Projected Outcomes of Reducing Federal Funding for Syringe Service Programs via Executive Order

Authors: Kirk B. Fetters, Pranav Padmanabhan, Kristina Yamkovoy, Xiaoyu Guan, Sarah E. Scott, Lauren Kerr, Kathleen Joseph, Gwenyth L. Day, Marina Plesons, Tyler S. Bartholomew, Hansel E. Tookes, Alia Al-Tayyib, Joshua A. Barocas

Abstract

Question

What are the potential mortality and overdose outcomes of halting federal funding for syringe service programs?

Findings

In this decision analytical model including 3 694 500 hypothetical persons who inject drugs, eliminating federal funding for programs that provide syringe exchange and other harm reduction interventions was projected to increase deaths among this population by 1100 to 39 600 in the next 5 years. During that same period, opioid overdose deaths are expected to increase by 500 to 15 600.

Meaning

These findings suggest that eliminating federal funding for syringe service programs may increase all-cause deaths and fatal opioid overdoses among persons who inject drugs.

Introduction

Since the 1990s, the US has experienced significant opioid-related morbidity and mortality. While there have been fluctuations in overdose deaths in the last quarter century, the first sustained decrease in overdose deaths has occurred from 2023 through 2025.^1,2,3^ Interventions such as medications for opioid use disorder (MOUD) and harm reduction services have been shown to decrease overdose risk and infectious complications.^4,5,6,7^

Harm reduction interventions are approaches to care that seek to reduce the harms associated with drug use while prioritizing patients’ autonomy. Interventions include services such as syringe exchange, overdose education, and naloxone distribution. In addition to syringe exchange, organizations known as syringe service programs (SSPs) provide sterile equipment, wound care services, drug checking, naloxone distribution, and linkage to MOUD and infectious disease prevention and care services.^4,8,9,10^

SSPs receive a substantial portion of funding from the federal government, but remain underfunded given the magnitude of the overdose epidemic and, thus, are unable to provide services to all in need.^11,12,13^ One systematic review noted that there is low coverage of syringe services for persons who inject drugs in North America, with only 41 syringes distributed per person who injects drugs per year.^14^ Increasing SSP budgets lead to increases in syringe and naloxone distribution and the size of the population they can reach.^11,12^ Elimination of funding could lead to closure or downscaling of many SSPs, particularly in rural and suburban areas where local or philanthropic funding may be limited but services are still needed.^11^

Until 2009, there was a ban on the use of federal funds for purchasing needles and syringes by SSPs.^15^ The Obama administration allowed federal funds to be used from 2009 to 2011, but the ban was reinstated until 2016, after an HIV outbreak among persons who inject drugs in Scott County, Indiana, catalyzed new attention to evidence-based solutions to prevent HIV acquisition in this population.^16,17,18^ Since 2016, there has been ongoing federal support for syringe exchange.^11,18,19,20^

On July 24, 2025, a federal executive order ended discretionary grants from the Substance Abuse and Mental Health Services Administration for “programs that fail to achieve adequate outcomes, including so-called ‘harm reduction,’” and threatened to bring civil or criminal cases against organizations that “knowingly distribute drug paraphernalia.”^21^ The executive order and a follow-up “Dear Colleague” letter from the Substance Abuse and Mental Health Services Administration stated that the US government would not provide funding for any other supplies to “promote or facilitate drug use, such as sterile water, saline, or ascorbic acid.”^21,22^

The impact that this executive order might have on the overdose epidemic remains uncertain. Simulation modeling offers a way to estimate the potential future impact of alternative policy environments.^20,23,24,25^ We used a validated microsimulation model to estimate the potential outcomes of reducing federal funding for SSPs on mortality among persons who inject drugs.

