Executive summary
Benefits benchmarking is a mature discipline with a blind spot. It compares plan design, premium contribution, offering rates and network breadth against a peer set, and it does all of that well. What it almost never reports is whether the people at the top of the organization used any of it. This paper benchmarks executive mental health support on that missing metric, sets out the published figures a benefits team can actually check its own program against, and gives a four-gate method for finding where senior utilization is being lost.
Employers have expanded mental health access faster than any other benefit category, and 73 percent now report rising utilization of mental health and substance use services.01
Reported use of the flagship access route, the employee assistance program, sits between 2.1 percent and 8 percent of employees at 47 percent of employers, and no standard benchmark breaks that figure out by seniority.04
Benchmark the benefit the way an epidemiologist would: define the population, instrument each barrier separately, and measure use rather than availability.
A benefits leader can say, with a number, whether the senior tier is covered in fact as well as on paper, and can direct spend at the gate that is actually closing.
The problemA benchmark that cannot see the top of the org chart§
Benefits benchmarking answers a specific question well: how does our plan compare with the plans our peers buy. The annual employer surveys that anchor the discipline are large and methodologically serious. The KFF Employer Health Benefits Survey interviewed 2,142 firms in 2024 and 1,862 in 2025.02, 03 The Bureau of Labor Statistics measures benefit access across the private workforce and reported that employee assistance programs were available to 40 percent of workers in establishments with fewer than 100 workers as of March 2025.05 Membership bodies publish benchmarking tools that sort every offering into categories a benefits committee can compare line by line.18 None of these instruments is weak. They are simply built to measure what an employer purchased, not what any particular tier of employee did with it.
That distinction stops being academic the moment the population of interest is senior. Gallup, drawing on 9,880 leaders in its 2026 workplace data, found that leaders report more daily stress than individual contributors, 46 percent against 39 percent, and more daily loneliness, 31 percent against 21 percent.07 The need is measurably higher at the top. The published use rate is a single organization-wide number that cannot show whether it was met. A benchmark that reports offering, cost and network breadth, and then reports one undifferentiated utilization figure, will show an executive tier and a call center tier as the same data point.
An employer can be at the ninetieth percentile on every benchmarked dimension of a mental health benefit and still have a leadership team that has never opened it. CEREVITY clinical whitepaper series, 2026
The evidenceWhat the research shows§
The figures below are all published, all recent, and all measure something an employer can benchmark. Read together they describe a market that has solved availability and has not yet started measuring reach. They also explain a pattern that benefits teams describe informally and rarely quantify, which is why executives do not use the EAP you already bought.
47%
of employers report EAP use between 2.1% and 8% of employees
Journal of Insurance Regulation, 202004
6x
higher intent to leave within three to six months among employees with burnout symptoms
McKinsey Health Institute, 202206
US$1T
lost each year in productivity to depression and anxiety worldwide
World Health Organization, 202415
46%
of leaders report daily stress, against 39% of individual contributors
Gallup, 202607
Taken together the four figures establish a pattern rather than four separate facts. Demand is high and concentrated at the top. The cost of unaddressed distress is large and mostly invisible on a benefits ledger. And the one number that would connect the two, the share of the senior population that actually entered care, is the number nobody publishes. That is the structural reason why benefits never reach the executive tier even when the benefit itself is generous.
04, 01, 02, 05 Journal of Insurance Regulation, 2020, reporting Compton and McManus (2015).
Business Group on Health, 2026 Employer Health Care Strategy Survey.
KFF Employer Health Benefits Survey, 2024, n=2,142 firms.
U.S. Bureau of Labor Statistics, Employee Benefits in the United States, March 2025.
