How to Vet a Retention Vendor For Clinical Trials

Patient retention in clinical trials looks better on paper than the number most vendors quote. In 151 publicly funded randomized trials, a median 89% of randomized participants had valid primary outcome data at their final follow-up, with an interquartile range of 79 to 97% [1]. Across 235 trials reporting a significant primary outcome, the median share lost to follow-up was 6%, yet plausible assumptions about what happened to those participants would have removed statistical significance in as many as 58% of them [2].

A single dropout percentage does not settle whether a trial's result holds up. Who leaves, when, and how the missing data gets handled matter more than the raw count. The companion piece on this site covers how often enrollment actually slips, and what a delay day costs; this one picks up after consent.

Refero does not run retention programs. It vets and matches the healthcare marketing and patient recruitment agencies that support them, so what follows is written for the person deciding whether to hire one, and what to ask before they sign.

Patient retention in clinical trials: statistics at a glance

Every figure below is sourced in full in Sources at the foot of the page.

• 89% median retention across 151 publicly funded trials, interquartile range 79 to 97% [1]
• 6% median loss to follow-up across 235 trials, interquartile range 2 to 14% [2]
• A 2021 Cochrane review of 81 retention trials found only four comparisons with moderate-certainty evidence [5]
• Barrier-reduction strategies were associated with a 10% higher retained proportion in longitudinal cohort studies [6]
• Recruitment and retention was the single biggest challenge for 36% of 852 research sites surveyed [7]

What patient retention in clinical trials means, and why "dropout" is three different things

Patient retention in clinical trials is the share of randomized participants who complete the protocol and provide the data the primary outcome needs, usually reported as the retained proportion at the primary endpoint. "Dropout" gets used to describe three distinct events: loss to follow-up (the participant stopped providing data and the study team cannot reach them), treatment discontinuation (the participant stopped the intervention but may still be measured), and withdrawal of consent (the participant asked to leave and data collection stops). ICH E9(R1) treats all three as intercurrent events the analysis plan has to anticipate, not just count after the fact [8]. A retention vendor that promises to "reduce dropout" has to say which of the three outcomes its work actually moves.

What the dropout benchmarks actually show

No single number describes patient retention across clinical trials. The studies that have tried to measure it report a spread wide enough to make any one figure misleading on its own.

The umbrella review is the widest look at the question, and it declined to report a single pooled figure, finding attrition running higher in cancer, obesity, and psychological-condition trials without collapsing that into one number [3]. That leaves the "acceptable dropout rate" question without a universal threshold. The number that matters is the attrition assumption already built into the trial's sample-size calculation, paired with a pre-specified plan for handling the missing data once dropout happens.

Why the dropout rate alone misleads: who leaves matters more than how many

A lower dropout number is not automatically a better trial. Differential dropout, meaning a bigger loss in one study arm than the other, can bias a trial's result regardless of the overall rate [10]. At a median loss of just 6%, plausible assumptions about what happened to the missing participants would have removed statistical significance in as many as 58% of 235 trials [2]. A retention program that lifts completion in one arm more than the other has not made the trial more reliable; it may have made the result harder to interpret. Any retention provider should be able to say how it monitors retention by arm, not only overall.

Why participants leave

Participants leave trials most often because the trial becomes a burden. In a 2025 survey of trial participants, biopsies (34%), travel to the clinic (28%), and diagnostic tests (24%) were named the most burdensome parts of taking part; 35% traveled more than an hour each way; and one in three received no compensation or reimbursement of any kind [11]. Protocol complexity compounds this: procedures required per participant visit are up 37% over the past 15 years [12], and recruitment and retention was the single biggest challenge for 36% of 852 surveyed research sites [7].

Retention interventions with evidence behind them

Most of what gets pitched as a retention strategy has not been tested against a control group. The interventions below are the exceptions.

Most of the evidence measures questionnaire return, not whether a participant stayed on treatment. The Cochrane review found only four comparisons at moderate certainty out of 81 retention trials [5], and reminder strategies were associated with worse retention in cohort data, likely because they get deployed on the hardest-to-retain participants [6]. A vendor that cannot say which evidence category its tactics fall into is asking a sponsor to take the effect on faith.

What retention service providers actually do

A retention service provider changes one or more sources of participant burden: travel and concierge logistics, stipend and reimbursement handling, home visits or mobile nursing, participant communication and reminders, site-support staffing, or advocacy partnerships. It does not change the protocol, and it cannot make a trial retain better than its design and its sites allow. The provider categories map to the evidence lines above: travel and concierge services rest on burden-reduction association data [6][11]; payment platforms rest on the moderate-certainty incentive evidence [4][5]; home-health services draw on decentralized-elements guidance with no effect size yet [14]; communication platforms face the counterintuitive finding that reminder-heavy programs correlate with worse retention in cohort data [6].

No published randomized trial tests a commercial retention vendor's own program against a control group; vendor completion-rate figures describe their own book of business, not a controlled comparison. There is also no sourced figure for what a replacement patient costs once one drops out (Refero's guide to what recruitment actually costs per enrolled patient covers the budget side that does have sourced figures, not yet live).

Three numbers to stop repeating

Retention content recycles figures that do not survive a look at their source. Two of them are covered on the page about the two dropout figures we already traced to their dead ends; this section covers three more.

"30% of participants drop out of clinical trials on average." The chain runs through a 2018 review [17] citing a 2013 feature article [16] that reports the figure as industry lore without a dataset. The measured picture looks nothing like a flat 30%: median retention was 89% [1], median loss to follow-up was 6% [2], and the widest review declined to pool a single number [3].

