The most consequential finding in the American Bar Association Commission on Women in the Profession’s new national study is not a prevalence number. It is a sentence about awareness. The women lawyers surveyed could generally name the attorney mental health resources their employers offered. They knew the employee assistance line existed. They knew counseling was covered. And they did not use it, because they did not believe it was confidential, and because they suspected that using it would be read as a lack of commitment in workplaces that reward constant availability. That finding appears in Mental Wellbeing in the Legal Profession: The Unique Challenges of Women Lawyers and a Path Forward, the first national study centered on the mental well-being of women lawyers, and it has been largely overshadowed by the report’s headline disparities.
It should not be. Read carefully, the report is less a description of a distressed population than an indictment of a delivery model. The profession has spent nearly a decade since the National Task Force on Lawyer Well-Being building infrastructure — assistance programs, resilience training, wellness committees, screening tools — on the assumption that the binding constraint was access. This study suggests the binding constraint is trust, and that the two are not the same problem with the same solution. Adding another resource to a system a lawyer does not trust produces one more unused resource.
What the study actually measured
The research rests on an anonymous survey of 2,915 lawyers, sixty-nine percent of whom were women, supplemented by twelve focus groups. Symptoms were screened with the four-item Patient Health Questionnaire, which asks about anxiety and depressive symptoms over the preceding two weeks. Among women who completed the instrument, 36.9 percent scored at or above the threshold that may warrant further clinical evaluation for anxiety, compared with 22.7 percent of men. For depressive symptoms the figures were 19.4 percent and 15.1 percent. Asked to rate how happy, calm, and peaceful they had been over the same two weeks on a hundred-point scale, women averaged 55.6 and men 62.9.
Those numbers deserve a caveat the report itself supplies and most coverage omits. The PHQ-4 is a screening instrument, not a diagnostic one. A score above threshold indicates that a conversation with a clinician is warranted, not that a disorder is present. This distinction matters more in this profession than in most, because lawyers are trained to read a number as a finding, and because the profession’s relationship to its own mental health data has been contested. Earlier this year, Benjamin Pyle and Clifford Rosky published a comparison of two high-quality national surveys in the Journal of Empirical Legal Studies and found that they disagree sharply about how many lawyers are unwell. The National Health Interview Survey places lawyers at or below the general public on psychological distress. The National Survey on Drug Use and Health places more than forty percent of lawyers in moderate or serious distress within the past year. Both surveys agree that lawyers report elevated alcohol misuse relative to the general public and to similarly educated peers.
The stressor data is where the ABA study becomes most useful. Women reported higher rates across nearly every category: financial strain at 40.0 percent against 29.5 percent for men, mental wellness concerns at 35.9 against 26.1, family relationship issues at 29.9 against 18.6. The single most frequently reported stressor among women lawyers was sleep disturbance, at 30.6 percent against 18.6 percent. Hold that finding for a moment; it turns out to be the load-bearing one.
Why attorney mental health resources go unused
The conventional frame for underused mental health benefits is stigma, and stigma is usually described as something irrational that education can dissolve. The lawyers in this study described something else. They described a confidentiality structure they had evaluated and found wanting, and a professional environment in which help-seeking carries an inference. Women in the focus groups tied their stress to workplace systems rather than to personal deficits — billable-hour requirements, workplace culture, caregiving load, the experience of being second-guessed and expected to prove competence — and they expressed particular distrust of employer-provided counseling.
Consider what an employer-sponsored well-being program asks a lawyer to accept. The organization providing the confidential clinical service is the same organization that sets the billable target, evaluates the work, determines compensation, and decides advancement. The lawyer is asked to believe that a firewall exists between the entity holding her clinical information and the entity holding her career. She has no way to audit that firewall. She has, however, spent her professional life learning to evaluate exactly this kind of arrangement on behalf of clients, and to advise them not to rely on assurances they cannot verify when the downside is severe and irreversible.
Declining to use the program is therefore not avoidance in the clinical sense. It is the correct application of professional judgment to a situation with asymmetric information and an unfavorable downside. The report’s own recommendations implicitly concede this, emphasizing that the effectiveness of a well-being policy depends far less on its existence than on whether leaders visibly use it and communicate that no professional penalty attaches. That is a recommendation about credibility, not about access.
There is a striking piece of convergent evidence in the Pyle and Rosky analysis. Attempting to explain why the two national surveys diverge so much, the authors point to methodological advantages of the survey that found more distress, and among them is the privacy of its data-gathering method. The plainest reading is that people, lawyers included, report their symptoms differently depending on the conditions under which they are asked. If that is true at the level of a federal health survey, it is emphatically true at the level of a firm-administered screening. The profession has been measuring lawyer distress and offering lawyer help through channels whose confidentiality the respondents doubt, and then treating the resulting numbers as a floor.
