Definition, deliverables, and adjacent boundaries
Participant recruitment is the professional process of reaching, informing, prescreening, referring, and supporting the enrollment of people who may meet an approved study protocol, while preserving a traceable record of each disposition. This entry sits in WSIC 6.5.2, “Participant and Sample Management.” It applies to drug, device, and behavioral intervention trials as well as observational cohorts, sleep experiments, and digital health studies. Its core deliverable is not a contact list. It is a defensible pathway through which the intended population has a fair opportunity to learn about research and an authorized study team determines eligibility without compromising voluntary choice.
Recruitment is not marketing acquisition, disease screening, clinical diagnosis, or the formal informed-consent process. A person who responds to an advertisement is only a potential lead; a high questionnaire score does not automatically make that person a patient, and completing a prescreen does not mean enrollment. Final eligibility must be established with the interviews, records, tests, and safety assessment specified by the protocol. Consent must be obtained by responsible, trained personnel through the ethics-approved process. A recruitment vendor must not substitute for an investigator’s medical judgment or promise assignment, efficacy, free treatment, or access to an investigational product.
Recruitment is also distinct from retention, biosample management, data management, and site management, although the interfaces must work together. A provider may operate a call center or website, manage community partnerships or digital advertisements, route site referrals, match candidates in a controlled database, and assist with appointments or travel. Agreements still need to state who approves materials, who may access health information, who confirms eligibility, who responds to safety concerns, and what happens to information after withdrawal. Sponsor, investigator, contract research organization, site, and technology platform remain accountable for the duties they assign and oversee.
Recruitability of the protocol and target-population design
Reliable recruitment begins with protocol feasibility, not media buying. The team should turn the condition definition, age, severity, prior treatment, comorbidities, medication rules, language, geography, device requirements, and follow-up burden into an executable eligibility pathway. De-identified aggregates, clinic volumes, and historical conversion rates can inform the candidate pool. Each estimate should then account for duplicates, lack of permission to contact, prescreen failures, diagnostic failures, and expected attrition. A database hit is not a consented participant, and neither is equivalent to a participant who can be randomized.
Eligibility criteria should serve the scientific question and safety rather than exclude people merely for convenience. Older adults, children, women, pregnant or breastfeeding people, people with disabilities, shift workers, rural residents, speakers of less common languages, and people with complex comorbidity should not be omitted without a scientific or ethical reason. The planned sample should resemble the population to which the result is intended to apply. Recruitment plans should set appropriate subgroup goals and compare the composition of inquiries, prescreens, enrollments, and withdrawals so that meeting a total count does not conceal dependence on one easily reached group.
Time and setting make recruitability especially important in sleep research. Overnight laboratory visits, fixed lights-out schedules, medication washout, sleep restriction, consecutive diaries, actigraphy, home-device setup, and morning transportation can disproportionately exclude people with inflexible work or caregiving duties. Potential participants, site staff, and community representatives should review the workflow during design. Unnecessary procedures should be removed, and reasonable scheduling, language, accessibility, remote, or transport options should be offered where they do not alter critical measurement conditions. Recruitment failure can be a protocol-design signal rather than simply a reason to spend more on advertising.
Channels, materials, and equitable reach
Channels may include sleep clinics and other clinical referrals, research registries, authorized health-record searches, community organizations, patient groups, schools or workplaces, public trial registries, search and social platforms, and traditional media. For each channel, the study should retain the material version, active dates, audience rules, cost, responses, and later disposition. Contact initiated by a treating clinician, targeting by a platform, and a list supplied by an outside vendor carry different expectations of trust and different data permissions. They should not be collapsed under a single generic authorization.
Recruitment materials should use language the intended audience can understand and accurately describe the study nature, purpose, main activities, duration, location, basic criteria, material burdens, investigational status, contact route, and general compensation arrangements. Headlines, images, short videos, search terms, landing pages, chat scripts, and automated replies form one information environment and should receive the review required in the governing jurisdiction with consistent version control. “May qualify” must not be turned into a diagnosis. Materials must not imply certain benefit, preferential clinical care, official endorsement, or zero risk, and a prominent payment figure must not eclipse time commitments and risks.
Equitable reach means removing participation barriers, not merely increasing impressions. Teams should identify language, digital access, transport, nighttime caregiving, time off work, disability accommodation, medical mistrust, and prior research harms, then work with community partners to revise explanations and support. Advertising algorithms should be monitored for repeatedly favoring the same high-response population. Sensitive health inferences, discriminatory exclusions, or unauthorized proxy profiles should not be used to chase conversion. Community relationships should disclose research goals, funding, and decision rights and should continue beyond a one-time request for names when enrollment is slow.
Prescreening, eligibility confirmation, and sleep phenotyping
Prescreening should collect only what is necessary to decide the next step and should begin by explaining that it is a research eligibility assessment, not clinical care. Telephone, web, or chat workflows should use approved scripts and retain the source and time of each response, the staff or system version, unresolved items, and the reason a path ended. A prescreen may identify an obvious mismatch, a need for verification, or a safety referral. It must not invent exclusions absent from the protocol, and staff must never reshape a candidate’s answers to improve conversion.
Sleep phenotyping must match the protocol. Insomnia symptoms, daytime sleepiness, snoring, and questionnaire scores may support risk stratification but are not necessarily clinical diagnoses. In particular, obstructive sleep apnea should not be diagnosed from a prediction tool or questionnaire alone. When a diagnosis, severity band, or circadian phenotype is required, the protocol should specify the appropriate clinical interview, sleep diary, actigraphy, home sleep apnea test, polysomnography, or other method, together with the device, scoring rules, observation window, and permitted age of prior tests.
