Picture a city that tells you, “We’re moving mental health crises away from police.” Then look a little closer and you see the uncomfortable math of implementation: even when the intent is humane, the machinery can still fail people in the moment they need help most. Personally, I think that gap—between political promise and operational reality—is where public trust goes to die.
Mayor Zohran Mamdani is staking his public safety legacy on overhauling how New York responds to mental health emergencies, but the hard truth is that he can’t redesign the whole system without first fixing a de Blasio-era program that has struggled for years. The Behavioral Health Emergency Response Division, or B-HEARD, was supposed to route trained mental health professionals and EMTs to certain calls instead of police. But an audit found that the teams did not respond to more than one-third of eligible mental health calls for reasons that weren’t clearly tracked. From my perspective, this is exactly what makes the debate so exhausting: we’re not arguing whether the goal is good—we’re arguing whether the execution is real.
A humane promise, a broken pipeline
B-HEARD’s mission sounds straightforward: treat specific mental health emergencies as a clinical and safety problem—not a law-enforcement default. The plan is to dispatch professionals and EMTs, which ideally reduces harm for people in crisis and helps officers focus on cases where police are truly needed. Yet the audit result—that over one-third of eligible calls didn’t get a B-HEARD team response—cuts against the narrative that simply “routing” calls solves the crisis. What makes this particularly fascinating is how quickly people confuse “new program” with “new outcomes.”
Personally, I think the most revealing element here is not the percentage by itself, but the phrase “untracked reasons.” When a system can’t explain failures in real time, it’s not just inefficient—it’s politically convenient. Critics can claim it’s a staffing issue; supporters can claim it’s a teething problem; officials can claim improvement is underway. But without transparent tracking, the public gets the worst of both worlds: neither accountability nor confidence.
There’s also an implied contradiction that many people don’t realize: the city rerouted more 911 calls to B-HEARD, but still sent police when B-HEARD didn’t show up. In other words, the “de-policing” story depends on a unit that itself can’t reliably operate at full coverage. This raises a deeper question for any reformer: are you building an alternative response—or building a fragile substitute?
The scaling problem no one can charisma-away
Mamdani’s current push is to center B-HEARD as the operational backbone of his Office of Community Safety. But experts caution that scaling it up won’t be easy, especially when the underlying constraints remain. One major complication mentioned is how B-HEARD is jointly run with the FDNY versus being fully operated through Health and Hospitals. Personally, I think this is where well-meaning public policy runs into the physics of bureaucracies.
Here’s my read: mental health emergency response isn’t just a clinical question; it’s a labor-operations question. If union rules require two EMTs be dispatched to every call, that can collide with an EMT staffing shortage and slow or block the very dispatch model the program needs. People often misunderstand this as “bureaucratic red tape,” as if it’s merely paperwork. But it’s actually workforce math—how many skilled people are available, how they’re contractually scheduled, and whether the system can sustain the volume.
If you take a step back and think about it, this is also a story about incentives. Police departments and public safety agencies can often meet demand because they already have dense coverage and established dispatch routines. Meanwhile, B-HEARD is trying to grow into a role that police used to cover—without having the same staffing depth, training pipeline, and institutional inertia. What this really suggests is that reform isn’t a switch; it’s a capacity-building campaign.
The Thrive 2.0 fear: when crises hijack the agenda
Dr. Gary Belkin, who oversaw policy for de Blasio’s Thrive from 2014 to 2018, warns about a recurring political pattern: headline-grabbing incidents can derail support for mental health training and care facilities. Personally, I think this is one of the most under-discussed mechanisms in public policy. Systems don’t fail only because of budget or logistics—they fail because the public conversation gets hijacked.
What’s interesting here is the “déjà vu” effect: when a crisis goes badly, it becomes a symbol, not a data point. Critics then use violent incidents involving people with mental illness as evidence that the entire approach is flawed. Supporters counter with “it’s about more funding” or “the model takes time.” But I’ve learned that time is not politically neutral; time can be weaponized.
