EDC for delirium clinical trials
Delirium comes and goes within a day, cannot be assessed in a deeply sedated patient, and the participant often cannot consent. A delirium trial needs shift-level assessments, honest handling of coma and sedation, and surrogate consent. Build it in Capture and test it free.
- Twice-daily bedside assessment
- Coma and sedation handled
- Surrogate (LAR) consent
Free sandbox · No credit card · 21 CFR Part 11 compliant
| Assessment | Baseline | Surgery | POD 1 | POD 2 | POD 3 | POD 7 | Day 30 |
|---|---|---|---|---|---|---|---|
| Delirium screen, morning and evening | |||||||
| Sedation / arousal level | |||||||
| Study drug and rescue sedatives | |||||||
| Pre-operative cognition and frailty | |||||||
| Length of stay and discharge destination | |||||||
| Adverse events |
x = form completed; POD = postoperative day
What a delirium trial needs from its EDC
- Repeated, short assessments: delirium fluctuates, so protocols assess at least daily and often every shift. The data model is many small assessments per participant, not a few large visits.
- "Not assessable" is a real answer: a patient who is deeply sedated or in coma cannot be rated for delirium. Forms must record that state explicitly, not leave a blank or a false negative.
- Outcomes built from the assessments: incidence, first-day onset, duration, delirium/coma-free days and severity are derived from the series, so every timestamp matters.
- Exposures that change the outcome: benzodiazepines, anticholinergics, opioids, dexmedetomidine and antipsychotics, with dose and time.
- Consent by someone else: many participants lack capacity, so the legally authorised representative route needs a documented, auditable process.
The endpoint model
Incidence, delirium-free days and severity: three different data problems
Delirium trials fall into two broad families. Prevention trials enrol patients at risk (older adults having surgery, patients admitted to intensive care) and ask whether an intervention lowers the incidence or shortens the duration of delirium. Treatment trials enrol patients who already have it and ask whether severity or duration improves. Registered ICU trials show the range. One pragmatic ICU study of a medication bundle named delirium severity on the Delirium Rating Scale-Revised-98 and delirium duration on the CAM-ICU as primary outcomes in its protocol, and its published report described delirium/coma-free days, measured with sedation scores and the CAM-ICU, together with severity as the primary outcomes. A different trial defined a response on the DRS-R-98 as a severity score below a cut-off with a minimum point reduction. Treat these as examples of shapes, not as a template for your own endpoint, and have your statistician fix the definition before the database is built.
Each endpoint puts a different load on the data. Incidence needs a clean denominator and a rule for patients who die or are never assessable. Delirium/coma-free days is built by counting days in a window, so a missing assessment is not a free day; the form must distinguish "assessed, negative" from "not assessed" from "unarousable". Severity scales are clinician-rated and need a named rater and a time, because the same patient can score very differently at 6 a.m. and 6 p.m. Capture stores each assessment as its own timed record, derives totals as read-only calculated fields where the scale allows it, and keeps an audit trail of any later change with reason.
Be careful about which tool you name in your protocol. The CAM and CAM-ICU, the 3D-CAM, the 4AT and the DRS-R-98 each have their own scoring, intended setting and licence or permission terms. Capture does not supply instrument wording or verify licences, and licence controls exist only for participant questionnaires, not for clinician-rated or site forms, so the licence for a bedside rating tool stays with your study documents.
Assessment
Arousal / sedation level
AROUSALAssessment status
DELSTATNot assessable: result and severity fields skipped
Reason not assessable
NOTASSRater
RATERLast benzodiazepine dose
LASTBZDCalculated from two date-times
5.0 h ago CalculatedProtocol to build
Delirium protocol elements and where they live in Capture
| Protocol element | What the data looks like | Where it lives in Capture |
|---|---|---|
| Eligibility and risk | Age, surgery type, baseline cognition, frailty, exclusion for existing dementia or delirium | Screening form; eligibility form; cognition screen can follow the MoCA or MMSE layouts if your protocol uses them |
| Consent | Patient consent where capacity exists, surrogate consent where not, later patient re-consent | LAR consent software; eConsent |
| Delirium assessment series | Timed result, rater, status, severity | Repeating-row or scheduled site form with skip logic for "not assessable" |
| Sedation and arousal | Score and time per assessment | Dropdown field on the same assessment record |
| Drug exposures | Benzodiazepines, anticholinergics, opioids, antipsychotics: dose and time | Concomitant medications form, repeating rows |
| Study intervention | Drug, dose, start and stop, or non-drug bundle fidelity | Intervention form per day |
| Hospital outcomes | ICU stay, hospital stay, discharge destination, death | Outcome form; export for time-to-event analysis |
| Follow-up cognition | Telephone or visit cognitive and function measures at 1 to 6 months | Site form or participant questionnaire if the participant can self-complete |
Capture does not connect to bedside monitors, pharmacy systems or ICU charting. Assessments are entered by site staff from the clinical record.
