Abstract
Objective
We sought to determine the incidence and outcomes of head and neck cancer (HNC) among patients with schizophrenia.
Study Design
Cohort study utilizing TriNetX, a database containing millions of deidentified clinical records.
Setting
Multicenter study utilizing records from 68 healthcare organizations.
Methods
A cohort of patients with schizophrenia was analyzed for the annual incidence of HNC diagnosis between 2011 and 2021. TriNetX was queried for adult patients with HNC with and without schizophrenia (or long‐term antipsychotic use, as a surrogate). Cohorts were 1:1 propensity‐matched based on sociodemographic variables to produce matched cohorts of 25,077 patients each. Outcomes included mortality, recurrence in lymph nodes and lung, systemic treatment, failure to thrive, and hospice enrollment. Outcomes are reported as hazard ratio (HR; Cox proportional hazards model) and odds ratio (OR) with 95% CI.
Results
Incidence of HNC in patients with schizophrenia peaked at 0.061% in 2012. HNC patients with schizophrenia have a significantly increased risk of mortality (HR 1.37, 95% CI 1.10‐1.72), locoregional recurrence (OR 1.36, 95% CI 1.30‐1.43), distant metastases to the lung (OR 1.72, 95% CI 1.59‐1.87), chemotherapy (OR 4.26, 95% CI 3.88‐4.69), radiation (OR 2.47, 95% CI 2.19‐2.78), failure to thrive (OR 2.41, 95% CI 2.32‐2.73), and hospice enrollment (OR 3.17, 95% CI 2.66‐3.76) compared to HNC patients without schizophrenia.
Conclusion
HNC patients with schizophrenia have a significant increase in risk of mortality, recurrence, and poor outcomes compared to those without schizophrenia. These findings support a renewed focus on ensuring safety nets for this vulnerable population to ensure appropriate cancer screening and care.
Keywords: head and neck cancer, mental illness, psychosis, psychotic disorders, schizophrenia
Schizophrenia affects an estimated 1% of the United States, or about 3 million people. 1 Nonelderly people with schizophrenia have an increased rate of all‐cause mortality compared to the general population, attributed to economic disadvantage, negative health behaviors, and difficulty accessing healthcare. 2 However, the link between schizophrenia and risk of cancer is unclear, with reports of increased, equivalent, and decreased risk. 3 , 4 , 5 The incidence of head and neck cancer (HNC) among people with schizophrenia is particularly nebulous, though this population is more likely to use alcohol and tobacco, known drivers of HNC. 6 Cancer outcomes among patients with schizophrenia are also poor, with increased risk of mortality. 7 This may be secondary to higher rates of comorbidities that alter a patient's ability to tolerate cancer treatment, decreased access to preventative screenings, and reduced likelihood of accessing and completing medical treatment. 8 , 9 , 10
In this study, we utilized a large national database to describe the incidence of HNC among patients with schizophrenia. Further, we describe the outcomes, including mortality, likelihood of receiving treatment, and likelihood of recurrence among patients with HNC and schizophrenia as compared to a matched cohort of HNC patients without schizophrenia.
Design and Methods
This study did not require approval from the University of Southern California institutional review board as it did not meet the criteria for Human Subjects Research.
Study Population
This retrospective cohort study used TriNetX, a globally federated health research network with deidentified electronic medical records. We used the US Collaborative Network, which contains data from 66 healthcare organizations and more than 120 million patients. The data reviewed are a secondary analysis of existing data, do not involve intervention or interaction with human subjects, and are deidentified per the deidentification standard defined in Section §164.514(a) of the HIPAA Privacy Rule. The process by which the data are deidentified is attested to through a formal determination by a qualified expert as defined in Section §164.514(b)(1) of the HIPAA Privacy Rule.
Incidence Study
A cohort of patients ≥18 years old with schizophrenia between January 1, 2011, and December 31, 2021, was created. A reference cohort included patients ≥18 years old with a least one visit in the same time period. Outcomes measured included all HNC and cancer of the oral cavity, oropharynx, larynx, and hypopharynx. The lookback window was set to any time before 1 day before the time window. Eleven time windows were set from January 1 to December 31 for each year between 2011 and 2021. Statistical analysis was conducted in February 2025 using the TriNetX “Incidence and Prevalence” function to assess the annual incidence proportion of HNC. Results were further stratified by ethnicity. In situations in which the case number is between 1 and 10, TriNetX automatically rounds to 10 for the protection of privacy, which may result in overestimation. Subsite analysis by ethnicity was excluded for hypopharyngeal cancer as all groups had case numbers rounded to 10.
Outcomes Analysis: HNC With Schizophrenia Versus HNC With No Schizophrenia
Two cohorts were created to compare outcomes of patients with HNC with and without schizophrenia (Figure 1). Cohort 1 (schizophrenia) included patients ≥18 years old with a diagnosis of schizophrenia preceding HNC diagnosis. Cohort 2 (no schizophrenia) included patients ≥18 years old with an HNC diagnosis and no history of schizophrenia. The index event was set to the diagnosis of HNC. As TriNetX only contains follow‐up data for a maximum of 20 years post‐diagnosis, the maximum follow‐up time was set to 15 years to avoid instability due to small numbers at risk. The outcome measured was mortality. The TriNetX built‐in propensity score matching function was used (1:1 matching using a nearest neighbor greedy matching algorithm with a caliper width of 0.1 pooled standard deviations). Characteristics used for matching included age at index event, sex, race, ethnicity, tobacco and alcohol use, and American Joint Committee on Cancer (AJCC) tumor stage, when available. Advanced analytics were used for analysis. 11 Patients with an outcome before the index event and patients that met index event criteria greater than 20 years ago were excluded. The Cox proportional hazards model was used to determine the hazard ratio (HR) and log‐rank P‐value. P‐values below .05 were defined as statistically significant.
