From Paper Targets to Patient Mix: Auditing Diversity Action Plans Against the Enrollment Record
In 2026, diversity action plans are no longer a novelty in US drug and device development. Sponsors know the vocabulary: underrepresented populations, enrollment goals, site selection, retention. The harder question is whether that vocabulary has migrated from the plan to the patient mix. A plan is a promise about future enrollment. An enrollment ledger is a record of what happened. The distance between the two is where accountability lives.
The Plan Is Not the Proof
Federal encouragement of diversity action plans has given sponsors a structured way to describe whom they intend to enroll. The existence of such a plan, however, does not establish that a trial was designed to reach the patients it names. Nor does a table of demographic targets prove that sites were staffed to enroll them. Sponsors can file a polished document describing aspirational enrollment while selecting sites by historical speed, investigator convenience, or prior relationship. Those choices are not necessarily improper, but they shape who gets asked to participate. A serious review therefore begins with the operational choices behind the plan: which sites were activated, how they were funded, what language capacity they had, whether reimbursement for travel was available, and how follow-up was structured.
Reading the Enrollment Ledger
Clinicians and researchers should ask for a simple table: planned enrollment by demographic group, actual enrollment, screen failure rates, and completion rates. The denominators matter. A trial that enrolls 20 Black patients out of 200 is not equivalent to one that enrolls 20 out of 2,000. Representation ratios should be calculated against the eligible disease population, not against the general US population alone. Age, sex, geography, disability, and comorbidity deserve the same scrutiny. FDA Drug Trials Snapshots and ClinicalTrials.gov results can help, but they are downstream artifacts. Site-level data are more revealing. A site that screens many Black patients but enrolls few may have eligibility criteria or consent friction that the plan never anticipated. A site that enrolls well but loses participants before the primary endpoint may be manufacturing a diverse baseline and an undiverse result.
Retention is the second ledger. A trial can enroll a diverse cohort and lose it by week 12. Sponsors should report follow-up completion by group, protocol deviations, and missing primary endpoint data. Without retention data, diversity is a recruitment photograph, not an outcome.
Three Ways Plans Become Paper
First, vague targets. A statement that a sponsor will increase enrollment of underrepresented groups is not a target. A target needs a denominator, a time frame, a responsible owner, and a budget line. Second, site theater. Adding sites in diverse areas without funding their research infrastructure is not inclusion; it is a map exercise. Third, unreported amendments. Plans change. When enrollment lags, sponsors may quietly relax or redefine goals. Those changes belong in the public record, not in a private slide deck.
Metrics That Would Make Plans Auditable
A credible diversity action plan should commit to a small set of auditable measures: disease-specific representation ratios; screen-to-enroll conversion by demographic group; site activation and accrual by catchment area; retention and data completeness by group; consent comprehension checks; and a plan-to-results concordance table at disclosure. The last item is essential. At the time results are posted, sponsors should show what they promised, what they achieved, and what they changed. Institutions should require that table during feasibility review. Journals should ask for it when trials make diversity claims. IRBs should see the operational plan, not just the consent form.
What Clinicians Can Do Now
When appraising a trial, clinicians should ask whether the enrollment composition matches the disease burden in the treated population. They should look for retention data before accepting subgroup claims. When serving on advisory boards or protocol review committees, they should demand that diversity action plans include named operational owners, budget lines, and stop or go criteria. A plan without a budget is a wish. A target without a denominator is a slogan. A commitment without a retention strategy is a recruitment campaign.
Diversity action plans can become instruments of accountability. They can also become another compliance document. The difference is not the eloquence of the filing. It is whether the enrollment ledger can be read, compared, and challenged.