Veterans Crisis Hotline Waste: Millions Paid, Zero Calls

I keep coming back to one number, and it still feels wrong every time I write it down. A backup veterans crisis line, built so someone would pick up if the main system went dark, logged an entire stretch of operations without answering a single veteran. Not a slow night. Not a quiet weekend. Zero calls taken, while the invoices kept clearing. If you have ever sat with a friend who was one bad hour away from doing something permanent, that silence is not an accounting footnote. It is the whole story.

A September review by the department’s inspector general put the price of that silence at about $11.2 million. The period that matters ran from July 2022 through June 2023. Contract staff were on the books. Phones were not being answered. Somewhere between a security-clearance assumption and a supervisor who never sent the stop-payment note, the money simply continued.

What the Backup Line Was Supposed to Do

The main Veterans Crisis Line is meant to run around the clock. Trained government staff take the call, stay with the person, and send a medical team if the risk is immediate. That is the plain version. The harder version is that veterans die by suicide at roughly twice the rate of other Americans, which is why a dropped call is not a customer-service complaint. It is a hole in the floor.

In 2019 a contractor was brought in to stand up a backup center. The idea was narrow and, on paper, sensible. If a power cut, a system failure, or some other emergency knocked the primary line offline, another room of people would already be trained and ready. Redundancy is not glamorous. In emergency work it is the difference between a plan and a hope.

Then the paperwork collided with the job. New staff came on in 2022. Officials assumed those workers would not need a formal security clearance. They were wrong. Recording calls and documenting a veteran’s identity and mental health history is not casual note-taking. It requires authorization that typically takes twelve to eighteen months to obtain through the federal clearance process. Without that authorization, the people hired to answer could not do the one thing they were hired to do.

A backup that cannot legally touch the call is not a backup. It is a rented room with headsets.

I’ve found that government projects fail in two familiar ways. Sometimes nobody is in charge. Sometimes everybody is in charge of a slice, and the slices never meet. This one looks like the second. Clearance was someone else’s timeline. Staffing was the vendor’s problem. Payment was a routine that kept running because no one wanted to be the person who halted it without a formal instruction.

A Year of Phones That Never Rang Through

From July 2022 until the contract ended in June 2023, the backup center did not answer a single call from a veteran. Read that again slowly. The center existed. People were assigned to it. Money moved. The public purpose of the contract did not happen.

Some employees later said they did not realize the contractor was not performing. Others said they were waiting for a supervisor to issue a formal notice before payments could stop. In March 2023, one employee even marked the contractor’s work as satisfactory. Satisfactory, while the call log sat at zero. That rating is the detail that sticks. It suggests the review process was measuring presence, not outcome.

Perhaps the most interesting aspect is how ordinary the failure sounds once you strip the jargon. A team assumed a rule did not apply. The rule did apply. The fix took longer than the contract had left. Nobody pulled the payment lever. If this were a private clinic, the board would be in the room by Thursday. Inside a large department, the same fact can live in a folder for months.

Staffing That Never Matched the Contract

Clearance was not the only gap. The company was required to staff the backup center with 92 employees. Fewer than 50 people were actually assigned. Even with proper clearance, the center would have been running at roughly half strength. A half-staffed emergency desk is not a contingency plan. It is a bet that the bad night will be a small one.

There is a second question sitting under the first. Before the contractor was hired, the department had already started building its own in-house backup call center. So why bring in an outside firm at all? Auditors raised that point directly. Redundancy you already intend to build yourself, plus a vendor you cannot clear, plus a headcount you never reach, is a strange way to protect a suicide-prevention line.

• Backup role: answer only if the main crisis line went down
• Reality from mid-2022 to mid-2023: no veteran calls answered
• Contract staffing target: 92 people
• People actually assigned: fewer than 50
• Payments continued: about $11.2 million
• A performance mark in March 2023: satisfactory

Those bullets are tidy. The lived version is not. A veteran in a parking lot at 1 a.m. does not care which office owns the backup contract. They care whether a human voice arrives before the thought finishes itself.