Figure 1.
Flow diagram of cohort development. Patients with head and neck cancer (HNC) were separated into groups of at least 1 year of antipsychotic use before HNC diagnosis or no history of antipsychotic use, then 1:1 propensity‐matched. HCOs, healthcare organizations.
Outcomes Study: HNC With Antipsychotics Versus HNC With No Antipsychotics
Due to the small size of the cohort of HNC patients with schizophrenia, we sought to develop a larger cohort that would allow for analysis of the incidence of rare outcomes. The methodology for this outcome analysis is similar to that used in the literature to evaluate outcomes of breast cancer in female patients with and without schizophrenia. 12 Two cohorts were created to compare outcomes of patients with HNC with and without antipsychotic use. Long‐term antipsychotic use was used as a surrogate for the diagnosis of schizophrenia in this study population. Cohort 1 (antipsychotics) included patients ≥18 years old with an HNC diagnosis and a history of antipsychotic use for ≥1 year before HNC. Cohort 2 (no antipsychotics) included patients ≥18 years old with an HNC diagnosis and no history of antipsychotic use. The index event was set to the diagnosis of HNC. Outcomes included mortality, recurrence in lymph nodes, metastasis in lungs, chemotherapy treatment (cisplatin or carboplatin), radiation therapy, adult failure to thrive, and hospice enrollment. Propensity score matching and analysis were conducted as above. In addition, measures of association, which include risk difference, risk ratio, and odds ratio (OR), were calculated. Billing and data codes used for analysis were selected based on clinical expertise and recently published literature. 13 , 14 , 15 , 16 All billing and data codes are listed in Supplemental Table S1, available online.
Results
Incidence Study
We identified 307,985 adult patients with schizophrenia between 2011 and 2021, and a total of 883 HNC cases diagnosed in this time. Incidence peaked in 2012 with an incidence proportion of 0.061% and a trough of 0.032% in 2011. HNC incidence among patients with schizophrenia was not globally increased compared to the general population, though incidence was higher in several years (2012, 2013, 2019, and 2021) (Figure 2). When stratified by subsite, incidence at the oral cavity, oropharynx, and larynx subsite peaked in 2019, with an incidence of 0.023%, 0.016%, and 0.019%, respectively. Incidence at the hypopharynx subsite peaked a 0.018% in 2012. When stratified by ethnicity, the incidence of all HNC was higher among Hispanic and Latino patients, but the incidence varied by ethnicity under subsite analysis (Supplemental Figure S1, available online).
Figure 2.
Incidence of head and neck cancer (HNC) among patients with schizophrenia from 2011 to 2021. Incidence of all HNC and subsites among patients with schizophrenia, with annual incidence of HNC among the general population for reference.
Outcomes Analysis: HNC With Schizophrenia Versus HNC With No Schizophrenia
Before propensity matching, the schizophrenia cohort contained 651 patients with a mean age of 57.9 (SD 10.5). The no schizophrenia cohort contained 172,058 patients with a mean age of 61.8 (SD 12.2). Both cohorts were predominantly male, white, and not Hispanic or Latino. After propensity matching, each cohort contained 625 patients (Table 1). Patients in the schizophrenia cohort had a 37% increased risk of mortality compared to those in the no schizophrenia cohort (HR 1.37, 95% CI 1.10‐1.72, log‐rank P < .001) (Figure 3). The risk of death within 1 year of HNC diagnosis was also significantly increased in the schizophrenia cohort compared to the no schizophrenia cohort (HR 1.65, 95% CI 1.25‐2.19, log‐rank P < .001) (Supplemental Figure S2, available online).
Table 1.