| Benchmark metric | Published figure | Source and year | What it cannot tell you |
|---|---|---|---|
| EAP offered, largest employers | More than 97% of companies with over 5,000 employees; 75% of midsize and small companies | Journal of Insurance Regulation, 202004 | Whether anyone above director level has ever opened it |
| EAP access, private industry | 40% of workers in establishments with fewer than 100 workers | Bureau of Labor Statistics, March 202505 | Access is counted at the establishment, never at the seniority tier |
| Reported EAP utilization | 2.1% to 8% of employees at 47% of employers; above 8% at only 19% | Compton and McManus, 2015, reported in Journal of Insurance Regulation04 | Case counts are not broken out by grade, role or reporting line |
| Mental health network breadth | 30% of firms call their network very broad; 24% call it somewhat or very narrow | KFF Employer Health Benefits Survey, 202402 | Breadth is a directory count, not an appointment a sitting CFO can take |
| Timely access to care | 70% of firms believe their largest plan provides timely access to mental health services | KFF Employer Health Benefits Survey, 202503 | This is the employer's belief, not a measured wait time by tier |
| Employer and employee perception gap | 22 points on average across the workplace factors tied to mental health | McKinsey Health Institute, 202206 | The gap is widest exactly where honest self-report carries the most career risk |
| Program return on investment | Median CA$1.62 per dollar per year, rising to CA$2.18 for programs running three years or more | Deloitte, 201913 | Return is computed on the population that used the program, not the population that was offered it |
One row deserves separate attention. The 22-point average gap between what employers believe about their own workplace conditions and what employees report is the largest single finding in the McKinsey Health Institute survey of 14,509 employees and 1,389 human resources decision makers across fifteen countries.06 A benchmark built from employer self-report inherits that gap. It also sits on top of a baseline treatment gap that has nothing to do with employers at all: 23.1 percent of United States adults had any mental illness in 2022 and 50.6 percent of them received any treatment in the year, 41.6 percent among men.16 The Gallup data below shows what the gap looks like when the same four experience measures are asked of leaders and of individual contributors separately.
07 Gallup, 2026. Leaders Have Better Lives but Worse Days. Leader subsample n=9,880, defined as managers of managers, from 2025 World Poll data.
The frameworkA model you can name and own§
A single utilization percentage cannot be acted on, because it does not say which barrier produced it. The model below breaks the path from eligibility to care into four gates, each of which can be measured on its own and each of which fails for a different reason. It is built for the population this paper is about, and it assumes the reader already accepts that senior distress is real and treatable through confidential burnout treatment for senior leaders rather than through resilience messaging.
CEREVITY model
The Four-Gate Utilization Model
Four sequential filters between an eligible senior leader and a first clinical session. Utilization is the product of all four, so a program can clear three gates and still report a use rate near zero. Each gate has its own diagnostic question and its own fix.
Awareness
The leader knows the benefit exists and knows it applies to them. Executive populations are routinely excluded from the internal communications that carry benefit reminders, because those campaigns are aimed at the general workforce.
Relevance
The offer matches the problem at this level. A short-course model built around workplace stress does not read as an answer to a leader carrying a restructuring, a board conflict and a marriage under strain at once.
Exposure
Using the benefit creates no record and no rumor. This is where senior utilization is most often lost, and it is the gate that stigma-reduction campaigns cannot open, because the concern is structural rather than attitudinal.
Depth
The format can hold the work. A leader who clears the first three gates and is offered six sessions of general counseling will attend one or two and conclude the benefit is not for them, which then depresses the next year's numbers as well.
SCHEMATIC Schematic, not measured data.
Structure derived from the barrier findings in references 09 and 10 and the utilization band in reference 04. Index values are illustrative.
The practical value of separating the gates is that each one has a different owner. Awareness belongs to internal communications. Relevance and depth belong to whoever designs the benefit. Exposure belongs to whoever chose the billing and record architecture, and it is usually the gate nobody has been assigned. A program that measures only the end-to-end rate will keep spending on the first gate, because that is the cheapest one to work on, while the third gate stays shut.
By professionHow it presents across roles§
The three groups below were chosen because each has published, profession-specific evidence on both distress and help-seeking, which makes them the only populations where a benefits team can benchmark senior utilization against anything external at all. In each case the barrier that closes the exposure gate is documented in the profession's own literature.
Corporate senior executives
Gallup's 2026 analysis of 9,880 leaders, defined as managers of managers, found that leading raises every negative daily experience it measures. Leaders reported daily stress at 46 percent against 39 percent for individual contributors, daily anger at 33 percent against 21 percent, daily sadness at 34 percent against 23 percent, and daily loneliness at 31 percent against 21 percent.07 The same body of work found manager engagement falling from 31 percent in 2022 to 22 percent in 2025, the steepest decline of any employee group, against a global backdrop in which only 20 percent of employees were engaged and 34 percent were thriving.08 What makes this population hard to benchmark is not the prevalence figure, it is the reporting geometry. A senior executive who uses an employer-sponsored program is often entering a system administered by people who report to them, or by a vendor whose contract they signed. The exposure gate closes before awareness or relevance are ever tested. Consider a hypothetical scenario in which a chief operating officer notices six weeks of broken sleep ahead of a board meeting. The benefit exists, the network is broad, and the intake line is staffed. The calculation that stops the call is not about the quality of the care. It is about who will eventually see a utilization report. The measurable consequence is that this person appears in the denominator of every benchmark and in the numerator of none.