Methods

Overview

In this decision analytical study, we used a validated closed cohort microsimulation model to assess the potential outcomes of reducing federal funding for SSPs among a cohort of persons who inject drugs in the US during a 5-year period (August 1, 2025, to August 30, 2030). We created hypothetical scenarios consisting of different levels of disruptions to SSPs due to federal funding cuts and compared them with a scenario in which federal funding remained stable with no disruptions. The Colorado Multiple Institutional Review Board deemed this study exempt from ethics review and informed consent because it did not involve human participants. We used the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline.

Given uncertainty and variability in potential service disruptions attributable to loss of federal funding, we modeled 3 hypothetical scenarios under 2 different (1) SSP services reduced by 11% (referred to as the low service disruption case), and (2) SSP services reduced by 80% (referred to as the high service disruption case). These reductions represent lower and upper bounds, respectively, of estimates of federally sourced SSP funding based on published literature,^11,26^ correspondences with SSP providers, and the understanding that some funding from state or local governments originates as federal block grants, which may also be affected by this executive order. We assumed that lost funding corresponds to an equivalent reduction in services provided. Within both cases, we assumed that disruptions began in August 2025 and modeled 3 potential scenarios (total 6 hypothetical scenarios based on each disruption case): (1) funding disruptions remain unchanged for the next 5 years (through August 2030); (2) funding returns to pre–executive order levels after 1 year (August 2026); and (3) 100% of pre–executive order funding returns in January 2029.

We divided the population of persons who inject drugs by SSP use (using and not using SSPs) and baseline injection practices (using and not using practices such as skin cleaning and avoiding syringe sharing) and simulated funding disruptions by varying the proportion of persons who inject drugs and access SSPs. In doing so, we assumed that funding disruptions would limit new injection equipment, naloxone, and MOUD referral services offered at SSPs, reducing the size of the population SSPs could reach. In all scenarios, we assumed a linear relationship between funding levels and proportions of persons who inject drugs receiving SSP services. The primary outcomes of this analysis were 5-year all-cause and overdose mortality, and the secondary outcome was nonfatal overdose.

We performed probabilistic and deterministic sensitivity analyses in which we varied uncertain parameters. We populated the model with national data and calibrated it to be representative of persons who inject drugs in the US in August 2025.

Model Description

In the model, simulated cohorts were assigned demographic and behavioral characteristics reflecting persons who inject drugs in the US. Among those who had a history of any injection drug use, the model assigned a current injection status and injection-related behaviors (low or high injection frequency, skin cleaning, and needle sharing), which governed the risk of severe injection-related infection and overdose (referred to as sequelae). Individuals progressed weekly through the model and encountered probabilities of sequelae, hospitalization, receiving outpatient care, transitioning between injection-related behavior states, and death from all causes. Full model details can be found in eAppendix 1 in Supplement 1.

The model cohort was stratified by sex, age, and injection behavior profile. Only individuals with current injection drug use were at risk of sequelae. Probabilities of sequelae risk were derived using age, sex, and injection behavior profile for people who use and do not use SSPs. Individuals who developed sequelae had a probability of being hospitalized and treated, and all individuals encountered probabilities of linking to outpatient MOUD (eAppendix 1 in Supplement 1). Individuals faced a risk of death from sequelae and competing causes of death, which were not explicitly modeled.

Simulated Cohort

We modeled a large, estimated population of 3 694 500 persons who inject drugs in the US (Table 1).^32^ We divided our simulated population into 4 (1) those who use SSP services at least once a year and regularly use safer injection practices (defined as cleaning skin prior to injection and not sharing needles); (2) those who use SSP services but do not regularly use safer injection practices; (3) those who do not use SSP services but regularly use safer injection practices; and (4) those who do not use SSP services or regularly use safer injection practices (Figure 1). Individuals were placed within a cohort at the beginning of the model simulation.

We derived cohort demographic characteristics and behavioral parameters primarily from the 2022 National HIV Behavioral Surveillance (NHBS) survey of persons who inject drugs in the US. To estimate differences in key parameters between SSP users and nonusers, we used the NHBS survey of persons who inject drugs in a single city in 2024, which was stratified by SSP use. We used published literature to confirm these data, estimate other model parameters and transition probabilities, and identify calibration and validation targets^27,28,29,30,31^ (Table 1 and eAppendixes 2-4 in Supplement 1).