Cohort Demographics
| Schizophrenia versus no schizophrenia | ||||||
|---|---|---|---|---|---|---|
| N (%) | ||||||
| Before propensity matching | After propensity matching | |||||
| Schizophrenia cohort N = 651 | No schizophrenia cohort N = 172,058 | Std. diff.a | Schizophrenia cohort N = 625 | No schizophrenia cohort N = 625 | Std. diff.a | |
| Age at index (SDb) | 57.9 (10.5) | 61.8 (12.2) | 0.342 | 57.9 (10.5) | 58.2 (10.8) | 0.022 |
| Sex | ||||||
| Male | 411 (65.7) | 120,278 (76.2) | 0.087 | 410 (65.6) | 416 (66.6) | 0.020 |
| Female | 215 (34.3) | 47,756 (30.3) | 0.087 | 215 (34.4) | 209 (33.4) | 0.020 |
| Unknown | 0 (0.00) | 48 (0.03) | 0.025 | 0 (0.00) | 0 (0.00) | 0 |
| Race | ||||||
| American Indian or Alaska Native | 10 (1.60) | 585 (0.37) | 0.125 | 10 (1.60) | 10 (1.60) | <0.001 |
| Asian | 10 (1.60) | 3291 (2.09) | 0.036 | 10 (1.60) | 10 (1.60) | <0.001 |
| Black or African American | 232 (37.1) | 15,585 (9.88) | 0.677 | 231 (37.0) | 237 (37.9) | 0.020 |
| Native Hawaiian or Other Pacific Islander | 0 (0.00) | 178 (0.11) | 0.048 | 0 (0.00) | 10 (1.60) | 0.180 |
| White | 336 (53.7) | 119,464 (75.7) | 0.474 | 336 (53.8) | 336 (53.8) | <0.001 |
| Other race | 16 (2.56) | 5262 (3.34) | 0.046 | 16 (2.56) | 14 (2.24) | 0.021 |
| Unknown | 33 (5.27) | 13,417 (8.50) | 0.128 | 33 (5.28) | 33 (5.28) | <0.001 |
| Ethnicity | ||||||
| Hispanic or Latino | 34 (5.43) | 13,417 (5.62) | 0.008 | 33 (5.28) | 30 (4.80) | 0.022 |
| Not Hispanic or Latino | 509 (81.3) | 120,278 (76.2) | 0.125 | 509 (81.4) | 508 (81.3) | 0.004 |
| Unknown | 83 (13.3) | 28,633 (18.1) | 0.135 | 83 (13.3) | 87 (13.9) | 0.019 |
| AJCCc tumor stage | ||||||
| Stage 0 | 10 (1.60) | 515 (0.33) | 0.131 | 10 (1.60) | 10 (1.60) | <0.001 |
| Stage 1 | 10 (1.60) | 1861 (1.18) | 0.036 | 10 (1.60) | 10 (1.60) | <0.001 |
| Stage 2 | 10 (1.60) | 1151 (0.73) | 0.081 | 10 (1.60) | 10 (1.60) | <0.001 |
| Stage 3 | 10 (1.60) | 1156 (0.73) | 0.081 | 10 (1.60) | 10 (1.60) | <0.001 |
| Stage 4 | 10 (1.60) | 2358 (1.49) | 0.008 | 10 (1.60) | 10 (1.60) | <0.001 |
| Alcohol use | 166 (26.5) | 7475 (4.74) | 0.629 | 165 (26.4) | 160 (25.6) | 0.018 |
| Nicotine dependence | 285 (45.5) | 18,523 (11.7) | 0.806 | 284 (45.4) | 287 (45.9) | 0.010 |
| Tobacco use | 109 (17.4) | 6076 (3.85) | 0.451 | 108 (17.3) | 105 (16.8) | 0.013 |
Std. diff. = standardized difference.
SD = standard deviation.
AJCC = American Joint Committee on Cancer.
Figure 3.
Survival curve of patients with head and neck cancer (HNC) with versus without schizophrenia. Patients with HNC and schizophrenia have significantly higher likelihood of mortality than patients with HNC without schizophrenia. HR, hazard ratio.
Outcomes Study: HNC With Antipsychotics Versus HNC With No Antipsychotics
Before propensity matching, the antipsychotic cohort contained 25,496 patients with a mean age of 62.5 years (SD 12.1). The no antipsychotics cohort contained 148,822 patients with a mean age of 62.7 years (SD 12.0). Both cohorts were predominantly male, white, and not Hispanic or Latino. Following propensity matching, each cohort contained 25,077 patients (Table 2).
Table 2.
Cohort Demographics
| Antipsychotics cohort versus no antipsychotics cohort | ||||||
|---|---|---|---|---|---|---|
| N (%) | ||||||
| Before propensity matching | After propensity matching | |||||
| Antipsychotics cohort N = 25,496 | No antipsychotics cohort N = 148,822 | Std. diff.a | Antipsychotics cohort N = 25,077 | No antipsychotics cohort N = 25,077 | Std. diff. | |
| Age at index (SDb) | 62.5 (12.1) | 62.7 (12.0) | 0.011 | 62.6 (12.1) | 62.8 (12.0) | 0.019 |
| Sex | ||||||
| Male | 16,522 (64.8) | 93,135 (62.6) | 0.070 | 16,343 (65.2) | 16,409 (65.4) | 0.006 |
| Female | 8380 (32.9) | 39,385 (26.5) | 0.089 | 8234 (32.8) | 8182 (32.6) | 0.004 |
| Unknown | 594 (2.33) | 16,302 (11.0) | 0.353 | 500 (1.99) | 486 (1.93) | 0.004 |
| Race | ||||||
| American Indian or Alaska Native | 96 (0.38) | 357 (0.24) | 0.022 | 96 (0.38) | 91 (0.36) | 0.003 |