How Payment Outran Performance

The bureaucratic sequence is almost dull, which is why it works. Someone assumes clearance is unnecessary. Clearance turns out to be mandatory. The queue for authorization runs a year or more. Meanwhile the contract vehicle is already live, so invoices look normal. Staff who might flag the problem either do not see it or decide they need written cover before they stop a payment. A satisfactory rating lands in the file. The contract runs out. The audit arrives later.

In my experience, the dangerous moment in any public contract is not the kickoff meeting. It is month eight, when the original justification has faded and the payment calendar has become muscle memory. Stopping money feels like an event. Letting money continue feels like administration. That asymmetry is how idle work gets funded.

None of this requires a villain with a yacht. It requires a process that treats “we are waiting on another office” as a reason to keep paying rather than a reason to pause. Waiting can be responsible. Waiting while the core deliverable is legally impossible is something else.

Tables flatten pain into columns. Still, the columns are useful. They show the gap between a contract’s sentence and a contract’s result. If you manage vendors for a living, you already know that gap. You just rarely see it priced this clearly, or attached to a line people call when they are done pretending they are fine.

Why the Clearance Assumption Mattered

Security clearance sounds like a spy-movie prop until you remember what the backup staff were supposed to handle. A crisis call is intimate. Names, locations, diagnoses, the shape of a plan someone may be forming. Recording that, and writing it into a system other clinicians can see, is sensitive work. The authorization exists so that information does not wander.

The mistake was treating that barrier as optional. Once the assumption failed, the timeline was no longer the department’s to compress. Twelve to eighteen months is a long time in a suicide-prevention calendar. It is also a normal time in a clearance calendar. Those two clocks do not share a rhythm, and nobody appears to have reconciled them before people were hired and paid.

Could the work have been redesigned so uncleared staff did something useful? Training, script review, drills that did not touch live records? Maybe. The review does not describe a serious pivot of that kind. What it describes is a waiting room that still sent bills.

The In-House Center Nobody Fully Explained

Here is the part I cannot shake. The department had already begun an internal backup center before the contractor was brought on. That does not automatically make the contract foolish. Surge capacity, a different location, a vendor who can scale faster than civil-service hiring: those are real reasons. They just need to be reasons you can say out loud after the fact.

If the in-house build was slow, say so and define what the vendor covers in the gap. If the vendor was meant to be temporary, write an end date tied to a working handoff. If both efforts were hedges against the same failure, explain why two incomplete hedges beat one finished desk. The audit’s question is fair. Duplication without a handoff plan is how agencies buy the feeling of preparedness.

Preparedness you cannot switch on is a press release with a budget line.

A blunt way of reading the staffing gap

I am not interested in scoring partisan points off a crisis line. The line serves people who wore the uniform under administrations of both parties. The failure mode here is older than any one secretary: a contract that measures activity, a clearance process that measures risk, and a payment system that measures whether the form was signed.

What Veterans Actually Need From a Line Like This

Strip the org chart away and the need is simple. A person answers. They know the script without sounding like a script. They can stay on the line. They can get a clinician or an emergency team moving if the moment requires it. They do not put the caller on hold while three offices argue about badges.

Public health researchers have said for years that veteran suicide risk is not one cause. Pain, isolation, traumatic injury, alcohol, the cliff that sometimes follows discharge, access to lethal means. A hotline does not fix those. It is the handrail on the stairs. You still want the handrail bolted down.

That is why a silent backup is worse than no backup announced. A silent backup lets leaders say the contingency exists. Families hear “we have a system.” The system, in this window, could not take the call. Language and capacity drifted apart, and the invoice did not notice.

