Women’s leadership programs fall short on gender equality, review finds

A major systematic review published in eClinicalMedicine has concluded that women’s leadership development programmes, despite their global popularity and their framing as tools for gender equality, are failing to demonstrate the organisational change they are routinely claimed to deliver. The review, led by Daisy Morris and colleagues including Sana Rauf, Kathleen Riach and Helena Teede, synthesised 57 evaluations of 52 distinct programmes across 23 countries and found a striking mismatch between the equity ambitions that justify these programmes and the evidence they actually generate. The findings arrive at a moment of intense scrutiny for diversity interventions worldwide, and they challenge healthcare systems, universities and other institutions to rethink how leadership development for women is designed, delivered and judged.

The scale of the gender disparity the review set out to interrogate is well documented. Women make up roughly 70 per cent of the global health workforce and 41 per cent of active researchers, yet they hold only about a quarter of leadership and senior roles. Against this backdrop, women’s leadership development programmes, or WLDPs, have proliferated across healthcare, academia and beyond, particularly since the United Nations adopted Sustainable Development Goal 5 on gender equality in 2015. These programmes typically combine skills training, workshops, coaching, mentoring, networking and leadership identity development, and they are increasingly positioned not merely as career boosts for individual women but as interventions capable of transforming the institutions that employ them.

To test that claim, the research team searched nine bibliographic databases covering literature published between 25 September 2015 and 5 July 2026, a window deliberately anchored to the adoption of SDG5. They included only peer-reviewed evaluations of structured programmes delivered to adult professionals in formal organisational roles, excluding student-only populations and interventions consisting solely of mentoring or coaching. From 2314 initial records, 57 studies were retained. Because the evidence was too heterogeneous for statistical pooling, the team conducted a narrative synthesis registered prospectively on PROSPERO and reported in line with PRISMA and the Synthesis Without Meta-analysis guideline. Their central analytical device was an extended version of the Kirkpatrick evaluation framework, which distinguishes subjective, self-reported outcomes from objective, externally verifiable ones across seven levels, ranging from participant satisfaction through learning and behavioural transfer to organisational results.

The headline result is stark. Every one of the 57 studies reported outcomes at the individual level, but only 32 studies, or 56 per cent, captured any organisational measure at all, and these were typically descriptive metrics that could not attribute change to the programme. Subjective outcomes dominated the lower rungs of the framework: 88 per cent of studies reported self-reported knowledge gains such as confidence and leadership self-efficacy, compared with just 30 per cent reporting objectively assessed learning. At the behavioural level, 65 per cent relied on participant or colleague accounts of applying programme lessons, while 44 per cent documented verifiable behaviour such as leadership appointments, grants or awards. At the organisational level, the pattern inverted, with 44 per cent of studies reporting objective institutional data such as promotion and retention figures, but these came overwhelmingly from a small number of mature, well-resourced programmes, most often in academic medicine, where administrative data systems made tracking possible.

Methodological fragility compounded the problem. No study used randomisation, and only four included a verified non-participant comparison group. Only 40 per cent of studies had a pre-programme baseline, and just 28 per cent followed participants for at least a year, although follow-up in some cases extended to a decade. Ninety-six per cent of studies relied on self-reported data, and 68 per cent used summative evaluation only. Studies with long-term follow-up were far more likely to report objective behavioural and organisational outcomes, and the six sensitivity analyses the team conducted suggested that the scarcity of organisational evidence reflects what programmes choose to measure rather than the inclusion of methodologically weaker studies, since reporting did not improve when the synthesis was restricted to lower-risk or better-designed evaluations.

Perhaps the most provocative finding is what the authors call an equity paradox. Where recruitment processes were described, 37 of 43 studies, or 86 per cent, used selective entry, most commonly nomination or invitation by senior leaders, supervisor endorsement or competitive selection. In other words, programmes explicitly framed as gender equity interventions frequently reached women who were already institutionally visible and embedded in sponsorship networks, potentially concentrating development opportunities among those already recognised as high-potential. Demographic transparency was correspondingly thin: race or ethnicity was reported in only 39 per cent of studies, most cohorts were described as predominantly White, and just a single study disaggregated advancement outcomes by race or ethnicity. Sexual orientation and disability status were almost never captured, and no study reported across all dimensions of the PROGRESS-Plus equity framework the team used to map access and participation.

The second recurring pattern is an implementation-evaluation gap. Programme rationales frequently invoked structural explanations for women’s underrepresentation, including gender bias, the leaky pipeline and systemic inequality, yet programme components in all 57 studies targeted individual participants, and only one programme included organisation-facing elements such as sexual harassment policy development. Explicit alignment between a programme’s stated rationale, its delivered components and its measured outcomes was identified in just four studies, or 7 per cent. The authors argue that this misalignment effectively places the burden of change on individual women navigating inequitable systems rather than on the organisations that create and sustain those systems, a dynamic consistent with sociological evidence that one-off diversity training without structural accountability tends to produce limited institutional movement.

The authors are careful to note that limited organisational evidence is not evidence of ineffectiveness. Organisational transformation unfolds over years and depends on enabling environments, long follow-up, administrative data infrastructure and dedicated evaluation funding that few programmes command, since most are resourced for delivery rather than rigorous longitudinal assessment. Encouragingly, studies reporting organisational embedding or support, such as executive sponsorship, protected time and dedicated funding, were significantly more likely to report organisational outcomes, and the handful of programmes with comparison groups all measured institutional results. The review also acknowledges that qualitative and subjective evaluations, which feminist and participatory traditions regard as legitimate evidence, may be methodologically appropriate for small, context-specific programmes; the concern is not that self-reported change lacks value, but that changes in confidence, voice or leader identity cannot demonstrate that the barriers participants face have themselves shifted.

The implications reach into the design of the Sustainable Development Goals themselves. Goal 5 commits governments to empowering all women and girls, and the authors argue that the current generation of WLDPs, as designed and evaluated, does not appear to function as an organisational equity strategy capable of contributing measurably to that target. They call for a reconceptualisation built on transparent, equitable and accessible recruitment; outcome reporting disaggregated by race, career stage and other intersectional dimensions; and alignment of design, implementation and evaluation with organisational accountability mechanisms. Institutions that commission these programmes, they contend, must accept shared responsibility for the structural conditions that determine whether individual gains can be translated into systemic change, rather than treating leadership development as a substitute for policy reform.

The review directly informs the large-scale, Australian National Health and Medical Research Council-funded international initiative Advancing Women in Healthcare and Academic Leadership and the National Women in the Health and Medical Sciences Decadal Plan. Its authors acknowledge limitations, including the exclusion of grey literature and non-English publications, the absence of randomisation in the underlying evidence and the geographic concentration of studies in high-income countries, with the United States alone accounting for 63 per cent of the included research. Even so, the central message is difficult to escape: women’s leadership programmes consistently benefit the women who attend them, but whether, and under what organisational conditions, they shift the inequitable patterns that produce gender leadership gaps remains, on the evidence assembled here, an open question that institutions can no longer afford to leave unmeasured. Echoing the 2025 UN gender snapshot’s call for better data and coordinated action, the study reframes leadership development not as a finished solution but as one component of a broader, harder accountability project whose success must be made empirically visible.