Why the sleep finding is not a symptom
Sleep disturbance is easy to file as a downstream consequence — stress arrives, sleep suffers, address the stress and sleep resolves. The neuroscience does not support treating it that way, and for lawyers specifically the misclassification is costly.
In a study that has become foundational to the field, Seung-Schik Yoo, Matthew Walker, and colleagues used functional imaging to examine the emotional brain after sleep deprivation. Sleep-deprived participants showed markedly amplified amygdala responses to aversive images, alongside a weakening of the functional connection between the amygdala and the medial prefrontal cortex — the region that ordinarily contextualizes a threat signal and modulates the response to it. The result is a brain that reacts more strongly to provocation and regulates that reaction less well.
Now place that finding inside a working day built on adversarial exchange. Opposing counsel writes something insulting. A client questions the strategy. A partner’s edits arrive with an edge. Legal work does not merely expose a lawyer to provocation; it requires the lawyer to metabolize provocation and respond with calibrated judgment rather than reactivity. That calibration is precisely the prefrontal function that insufficient sleep degrades. The most frequently reported stressor in the ABA study is therefore not only a quality-of-life matter. It is a direct tax on the cognitive capacity the profession is paid for.
The relationship also runs in both directions, which is what makes it a self-sustaining loop rather than a linear sequence. Anxiety fragments sleep; fragmented sleep amplifies the reactivity that generates anxiety; the amplified reactivity produces more of the conflict and rumination that fragment the next night. A lawyer inside that loop experiences it as evidence of personal deterioration — she is snapping at people, she cannot let a hostile email go, she is drafting at eleven at night because the day’s concentration failed — when it is better understood as a physiological cycle with an identifiable entry point.
The data on how common this is extends well beyond one report. Bloomberg Law’s 2024 attorney well-being survey found disrupted sleep the single most commonly reported challenge at 56 percent, with anxiety immediately behind it at 55 percent. Nor is the pressure easing. Writing in the ABA Journal this spring, Jason Ward argued that generative artificial intelligence is quietly compressing deadlines and adding a new species of labor — the continuous verification of machine output, which converts saved time into sustained vigilance rather than into recovery.
What this means for the lawyer reading it
Three things follow, and none of them is a wellness tip.
The first concerns what to watch. Sleep is worth treating as a leading indicator rather than as wreckage to be cleaned up after a hard stretch. A month of broken sleep is not the residue of a difficult trial; it is a condition that will shape how the next difficult matter is handled. The two-week window the PHQ-4 uses is a reasonable interval for honest self-observation — not to score oneself, and certainly not to self-diagnose, but to notice whether a bad fortnight has quietly become the baseline. The women in the study reported that talking with family and friends, exercising, and spending time outdoors were real and valuable, and also insufficient when workload expectations were unchanged. That is an important pairing to hold intact. These practices are worth keeping, and they will not resolve a structural problem on their own.
The second concerns the arithmetic of confidentiality. If the reason a lawyer has not sought care is a rational assessment that the employer-provided channel cannot be verified as safe, then the answer is not to talk herself out of the assessment. The answer is to change the facts it applies to. A clinician in independent practice, retained privately and situated outside the firm’s reporting and reimbursement structures, is not subject to the conflict that made the calculation come out the way it did. This is an unglamorous, structural observation, and it resolves more cases of non-use than any amount of anti-stigma messaging.
The third concerns fit. Legal culture is specific, and a great deal of clinical time can be lost explaining why a two-week turnaround is not negotiable, what a partnership vote does to a person, or why “just set a boundary with the client” is not responsive advice. Working with a clinician who already understands the structure of legal practice is not a luxury; it removes a translation burden from the person least able to carry it.
The question about to be reopened
Within the coming months, the profession will receive its most comprehensive picture in a decade. The American Bar Association, through its Commission on Lawyer Assistance Programs, and Krill Strategies have surveyed a nationally representative sample of roughly 36,000 lawyers, producing a ten-year update to the landmark 2016 study conducted with the Hazelden Betty Ford Foundation and the first national estimate of work-related burnout in the profession. The findings are moving through peer review.
Those numbers will be reported as a verdict. They will be more useful read as a question, because the lesson of the Commission on Women’s report is that the conditions of asking are not incidental to the answer. A profession that measures its members through channels those members do not trust, and then offers help through the same channels, will keep producing data that undercounts the problem and programs that go unused. The remedy is not another resource. It is a channel a lawyer has reason to believe in.
AttorneyTherapists.com maintains a directory of licensed clinicians who work specifically with attorneys, in independent practice and outside any employer’s structure.
