You’re told to “do something about mental health” before renewal—what counts as success?
Three months before renewal, the request sounds simple: “Do something about mental health.” What usually follows is a scramble for a new benefit, a launch email, and a hope that utilization equals impact. Finance won’t accept that. Success has to mean a defensible change in costs you already pay—fewer avoidable ER visits, fewer inpatient admits tied to crises, shorter disability durations, or better follow-through on chronic care when anxiety or depression is in the way.
That definition forces a harder question up front: are you buying a program to treat need, or fixing how people find and use what you already have? The answer determines whether “doing something” reduces claims or just adds another line item.
Where are your costs actually coming from (and which of those can mental health move)?
That “another line item” risk is easiest to see when you map where your spend is actually coming from. In most mid-sized plans, a small group drives a big share of claims: inpatient stays, repeated ER visits, high-cost specialty medicines, and complex chronic conditions with lots of follow-up care. Mental health can touch those costs, but usually indirectly—by reducing crisis episodes, improving follow-through on diabetes or cardiac care when depression disrupts routines, or shortening leave when people get timely, appropriate treatment.
What it rarely moves is the base cost of routine care. A mindfulness app won’t change your negotiated rates, and it may increase outpatient visits at first as people finally seek care. That’s not failure, but you need to know it before you promise savings.
Once you know which buckets matter for your plan, you can decide whether you need a new vendor—or better access and navigation into what you already have.
The first fork in the road: adding a vendor vs. fixing access and navigation
That choice—new vendor or better access—shows up fast when you ask a simple question: where would an employee go today if they needed a therapist this week? If the real problem is “I called three numbers and gave up,” adding another program often just adds another place to get stuck. In that case, wins come from tightening the front door: clear entry points, fewer handoffs, faster appointments, and help matching people to in-network options that are actually taking new patients.
A vendor can still make sense when the gap is true capacity or clinical scope. If your network can’t deliver timely therapy, psychiatry, or higher-acuity care, navigation alone won’t create appointments. But expect real work: eligibility rules, data feeds, and member communications, plus overlap with EAP, telehealth, and behavioral carve-outs that confuses employees and dilutes use.
Before you pick a path, require a plain-language explanation of how it changes ER, inpatient, and disability—not just utilization.
What cost-reduction mechanisms should a credible program be able to explain?
That plain-language explanation should start with what happens on a bad week: someone can’t get an appointment, symptoms spike, and the first “access point” becomes the ER or an inpatient admit. A credible program can walk you through how it prevents that sequence—same-week triage, fast psychiatry when meds are involved, and a clear step-up path to higher-acuity care before things break.
The second mechanism is avoiding repeat events. If a member bounces between urgent care, the ER, and short outpatient visits, you’re paying for churn. Programs that reduce churn usually do two things: keep people in care long enough to stabilize, and coordinate with medical care so depression or anxiety doesn’t derail diabetes visits, cardiac rehab, or physical therapy. The cost signal isn’t “more sessions”; it’s fewer crises and fewer abandoned care plans.
These mechanisms require data sharing and tight eligibility rules, and both can be slow to implement. That’s where vendor ROI claims start to get fuzzy—unless you demand proof early.
When vendors promise ‘ROI,’ what evidence should you demand in week one of evaluation?
“Proof early” usually turns into a polished ROI deck and a request for a census file. Before you get pulled into pricing models, ask for the exact claim-level outcomes their program expects to move in your plan: ER visits tied to behavioral crises, inpatient admits with behavioral diagnoses, readmits, and disability days. If they can’t name the pathways and the codes or episode definitions they use, you’re not evaluating ROI—you’re evaluating marketing.
In week one, require three things in writing: (1) an evaluation design you can reproduce (comparison group approach, pre/post windows, and how they handle regression to the mean), (2) a data map that shows what they need from you versus your carrier/TPA and when you’ll see the first signal, and (3) a leakage plan (how they keep members in-network and prevent duplicate care with EAP, telehealth, or a behavioral carve-out). Then ask for a de-identified example report, not screenshots.
Most vendors can’t see inpatient and ER claims fast, and some never get them at all. If their “ROI” depends on data they won’t reliably receive within your renewal cycle, shift the conversation to metrics you can defend anyway.
Pick the metrics you can defend at the budget meeting—before you roll anything out
If their ROI depends on claims you won’t see until after renewal, you still need numbers you can stand behind in a budget meeting. Start with a small set of “can’t argue” operational metrics that connect to the cost pathways you already identified: time to first appointment, percent of members who get to the right level of care within 7–14 days (therapy vs. psychiatry vs. higher-acuity), and engagement through a clinically reasonable window (for example, 4+ sessions or a med follow-up within 30 days). If access and churn don’t improve, savings won’t either.
Then add two outcomes you can track with imperfect data: behavioral-crisis ER visits and behavioral-related inpatient admits per 1,000 members, even if you have to accept a lag. Pair that with disability incidence and average duration if you control leave data. So lock your baseline window, define eligibility, and decide your comparison group before the first email goes out. Those decisions also determine what rollout choices you can afford to make.
Rollout choices that determine whether savings show up: eligibility, incentives, and integration
Those baseline and comparison decisions turn into rollout choices fast, starting with who you let in. If you open eligibility to everyone with no pathway, you may drive helpful use but drown your evaluation in low-acuity demand and longer waits for the members most likely to hit ER or inpatient. If you target too tightly (recent crisis, high risk scores), you can show movement but miss the people sliding toward a costly episode. Write the rules in plain terms: what triggers outreach, what gets fast-tracked, and what happens after the first visit.
Incentives change who shows up and how long they stay. A gift card for “sign up” can inflate counts without improving stabilization; a small reward for completing an intake plus a follow-up visit is harder, but closer to the cost mechanisms. Expect pushback from payroll, tax, and legal on how you deliver it.
Integration is the slow, unglamorous part that determines whether savings are even visible. If EAP, telehealth, and your behavioral network all send separate messages and can’t share basic referral status, you’ll pay for duplicate care and still hear “I couldn’t find help.”
Making the case during renewal: a realistic narrative finance will accept
If EAP, telehealth, and your behavioral network all send separate messages, finance will assume you paid for overlap. Your renewal story needs to start with a clean map: one front door, what changed operationally (wait times, correct level-of-care placement, sustained engagement), and which cost buckets you expect to move next (crisis ER, behavioral inpatient, disability days). Put dates on it: “Q2 access fixes, Q3 engagement lift, claims signals lag into Q4–Q1.”
If you can’t get timely ER/inpatient claims, say it and use leading indicators you control. If the vendor can’t prevent out-of-network leakage, show the expected extra spend. Ask finance to fund the mechanism, not the slogan—and agree now on what would trigger a pivot.