Eligibility review should also resolve dates that cross midnight, time zones, shift work, travel, medication and substance use, caffeine and alcohol, prior treatment, cardiopulmonary or mental-health comorbidity, pregnancy, and access to required technology. Consumer-wearable summaries, screenshots, and self-reported diagnoses can create verification leads but should not replace source reports when the evidence is insufficient. Borderline cases, waivers, and rescreening must be handled by authorized personnel under predefined rules with an audit trail. Recruitment targets must never pressure sites to relax a safety boundary.
Informed choice, compensation, and participation burden
The first approach already shapes a prospective participant’s decision, so recruitment communication should be staged, optional, and open to questions. Permission for prescreening, authorization to use health information, and informed consent to join a study are separate matters; a signature is not proof of understanding. Before formal consent, people need adequate time and clear information about randomization, placebo or comparator, experimental procedures, foreseeable risks, alternatives, uses of data and samples, the right to withdraw, and arrangements for injuries or important new information. Minors, people with impaired decision-making capacity, and people with language needs require the safeguards applicable to them.
Compensation may recognize time, inconvenience, and other reasonable burdens, while travel, lodging, caregiving, or connectivity expenses may be reimbursed under the applicable plan. Payment is not a clinical benefit and cannot offset risk. The ethics review should examine the amount, method, and schedule, and the approved materials should explain them clearly. Compensation earned for completed activities should not be forfeited because a person withdraws, and a completion payment should not be so large that it pressures someone to remain. Referral rewards, per-enrollment fees, and vendor performance incentives need controls so that they do not promote concealment of exclusions, repeated solicitation, or disproportionate targeting of economically vulnerable people.
The burden of sleep research is more than a visit count. Overnight monitoring can affect employment, caregiving, and transport. Sleep restriction, treatment pauses, or medication changes can increase sleepiness or symptoms, while adhesive sensors, audio or video, and bedroom data can create physical or privacy concerns. These burdens should be disclosed during recruitment, with plans for safe travel, urgent contact, device failure, skin reactions, unexpected findings, and withdrawal. Recruiters should not give individualized instructions to stop medication or alter treatment merely to meet an enrollment target.
Digital recruitment, privacy, and authenticity controls
When electronic health records, registries, or platform data are used for feasibility, identification, or contact, the team must separately establish the legal basis, ethics or privacy approval, data minimization, and the person permitted to make contact. Permission to review records in preparation for research does not automatically authorize exporting identities or contacting individuals. A landing page should identify the collector, purpose, retention period, recipients, and deletion route. Information about people who never enroll remains protected and must not become an indefinite marketing list or be repurposed for unrelated model training.
Social media and decentralized research can expand access, but they also attract bots, duplicate submissions, identity misuse, fabricated criteria, forwarded screenshots, and incentive fraud. Proportionate controls may combine deduplication of contact information, temporal and logical consistency checks, controlled callbacks, verification against documents or clinical records, human review, and audit trails. Identity checks should offer alternatives so that legitimate candidates without a stable address, formal identification, credit history, or fast internet are not systematically excluded. A risk flag should not become a permanent blacklist without review.
Automated matching, chatbots, and generative text may help route inquiries or draft communications, but their intended use must be limited, accuracy validated, bias monitored, and human takeover available. A system should not make the final medical eligibility decision, hide the reason for rejection, or fabricate research answers. Access control, protected transmission and storage, vendor access, cross-border flows, incident response, and exit export must be designed before launch. High-risk findings, statements of self-harm, severe sleepiness, or acute symptoms should enter an approved escalation pathway rather than remain in an unattended inbox.
Governance, metrics, procurement acceptance, and closure
Governance records should assign the responsibilities, training, approval chain, and escalation times of the sponsor, investigator, site, contract research organization, recruitment provider, advertising platform, translator, and call center. Before launch, acceptance testing should cover script branches, language versions, mobile and accessible use, telephone and appointment routing, time zones, duplicate leads, withdrawal, deletion, and urgent scenarios. A change to materials, algorithms, eligibility criteria, site capacity, compensation, or contact details should trigger the appropriate review, approval, and version control.
Performance cannot be reduced to lead volume or cost per enrollment. At minimum, the funnel should be examined by channel and population from impressions to valid inquiries, successful contact, prescreen, consent, formal screening, enrollment, and randomization, together with elapsed time, screen-failure reasons, duplicates, no-shows, early withdrawal, complaints, safety events, and representation gaps. Definitions, denominators, and observation windows must remain stable. A higher conversion rate accompanied by misleading claims, unnecessary screening, privacy complaints, or systematic loss of a population is not success.
Procurement terms should prohibit guaranteed efficacy or the use of noncompliant methods to guarantee enrollment. They should address data ownership, subcontractors, audit rights, service interruption, incident notification, candidate access and deletion requests, return, and destruction. At closure, public recruitment status should be updated, advertisements and automated messages stopped, expired links disabled, and candidates still in the pathway notified. Screen-failure and withdrawal records should be handled under the approved plan. A defensible delivery shows that the right people were approached in an understandable and non-manipulative way, eligibility decisions can be reconstructed, non-enrolled people were treated with equal respect, and recruitment bias has not made the study uninterpretable.