Individual confidential therapy for chief executives
Organizational the partnership model for corporate organizations
Attorneys and law firm partners
The 2016 study of 12,825 licensed, employed attorneys remains the largest profession-specific dataset of its kind. It found that 28 percent, 19 percent and 23 percent of respondents reported mild or higher levels of depression, anxiety and stress respectively, and that 21 percent screened at levels consistent with problematic drinking. Problematic drinking was highest early in a career, at 32 percent of those aged 30 or younger, 31 percent of junior associates and 26 percent of senior associates.09 The finding that matters for benchmarking is not the prevalence, it is the barrier ranking. The two most commonly reported reasons for not seeking treatment were not wanting others to find out, and concerns about confidentiality.09 Those are exposure-gate failures in the exact sense used in the previous section, and they are structural: a licensed professional whose fitness can be questioned is making a rational calculation about a record, not an emotional one about stigma. A firm that benchmarks its wellness spend against peer firms and finds itself well provisioned may still be measuring a program its equity partners will never touch. The useful benchmark for a legal organization is not spend per head. It is the share of partners who could use the program without any part of the firm learning that they did.
Individual therapy where nothing reaches the bar
Organizational EAP alternative for law firms
Physicians and physician leaders
The American Medical Association reported that 48.2 percent of physicians experienced at least one symptom of burnout in 2023, down from 53 percent in 2022 and 62.8 percent at the 2021 peak, drawn from more than 12,400 responses across 81 health systems in 31 states. In the same data 50.7 percent reported significant job stress and 50.4 percent felt valued by their organization.11 Prevalence is therefore measured with unusual precision in this population, which makes the utilization gap easier to see. A national survey of more than 5,800 physicians found that close to 40 percent would hesitate to seek professional help for a mental health condition because of concerns about medical licensure, with roughly two-thirds of states asking about past mental health diagnoses or treatment on licensing documents, and physicians in states with broader questions significantly more inclined to avoid care.10 A hospital system can benchmark its wellness program favorably against peer systems on every dimension a benefits survey measures, and the licensing question will still close the exposure gate for a large share of its medical staff. Any benchmark of a physician benefit that does not report what the care leaves behind in a record is measuring the wrong variable.
Individual therapy that leaves no insurance record
Organizational a physician wellness program for medical groups
The stakesThe cost of inaction§
The cost of an unused benefit is not the premium. It is the difference between what the organization paid for and what it received, plus the consequences that the benefit was bought to prevent and did not. The Office of the U.S. Surgeon General reports that 76 percent of workers described at least one symptom of a mental health condition and 84 percent said workplace conditions had contributed, while 81 percent said they would look for workplaces that support mental health in future job decisions.17 Three cost categories are large enough to change a benefits decision, and all three are measurable with published figures.
Replacement at the top
A review of 31 case studies drawn from eleven research papers put the median cost of replacing an employee at 21 percent of annual salary, with executive positions running as high as 213 percent because of the skills and search involved.14 Set that against the McKinsey finding that employees with burnout symptoms were six times more likely to report intending to leave within three to six months.06 One prevented senior departure funds a great deal of benefit design.
Output that never registers as absence
Deloitte put the annual cost of poor mental health to employers in the United Kingdom at 53 to 56 billion pounds for 2020 to 2021, a 25 percent rise since 2019, with presenteeism the largest single category and close to 40 percent of turnover costs attributable to mental health.12 Globally the World Health Organization estimates 12 billion working days lost each year to depression and anxiety, at US$1 trillion in lost productivity.15 None of this appears on a benefits invoice.
Spend measured on the wrong clock
Deloitte's analysis of seven Canadian employers with three or more years of data found a median annual return of CA$1.62 per dollar invested, rising to CA$2.18 for programs running three years or more, and noted that reaching a positive return can take three years or longer.13 The benchmarking cycle that judges these programs is annual.03 A program cancelled at month twelve is not a program that failed. It is a program measured before it could report.
13, 03 Deloitte, 2019. The ROI in workplace mental health programs. Seven Canadian employers with three or more years of data.
KFF Employer Health Benefits Survey, 2025, an annual instrument.
The solutionWhat effective care looks like§
Good care for this population has three requirements that a general benefit is not built to meet. It has to be clinically serious rather than supportive, because the presenting problems are usually compound. It has to be structurally private, meaning the privacy comes from the architecture of the arrangement rather than from a promise. And it has to offer enough time in one sitting to be worth the calendar cost of attending, which is why weekly 50-minute work suits some presentations and not others. The approach behind this model is the part that determines whether the exposure gate opens.