Sequelae of Drug Use and Hospitalization

We estimated rates of fatal and nonfatal overdose from a meta-analysis of national data and local surveillance data.^29,30^ Severe bacterial injection-related infections such as infective endocarditis and severe skin and soft tissue infections were also modeled, but results were not reported herein for clarity.^33^

Outpatient and Harm Reduction Services

We derived probabilities of linkage to outpatient primary and specialty care from a combination of cohort studies.^34,35,36,37^ We modeled the effect of SSP use in 3 key the probability that an overdose is fatal, the probability of MOUD initiation, and the probability of using safer injection practices. These effects are supported by published data and notably reflect delivery of more services than just syringe exchange.^33,38,39^ We assumed that SSP use decreased the probability that a given overdose would be fatal and increased the probability of MOUD initiation, reflecting naloxone distribution and MOUD prescription or referral, respectively. We assumed MOUD retention was equal in both groups.^40^ We did not explicitly model protective effects of MOUD on overdose from acute opioid exposure, although MOUD increases the probability of transitioning to injecting at a lower frequency, which reduces overdose risk. In addition, SSP users were more likely to be categorized as using safer injection practices (cleaning skin prior to injection and not sharing needles) at model initiation and transitioning to using safer injection practices thereafter.

Calibration

We calibrated the model to national 1-year skin and soft tissue infection and infective endocarditis hospitalization incidence, fatal and nonfatal overdose incidence, and all-cause mortality. We assessed external validity by comparing the 1-year simulated fatal overdose incidence and overdose reduction associated with MOUD with the latest national data (eAppendix 3 in Supplement 1).^3^

Statistical Analysis

We simulated funding disruption by varying the proportion of persons who inject drugs who comprised the SSP-using cohorts. In a particular scenario, the proportions of persons who inject drugs in each cohort were determined initially by baseline funding and remained constant throughout periods in which funding remained consistent. If funding was reinstated during the study period, a fraction of individuals not using SSPs were reassigned to SSP-using cohorts at the time of funding reinstatement. We assumed direct linear relationships between the magnitude of funding changes and the number of persons who inject drugs who access SSPs.

For all scenarios, we generated separate model output for each cohort, then calculated weighted means of cohort outputs to create a composite output, with weights corresponding to proportions of persons who inject drugs in each cohort (Table 1). Reported outcomes represented composite outputs (Table 2). For scenarios 2 and 3 in both low and high service disruption cases, composite outputs represented the sum of cohort-specific weighted means in 2 separate periods.

We conducted probabilistic sensitivity analyses in which uncertain parameters were randomly drawn from distributions (eAppendix 5 in Supplement 1). We generated 95% credible intervals (CrI) by calculating weighted means of 1000 iterations of each cohort with a population of 1000 persons who inject drugs each.^33^ We calculated weighted means of outputs corresponding to paired randomly drawn sets of uncertain parameters for each cohort. Pearson correlation coefficients between cohorts ranged from 0.77 to 0.83 in probabilistic sensitivity analysis runs for all outcomes (eAppendix 5 in Supplement 1). The 95% CrIs represent the 2.5th and 97.5th percentiles of weighted composite outputs.

We performed deterministic sensitivity analyses to assess the estimated impact of key parameters on model outcomes (eAppendix 5 in Supplement 1). We focused our deterministic sensitivity analyses on scenario 1 in both high and low service disruption cases, as this is the worst-case scenario, but deterministic sensitivity analyses for scenarios 2 and 3 are reported in eAppendix 5 in Supplement 1. First, we varied relative differences in overdose risk and MOUD initiation probability between those who did and did not use SSPs.^27,41^ Next, we varied the relative proportions of persons who inject drugs who regularly used safer injecting practices within groups of those who did and did not utilize SSPs.^27^

Results

The hypothetical study population consisted of 3 694 500 persons who inject drugs (43.0% female and 57.0% male; mean [SD] age, 49.5 [17.5] years). In the base case scenario in which baseline levels of federal funding prior to August 2025 were unchanged through August 2030, we estimated there would be 784 900 (95% CrI, 758 000-952 800) all-cause deaths, 226 900 (95% CrI, 178 200-288 400) overdose deaths, and 3 722 000 (95% CrI, 3 127 500-4 953 700) nonfatal overdoses. Table 2 details rates per 1000 persons who inject drugs.

