Individual Submission Summary
Share...

Direct link:

Assessing the Impact of the National 988 Emergency Hotline on Youth Suicide

Saturday, November 7, 10:15 to 11:45am, Property: Boston Marriott Copley Place, Floor: 3rd Floor, Room: Brandeis

Abstract

Youth suicide is a national emergency. Between 2001 and 2024, suicide deaths among 15–24-year-olds more than doubled and the rate rose nearly 60 percent. In July 2022, the federal government launched 988 as a national three-digit crisis hotline. This was the most significant investment in suicide prevention infrastructure in U.S. history. Yet no study has produced credible causal estimates of its effect on suicide. The central challenge is that 988 launched everywhere simultaneously, leaving no obvious comparison group. I exploit two sources of variation to provide a solution to this identification problem. First, states entered the 988 era with dramatically different crisis system capacity: Some with well-funded, high-performing hotlines where 988 was largely a rebrand, and others where launch triggered sharp improvements in call answer rates. For example, less than 20 percent of calls to suicide hotlines were answered in Illinois before 988, but this increased to more than 80 percent in the 12 months starting July 2024. Similarly, large increases occured in Texas, New York, Florida, and Georgia. But, with robust systems in place, California, North Carolina and Pennsylvania were already answering over 80 percent of calls in-state before 988. This cross-state variation in treatment intensity, driven by pre-existing infrastructure is the foundation of my identification strategy. My primary approach treats system quality as a continuous treatment and estimates dose-specific average treatment effects following Callaway, Goodman-Bacon, and Sant'Anna (2024). Using in-state call answer rates and abandoned call rates as measures of crisis system quality, I compare within-state changes in youth suicide mortality against counterfactual trends implied by states on different treatment paths to isolate the effect of policy-induced improvements in hotline access.

A second strategy addresses the concern that states investing in 988 infrastructure may differ on unobserved dimensions that also affect suicide trends. I instrument for call answer rates using state-level policy changes that expanded 988 capacity. I compile data on funding legislation, staffing mandates, and designated appropriations at the state-year level. I use an IV-DiD design to estimate local average treatment effect for states whose hotline quality responded to these policy levers, sharpening the causal interpretation and providing a robustness check on the continuous DiD estimates.The primary outcome is suicide mortality among 15–24-year-olds from CDC WONDER state-month vital statistics data. I also examine suicide attempts using the Youth Risk Behavior Surveillance System. Where I detect effects, I translate estimates into an implied cost per life-year saved, a metric directly actionable for state legislatures weighing appropriations for crisis services. This paper will produce causal evidence on whether 988 has reduced youth suicide mortality. The convergence of two identification strategies, one leveraging continuous cross-state variation in system quality, the other exploiting discrete policy-induced shifts, permits bounding the range of effects and assesses robustness to alternative identifying assumptions. This is among the most consequential unanswered questions in U.S. public health policy, and the answer has direct implications for the states and policymakers still deciding how much to invest in crisis response infrastructure.

Author