Patient recruitment remains the single biggest cause of clinical trial delay because the pool narrows at several independent points: eligibility criteria written for statistical cleanliness, site selection that concentrates capacity in a few academic centers, and patient awareness gaps that keep eligible people from ever hearing about the trial.
Why is recruitment still the biggest cause of trial delay?
Every other part of a clinical trial can be planned to a schedule. Manufacturing runs on a timeline, regulatory review has predictable windows, and data management scales with headcount. Recruitment does not behave the same way, because it depends on a population that has to exist, be found, be eligible, and agree, all before a single data point is collected.
That dependency is why recruitment absorbs delay from everywhere else in the trial without ever being the stated cause. A protocol amendment that tightens an inclusion criterion, a site that opens three months late, a competing trial recruiting from the same clinic list, all of these show up downstream as a slow enrollment curve, even though the actual problem sits somewhere else in the design.
The result is a category of delay that is hard to budget for because it looks different in every trial and shares no single root cause across programmes. A sponsor can hit every internal milestone for study startup and still watch enrollment stall for a reason nobody flagged during protocol review, because the constraint was never in the plan itself, it was in the population the plan assumed existed.
- 01
Model criteria compounding
Estimate the eligible pool after every exclusion is applied together, not one at a time.
- 02
Map real site capacity
Check current investigator caseload against site history before committing to a site list.
- 03
Build the referral path
Give treating physicians a concrete, low-effort way to flag eligible patients.
- 04
Add decentralized options
Layer remote visits and local labs onto site-based recruitment rather than replacing it.
How does protocol design narrow the eligible pool?
Eligibility criteria are written to produce a clean, interpretable dataset, not to reflect how the condition actually presents in practice. Each additional exclusion, a prior treatment, a comorbidity, a lab value outside a narrow band, removes a slice of the population that would otherwise qualify, and these exclusions compound rather than add.
A protocol with ten criteria that each exclude twenty percent of an otherwise eligible population does not leave eighty percent of that population available. It can leave a small fraction, because the exclusions are not evenly distributed across the same people. Sponsors that model this compounding effect before locking the protocol tend to find the eligible pool is a tenth the size the epidemiology suggested.
The criteria are rarely arbitrary. Each one usually protects against a real confound, a prior therapy that would muddy the efficacy signal, a comorbidity that raises safety risk, a lab value tied to disease severity. The problem is that these decisions are made criterion by criterion, in isolation, without anyone modelling how they interact once applied to the same real-world population at the same time.
What role does site selection and investigator burden play?
Trials are typically recruited through a limited set of sites, and site selection tends to favor academic medical centers with existing research infrastructure over the community clinics where most patients with a given condition are actually treated. That concentrates competition: several sponsors running similar trials draw from the same short list of principal investigators and the same regional patient base.
Investigator burden compounds the constraint. A site coordinator juggling six active protocols has limited time to screen charts for a seventh, and screening effort, not patient scarcity, is often the actual limiting factor. Programmes that commission structured patient and physician research before site selection tend to identify which sites have real, uncommitted capacity rather than relying on prior trial history alone.
Site history is a poor proxy because it measures whether a site has run trials before, not whether it currently has bandwidth or an actively engaged coordinator. A site with a strong enrollment record from two years ago may now be running four other protocols and screening none of them thoroughly, while a site with a thinner track record but current capacity gets passed over on paper alone.
| Site-based recruitment | Decentralized/hybrid | |
|---|---|---|
| Patient travel burden | High, limits radius around each site | Reduced, widens practical catchment |
| Screening dependency | Relies on site coordinator time | Can supplement with remote screening |
| Best suited to | Common conditions, dense patient clusters | Rare or geographically scattered populations |
Why don't eligible patients hear about the trial?
A meaningful share of patients who would qualify for a trial never learn it exists, because awareness depends on their treating physician mentioning it, and physicians are not incentivized or resourced to track every open protocol relevant to their patient panel. The referral pathway from diagnosis to trial enrollment has several points where it can simply not happen.
Decentralized and hybrid trial designs, remote monitoring, local lab draws, telehealth visits, address part of this by reducing the travel burden that keeps otherwise willing patients from enrolling. They do not fix the awareness gap itself; they widen the radius around a site that a newly aware patient can practically join from.
Digital outreach, targeted advertising, patient community partnerships, social platforms, extends reach further, but it runs into a filtering problem of its own. Broad awareness campaigns generate large volumes of interest from people who turn out to be ineligible on screening, which shifts the bottleneck from finding candidates to processing inquiries, without necessarily shortening the time to a full enrollment.
Why is rare disease recruitment the extreme case?
Rare disease trials compress every recruitment problem into a population that is already small. A condition affecting a few thousand people globally leaves little room for exclusion criteria to remove further candidates, and the remaining patients are often geographically scattered rather than concentrated near a trial site.
This is also where the trade-offs between the biosimilar market and originator recruitment become visible: sponsors developing a therapy in a space where a biosimilar competitor already treats part of the eligible population face an even smaller pool of treatment-naive candidates. Patient advocacy groups and disease registries become the primary recruitment channel precisely because conventional site-based screening cannot find enough patients on its own.
The economics follow from the same scarcity. A rare disease trial can justify recruitment spend per patient that would be indefensible in a common indication, because the alternative is not a slower trial, it is a trial that never reaches a statistically usable cohort. That changes the calculus around decentralization, patient registries and international site networks from a nice-to-have into the only realistic path to enrollment.