1. Confirm the worker can legally handle the record before the first paycheck.
2. Tie every invoice to a measurable event, even if the event is a documented drill.
3. Staff to the number in the contract or rewrite the number.
4. Give a named official the authority to pause payment without a committee.
5. Do not run a parallel in-house build unless the overlap is written down.

Those five steps are not radical. They are what a careful operations lead would put on a whiteboard. The fact that they need saying, after an audit, is the tell.

The Satisfactory Rating, and What It Reveals

Performance ratings are supposed to be boring proof that someone looked. A satisfactory mark in March 2023, months into a stretch with no answered veteran calls, suggests the look was aimed at the wrong object. Maybe the rater saw timesheets. Maybe they saw a staffing chart. Maybe they saw a vendor who replied to email. None of that is the job.

Ratings drift when the metric is comfort. Comfort says the vendor is responsive, the meetings happen, the issues are “in process.” A crisis line does not get credit for process. It gets credit for a voice. If the metric cannot be faked by a calendar invite, use that metric. If it can, throw it out.

Would I have caught it sooner if I had been in that office? I hope so. I also know how easy it is to trust a status slide. Slides are calm. Call logs are not. The discipline is to open the log.

Money, Trust, and the Cost of Looking Away

$11.2 million is not the largest contract in a department that size. That is exactly why it matters. Waste at a scale people tune out becomes a habit. A habit of paying for unreadiness spreads. The next backup, the next clinic scheduling desk, the next records project, inherits the same permission structure: we are waiting, so we pay.

Taxpayers are not a mood. They are the people funding the handrail. Veterans are not a slogan. They are the people the handrail is for. When those two groups are told a system exists, and the system spent a year unable to pick up, trust thins in a way no rebrand repairs quickly.

There is a temptation to treat this as a one-off comedy of errors. I do not buy that. The ingredients are standard issue. Unclear ownership of a prerequisite. A vendor paid on a schedule rather than a result. An internal alternative that muddies the point of the contract. A culture that waits for permission to stop. You can find that recipe in buildings that have nothing to do with veterans. The stakes are just higher here.

A simple readiness test:
  Can they take the call today?
  If no, why is the invoice still moving?
  Who is allowed to say stop?

Pin that above a contracting desk. It will not fix clearance timelines. It will stop a few quiet disasters.

What a Working Contingency Actually Looks Like

A real backup is dull on purpose. People are cleared before they sit down. Rosters match the contract. Drills happen on a calendar, and the drill log is the invoice attachment. If the main line fails at 3 a.m., the failover is a switch, not a meeting. Supervisors can halt pay when the switch cannot be thrown. None of that is heroic. Heroic is what you need when the dull work was skipped.

Compare that with the version in the review. Uncleared staff. Half the promised bodies. A parallel in-house effort. A satisfactory rating. Nearly a year of silence. Millions paid. You do not need a conspiracy to get from the first version to the second. You need a few missed handoffs and a payment system that defaults to yes.

Smaller agencies sometimes do this better, oddly enough, because they cannot hide a dead contract inside a huge budget. Everyone can see the empty chairs. Scale is not an excuse. Scale is the reason you write the stop rule down.

Questions the Audit Leaves on the Table

Audits answer what happened. They rarely settle what should happen to the people who let it happen. A few questions still feel open, and they are worth asking without turning them into a cartoon.

Who owned the clearance assumption, and why was it not tested against the actual task of recording clinical calls? Who had the authority to freeze invoices, and why did “waiting for a formal notice” become the default? Why did a satisfactory rating survive contact with a zero-call record? What, exactly, was the in-house center supposed to cover that the vendor was not, and the reverse? Were any live calls misrouted toward a desk that could not take them, or did the main line simply never fail in that window? The last question matters. A backup that is never needed can still be a waste if it was never capable. It is a different waste from a backup that was needed and absent.

I would also want the after-action in plain language, posted where veterans’ families can read it without a law degree. Not a victory lap. A list of controls that changed. If the controls did not change, the audit is a scrapbook.