CEREVITY is a nationwide network of independent licensed clinicians, matched to the person and delivered by secure video on a private-pay basis, with no insurance claim, no diagnosis code submitted to a plan and no shared employer record. Sessions run in three formats. The weekly 50-minute session carries ongoing work. The extended session format runs 90 minutes and gives room to material that an hour keeps interrupting. The 3-hour intensive format concentrates work that would otherwise take a quarter into a single block, which is often the only shape that fits a senior calendar.
ImplementationHow to put it into practice§
The four steps below turn the model in section 03 into a measurement plan a benefits team can run inside one plan year. None of them requires a new vendor to start. The first two are diagnostic, the third is structural, and the fourth is what makes the result comparable with a peer set.
- 01
Define the population you are actually benchmarking
Write down the denominator before you look at any number. Senior leaders, partners or attending physicians are a distinct population with a distinct barrier profile, and a utilization rate computed across the whole workforce cannot be disaggregated after the fact. Ask your current vendor whether the data can be cut by band at all. Many contracts cannot produce it, and finding that out is itself a benchmarking result.
- 02
Instrument the four gates separately
Awareness is testable with a two-question pulse survey. Relevance is testable by asking whether the offer would fit the last hard month the respondent had. Exposure is testable by asking who would know. Depth is testable from vendor data on completed episodes against opened ones. Four small measurements locate the failure that one utilization percentage only reports.
- 03
Remove the record from the decision
The exposure gate does not respond to communication campaigns because the concern is not attitudinal. It responds to arrangements in which no claim is filed, no diagnosis code reaches a plan, and the clinical record sits with the treating clinician rather than with any employer system. Design that first, then communicate it plainly, in that order.
- 04
Re-benchmark on use, then on depth
Once the tier-level rate exists, the peer comparison becomes possible for the first time. Report it alongside the conventional offering and cost benchmarks rather than instead of them, and hold the program for the three-year horizon that the published return figures actually require.13
RecommendationsWhere to start§
Clinical
Match depth to the problem, not to the plan year
Session count caps are a procurement artifact, not a clinical one. Where the presenting picture is compound, a benefit that can offer a longer single block is more likely to be used at all than one that offers six short ones.
Clinical
Treat isolation as a measured variable
Gallup measured daily loneliness at 31 percent among leaders against 21 percent among individual contributors.07 That is a benchmarkable clinical variable, and the clinical cost of having no peer to talk to is what it describes.
Structural
Report utilization by tier, and publish the denominator
A rate without its denominator cannot be compared with anything. Ask for the senior-tier numerator and the senior-tier denominator as separate fields, which is also the only honest basis for how return on executive wellness spend is actually calculated.
Structural
Buy the executive layer separately
A general program and a senior-tier program fail at different gates, so one vendor contract rarely clears both. Treating the second as its own line item is what makes an executive mental health benefit designed as its own layer measurable rather than aspirational.
FAQCommon questions§
What is benefits benchmarking?
How do you benchmark mental health benefits?
Do executives use standard benefits?
How does private-pay billing work?
How is my privacy protected?
MethodologyHow this paper was built§
Methodology
This paper was assembled between June and August 2026 from published sources retrieved directly from the publishing organization in every case. Four categories of source were searched. First, annual employer benefit surveys: the KFF Employer Health Benefits Survey for 2024 and 2025, which interviewed 2,142 and 1,862 firms respectively; the Bureau of Labor Statistics Employee Benefits Survey for March 2025; and the Business Group on Health employer strategy surveys covering the 2026 plan year. Second, workforce research at scale: the McKinsey Health Institute Employee Mental Health and Wellbeing Survey fielded February to April 2022 across fifteen countries, with 14,509 employee respondents and 1,389 human resources decision makers; and Gallup World Poll data collected through 2025 and reported in the 2026 State of the Global Workplace series, including a leader subsample of 9,880. Third, profession-specific studies: the 2016 survey of 12,825 licensed attorneys published in the Journal of Addiction Medicine and summarized in The Bar Examiner; a national survey of more than 5,800 physicians on licensure questions and help-seeking published in Mayo Clinic Proceedings; and American Medical Association organizational data drawn from more than 12,400 physician responses across 81 health systems. Fourth, cost and return studies: Deloitte analyses published in 2019 and 2022, and a 2012 Center for American Progress review of 31 turnover case studies drawn from eleven research papers published between 1992 and 2007. Government and multilateral prevalence figures come from the World Health Organization, the National Institute of Mental Health and the Office of the U.S. Surgeon General. Four limitations should be read alongside the figures. The employee assistance program utilization band of 2.1 percent to 8 percent originates in a 2015 study of 44 organizations and is reported here as it appears in a 2020 review; it is a decade old and predates the expansion of digital vendors. Two of the largest workforce datasets, the McKinsey and Gallup surveys, are global rather than United States only, so national figures will differ. The Deloitte cost figures are United Kingdom pounds and the Deloitte return figures are Canadian dollars; neither has been converted, because a converted figure would imply a precision the source does not carry. And the 2012 turnover review draws on case studies published between 1992 and 2007, which makes the 213 percent executive figure an upper bound from an older literature rather than a current market rate. No CEREVITY internal intake data is used in this paper. Every numeric claim above carries a numbered citation to the reference list below, and every reference was opened and read rather than cited from a secondary summary, with the two exceptions noted in the list itself where a reporting organization is cited for a study published elsewhere.