| Asian | 567 (2.22) | 4708 (3.16) | 0.074 | 565 (2.25) | 551 (2.20) | 0.004 |
| Black or African American | 3036 (12.0) | 12,400 (8.33) | 0.093 | 2936 (11.7) | 2951 (11.8) | 0.002 |
| Native Hawaiian or Other Pacific Islander | 92 (0.35) | 1500 (1.01) | 0.087 | 92 (0.37) | 81 (0.32) | 0.007 |
| White | 19,438 (76.5) | 96,077 (70.5) | 0.137 | 19,217 (76.6) | 19,296 (76.9) | 0.007 |
| Other race | 667 (2.60) | 4606 (3.09) | 0.044 | 667 (2.70) | 658 (2.62) | 0.002 |
| Unknown | 1506 (5.91) | 16,639 (12.2) | 0.221 | 1504 (5.99) | 1449 (5.77) | 0.009 |
| Ethnicity | ||||||
| Hispanic or Latino | 943 (3.70) | 5945 (3.99) | 0.033 | 936 (3.73) | 907 (3.62) | 0.006 |
| Not Hispanic or Latino | 19,960 (78.6) | 93,717 (68.8) | 0.224 | 19,655 (78.4) | 19,669 (78.4) | 0.001 |
| Unknown | 4501 (17.7) | 36,625 (26.9) | 0.221 | 4486 (17.9) | 4501 (17.9) | 0.002 |
| AJCCc tumor stage | ||||||
| Stage 0 | 60 (0.24) | 443 (0.30) | 0.012 | 60 (0.24) | 65 (0.26) | 0.004 |
| Stage 1 | 267 (1.05) | 1479 (0.99) | 0.005 | 264 (1.05) | 209 (0.83) | 0.023 |
| Stage 2 | 246 (0.96) | 846 (0.57) | 0.045 | 234 (0.93) | 217 (0.87) | 0.007 |
| Stage 3 | 228 (0.89) | 793 (0.53) | 0.043 | 217 (0.87) | 203 (0.81) | 0.006 |
| Stage 4 | 328 (1.29) | 1498 (1.01) | 0.026 | 318 (1.27) | 281 (1.12) | 0.014 |
| Alcohol use | 3205 (12.6) | 3721 (2.50) | 0.378 | 2879 (11.5) | 2883 (11.5) | 0.001 |
| Nicotine dependence | 6493 (25.6) | 10,978 (7.38) | 0.481 | 6166 (24.6) | 6148 (24.5) | 0.002 |
| Tobacco use | 2384 (9.35) | 3242 (2.18) | 0.301 | 2141 (8.54) | 2128 (8.49) | 0.002 |
Standardized difference.
Standard deviation.
American Joint Committee on Cancer.
Patients in the antipsychotics cohort had a 34% increased risk of mortality compared to those without antipsychotics use (HR 1.34, 95% CI 1.29‐1.39, log‐rank P < .001) (Supplemental Figure S3, available online). Patients in the antipsychotics cohort were also significantly more likely to receive radiation therapy (OR 2.47, 95% CI 2.19‐2.78) and chemotherapy (OR 4.26, 95% CI 3.88‐4.69) compared to those in the no antipsychotics cohort. Patients in the antipsychotics cohort had an increased risk of locoregional recurrence (OR 1.36, 95% CI 1.30‐1.43), distant metastasis (OR 1.72, 95% CI 1.59‐1.87), failure to thrive (OR 2.41, 95% CI 2.32‐2.73), and hospice enrollment (OR 3.17, 95% CI 2.66‐3.76) compared to patients in the no antipsychotics cohort (Table 3).
Table 3.
Head and Neck Cancer Outcomes in Patients on Antipsychotics Versus Patients Not on Antipsychotics
| All time post‐HNCa diagnosis | Within 1 y of HNC diagnosis | Within 3 y of HNC diagnosis | |||||||
|---|---|---|---|---|---|---|---|---|---|
| N (%) | N (%) | N (%) | |||||||
| Outcome | Antipsychotics cohort N = 25,077 | No antipsychotics cohort N = 25,077 | Odds ratio (95% CI) | Antipsychotics cohort N = 25,077 | No antipsychotics cohort N = 25,077 | Odds ratio (95% CI) | Antipsychotics cohort N = 25,077 | No antipsychotics cohort N = 25,077 | Odds ratio (95% CI) |
| Recurrence in LNsb | 4059 (18.1) | 3392 (13.9) | 1.36 (1.30‐1.43) | 3443 (15.3) | 2984 (12.3) | 1.30 (1.20‐1.31) | 3825 (17.0) | 3228 (13.3) | 1.34 (1.28‐1.41) |
| Recurrence in lung | 1561 (6.40) | 950 (3.80) | 1.72 (1.59‐1.87) | 938 (3.80) | 593 (2.40) | 1.64 (1.48‐1.82) | 1366 (5.60) | 824 (3.30) | 1.73 (1.59‐1.89) |
| Radiation therapy | 931 (3.80) | 392 (1.60) | 2.47 (2.19‐2.78) | 655 (2.70) | 297 (1.20) | 2.27 (1.98‐2.61) | 822 (3.30) | 344 (1.40) | 2.48 (2.18‐2.81) |
| Chemotherapy | 2111 (9.00) | 563 (2.30) | 4.26 (3.88‐4.69) | 1836 (7.80) | 496 (2.00) | 4.17 (3.77‐4.61) | 2014 (8.50) | 539 (2.20) | 4.23 (3.84‐4.66) |
| Failure to thrive | 2031 (8.30) | 864 (3.50) | 2.41 (2.32‐2.73) | 1287 (5.20) | 549 (2.20) | 2.46 (2.22‐2.72) | 1674 (6.80) | 699 (2.80) | 2.54 (2.32‐2.78) |
| Hospice | 533 (2.10) | 172 (0.70) | 3.17 (2.66‐3.76) | 286 (1.10) | 92 (0.40) | 3.13 (2.48‐3.96) | 391 (1.60) | 117 (0.50) | 3.36 (2.74‐4.13) |
Head and neck cancer.
Lymph nodes.
