In addition to standard law enforcement, New Jersey now provides a range of additional responses to mental health crises, all operating under the ARRIVE Together program—a statewide initiative originally designed to pair law enforcement officers with mental health professionals. Today, the program is present across all counties and has expanded rapidly. Starting with 1,280 ARRIVE reports in 2023, when standardized reporting began, the program grew to 4,085 reports in 2024 and 7,735 in 2025. So far this year, there have been 696 reports.[1]
State funding for ARRIVE programs is typically directed to and managed by local police departments and prosecutors’ offices and they decide which models will be applied. For professional mental health support, they partner with healthcare providers. Basic information about ARRIVE interactions is available on the public dashboard created by the New Jersey Office of Attorney General (NJOAG). Below is what we learned from analyzing over 13,000 reports documented from May 2023 until March 5, 2026.
In terms of response models applied, 76.6% of ARRIVE interactions included only follow-up (40.2% or 5,549 follow-up + 36.4% or 5,017 close follow-up), and 17.7% (2,446) included only co-response. The two other models were applied much less frequently: only telehealth was reported in 2.7% (375) cases and only non-law enforcement response in 2.2% (307) cases. Less than 1% involved multiple models applied at the same time. All but two counties have multiple models available.
Of the 9,962 (72.2%) interactions for which law enforcement outcome data was recorded, more than half (54.7%) ended with community-based outcomes—either the situation was resolved on scene, or the individual was referred to mental health services, or both. A quarter resulted in voluntary hospital transport, meaning the person consented to additional care, and 22.8% resulted in involuntary hospital transport. Rarely, these categories can overlap because the outcome field allows multiple selections per report.
The data reveal important patterns about populations that ARRIVE teams typically encountered. Fourteen percent (1,929) were currently experiencing homelessness, highlighting the connection between housing instability and mental health crises. Forty-one percent (5,650) involved a person not fluent in English, underscoring the need for multilingual resources and culturally competent care.
In terms of behaviors indicated prior to arrival, the most common were welfare checks (4,851 or 35.2%), emotional dysregulation (4,633 or 33.6%), suicidal ideation (2,945 or 21.3%), and reports of the person experiencing confusion or disorientation (2,627 or 19.0%)—situations where having a mental health professional present can make a critical difference. The majority of reports (56.6%) indicated multiple behaviors.
While these findings are promising and the public dashboard is valuable, the publicly available demographic data cannot be linked to individual interactions due to privacy protections, leaving equity questions unanswered: how outcomes differ by race, age, or gender. Furthermore, granular geographical analysis is not possible because the data are not linked to specific locations including date and time where the encounters took place. Lastly, no comparison group exists. Without data about traditional law enforcement responses to similar situations, we cannot know conclusively the extent to which ARRIVE responses lead to more appropriate outcomes.
References:
[1] Relevant definitions from the ARRIVE dashboard: (1) “A single ARRIVE report may include the use of more than one ARRIVE model,” (2) “An ARRIVE Interaction occurs when an ARRIVE mental health professional responds to a potential mental health incident, either in person, virtually/telephonically, or by following up after the incident,” and (3) “An ARRIVE Interaction begins when the ARRIVE team is requested to respond to a potential mental health incident.”