INTRODUCTION

Data on how and when homeless individuals die are limited. Knowledge about cause-specific mortality in the general population has led to policy changes that have increased Americans’ life expectancy. This is not true for people experiencing homelessness, despite their increased risk for premature mortality.1 We examined deaths among people experiencing homelessness in San Francisco (SF)—which has one of the highest rates of homelessness in the USA2—from 2016 to 2018.

METHODS

We used mortality data from the SF Office of the Chief Medical Examiner (OCME).3 OCME reports include manner of death—accident, natural, homicide, suicide or undetermined; cause, location, and date of death; demographics; and an investigative narrative incorporating medical records review and forensic death investigation. We include autopsy and toxicology reports when available.

We linked OCME records to a SF Department of Public Health (SFDPH) database—the Coordinated Care Management System (CCMS). SFDPH creates a record for anyone who uses county behavioral health, housing, urgent/emergent medical, mental health, or substance use services, who have a jail visit, or who are identified as homeless. CCMS defines homelessness based on homeless service use (e.g., shelter, navigation services) and self-reported housing status collected during healthcare encounters.4 CCMS linkage allowed us to obtain homelessness history, and medical, behavioral, county jail, and shelter use prior to death for each decedent.

We categorized an SF OCME death as homeless if it met at least one criterion for the US Department of Housing and Urban Development definition of homelessness death based on our data.5 Using the linked 2016–2018 OCME-CCMS data set, we created a homeless death cohort for each study year. We calculated yearly homeless death estimates and used descriptive statistics to understand trends in demographics, causes and locations of death, toxicology findings, and use of health and social services before death.

We obtained institutional review board approval and adhered to the Protected Health Information and Code of Federal Regulations (CFR) 42 part 2 protocols governing the use of substance use disorder data.

RESULTS

In SF, there were 128 homeless deaths in 2016, 128 homeless deaths in 2017, and 135 homeless deaths in 2018. Most decedents were male, White, and between ages 40–60 (Table 1).

Table 1 Characteristics of Deaths Among People Experiencing Homelessness in San Francisco, by year, 2016–2018
Full size table

The most common cause of death in each year was acute drug toxicity (i.e., unintended overdose, including alcohol). Traumatic injuries (e.g., gunshot wounds, blunt force injuries, falls) were the second leading cause of death each year, which included homicides (10% of all deaths) and suicides (5% of all death). Cardiovascular disease was the 3rd leading cause (Table 2).

Table 2 Causes of Death Among People Experiencing Homelessness in San Francisco, by year, 2016–2018
Full size table

Complete toxicology reports were available for 76% of all cases. Over half were positive for methamphetamine (54%), followed by opioids (46%) and alcohol (30%). From 2016 to 2018, there was a five-fold increase in reports positive for fentanyl. Most (76%) of the toxicology samples were positive for multiple substances.

We found 42% of participants had been homeless for over 10 years (continuously or intermittently) at their time of death. Most (70%) had been unsheltered in the year before death. Four percent had lived in county permanent supportive housing at some point prior to death, most often for under a year. Death locations were concentrated in three SF neighborhoods. Over 25% of all deaths occurred in public view.

In the year before death, 63% used medical services including 27% who had at least one encounter with county jail health services. About 20% had used a mental health service and 16% used a substance use treatment service. While 10% had no service use documented in CCMS, a third (32%) had been among the top 5% of SF’s highest users of urgent and emergency health services.

DISCUSSION

We identified risk factors associated with death among people experiencing homelessness in SF via a standardized methodology that allowed us to track deaths over multiple years.

While heart disease and cancer are the first and second causes of death for California’s general population6, acute intoxication (unintended overdose) and trauma (including homicide and suicide rates over triple that of the US general population) were the leading causes of homeless deaths. Most died unsheltered, and in three specific neighborhoods. These stark differences in cause-specific mortality should inform death prevention efforts within the homeless population differently than the general population.