Discussion
There are approximately 60,000 new cases of HNC each year in the United States resulting in about 12,000 deaths. 17 The link between schizophrenia and cancer risk is unclear, with the literature mixed on whether or not schizophrenia is associated with an increased risk of developing cancer at multiple subsites. 18 , 19 In this study, we identified a consistent incidence of HNC among patients with schizophrenia, with a slight peak in 2012. There were no major changes in HNC diagnostic criteria, data acquisition, or screening recommendations during this time that explain this increase in incidence, though it has been previously noted in the literature. 20 , 21 Though there are years (2012, 2013, and 2021) in which the incidence of HNC appears to be increased among patients with schizophrenia compared to the general population, this study did not identify an apparent overall difference in incidence.
This study did, however, identify a significantly increased risk of mortality among HNC patients with schizophrenia. Patients with schizophrenia have an increased risk of premature death compared to the general population. 22 Such increased mortality risk is likely a by‐product of higher rates of comorbidities such as metabolic syndrome and diabetes, increased likelihood of completing suicide, cardiovascular risks associated with antipsychotic use, and decreased access to regular medical care. 23 , 24 , 25 This study found a significant increase in risk of death even within 1 year of HNC diagnosis, indicating that those with schizophrenia may be more likely to die of cancer. Indeed, patients with schizophrenia have been found to have an increased risk of mortality due to lung, colon, and breast cancer. 10 , 18 Some of this increased mortality risk may be secondary to reduced likelihood of accessing and completing treatment. 26 However, our analysis found that HNC patients with chronic usage of antipsychotics were more likely than those without to receive radiation and chemotherapy. This finding, while initially counterintuitive, may be because patients with schizophrenia present with later‐stage disease than those without. Despite propensity matching based on AJCC cancer stage, few patients had a stage available in the database; more than 95% of patients in both cohorts had an unknown cancer stage. Additionally, the TriNetX database lacks information regarding completion or duration of treatment. It is likely that patients with schizophrenia were more likely to present with late‐stage HNC and thus qualify for multimodal treatment. The increased risk of locoregional recurrence and distant metastases in this cohort further supports the notion that patients with schizophrenia are more likely to present with later‐stage disease and less likely to complete treatment. More research is needed to clarify the tumor stage at presentation, treatment modalities offered, and treatment completion to fully understand the landscape of treatment efficacy, or lack thereof, in this population.
Conversely, it is possible that patients with schizophrenia are less likely to be offered or accept surgical treatment and thus receive other cancer‐directed treatments. Several studies have found that cancer patients with schizophrenia are less likely to undergo surgical treatment. 27 , 28 Surgeons may feel that the risks of surgery outweigh the benefits in this population. Indeed, patients with schizophrenia have worse surgical outcomes compared to the general population in terms of postoperative complications and mortality. 29 Given the heavy burden of perioperative care in HNC patients (ie, long‐term management of gastronomy and tracheostomy tubes and extensive rehabilitation services), surgeons may deem patients with significant psychiatric burden to be poor surgical candidates. Several studies have shown patients with significant psychiatric comorbidities to have acceptable surgical outcomes when combined with proper inpatient perioperative psychiatric support. 30 , 31 Appropriate surgical treatment likely requires a specialized multidisciplinary team to deliver oncologic, surgical, social, and psychiatric care while inpatient. A benefit of surgical treatment in this population is the ability to complete treatment before discharge, as opposed to systemic treatments that require frequent follow‐up. High‐quality studies are needed to show how such a multidisciplinary team can effectively deliver care to this vulnerable patient population. Nevertheless, for patients with schizophrenia with potentially curable cancers, surgical treatment with significant inpatient psychiatric support should be pursued, as this avenue increases the likelihood of treatment completion and thus improved outcomes.
Patients with chronic usage of antipsychotics were also significantly more likely to be diagnosed with failure to thrive. This could be a factor of poor participation in cancer treatment leading to more advanced disease. Further, patients with severe psychiatric disease may be less likely to access other forms of supportive care, such as nutritional counseling, which is often key for HNC patients. HNC and its treatments have a significant impact on swallow function, often requiring dietary modifications and supplementary nutrition via a gastronomy tube. Patients with schizophrenia are less likely to have access to medical care and less likely to be compliant with medical care, leading to worse health outcomes and more adverse events. 32 , 33 Thus, patients with HNC and schizophrenia may be less likely to participate in speech and swallow therapy and be able to properly care for feeding tubes, increasing the risk for malnutrition and failure to thrive.
Finally, our study found an increased rate of hospice enrollment among HNC patients with chronic usage of antipsychotics. This, again, could be associated with an increased risk of presenting with late‐stage disease and decreased compliance with medical treatments, leading to poor outcomes. A prior study in France found that cancer patients with schizophrenia were significantly more likely to receive palliative care and less likely to receive aggressive cancer treatment in the final month of life compared to cancer patients without schizophrenia. 34 There may be a provider preference for transition to palliative care rather than high‐intensity end‐of‐life care due to the risks of toxicity and difficulty with completing aggressive treatment in this population. 26 Palliative care centers are often designed to deal with complex patients lacking the capacity to make medical decisions and/or social support, so providers may be more likely to involve palliative care earlier for patients with psychotic disorders. 35 Additionally, patients with schizophrenia may be more likely to be hostile towards receiving aggressive treatments at the end of life. Finally, one study found an increased rate of do‐not‐resuscitate orders among cancer patients with schizophrenia compared to those with cancer without schizophrenia. 36 As such, patient preference may guide the higher likelihood of hospice enrollment in this population.
