An efficacy result means little if participants did not take the drug. Capture records what was dispensed and returned at each visit, kit by kit and tablet by tablet, with an audit trail on every count, so adherence is calculated from data rather than estimated by the site.
Free sandbox · No credit card · 21 CFR Part 11 aligned
Bottle returned at this visit
Bottle or kit number
DAREFIDDate dispensed
DADTCTablets dispensed
DAORRESDate returned
Tablets returned
Auto-query: returned count is above the dispensed count
Tablets reported lost
Missed doses
Reason for any missed doses
At a glance
Why it matters
A participant who takes half the doses dilutes the treatment effect, and in a non-inferiority trial poor adherence can make two treatments look more alike than they are. Adherence data also feeds per-protocol population definitions, dose-response analyses and safety interpretation. Regulators and reviewers expect a trial to say how adherence was measured and what it was.
The ESPACOMP Medication Adherence Reporting Guideline (EMERGE, Annals of Internal Medicine, 2018) asks authors to separate three phases: initiation (the first dose), implementation (how closely dosing matched the regimen) and persistence (how long the participant kept taking it). A pill count mostly measures implementation over each dispensing interval; a discontinuation date on the end-of-treatment form measures persistence. Designing the forms around those phases makes the reporting straightforward later.
Disease-specific trial guidelines say the same thing in practical terms. The International Headache Society guidelines for preventive migraine trials (Cephalalgia, 2020), for example, recommend monitoring adherence through medication counts, inspection of injection devices, electronic diary reminders and smart packaging.
Methods compared
No method is perfect, and the honest choice depends on the drug, the endpoint and the budget. Returned pill counts are cheap, objective about the bottle and easy to audit, but they cannot tell you when tablets were taken, and a participant who discards tablets before a visit looks perfectly adherent. Unreturned bottles leave gaps that have to be handled by rule.
Participant diaries record each dose with a date and time, show patterns such as weekend gaps, and capture reasons for missed doses. They depend on honesty and memory, so they tend to overstate adherence, which is why a timed phone diary with a completion window beats a paper sheet filled in the night before the visit. A ready-made medication adherence and response diary covers dose timing, missed doses and reasons.
Electronic caps, such as MEMS caps, record each bottle opening with a timestamp. They show dosing patterns well, but an opening is not an ingestion, they add cost and logistics, and some participants decant tablets into pill organisers. Drug or metabolite levels in blood or urine, where an assay exists, give biological confirmation at the price of a sample and a laboratory.
| Method | Shows dose timing | Objective | Main weakness | |
|---|---|---|---|---|
| Returned pill count | Tablets discarded before the visit | |||
| Phone diary | Self-report overstates adherence | |||
| Electronic cap | Opening is not swallowing | |||
| Drug level in blood or urine | Cost, assay, short window |
The calculation
Twice-daily dosing, one bottle of 60 tablets dispensed on day 0 and returned on day 28.
| Quantity | Value | How it is defined |
|---|---|---|
| Tablets dispensed | 60 | From the dispensing record |
| Tablets returned | 9 | Counted by the site with the participant |
| Tablets reported lost | 1 | Participant report, recorded separately |
| Tablets presumed taken | 50 | 60 − 9 − 1 |
| Days in interval | 28 | Return date minus dispensing date, rule for the return day stated in the protocol |
| Tablets expected | 56 | 2 per day × 28 days |
| Adherence | 89% | 50 ÷ 56 × 100, inside an 80% to 120% range |
Some protocols exclude lost tablets from the numerator, as here; others treat them as not taken. Decide once and apply it in analysis. Values above 100% usually mean extra doses, a counting error or tablets taken from another bottle.
eCRF design
Record counts, not percentages. The site enters what it can verify (bottle or kit number, dispensing date, number dispensed, return date, number returned, number reported lost) and a coded reason for missed doses. The percentage is then calculated the same way for every participant, instead of each coordinator working it out on a calculator with slightly different rules for the return day.