One thing that immediately stands out is that Thrive faced criticism for transparency and unclear goals while spending more than $1 billion over six years. If you’re Mamdani, the temptation would be to move fast and promise big—because that’s how you win headlines. But Belkin’s warning implies the real risk isn’t the intent; it’s the narrative fragility. If your program doesn’t produce measurable improvements quickly and credibly, the next incident can erase years of groundwork.
Funding gaps: the quiet villain
Supporters suggest the fix is funding, and Mamdani has argued B-HEARD was “kneecapped” by lack of resources. At the same time, the city faces a budget gap, which makes expansion harder than campaign language. From my perspective, this is the part everyone argues about in theory but rarely confronts in its full form: mental health care is not a single program—it’s an ecosystem.
If B-HEARD is the front door, someone still has to fund the rooms behind it: follow-up services, short-term support, ongoing treatment pathways, and beds or alternatives to hospitalization. When those downstream parts are underbuilt, the system can only respond to crises, not prevent them. That is how “upstream” visions become downstream realities.
And this is where the Office of Community Safety opening details feel telling: it begins with a tiny staff and a modest budget compared to the much larger plan Mamdani proposed as a candidate. Personally, I think smaller initial budgets can be strategic, but they also invite skepticism—especially from political adversaries who will label it as “Thrive 2.0.” The political battlefield matters because it determines whether officials are allowed to iterate or forced to defend every step.
Upstream vs. crisis response: the deeper test
Elizabeth Glazer, who previously ran the Mayor’s Office of Criminal Justice, argues Mamdani must address what happens before and after a crisis—not just the moment of emergency response. She’s basically saying: if you keep operating like an ambulance, you’ll never build a healthcare system. In my opinion, this is the clearest conceptual distinction in the entire debate.
B-HEARD and related centers are framed as alternatives to hospitals, including support and connection centers that allow short stays. Under de Blasio, the city introduced the idea, but only one center had opened in East Harlem by a certain point, with another center in the Bronx later closing due to low traffic. What many people don’t realize is that “access” is not the same as “usage.” People need trust, continuity, and clear pathways—otherwise even a well-designed facility becomes underutilized.
This implies that success metrics can’t stop at dispatch rates. If the city reroutes calls to non-police responders but fails to connect people afterward, the system may simply swap one kind of failure for another. Personally, I think the most important question is whether the city is building a durable care pathway or just optimizing crisis throughput.
Metrics and accountability: promise depends on measurement
Mamdani’s chief of staff says the administration will be “vigorously concerned with new metrics,” and that City Hall is consulting experts across the country. I take that claim seriously—not because I trust any administration’s messaging by default, but because measurement is the only language power can’t fully dodge.
However, I’m also wary of metrics theater. It’s easy to publish numbers that look good while hiding the operational truth—like counting dispatches without explaining refusals, staffing shortages, or system downtime. If “untracked reasons” caused one-third of eligible calls to miss a response, then the next reform must include transparent logging and real-time accountability. What this really suggests is that the reform’s credibility will live or die on auditability.
There’s also a cultural dimension: New Yorkers have learned to distrust broad public safety rhetoric because prior initiatives were complex, expensive, and vulnerable to political blame. Thrive’s controversy shows how quickly big plans become political liabilities if transparency and outcomes don’t keep pace. Personally, I think Mamdani can’t afford just a moral vision—he needs an operational one.
Where this goes next
If Mamdani wants this to be more than a rebranding exercise, he will likely need to tackle at least three constraints at once: staffing capacity, care continuity, and transparent accountability. Anything less risks creating a system that looks compassionate in press conferences but behaves inconsistently on the ground. From my perspective, the real “Office of Community Safety” test is whether it can maintain coverage during the worst moments, not just during managed pilot phases.
In a city that runs on emergency response, the public deserves more than intentions—they deserve reliability. And if the city can’t reliably reduce police involvement in eligible mental health calls, then the reform will be interpreted not as progress, but as another broken promise.
Ultimately, the question isn’t whether mental health emergencies should receive clinical support. The question is whether New York can build a care infrastructure sturdy enough to outlast headlines, staffing realities, and political cycles.