People with schizophrenia are particularly vulnerable, facing significant socioeconomic and health disparities. Recognizing that HNC patients with schizophrenia may have difficulty accessing and complying with appropriate medical care is an important first step to ensuring equal care. Improving outreach to patients with severe mental illness may help identify cancer earlier and link patients with appropriate care. 37 Additionally, taking an integrated person‐centered approach involving collaboration of psychiatrist, social workers, oncologists, and caregivers can build a support network to ensure appropriate mental health and cancer care. 38 Additionally, ensuring safe housing, access to food, mental healthcare, and psychosocial support is an important aspect of care for this population. 39 Future research should focus on developing multidisciplinary teams that are able to provide well‐rounded cancer care to patients with schizophrenia. Additionally, given the increased rate of smoking and alcohol use in this population, improved screening guidelines may be appropriate.
This study is limited by its retrospective nature. The inability to match most patients based on disease severity is a significant limitation, as we are unable to determine if patients with schizophrenia have worse HNC outcomes due to late presentation or due to issues accessing and complying with treatment. Patients with schizophrenia are significantly more likely to present with later‐stage or advanced cancer than those without schizophrenia, in part due to difficulty accessing and navigating healthcare. 40 , 41 , 42 Nevertheless, both are important considerations for this vulnerable patient population who likely lacks access to preventative care and multidisciplinary treatment. Future studies should focus on clarifying the disease stage at presentation and treatment compliance to further elucidate this gap. Another limitation of this study is the utilization of antipsychotic use as a surrogate for schizophrenia. The cohort of HNC patients with schizophrenia was small, and thus it would not have been possible to look at differences in the incidence of rare outcomes. The use of a cohort of patients with antipsychotic use as a surrogate for a cohort of patients with schizophrenia was previously published. 12 Furthermore, we analyzed risk of mortality, a relatively common outcome, among HNC patients with and without antipsychotic use and among HNC patients with and without schizophrenia and found the results to be similar, lending credence to the use of an antipsychotics use cohort as a surrogate. Additionally, the TriNetX database relies upon billing codes; the accuracy of our analysis is reliant on appropriate documentation in the electronic health record. Though cohorts were 1:1 propensity‐matched, missing data introduces another avenue for bias. Most patients lacked documentation of tumor stage, and information regarding tumor size, nodal involvement, and presence of metastases was unavailable, which limits the ability to fully match based on stage at presentation. Though the TriNetX database contains data from 68 healthcare organizations, including inpatient and outpatient sites, it is not necessarily representative of the entire population of the United States. Additionally, TriNetX only contains follow‐up data for up to 20 years, so findings regarding long‐term outcomes may be limited. Despite these limitations, this is the first large‐scale analysis to our knowledge that studies the incidence and outcomes of HNC among patients with schizophrenia in the United States.
Conclusion
Patients with HNC and schizophrenia are at significant risk of increased mortality, disease recurrence, and failure to thrive post‐HNC diagnosis. This increased risk of poor outcomes among patients with schizophrenia highlights the vulnerability of this population and the need for a greater safety net. Appropriate oncologic care for patients with HNC and schizophrenia likely involves a multidisciplinary team able to focus on the psychosocial, supportive, psychiatric, and oncologic needs of these patients.
Author Contributions
Talia A. Wenger, conceptualization, methodology, formal analysis, investigation, data curation, writing—original draft, writing—review and editing, visualization; Stephanie Wong, formal analysis, investigation, writing—original draft, writing—review and editing; Shu‐Yun Cheng, formal analysis, investigation, writing—original draft, writing—review and editing; Liyang Tang, investigation, writing—original draft, writing—review and editing; Daniel Kwon, investigation, writing—original draft, writing—review and editing; Niels Kokot, investigation, writing—original draft, writing—review and editing; Yang Chai, investigation, visualization, writing—original draft, writing—review and editing; Uttam Sinha, investigation, writing—original draft, writing—review and editing; Albert Y. Han, conceptualization, methodology, formal analysis, investigation, writing—original draft, writing—review and editing, supervision, project administration.
Disclosures
Competing interests
The authors declare no conflicts of interest.
Funding source
None.
Supporting information
Supporting Information.
This article was presented as a poster presentation at the Medical Oncology Association of Southern California's Annual Oncology Summit & Research Symposium; March 22, 2025; Newport Beach, California.