Participants often hold more than one bottle at a time, return bottles at different visits, or bring back an empty one and forget another. A repeating table section, with one row per bottle or kit, handles this cleanly, and each row carries its own field-level audit trail with timestamp, user, old value, new value and reason for change. When a monitor finds that a count was wrong, the correction is visible rather than overwritten.
Edit checks catch the errors that matter while the participant is still in the room. A rule on returned tablets (no more than the number dispensed) and a rule that the return date falls after the dispensing date each raise an auto-query when broken, and a reason field that appears when tablets are reported lost keeps the explanation on the form. See edit checks software for clinical trials for how the rules and query messages are set up.
Capture's calculated fields cover a fixed set of clinical formulas (BMI, eGFR, QTc and similar), and adherence % is not one of them. If the protocol needs the site to act on adherence at the visit, for example to counsel a participant below 80%, have the site enter the percentage from its compliance worksheet and attach a range edit check, so a value outside 80% to 120% raises an auto-query asking for the reason. The analysis still recalculates it from the counts.
Add the count fields and edit checks in the free sandbox, then enter a returned count above the dispensed count and watch the query appear.
In Capture
Capture's inventory module follows each kit from shipment to destruction. Kits are proposed by earliest expiry, each dispensing is linked to a visit and committed under an electronic signature, and dispensing units per kit are supported. Returned kits are recorded as used or unused and reconciled by the monitor, and the drug accountability report reconciles what was shipped, dispensed, returned and destroyed. Blinded roles never see the treatment arm. The IMP inventory and drug accountability page describes the full kit lifecycle.
That tells you which kit went to which participant and whether it came back. How many tablets were inside a returned bottle is a clinical observation, so it lives on the compliance eCRF next to the visit data, linked by kit number. Keeping the two together in one system and one audit trail removes the usual reconciliation between a pharmacy spreadsheet and the EDC.
Between visits, a short phone diary can record each dose, with reminders by email or SMS and a completion window so doses are logged on the day rather than reconstructed. At the visit, the site compares the diary with the returned count; large differences are worth a conversation and a note. For randomisation and supply together, see RTSM software for clinical trials.
Before first dispensing
Numerator, denominator, return-day rule and treatment of lost tablets.
For example 80% to 120%, and what happens outside it.
Rule for missing returns stated before data arrive.
Dispensed, returned and lost as numeric fields per bottle.
Returned above dispensed, return before dispensing.
Diary, caps or drug levels where the endpoint needs dose timing.
Kit number on the compliance form matches the dispensing record.
Tablets dispensed minus tablets returned (and, by protocol rule, minus tablets lost), divided by the tablets expected over the interval, times 100. Expected tablets are the prescribed daily dose times the number of days.
Many protocols treat roughly 80% to 120% as adequate, but the range is a protocol decision. Define it, and what happens outside it, before the first dispensing visit.
They are objective about what left the bottle but cannot show when doses were taken, and discarded tablets look like adherence. Pairing a count with a diary or electronic cap gives a fuller picture.
No. Adherence % is not one of Capture's calculated fields. Record the counts and derive adherence in analysis, or have the site enter its worksheet value with a range edit check that raises an auto-query.
With a repeating table section: one row per bottle or kit, each with its own field-level audit trail.
This page describes no cap integration. If you use caps, enter the summary values the cap vendor reports on a site form, or analyse the vendor file alongside the study export.
Yes. The free sandbox includes every feature, including inventory with sample kits. No credit card; you pay only when you go live.
Keep exploring
IMP inventory and drug accountability
Kits from depot to destruction.
Medication adherence diary
Participant dose diary template.
Edit checks software
Range rules and auto-queries.
RTSM software
Randomisation and supply together.
Patient diary software
Daily dose diaries.
Source data verification
Checking counts against source.
Counts, kits and diaries in one study with one audit trail. Free sandbox, no credit card, pay only when live.