References
- 01Business Group on Health. (2025). 2026 Employer Health Care Strategy Survey: Executive Summary. businessgrouphealth.org
- 02KFF. (2024). 2024 Employer Health Benefits Survey. Kaiser Family Foundation. kff.org
- 03KFF. (2025). 2025 Employer Health Benefits Survey. Kaiser Family Foundation. kff.org
- 04Brooks, C. D., and Ling, J. (2020). Are We Doing Enough: An Evaluation of the Utilization of Employee Assistance Programs to Support the Mental Health Needs of Employees During the COVID-19 Pandemic. Journal of Insurance Regulation, 39(8). Reports Compton and McManus (2015) and Amaral (2007). content.naic.org
- 05U.S. Bureau of Labor Statistics. (2025). Employee Benefits in the United States, March 2025. News release, 25 September 2025. bls.gov
- 06McKinsey Health Institute. (2022). Addressing employee burnout: Are you solving the right problem? mckinsey.com
- 07Gallup. (2026). Leaders Have Better Lives but Worse Days. gallup.com
- 08Gallup. (2026). State of the Global Workplace. gallup.com
- 09Albert, L., and Krill, P. R. (2016). Wellness and the Legal Profession: Implications of the 2016 Landmark Study. The Bar Examiner, March 2016. Reports Krill, Johnson and Albert, Journal of Addiction Medicine, 2016. thebarexaminer.ncbex.org
- 10Mayo Clinic News Network. (2017). Physician licensing laws keep doctors from seeking care. Reports Dyrbye et al., Mayo Clinic Proceedings, 2017. newsnetwork.mayoclinic.org
- 11American Medical Association. (2024). Physician burnout rate drops below 50% for first time in 4 years. ama-assn.org
- 12Deloitte. (2022). Mental health and employers: The case for investment, pandemic and beyond. deloitte.com
- 13Deloitte. (2019). The ROI in workplace mental health programs: Good for people, good for business. deloitte.com
- 14Boushey, H., and Glynn, S. J. (2012). There Are Significant Business Costs to Replacing Employees. Center for American Progress. americanprogress.org
- 15World Health Organization. (2024). Mental health at work. Fact sheet, 2 September 2024. who.int
- 16National Institute of Mental Health. (2024). Mental Illness. Statistics page, National Survey on Drug Use and Health 2022 data. nimh.nih.gov
- 17Office of the U.S. Surgeon General. (2025). Workplace Mental Health and Well-Being. U.S. Department of Health and Human Services, last reviewed January 2025. hhs.gov
- 18Society for Human Resource Management. (2026). SHRM Employee Benefits Survey. shrm.org
PsyD, Licensed Psychologist
Benjamin Rosen, PsyD is a clinical psychologist licensed in California, seeing clients by telehealth through CEREVITY's nationwide network of independent licensed clinicians. He works with startup founders, AI and tech executives, venture-backed leaders and innovation leaders on founder and executive burnout, on the ethical complexity of emerging technology, on identity shifts during scaling, on anxiety and high-stakes decision-making, and on isolation in senior leadership. His clinical work draws on acceptance and commitment therapy, cognitive behavioral therapy, existential therapy and solution-focused therapy.
Talk to someone who gets it.
If this paper describes something you recognize, a confidential conversation is the next step. CEREVITY matches you to an independent licensed clinician who works with people in your position.
Schedule consultationPrivate-pay, telehealth, nationwide. Questions: (562) 295-6650