References
- 1. Jablensky A. The 100‐year epidemiology of schizophrenia. Schizophr Res. 1997;28(2):111‐125. 10.1016/S0920-9964(97)85354-6 [DOI] [PubMed] [Google Scholar]
- 2. 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. 10.1001/jamapsychiatry.2015.1737 [DOI] [PubMed] [Google Scholar]
- 3. Lichtermann D, Ekelund J, Pukkala E, Tanskanen A, Lönnqvist J. Incidence of cancer among persons with schizophrenia and their relatives. Arch Gen Psychiatry. 2001;58(6):573‐578. 10.1001/archpsyc.58.6.573 [DOI] [PubMed] [Google Scholar]
- 4. Ji J, Sundquist K, Ning Y, Kendler KS, Sundquist J, Chen X. Incidence of cancer in patients with schizophrenia and their first‐degree relatives: a population‐based study in Sweden. Schizophr Bull. 2013;39(3):527‐536. 10.1093/schbul/sbs065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Catts VS, Catts SV, O'Toole BI, Frost ADJ. Cancer incidence in patients with schizophrenia and their first‐degree relatives—a meta‐analysis. Acta Psychiatr Scand. 2008;117(5):323‐336. 10.1111/j.1600-0447.2008.01163.x [DOI] [PubMed] [Google Scholar]
- 6. Jane‐Llopis E, Matytsina I. Mental health and alcohol, drugs and tobacco: a review of the comorbidity between mental disorders and the use of alcohol, tobacco and illicit drugs. Drug Alcohol Rev. 2006;25(6):515‐536. 10.1080/09595230600944461 [DOI] [PubMed] [Google Scholar]
- 7. Kisely S, Crowe E, Lawrence D. Cancer‐related mortality in people with mental illness. JAMA Psychiatry. 2013;70(2):209‐217. 10.1001/jamapsychiatry.2013.278 [DOI] [PubMed] [Google Scholar]
- 8. Guan NC, Termorshuizen F, Laan W, et al. Cancer mortality in patients with psychiatric diagnoses: a higher hazard of cancer death does not lead to a higher cumulative risk of dying from cancer. Soc Psychiatry Psychiatr Epidemiol. 2013;48(8):1289‐1295. 10.1007/s00127-012-0612-8 [DOI] [PubMed] [Google Scholar]
- 9. Konishi T, Fujiogi M, Michihata N, et al. Breast cancer surgery in patients with schizophrenia: short‐term outcomes from a nationwide cohort. Br J Surg. 2021;108(2):168‐173. 10.1093/bjs/znaa070 [DOI] [PubMed] [Google Scholar]
- 10. Nordentoft M, Plana‐Ripoll O, Laursen TM. Cancer and schizophrenia. Curr Opin Psychiatry. 2021;34(3):260‐265. 10.1097/YCO.0000000000000697 [DOI] [PubMed] [Google Scholar]
- 11.TriNetX advanced analytics. TriNetX. July 17, 2024. Accessed August 27, 2024. https://trinetx.com/solutions/live‐platform/features/advanced‐analytics/
- 12. Den J, Nelson N, Klimberg VS. Analysis of the incidence and outcomes of breast cancer in women with schizophrenia. Am J Surg. 2025;239:116050. 10.1016/j.amjsurg.2024.116050 [DOI] [PubMed] [Google Scholar]
- 13. Parker M, Kalluri A, Jiang K, et al. Prevalence, treatment patterns, and survival of patients with brain metastases from small cell lung cancer: a retrospective study using the TriNetX Oncology Database. Neurooncol Pract. 2025;12(2):257‐270. 10.1093/nop/npae095 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Wüster J, Heene S, Brandenburg L, et al. Acetylsalicylic acid and head and neck cancer: risk, five‐year survival, and recurrence in over 11,000,000 patients—a retrospective case–control study. Cancers. 2025;17(13):2065. 10.3390/cancers17132065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lorenz FJ, Mahase SS, Miccio J, King TS, Pradhan S, Goyal N. Update on adherence to guidelines for time to initiation of postoperative radiation for head and neck squamous cell carcinoma. Head Neck. 2023;45(7):1676‐1691. 10.1002/hed.27380 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Wenger TA, Gao J, Nurimba M, Phung PG, Sinha UK. Palliative care utilization among head and neck cancer patients: a population‐based analysis. Oral Oncol. 2025;162:107205. 10.1016/j.oraloncology.2025.107205 [DOI] [PubMed] [Google Scholar]
- 17. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer J Clin. 2015;65(1):5‐29. 10.3322/caac.21254 [DOI] [PubMed] [Google Scholar]
- 18. Casanovas F, Dinamarca F, Posso M, et al. Cancer characteristics in patients with schizophrenia: a 25‐year retrospective analysis. Psychiatry Res. 2024;342:116206. 10.1016/j.psychres.2024.116206 [DOI] [PubMed] [Google Scholar]
- 19. Zhuo C, Zhuang H, Gao X, Triplett PT. Lung cancer incidence in patients with schizophrenia: meta‐analysis. Br J Psychiatry. 2019;215(6):704‐711. 10.1192/bjp.2019.23 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Suh JD, Cho JH. Trends in head and neck cancer in South Korea between 1999 and 2012. Clin Exp Otorhinolaryngol. 2016;9(3):263‐269. 10.21053/ceo.2015.01123 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.NC SCHS: Statistics and reports: cancer incidence in North Carolina 2015. Accessed February 28, 2025. https://schs.dph.ncdhhs.gov/data/cancer/incidence/2015.htm