Test the not-assessable logic before the first patient
Build the assessment form, enter a week of sample shifts including a sedated day, and check that the derived days are right. Free sandbox, no credit card.
Operations
Surrogate consent, shift coverage and missing assessments
Consent is the first hard problem. Patients at risk of delirium may have cognitive impairment at baseline, and those with delirium by definition have impaired attention. Protocols therefore use a mix of patient consent where the person has capacity, consent from a legally authorised representative (LAR) where not, and sometimes deferred consent or consent to continue once capacity returns. Which routes are lawful depends on country, state and ethics committee, so settle that first. In Capture, eConsent supports an LAR signature and an optional witness; the document is fully executed only after an authorised site staff member countersigns, and the signed PDF carries signature data, timestamps and audit trail. If a participant regains capacity, a re-consent step belongs in your protocol and workflow; do not assume the platform will prompt it for you. See impartial witness eConsent for the witness routes.
Coverage is the second. A twice-daily or per-shift assessment is a promise that someone is there at 7 a.m. on a Sunday. Build a daily completion view (the dashboard and status overview show what is outstanding) and agree who owns a missing assessment. When a day is missed, record it as missed, with the reason, rather than back-filling a negative. Analysis rules for missing days are the statistician's call, but they cannot be applied if the data silently say "negative".
Third, keep the intervention honest. In multicomponent non-drug trials (sleep protection, early mobilisation, reorientation), what matters is fidelity: what was actually delivered on each shift. A short daily fidelity form with checkboxes is cheap to build and often decides whether a null result is a failed intervention or a failed delivery. For the wider neuro-cognitive picture see EDC for CNS and neurology trials; for surgery-based populations see EDC for surgical trials.
Before first participant
Delirium study readiness checklist
Assessment tool and licence confirmed
Tool, version, who may administer it, and the rights-holder terms filed in study documents.
Definition of onset and duration written
Day boundaries, handling of coma and death agreed with the statistician.
"Not assessable" built in
Explicit status field with reason and skip logic on result fields.
Consent routes approved
Patient, LAR and deferred routes as approved by the ethics committee.
Exposure list defined
Which sedatives, opioids and anticholinergics are captured, with dose and time.
Weekend and night coverage planned
Named assessors per shift and a daily completeness check.
Can Capture record a patient who cannot be assessed?
Yes. Add an assessment status field with a reason and use skip logic to hide the result fields. This keeps a deeply sedated shift from being entered as a negative screen.
Does it calculate delirium-free days?
Capture can show derived values as read-only calculated fields, and exports give your statistician each timed assessment. The exact counting rule for delirium/coma-free days belongs in your analysis plan.
Can a family member consent for the patient?
eConsent supports a legally authorised representative signature and an optional witness, with countersignature by authorised site staff. Whether surrogate or deferred consent is allowed is set by law and your ethics committee.
Which delirium tools can I use?
Any tool your protocol names. Capture does not supply instrument wording or verify licences, and licence controls exist only for participant questionnaires, so keep the licence for a clinician-rated tool with your study documents.
Can it pull data from the ICU chart?
No integration is claimed. Site staff enter assessments and drug doses from the clinical record.
Does it work for post-operative delirium in older adults?
Yes. Build the pre-operative baseline, the daily assessment series, exposure log and 30-day follow-up as one study.
What does it cost?
The sandbox is free with every feature. Plan rates for live studies are on the pricing page.
Keep exploring
Related pages
Start building your delirium study free
Free sandbox with every feature. No credit card.