Lyra Says Its Mental Health Benefit Reduces Employer Health Care Costs by 26% Per Year

Lyra says employers offering its mental health benefit experienced an average 26% reduction in annual health-care costs, sustained over four years. Its announcement goes further, saying Lyra drove reductions in employer health-plan spending across mental health, physical health, and pharmacy costs.

The evidence supports a narrower finding: in each of four years, members who used Lyra had substantially lower health-plan claims than matched members who did not. It does not establish that Lyra caused employer health-care costs to fall 26% per year.

Where the 26% comes from

The May 2024 study, commissioned by Lyra and conducted by Aon, compared members who used Lyra with eligible members from the same three employers who did not. The groups were matched on demographics, geography, and diagnosed medical and mental-health conditions.

Participant claims were lower in every year studied. The annual differences were $2,271, $1,736, $1,574, and $2,330 per participant, averaging $1,978 per year.

That is meaningful evidence of a recurring annual participant-versus-nonparticipant difference.

But it is important to distinguish that finding from a 26% decline in costs. Participant spending itself increased from $4,993 in 2018 to $6,547 in 2021. The 26% describes how much lower participants' claims were, on average, relative to matched nonparticipants. It does not describe participants' costs falling 26% after receiving Lyra.

How would Lyra have to produce the savings?

Part of the mechanism is visible.

Members using Lyra received much of their mental-health care through the Lyra benefit rather than through health-plan providers. Their professional mental-health claims were consequently much lower. Across the four years, that category alone accounted for roughly $1,300 to $1,500 per participant of the annual claims difference.

That can be economically valuable. But moving an encounter from the medical plan to a separately purchased mental-health benefit initially changes where the service is paid, not necessarily the purchaser's total cost.

The further economic proposition requires something more:

Lyra care → better or more efficient treatment → less downstream utilization and spending → savings greater than the cost of providing Lyra.

There are signals consistent with that pathway. Participants had lower pharmacy spending and some lower emergency and mental-health facility utilization. But broader physical-health medical spending was much less consistent.

The main unresolved transition is attribution

The study uses substantial matching, but participation was not randomized. Members entered the participant group because they chose to use Lyra.

The reports themselves identify remaining differences that matching cannot observe, including socioeconomic and job differences, reasons for seeking Lyra care, care-seeking preferences, condition severity, and diagnostic differences.

There is also no untreated pre-Lyra period in the four-year analysis. The three employers had implemented Lyra before or at the beginning of the 2018 study period.

The evidence therefore cannot determine how much of the participant/nonparticipant spending difference was caused by Lyra and how much reflects differences between people who used the benefit and people who did not.

That distinction matters because Lyra's public claim moves from participants had lower matched claims to Lyra reduced employer health-care costs.

The first is demonstrated. The full transition to the second remains unresolved.

Does the later ROI study close the gap?

The May 2024 study did not include what employers paid for Lyra and explicitly says its results should not be interpreted as an estimate of return on investment.

The July 2024 study addresses part of that limitation. For one additional employer, it reports $4,138 lower annual claims among matched participants and includes $1,162 per participant in Lyra session and support fees, producing a reported ratio of $3.04 in claims difference for every dollar spent. Lyra now presents that result publicly as approximately a 3:1 ROI.

That is useful because it puts program cost into the equation.

It does not, however, resolve the prior attribution problem. The July study still compares people who chose to use Lyra with matched people who did not. If some portion of the $4,138 claims difference would have existed without Lyra, that portion cannot properly be treated as savings produced by the program.

What should a purchaser ask?

Before underwriting the 26% claim, the most useful questions are straightforward:

• What was total employer spending before and after Lyra implementation, including Lyra fees?

• How much of the reported claims difference represents mental-health services moving from the medical plan into Lyra, versus utilization that was avoided?

• What evidence shows that the participant/nonparticipant difference was caused by Lyra rather than unmeasured differences in who chose to participate?

• What specific downstream utilization changes account for savings outside professional mental-health claims?

• What participant-level effect should an employer reasonably translate into savings across its entire eligible population?

Evidentiary Ceiling

The studies provide credible observational evidence that Lyra users have lower health-plan claims than matched nonusers and identify plausible ways the benefit could reduce spending. They do not establish that Lyra causes employer health-care costs to fall 26% per year.

For a purchaser, that difference is material. The 26% figure can support further investigation of Lyra's economic value. It is not, by itself, a 26% savings assumption to underwrite.