"85% of clinical trials fail to retain enough patients." This traces to trade-magazine articles and vendor infographics. No dataset defines what "enough" retained patients means, and no primary source could be located.

"Up to 20% dropout is acceptable." This comes from the "5 and 20" heuristic in a 2002 methods paper whose authors framed it as a rough guide, not a threshold [18]. At a median loss of 6%, worst-case assumptions would have removed statistical significance in 58% of trials [2], and whether dropout biases a result depends on who leaves, not the count against a fixed cutoff [10].

How to choose a patient retention vendor

These questions come from the failure modes the evidence above actually shows.

• Which of the three outcomes do you move: loss to follow-up, treatment discontinuation, or withdrawal of consent? A provider that cannot name the outcome cannot have measured it.
• What is your retained proportion at the primary endpoint on your last three studies, broken out by arm? Overall completion hides differential dropout, which is what actually biases a result [10].
• Which of your tactics has evidence above the case-study level, and where does it come from? Only four retention comparisons reach moderate certainty in the Cochrane review [5]; ask which of them the program actually uses.
• What do you change about visit burden, and what do you leave to the protocol? Travel and procedures are what participants themselves name as the burden [11], and a communication layer alone does not touch either one.

If you would rather not run that evaluation alone, Refero screens healthcare marketing and patient recruitment agencies against five published criteria and introduces up to three that fit your brief. It is free for buyers, and there is no obligation to hire anyone we introduce. Tell us what you need.

Frequently asked questions

What is patient retention in clinical trials?

Patient retention in clinical trials is the share of enrolled participants who stay in the study through their final visit and provide the data the primary outcome needs. In 151 publicly funded randomized trials, the median was 89%, with half falling between 79 and 97% [1]. The term is not the same as a clinic keeping patients coming back or record retention.

What is an acceptable dropout rate in a clinical trial?

There is no universal threshold. The real target is the attrition assumption built into the trial's sample-size calculation, plus a plan for handling the missing data. The often-quoted "5 and 20" rule comes from a 2002 methods paper whose authors framed it as a rough heuristic [18], and even at a median loss of 6%, worst-case assumptions would have removed statistical significance in 58% of 235 trials [2].

What is the average dropout rate in clinical trials?

No credible pooled average exists. The measured picture: 89% median retention across 151 publicly funded trials [1], 6% median loss to follow-up across 235 trials [2], and typical losses of up to 12% with rates as high as 70% in some studies [3]. The often-repeated 30% average traces to a 2013 feature article that offered no dataset [16][17].

Why do patients drop out of clinical trials?

Participants leave most often because the trial becomes a burden. In a 2025 survey, biopsies (34%), travel (28%), and diagnostic tests (24%) were named the most burdensome parts; 35% traveled more than an hour each way; and one in three received no compensation [11]. Procedures per visit rose 37% over 15 years [12].

Which patient retention strategies in clinical trials have evidence behind them?

Very few, and most evidence measures questionnaire return rather than staying on treatment. A 2021 Cochrane review of 81 retention trials found only four comparisons with moderate-certainty evidence [5]. A monetary incentive with a postal questionnaire raised returns by a relative risk of 1.18 [4], and reminder-heavy programs were associated with worse retention in cohort studies [6].

Sources

• Walters SJ, et al. Recruitment and retention of participants in randomised controlled trials: a review of trials funded and published by the United Kingdom Health Technology Assessment Programme. BMJ Open, 2017. Link
• Akl EA, et al. Potential impact on estimated treatment effects of information lost to follow-up in randomised controlled trials (LOST-IT): systematic review. BMJ, 2012. Link
• McChrystal R, et al. Participant and trial characteristics reported in predictive analyses of trial attrition: an umbrella review of systematic reviews. Trials, 2025. Link
• Brueton VC, et al. Strategies to improve retention in randomised trials: a Cochrane systematic review and meta-analysis. BMJ Open, 2014. Link
• Gillies K, et al. Strategies to improve retention in randomised trials. Cochrane Database of Systematic Reviews, 2021. Link
• Teague S, et al. Retention strategies in longitudinal cohort studies: a systematic review and meta-analysis. BMC Medical Research Methodology, 2018. Link
• WCG Clinical. 2024 Clinical Research Site Challenges Report. Link
• ICH. E9(R1) Addendum on Estimands and Sensitivity Analysis in Clinical Trials, 2019. Link
• Poongothai S, et al. Strategies for participant retention in long term clinical trials: a participant-centric approaches. Perspectives in Clinical Research, 2023. Link
• Bell ML, Kenward MG, Fairclough DL, Horton NJ. Differential dropout and bias in randomised controlled trials: when it matters and when it may not. BMJ, 2013. Link
• CISCRP. 2025 Perceptions and Insights Study. Link
• Getz K, et al. Insights Informing Strategies for Optimizing the Collection of Clinical Trial Data. Therapeutic Innovation and Regulatory Science, 2025. Link
• US Food and Drug Administration. Payment and Reimbursement to Research Subjects: Information Sheet, 2018. Link
• US Food and Drug Administration. Conducting Clinical Trials With Decentralized Elements: Guidance for Industry, 2024. Link
• US Food and Drug Administration. Enhancing Participation in Clinical Trials: Eligibility Criteria, Enrollment Practices, and Trial Designs. Guidance. Link
• Alexander W. The uphill path to successful clinical trials: keeping patients enrolled. P and T, 2013. Link
• Fogel DB. Factors associated with clinical trials that fail and opportunities for improving the likelihood of success: a review. Contemporary Clinical Trials Communications, 2018. Link
• Schulz KF, Grimes DA. Sample size slippages in randomised trials: exclusions and the lost and wayward. Lancet, 2002. Link