How Oversight Usually Misses This

Inspectors general catch a sample, not the ocean. That is not a knock. It is the job. The rest of the ocean depends on program staff who treat a zero in the outcome column as an alarm, not as a data quirk. Contract officers who read the statement of work against the timesheet. Finance staff who are allowed to ask, “What did we buy this month?”

Whistleblowing rules exist for a reason, and so does ordinary courage in a staff meeting. Someone in that chain knew the phones were quiet. Several people, according to the review, knew enough to explain later why they did not stop the money. Knowledge without a lever is how these stories age into footnotes.

Outside auditors who follow federal spending have made a cottage industry of this pattern: pay first, purpose later, purpose sometimes never. You do not have to love their politics to use their method. Follow the invoice until it hits a human result. If the result is missing, stay there.

A Fair Reading, Without the Easy Outrage

It is possible to hold two thoughts. Crisis-line work is hard, understaffed in many places, and staffed by people who absorb other people’s worst nights. And a backup contract that cannot take calls should not be paid as if it can. Those thoughts do not cancel each other. Blaming every clinician for a contracting failure is lazy. Excusing the contracting failure because the mission is noble is also lazy.

The clearance delay is real. Federal authorization is slow, and slow can be justified when the data is this sensitive. The unjustified part is hiring into a role the clearance makes impossible, then paying through the impossibility. You can respect the privacy rule and still refuse the invoice.

Same with staffing. Recruiting clinicians and trained responders is not like filling a warehouse shift. Falling short of 92 does not prove bad faith. Falling short of 50, while billing a 92-person contingency, does prove the contract and the floor were telling different stories. Rewrite one of them.

What Should Change Before the Next Contract

If I were rewriting the next backup agreement, I would start with a gate, not a goal. No invoice until a named set of staff hold the authorization the work requires. Not a plan to get authorization. The authorization. Second gate: a monthly drill with a logged outcome, reviewed by someone who does not work for the vendor. Third: a staffing floor that, if missed by a set margin, automatically cuts the fee. Fourth: a single official who can suspend payment in writing within five business days of a failed gate. Fifth: a one-page public note, quarterly, on whether the backup can take a live call that day.

That last item will annoy people. Good. Annoyance is cheaper than another year of silence. Public notes also force the in-house question into the open. If the internal center is ready, say the vendor’s role is over. If it is not ready, say what the vendor is covering. Ambiguity was the fertilizer here.

Training dollars should follow the same logic. Pay for practiced handoffs, not for seats. A responder who has never touched the failover procedure is not a reserve. They are a name on a slide.

The Human Scale Hiding in the Audit

Numbers make this easy to file under waste and move on. The harder picture is a veteran who does not know any of this, and should not have to. They dial a number they were told would work. On the nights the main line holds, none of the backup mess touches them. On the night it does not, the mess is the whole world.

Families who have already lost someone to suicide do not need a lecture about efficiency. They need the next family to get a voice. That is the standard. Everything else, the clearance queue, the vendor margin, the satisfactory box, is downstream of that standard. When downstream processes contradict the standard, the processes are wrong.

I keep thinking about the employee who marked the work satisfactory. Maybe they were tired. Maybe the form offered three choices and satisfactory felt polite. Politeness is not a control. A control would have asked a rude question: how many veterans did you speak to? The answer was available. It was zero.

Lessons That Travel Past One Department

You can lift this case out of veterans’ health and drop it onto almost any emergency-adjacent contract. Disaster call centers. Benefits hotlines. Child-welfare after-hours desks. The pattern repeats because the incentives repeat. Vendors are paid to be available. Agencies are graded on whether a contract was awarded. Availability and award are not the same as a completed call.

Private companies fail this way too, then go broke or lose the client. Public agencies fail this way and write a lessons-learned memo. The memo is useful if the next solicitation changes. If the next solicitation copies the last one, the memo is décor.