Low Service Disruption

First, we reported the estimated changes to the primary and secondary outcomes if federal funding accounts for 11% of SSP total budgets and all of that is lost. In scenario 1, in which the funding disruption lasts until 2030, we estimated there would be 5400 excess all-cause deaths (0.7% increase; 95% CrI, 0-1.2%), 2200 excess overdose deaths (1.0% increase; 95% CrI, −0.6% to 2.0%), and 21 300 fewer nonfatal overdoses (0.6% decrease; 95% CrI, −1.1% to 0) (Figure 2 and Table 2). In scenario 2, in which funding disruptions last 1 year and baseline funding is reinstated in August 2026, we estimated there would be 1100 excess all-cause deaths (0.1% increase; 95% CrI, 0-0.2%), 500 excess overdose deaths (0.2% increase; 95% CrI, −0.1% to 0.4%), and 4200 fewer nonfatal overdoses (0.1% decrease; 95% CrI, −0.2% to 0). In scenario 3, in which 100% of cancelled federal funding is reinstated in January 2029, we estimated there would be 3700 excess all-cause deaths (0.5% increase; 95% CrI, 0-0.8%), 1800 excess overdose deaths (0.8% increase; 95% CrI, −0.4% to 1.4%), and 14 700 fewer nonfatal overdoses (0.4% decrease; 95% CrI, −0.7% to 0).

High Service Disruption

We also reported the estimated changes to the primary and secondary outcomes if federal funding for SSPs accounts for 80% of their total budgets. In scenario 1, in which the initial funding disruption lasts for 5 years, we estimated there would be 39 600 excess all-cause deaths (5.0% increase; 95% CrI, −0.2% to 8.9%), 15 600 excess overdose deaths (6.9% increase; 95% CrI, −4.3% to 14.4%), and 155 000 fewer nonfatal overdoses (4.2% decrease; 95% CrI, −7.8% to 0.1%). In scenario 2, in which funding disruptions last 1 year and baseline funding is reinstated in August 2026, we estimated there would be 7900 excess all-cause deaths (1.0% increase; 95% CrI, 0-1.8%), 3200 excess overdose deaths (1.4% increase; 95% CrI, −0.9% to 2.9%), and 31 000 fewer nonfatal overdoses (0.8% decrease; 95% CrI, −1.6% to 0). In scenario 3, in which 100% of cancelled federal funding is reinstated in January 2029, we estimated there would be 27 400 excess all-cause deaths (3.5% increase; 95% CrI, −0.2% to 6.2%), 10 800 excess overdose deaths (4.8% increase; 95% CrI, −3.0% to 10.0%), and 107 300 fewer nonfatal overdoses (2.9% decrease; 95% CrI, −5.4% to 0.1%).

In the best-case alternate scenario, with low service disruption and funding reinstatement in August 2026, the executive order resulted in 1100 additional deaths and 500 additional overdose deaths during the 5-year study period compared with the base case. The worst-case alternate scenario, with sustained high service disruption, resulted in 39 600 additional deaths and 15 600 additional overdose deaths. In both funding disruption cases, funding would need to be raised to approximately 320% of pre–executive order levels in January 2029 to achieve zero excess deaths compared with the base case by August 2030.

Sensitivity Analyses

We performed deterministic sensitivity analyses in which we varied uncertain parameters and assumptions under both the low and high service disruption cases. Results are included in Table 3 for scenario 1 and eAppendix 5 in Supplement 1 for scenarios 2 and 3. All changes in deterministic sensitivity analyses led to increases in all-cause mortality.

