.2022.10.001. [DOI] [PubMed] [Google Scholar]
- 17.Milanovic S, Dedic N, Lew R, Burton D, Koblan KS, Camilleri M, et al. TAAR1 agonist ulotaront delays gastric emptying of solids in patients with schizophrenia and concurrent metabolic syndrome with prediabetes. Diabetes Obes Metab. 2024;26(6):2466–75. 10.1111/dom.15569. [DOI] [PubMed] [Google Scholar]
- 18.Milanovic S, Lew R, Xiong K, Casavant E, Galluppi G, Chen YL, et al. A double-blind study on ulotaront’s impact on weight-related parameters in schizophrenia patients with metabolic syndrome and prediabetes: Part II. Diabetes Obes Metab. 2025. 10.1111/dom.70110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Milanovic S, Lew R, Xiong K, Casavant E, Galluppi G, Chen YL, et al. An open-label study on ulotaront’s effects on insulin-glucose regulation in schizophrenia patients with metabolic syndrome and prediabetes: Part I. Diabetes Obes Metab. 2025. 10.1111/dom.70136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Jung E, Kong SY, Ro YS, Ryu HH, Shin SD. Serum cholesterol levels and risk of cardiovascular death: a systematic review and a dose-response meta-analysis of prospective cohort studies. Int J Environ Res Public Health. 2022;19(14):8272. 10.3390/ijerph19148272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Harvey R, Jankus DD, Mosley D, Healthcare U. Random assignment of proxy event dates to unexposed individuals in observational studies: an automated technique using SAS®. Midwest SAS Users Group (MSUG). 2012:16–8.
- 22.Ho DE, Imai K, King G, Stuart EA. MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Softw. 2011;42(8):1–28. 10.18637/jss.v042.i08. [Google Scholar]
- 23.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28(25):3083–107. 10.1002/sim.3697. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Awasthi S, Hadzi Boskovic D, Kapadia S. Association between timing of adjunctive brexpiprazole initiation and healthcare resource use and costs in major depressive disorder. J Med Econ. 2025;28(1):1348–56. 10.1080/13696998.2025.2549653. [DOI] [PubMed] [Google Scholar]
- 25.R Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2024. [Google Scholar]
- 26.Curtis LH, Hammill BG, Bethel MA, Anstrom KJ, Gottdiener JS, Schulman KA. Costs of the metabolic syndrome in elderly individuals: findings from the Cardiovascular Health Study. Diabetes Care. 2007;30(10):2553–8. 10.2337/dc07-0460. [DOI] [PubMed] [Google Scholar]
- 27.Sapienza J, Agostoni G, Repaci F, Spangaro M, Comai S, Bosia M. Metabolic syndrome and schizophrenia: adding a piece to the interplay between the kynurenine pathway and inflammation. Metabolites. 2025. 10.3390/metabo15030176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Bosia M, Spangaro M, Sapienza J, Martini F, Civardi S, Buonocore M, et al. Cognition in schizophrenia: modeling the interplay between interleukin-1β C-511T polymorphism, metabolic syndrome, and sex. Neuropsychobiology. 2021;80(4):321–32. 10.1159/000512082. [DOI] [PubMed] [Google Scholar]
- 29.Moniem I, Kafetzopoulos V. Sex differences in schizophrenia: symptomatology, treatment efficacy and adverse effects. Front Psychiatry. 2025;16:1594334. 10.3389/fpsyt.2025.1594334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Zhou Y, Song X, Guo Y, Lang X, Li Z, Zhang XY. Sex differences in metabolic disorder patterns of first-episode drug-naive patients with schizophrenia. Psychoneuroendocrinology. 2021;124:105061. 10.1016/j.psyneuen.2020.105061. [DOI] [PubMed] [Google Scholar]
- 31.Komuro J, Kaneko H, Suzuki Y, Okada A, Fujiu K, Takeda N, et al. Sex differences in the relationship between schizophrenia and the development of cardiovascular disease. J Am Heart Assoc. 2024;13(5):e032625. 10.1161/jaha.123.032625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.American Psychiatric Association. The American Psychiatric Association practice guideline for the treatment of patients with schizophrenia: guideline statement summary. 3rd Edition ed. 2020. p. 5–6. [Google Scholar]
- 33.Akinola PS, Tardif I, Leclerc J. Antipsychotic-induced metabolic syndrome: a review. Metab Syndr Relat Disord. 2023;21(6):294–305. 10.1089/met.2023.0003. [DOI] [PubMed] [Google Scholar]
- 34.Pillinger T, McCutcheon RA, Vano L, Mizuno Y, Arumuham A, Hindley G, et al. Comparative effects of 18 antipsychotics on metabolic function in patients with schizophrenia, predictors of metabolic dysregulation, and association with psychopathology: a systematic review and network meta-analysis. Lancet Psychiatry. 2020;7(1):64–77. 10.1016/S2215-0366(19)30416-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Achtyes ED, Hopkins SC, Dedic N, Dworak H, Zeni C, Koblan K. Ulotaront: review of preliminary evidence for the efficacy and safety of a TAAR1 agonist in schizophrenia. Eur Arch Psychiatry Clin Neurosci. 2023;273(7):1543–56. 10.1007/s00406-023-01580-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Hsu YC, Hung TY, Chen YB, Hung KC, Liang CS, Tseng PT, et al. Trajectory of efficacy and safety across Ulotaront dose levels in schizophrenia: a systematic review and dose-response meta-analysis. Int J Neuropsychopharmacol. 2025. 10.1093/ijnp/pyaf059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Hussain JM, Arif W, Humayun S, Shoaib N, Khan SM. Xanomeline‐trospium chloride (Cobenfy): a new era in managing schizophrenia—comparative effectiveness and economic challenges. Medicine Advances. 2025;3(2):116–21. 10.1002/med4.70013. [Google Scholar]
- 38.Kaul I, Sawchak S, Walling DP, Tamminga CA, Breier A, Zhu H, et al. Efficacy and safety of xanomeline-trospium chloride in schizophrenia: A randomized clinical trial. JAMA Psychiat. 2024;81(8):749–56. 10.1001/jamapsychiatry.2024.0785. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Koblan KS, Kent J, Hopkins SC, Krystal JH, Cheng H, Goldman R, et al. A non-D2-receptor-binding drug for the treatment of schizophrenia. N Engl J Med. 2020;382(16):1497–506. 10.1056/NEJMoa1911772. [DOI] [PubMed] [Google Scholar]
- 40.United States Food and Drug Administration (FDA). COBENFY: Prescribing information. https://www.accessdata.fda.gov/drugsatfda_docs/label/2024/216158s000lbl.pdf. (Accessed Oct 24, 2025). 2024.
- 41.Cuoco F, Agostoni G, Lesmo S, Sapienza J, Buonocore M, Bechi M, et al. Get up! functional mobility and metabolic syndrome in chronic schizophrenia: effects on cognition and quality of life. Schizophr Res Cognit. 2022;28:100245. 10.1016/j.scog.2022.100245. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated during and/or analyzed during the current study are not publicly available due to the commercially owned, proprietary nature of the datasets but are available from the corresponding author on reasonable request.