Site data monitoring report

Synthetic Adaptive Platform Trial in Critical Illness

Søndre Universitetshospital (DK-01), Denmark. Initiated 08 January 2024, 31 months ago. Report generated 28 August 2026 from data version working, rule set 455d0fba76e0.

What to do this week

Resolve RNG-002: heart rate plausible - 8 records
Heart rate must lie between 25 and 220 beats per minute.
Resolve RNG-003: temperature plausible - 7 records
Core temperature must lie between 30 and 43 degrees Celsius.
Resolve TMP-004: ae onset before randomisation - 4 records
Adverse event onset must not precede the participant’s first randomisation datetime in any domain.
Resolve LOG-002: not alive after death - 3 records
A daily record must not report the participant alive on a date after the recorded date of death.
Resolve STR-004: single randomisation per domain - 2 records
A participant must be randomised at most once within any single domain.
Resolve TMP-001: discharge not before admission - 1 record
Hospital discharge must not precede ICU admission.
Supply the missing randomisation date and time - 9 records
The primary endpoint cannot be calculated for these participants at all, so they are absent from the analysis.
Complete the missing daily ICU records - 71 records
Days with no record are counted as unknown, not as days free of life support, so the primary endpoint is understated for these participants.

This list is the report. Everything below it is the evidence behind these items, and exists so that a query can be answered with “here is why we are asking” rather than “the system flagged it”.

Open findings

Severity	Rule	Name	Findings	Participants
critical	RNG-002	heart_rate_plausible	8	8
critical	RNG-003	temperature_plausible	7	7
critical	TMP-004	ae_onset_before_randomisation	4	4
critical	LOG-002	not_alive_after_death	3	1
critical	STR-004	single_randomisation_per_domain	2	1
critical	TMP-001	discharge_not_before_admission	1	1
major	STR-001	required_field_present	177	104
major	LOG-001	no_gap_in_daily_records	14	14
major	RNG-001	weight_plausible	1	1

Participants to follow up

Critical findings only, with the identifiers needed to act on them.

Completeness

Proportion of required, human-entered fields carrying a value, by the month of the clinical event. Identifiers and system-stamped dates are excluded: they are always present and would flatter the measure.

form	Aug 2025	Oct 2025	Nov 2025	Dec 2025	Sep 2025	Jan 2026
adverse_events	100.0%	100.0%	100.0%	100.0%	NA	NA
daily_icu	99.7%	98.7%	99.5%	99.1%	98.4%	98.4%
outcome_30d	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
randomisation	100.0%	98.3%	98.5%	97.2%	100.0%	100.0%
screening	100.0%	99.4%	98.9%	97.5%	99.0%	100.0%

Timeliness

Days from the clinical event to entry into the EDC. Read this against the holiday column: entry slows around local public holidays in every country, and a report that ignores that generates queries nobody can act on.

This site against all sites

Compared at the same stage of participation, not on a calendar axis. A site in its third month is compared against every other site’s third month. Comparing a young site against an established one on calendar time penalises it for being young, which helps nobody.

Primary endpoint completeness

Days alive without life support at 30 days is the outcome every adaptive analysis reads. A day with no daily record counts as unknown, not as a day free of life support, so missing records understate the endpoint rather than inflating it.

Measure	Value
Participant-domain records	258.00
Not evaluable (no randomisation date)	9.00
Complete	178.00
Incomplete (some days unknown)	71.00
Died within 30 days	79.00
Total unknown days	149.00
Mean, all records	16.06
Mean, complete records only	14.38

All data in this report is synthetic, generated by the code in this repository from a fixed seed. The site names are invented. This is an independent educational project with no affiliation to any real trial, hospital or research group.