You’re asked to cut behavioral health spend—without breaking trust
It usually starts with a simple ask: “Bring behavioral health costs down.” But the people who use therapy, medication support, or leave options will notice if you tighten access or add hurdles. One poorly explained change can turn a cost conversation into a trust problem.
What makes this hard is that “behavioral health spend” isn’t one line item. Some of it sits in medical claims, some in pharmacy, some in EAP, and some shows up later as disability duration, absenteeism, or turnover. Vendor dashboards rarely add up cleanly, and chasing every metric burns weeks you don’t have.
The way forward is to tell a defensible story with a few buckets you can measure and manage.
Where is the money actually showing up this year? Start with four buckets you can defend
Those buckets work best when they match how your finance team already thinks about spend: paid claims, paid programs, and paid time not worked. Start by sorting everything into four places you can defend in a meeting.
Bucket 1: Medical claims tied to behavioral health and common comorbidities. This includes outpatient therapy and psychiatry, but also the “shadow spend” where depression, anxiety, or substance use shows up alongside diabetes, chronic pain, or ER use. Bucket 2: Pharmacy. Antidepressants and stimulants are visible, but so are opioid-related costs and meds used for sleep or anxiety.
Bucket 3: Point solutions and EAP. Treat vendor fees separately from any claims they generate so you don’t double count. Bucket 4: Work impact. Short-term disability, leave duration, and absenteeism often move more dollars than utilization reports suggest. The catch: this bucket is messier and slower, so set expectations before you promise savings.
The minimum data pull: what to request (and what to ignore) from each vendor
Once you’ve sorted spend into buckets, the usual snag is that every vendor hands you a different “engagement” story, and none of it ties back to dollars. So ask for the smallest set of fields that lets you reconcile counts, timing, and cost—then ignore the rest unless it changes a decision.
From your medical and pharmacy partners, request paid amounts by month, allowed amounts, member counts, and basic category groupings (BH diagnoses, BH CPT ranges, and key drug classes), plus high-cost claimant flags and site of care. From EAP and point solutions, request eligibility, unique users, sessions/visits, referral-out rates, average wait time to first appointment, and total fees paid—reported monthly and with a clear definition of “user.” For disability/leave, request BH-related leaves, average duration, recurrence within 6–12 months, and wage replacement paid.
Skip NPS, webinar attendance, and “minutes in app” unless you can show they predict fewer claims or shorter leaves. If a vendor can’t export at this level, treat that as a constraint when you decide what to renew.
When the reports don’t line up, how do you triangulate without getting stuck?
That export constraint is also your first triangulation move: pick one “source of truth” per bucket, then force everything else to match its definitions. In practice, misalignment usually comes from different time windows (incurred vs paid, rolling 12 vs calendar), different denominators (eligible lives vs registered users), or different counting rules (“users” vs “sessions”). If you don’t lock those down, every comparison turns into a debate about labels instead of dollars.
Start with a one-page crosswalk. For each feed, write the period, population, and unit, then map it to your four buckets. Reconcile top-down to bottom-up: total medical+pharmacy paid should tie to your carrier’s paid totals for the same months; EAP/point-solution fees should tie to AP; leave wage replacement should tie to payroll or disability invoices. Then do three sanity checks: month-to-month spikes (what changed), duplicate counting (referrals that later become claims), and impossible conversion rates (more “users” than eligible lives).
You won’t get perfect matches—lags, carve-outs, and small counts make that unrealistic—so use ranges and focus on what would change a renewal or access decision. Once the numbers are stable enough to trust, the real question becomes what’s driving the acceleration.
The cost accelerators hiding in plain sight: access delays, comorbidity, high-cost claimants, manager practices
That acceleration usually shows up as a familiar pattern: more people trying to get care, but the expensive events happen later. Long waits to a first appointment push employees toward urgent care, ER, or stop-and-start treatment that doesn’t stick. If your EAP shows rising referral-outs and your medical claims show more high-acuity visits a month or two later, that’s not “higher demand” in the abstract—it’s delayed access turning into higher-cost sites of care.
Comorbidity is the second multiplier. Depression alongside chronic pain, insomnia, or diabetes often looks like “medical trend” until you pull a simple overlap count: members with BH diagnoses plus 2+ chronic conditions, and their paid amounts. Then isolate the high-cost tail: the top 1–5% of members driving BH-related spend, and whether costs cluster around inpatient, residential, or intensive outpatient episodes.
Finally, don’t ignore manager practices because they don’t sit in a claims file. Poor role clarity, inconsistent schedules, and missed accommodations can extend leaves and increase repeat episodes. You may need proxy signals like leave duration by department, or recurrence within 6–12 months, before you can act on it.
Choosing interventions that reduce spend without “taking away support”
Those proxy signals are enough to choose actions that lower cost while still feeling like support. If wait time and referral-outs lead the story, start by fixing access: require network partners to meet a first-visit SLA, add a fast-track for medication management, and steer to higher-value sites of care before an ER visit happens. It costs time in contracting and eligibility files, and you may need to accept a higher unit rate to avoid much higher acuity later.
If the overlap count shows comorbidity, don’t launch a new “wellness” program. Tighten care pathways: identify members with BH plus 2+ chronic conditions, offer care navigation that coordinates PCP, therapy, and meds, and make sure referrals don’t die in voicemail. For the high-cost tail, put guardrails around inpatient/residential with strong discharge follow-up; otherwise readmissions erase savings.
If leave duration varies by department, pair benefits changes with manager habits: a short checklist for accommodations, predictable scheduling, and a return-to-work plan that starts in week one. Your next step is deciding what you can prove in 3–12 months.
Your 3–12 month scoreboard: the story you can tell before renewal season
Proving it in 3–12 months usually means avoiding “total BH trend” and showing movement in the drivers you can influence. Build a single scoreboard that tracks, by month: first-appointment wait time, EAP referral-out rate, percent of members reaching a second visit within 30 days, and ER/urgent care visits with a BH diagnosis.
Then tie it to dollars your CFO recognizes: paid medical+pharmacy for BH and for BH+2 chronic conditions, the top 1–5% high-cost claimant spend, and BH-related leave incidence and average duration. Expect messy attribution—carriers lag, vendors redefine “users,” and small counts swing—so pair each metric with a threshold you’d act on, and bring that to renewal conversations.
