Suicide incidence and risk among head and neck cancer survivors based on urban vs. rural residence status

Presentation: S009
Topic: Functional Outcomes / Quality
Type: Oral
Date: Friday, July 23, 2021
Session: 2:10 PM - 3:00 PM Function / Quality
Authors: Nosayaba Osazuwa-Peters, BDS, PhD, MPH, CHES1; Justin M Barnes, MD, MS2; Somtochi Okafor, MD1; Derian B Taylor, BS3; Adnan S Hussaini, MD4; Eric Adjei Boakye, PhD5; Matthew C Simpson, MPH3; Evan M Graboyes, MD, MPH6; Trinitia Cannon, MD1; Walter T Lee, MD, MS1
Institution(s): 1Duke University School of Medicine; 2Washington University in St. Louis; 3Saint Louis University School of Medicine; 4George Washington University; 5Southern Illinois University School of Medicine; 6Medical University of South Carolina


Head and neck cancer (HNC) care is often centralized, and a proportion of patients need to travel to receive care. Additionally, survivors have rate of suicide that is almost four times that of the general population, and it is unknown whether rate and/or risk differ based on whether patients reside in rural areas vs. urban or metropolitan areas. We sought to determine whether the risk of suicide differed by rural-urban-metropolitan residence status.


We queried the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute for patients with HNC aged 18-74 years, who were diagnosed between 2000 and 2016. Death due to suicide was determined by ICD-10 codes (U03, X60-X84, Y87.0) and the cause of death recode (50220). Residence status was determined using the 2013 Rural Urban Continuum Codes (RUCC). Standardized incidence ratios (SIRs) of suicide were also obtained using SEER*Stat. To account for competing risks, we compared risks between groups using Fine and Gray proportional hazards regression.


We identified 416 cases of suicide among 134,510 HNC survivors. Suicide rates were 59.2, 64.0, and 126.7 (per 100,000 person-years) in metropolitan, urban, and rural counties, respectively. Incidence rate of suicide was significantly higher among HNC survivors compared to the general United States population (SIR for metropolitan residents: 2.8; 95% CI: 2.5, 3.1; SIR for urban residents: 2.8; 95% CI: 2.1, 3.7; and SIR for rural residents: 5.5; 95% CI: 3.1, 9.0). In Fine and Gray analyses, relative to metropolitan counties, the risk of dying by suicide was double among HNC survivors who reside in rural counties compared with those residing in metropolitan counties (sdHR = 2.09; 95% CI: 1.22, 3.57; p for urban*rural difference = .024). However, there was no significant difference between metropolitan and urban residents (sdHR = 1.06; 95% CI: 0.78, 1.42).


The risk of suicide is significantly elevated among HNC patients living in rural counties, which may be related to socioeconomic and demographic differences in this group. Further studies exploring the mechanisms underlying these trends and targeted interventions to mitigate these risks are warranted.