A practical test I like: imagine explaining the invoice to a veteran who used the main line last Tuesday and does not follow procurement. If the explanation requires the phrase “pending a formal notice,” you are not done. Try again until the sentence is about a person answering.

Where the Money Could Have Pointed

This is not an argument that the department should spend less on suicide prevention. It is an argument that prevention money should buy prevention. Counselors, peer lines that actually connect, faster follow-up after a crisis call, tighter links to local emergency care. Those are imperfect too. They are at least aimed at the night in question.

Eleven million dollars will not end a national pattern of veteran suicide. Spent on a desk that cannot pick up, it will not even dent that pattern. Spent on cleared, drilled, measured backup capacity, it might shorten the distance between a failing main line and a human voice. That is a modest claim. Modest claims are the ones worth funding.

There is also a morale cost inside the building. Staff who do answer the hard calls watch a parallel contract get paid for silence. That contrast is corrosive. People stay in this work because the call matters. Paying for the absence of the call tells them the institution is counting something else.

Reading the Timeline Without the Fog

Put the years in a row. A contractor is hired in 2019 to build backup capacity. An in-house backup effort is already underway. In 2022 new contractor staff arrive under a mistaken belief that clearance is unnecessary. The belief collapses. From July 2022 to June 2023 the center answers no veteran calls. Headcount stays under half of 92. In March 2023 someone still rates the work satisfactory. The contract ends that June. A September inspector general report prices the stretch at roughly $11.2 million.

Nothing in that sequence requires secret knowledge. It requires reading the contract against the call log, once, early. The fog arrives later, when everyone explains the piece they did not control. Clearance was slow. The supervisor had not written the memo. The vendor was short-staffed. The in-house center was coming. Each sentence can be true. Together they funded a year of unreadiness.

If you work near budgets, you have heard the cousin of this story. A software module nobody logged into. A training vendor whose completion rate was a rumor. A leased space that outlived the team. The crisis line version is the same shape with a worse downside. That is why it belongs in public, in plain words, without a throat-clearing paragraph about how complicated government is. Complicated is not a defense. Complicated is the environment the controls have to survive.

What I Would Tell a Family Reading This

If you are supporting a veteran right now, the main crisis line is still the number you have been given, and the people on it are still the point of the system. This audit is about a backup that did not function, not a claim that every call goes unanswered. Use the line. Stay on it. Ask for the follow-up. The existence of a bad contract does not make the working desk fictional.

If you are a taxpayer who is angry, aim the anger at the gate that failed, not at the clinician who picks up on a Tuesday. Ask your representatives whether payment can be paused when a backup cannot legally operate. Ask whether the in-house center is actually live. Ask for the drill log. Anger that stops at a headline does not move a contracting manual. Anger that asks for a gate sometimes does.

And if you are inside an agency reading this with a familiar knot in your stomach, you already know the sentence to send. Not a task force. A pause. “We cannot do the work yet, so we should not be paying as if we can.” That sentence is shorter than the audit. It would have been enough.

A Closing Look at the Silence

The image I cannot decorate away is a room set up for emergencies that never took the emergency. Headsets, chairs, a contract, a rating that said things were fine. Outside that room, a population that dies by suicide at about twice the civilian rate, and a promise that if the main line failed, someone else would be there.

They were not there. Not because the night was kind. Because the prerequisites were missed and the money was not. That is the whole finding, once you take the stationery off it.

Public money is a trust, and crisis response is a narrower trust inside it. You can miss a facilities contract and anger people. You can miss this one and fail someone on the worst night of their life. The standard has to match the downside. In this stretch, it did not. The next contract does not get to inherit that excuse.

I will keep the number where I can see it. About $11.2 million. Zero veteran calls answered. Fewer than half the staff the paper required. A satisfactory mark in the middle of the quiet. An in-house backup that made the hire harder to justify, not easier. If a future status slide cannot survive those five facts, it is not a status. It is a wish.