Discussion

Injection drug use is common in the US and affects every region.^32,42,43^ More than 30% of US residents have been affected by overdose.^44,45^ After 2 decades of consistent increases, the crisis of opioid-related overdose deaths in the US started slowing in 2023. While causal inference is difficult to discern, the recent decrease in opioid overdose fatalities was partially attributed to increased access to harm reduction services, including naloxone, and increased uptake and retention of MOUD.^46^ Our analysis suggests that eliminating federal support for SSPs may reverse progress in reducing overdose deaths. In every scenario modeled, eliminating federal funding led to excess deaths compared with maintaining funding at its level prior to August 2025.

Our findings suggest substantial expected increases in fatal overdose deaths. In the worst-case funding scenario, we anticipated nearly 40 000 additional deaths and 16 000 additional preventable overdose deaths among persons who inject drugs in the US in the next 5 years compared with a scenario in which funding was not cut. Even in the best-case scenarios in which funding is temporarily reduced and reinstated 1 year later, we modeled excess mortality among persons who inject drugs. We projected that federal funding and SSP capacity would need to more than triple in January 2029 to net 0 excess deaths compared with the base case of no funding disruptions.

Notably, the model estimated decreases in nonfatal overdoses in funding disruption scenarios. This finding may be explained by decreased access to naloxone in funding disruption scenarios, which rendered some reversible, nonfatal overdoses fatal; SSPs are a major provider of naloxone and distributed more than 700 000 naloxone kits in 2019.^47^ Evidence suggests that SSP use may decrease both fatal and nonfatal overdose risk.^48^ However, we conservatively assumed that SSPs do not directly decrease probability of nonfatal overdose, only that overdoses are less likely to be fatal for an individual with access to an SSP due to provision of naloxone.

Limitations

This study has some limitations. First, the extent of implementation of this executive order is uncertain, and it is unclear whether there will be any contravening actions by Congress or state or local governments. Also, we are unable to determine direct correlations between loss of funding and potential SSP closures, and nonlinear or threshold effects are plausible. These uncertainties are impossible to capture completely in this model, although we mitigated this limitation by performing extensive deterministic and probabilistic sensitivity analyses. Second, there are data limitations for some key input parameters. Specifically, we assumed that differences in fatal overdose risk and MOUD initiation between SSP users and nonusers would follow similar patterns as described in a single city’s NHBS data, as these are what were available, although sensitivity analyses show that this does not have a large impact on the estimates. Third, our results do not account for the cumulative effects of reduced access to SSP services. We assumed equal risk of sequelae among individuals in each SSP use cohort regardless of past SSP use, which may underestimate total deaths and fatal overdoses in scenarios 2 and 3 (ie, funding reinstated during the study period). For example, persons who inject drugs who survived until some funding was reinstated may have accumulated more nonfatal overdoses compared with the base case, which puts them at greater risk of fatal overdose in subsequent periods.^49^ An alternative modeling approach to address this showed no major changes in projected outcomes (eAppendix 6 in Supplement 1). Last, we were unable to explicitly model people who use drugs through routes of administration other than injection or infections such as HIV or hepatitis C virus, and SSPs are often the only location many persons who inject drugs are able to access screening and treatment services. Thus, funding cuts to SSPs may lead to more downstream morbidity and mortality than our model estimates. Other modeling studies have projected an increase in HIV incidence of 35% to 60% attributed to full or partial closures of SSPs. Future modeling work should include other transmissible infections such as HIV and hepatitis C virus.^13,20^

Conclusions

In recent years, harm reduction interventions provided by SSPs have mitigated overdose deaths, saving many individuals the experience of losing a loved one. The findings of this decision analytical model study suggest that eliminating federal support for SSPs may result in excess overdose deaths and reverse recent progress in this field. Future studies should investigate the clinical effects of funding changes.