- 22. Correll CU, Solmi M, Croatto G, et al. Mortality in people with schizophrenia: a systematic review and meta‐analysis of relative risk and aggravating or attenuating factors. World Psychiatry. 2022;21(2):248‐271. 10.1002/wps.20994 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Levinson Miller C, Druss BG, Dombrowski EA, Rosenheck RA. Barriers to primary medical care among patients at a community mental health center. Psychiatr Serv. 2003;54(8):1158‐1160. 10.1176/appi.ps.54.8.1158 [DOI] [PubMed] [Google Scholar]
- 24. De Hert M, Detraux J, van Winkel R, Yu W, Correll CU. Metabolic and cardiovascular adverse effects associated with antipsychotic drugs. Nat Rev Endocrinol. 2011;8(2):114‐126. 10.1038/nrendo.2011.156 [DOI] [PubMed] [Google Scholar]
- 25. Vancampfort D, Stubbs B, Mitchell AJ, et al. Risk of metabolic syndrome and its components in people with schizophrenia and related psychotic disorders, bipolar disorder and major depressive disorder: a systematic review and meta‐analysis. World Psychiatry. 2015;14(3):339‐347. 10.1002/wps.20252 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Irwin KE, Henderson DC, Knight HP, Pirl WF. Cancer care for individuals with schizophrenia. Cancer. 2014;120(3):323‐334. 10.1002/cncr.28431 [DOI] [PubMed] [Google Scholar]
- 27. Protani MM, Alotiby MKN, Seth R, et al. Colorectal cancer treatment in people with severe mental illness: a systematic review and meta‐analysis. Epidemiol Psychiatr Sci. 2022;31:e82. 10.1017/S2045796022000634 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Lin J, Sato S, Aso S, Fushimi K, Matsui H, Yasunaga H. Association of comorbid schizophrenia with cancer stage at admission, treatments, length of stay, and 30‐day in‐hospital mortality in patients with pancreatic cancer: a retrospective matched‐pair cohort study in Japan. Eur J Cancer. 2025;222:115468. 10.1016/j.ejca.2025.115468 [DOI] [PubMed] [Google Scholar]
- 29. Liao CC, Shen WW, Chang CC, Chang H, Chen TL. Surgical adverse outcomes in patients with schizophrenia: a population‐based study. Ann Surg. 2013;257(3):433‐438. 10.1097/SLA.0b013e31827b9b25 [DOI] [PubMed] [Google Scholar]
- 30. Obuchi T, Okabayashi K, Imakiire T, Yoneda S, Iwasaki A. The outcomes of surgery in lung cancer patients with schizophrenia. Surg Today. 2014;44(5):855‐858. 10.1007/s00595-013-0599-0 [DOI] [PubMed] [Google Scholar]
- 31. Miñambres I, Rubio‐Herrera MÁ, Nicolau J, et al. Outcomes of bariatric surgery in patients with schizophrenia. Nutrients. 2024;16(15):2487. 10.3390/nu16152487 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Daumit GL, Pronovost PJ, Anthony CB, Guallar E, Steinwachs DM, Ford DE. Adverse events during medical and surgical hospitalizations for persons with schizophrenia. Arch Gen Psychiatry. 2006;63(3):267‐272. 10.1001/archpsyc.63.3.267 [DOI] [PubMed] [Google Scholar]
- 33. Brugha TS, Wing JK, Smith BL. Physical health of the long‐term mentally ill in the community. Br J Psychiatry. 1989;155:777‐781. 10.1192/bjp.155.6.777 [DOI] [PubMed] [Google Scholar]
- 34. Fond G, Salas S, Pauly V, et al. End‐of‐life care among patients with schizophrenia and cancer: a population‐based cohort study from the French national hospital database. Lancet Public Health. 2019;4(11):e583‐e591. 10.1016/S2468-2667(19)30187-2 [DOI] [PubMed] [Google Scholar]
- 35. Donald EE, Stajduhar KI. A scoping review of palliative care for persons with severe persistent mental illness. Palliat Support Care. 2019;17(4):479‐487. 10.1017/S1478951519000087 [DOI] [PubMed] [Google Scholar]
- 36. Ganzini L, Socherman R, Duckart J, Shores M. End‐of‐life care for veterans with schizophrenia and cancer. Psychiatr Serv. 2010;61(7):725‐728. 10.1176/ps.2010.61.7.725 [DOI] [PubMed] [Google Scholar]
- 37. Bellman V, Russell N, Depala K, et al. Challenges in treating cancer patients with unstable psychiatric disorder. World J Oncol. 2021;12(5):137‐148. 10.14740/wjon1402 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Irwin KE, Park ER, Fields LE, et al. Bridge: person‐centered collaborative care for patients with serious mental illness and cancer. Oncologist. 2019;24(7):901‐910. 10.1634/theoncologist.2018-0488 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Shalev D, Brewster KK, Levenson JA. End‐of‐life care for patients with schizophrenia #332. J Palliat Med. 2017;20(7):787‐788. 10.1089/jpm.2017.0164 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Wootten JC, Wiener JC, Blanchette PS, Anderson KK. Cancer incidence and stage at diagnosis among people with psychotic disorders: systematic review and meta‐analysis. Cancer Epidemiol. 2022;80:102233. 10.1016/j.canep.2022.102233 [DOI] [PubMed] [Google Scholar]
- 41. Suvisaari J, Manderbacka K, Keskimäki I, Arffman M, Ahlgren‐Rimpiläinen A, Pukkala E. 152. Cancer care and mortality in patients with schizophrenia, substance use, and mood disorders in Finland. Schizophr Bull. 2017;43(suppl 1):S79. 10.1093/schbul/sbx021.210 [DOI] [Google Scholar]
- 42. Davis LE, Bogner E, Coburn NG, et al. Stage at diagnosis and survival in patients with cancer and a pre‐existing mental illness: a meta‐analysis. J Epidemiol Community Health. 2020;74(1):84‐94. 10.1136/jech-2019-212311 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information.