This is a ready-to-use escalation log for protocol deviations that originate in a decentralized or hybrid trial’s remote layer, a telehealth visit, a home-health vendor visit, or a wearable or connected-sensor data stream, rather than at a physical site. These deviations cannot be fixed by retraining a site coordinator, because the site never controlled the failing step; the log is built to route each one to whoever actually owns the fix. Replace every <<FILL: ...>> placeholder, and maintain it in a queryable system rather than prose. Use it alongside, not instead of, the site-based protocol deviation trending register.
Why a separate log
The site-based trending register assumes a coordinator who witnessed the failure and can be retrained. A decentralized-trial deviation often has no witness: a dropped video call, a home nurse who was never trained on this protocol’s specific handling steps, a device that silently stopped transmitting. Mixing these into the site register buries a vendor-level or platform-level pattern inside per-site noise, and routes the CAPA to the wrong owner, the site, when the fix actually belongs to a vendor’s training program or a device’s monitoring configuration.
Part 1: Escalation register (one row per event)
| Field | Format | Required | Notes |
|---|---|---|---|
| Event ID | Text | Yes | Links to the deviation record |
| Source | Controlled list | Yes | Telehealth visit / home-health vendor / wearable-connected sensor / ePRO or eCOA platform |
| Subject ID | ID only | Yes | No names |
| What happened | Controlled list plus free text | Yes | e.g. dropped call before required assessment completed, sample drawn or handled outside protocol tolerance, transmission gap, missed device sync |
| Discovered by | Controlled list | Yes | Platform alert / vendor self-report / site review / monitor SDV |
| Discovery date | Date | Yes | For time-to-detect |
| Root owner | Controlled list | Yes | Vendor training / device configuration / platform design / sponsor oversight gap / site (if genuinely site-caused) |
| Safety or endpoint data affected? | Y/N | Yes | Drives escalation urgency per the parent SOP |
| Escalated to | Controlled list | Yes | Vendor quality contact / sponsor clinical operations / sponsor clinical quality / investigator |
| CAPA reference | Text | Conditional | Where raised, and against which owner |
| Close date | Date | Yes (to close) |
Register table
| Event ID | Source | Subject ID | What happened | Discovered by | Root owner | Safety/endpoint affected? | Escalated to | CAPA | Close |
|---|---|---|---|---|---|---|---|---|---|
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
Part 2: Vendor and platform trending
Compute on a defined cadence (commonly monthly), because a single vendor or device platform usually serves multiple sites and subjects, and the pattern only becomes visible pooled across all of them.
| Metric | Definition | Threshold / action |
|---|---|---|
| Rate by vendor / platform | Events / total encounters through that vendor or platform | <<FILL: e.g. > 1.5x program mean = vendor oversight review>> |
| Rate by source type | Events per source category (telehealth / home-health / wearable / ePRO) | <<FILL: top category = candidate process or device fix>> |
| Repeated identical event, same vendor | Same “what happened” category, same vendor, count | <<FILL: >= N = formal vendor CAPA request>> |
| Safety/endpoint-affecting share | Events affecting safety or endpoint data / total events | <<FILL: any nonzero rate = immediate sponsor clinical quality review>> |
| Time-to-detect by source | Discovery date minus event date, by source category | <<FILL: consistently long for one source = detection-channel gap>> |
Part 3: Vendor oversight review record
| Field | Entry |
|---|---|
| Review period | <<FILL>> |
| Vendor(s) / platform(s) reviewed | <<FILL>> |
| Outliers identified | <<FILL>> |
| Decisions (vendor retraining / device reconfiguration / oversight plan update / contract action) | <<FILL>> |
| Sponsor clinical quality reviewer (name, date) | <<FILL>> |
Acceptance criteria
- Every event is attributed to a root owner outside the site by default; a “site” root-owner designation is justified in the free-text field, not assumed.
- Events affecting safety or endpoint data are escalated the same working day the log entry is created, per the parent protocol deviation management procedure.
- Vendor and platform metrics are computed pooled across sites, not buried inside per-site reporting.
- A repeated identical event at the same vendor triggers a formal vendor CAPA request, not a series of individual event closures.
References
ICH E6(R3), Good Clinical Practice, Section 3.8 (sponsor oversight of outsourced trial activity, risk-proportionate) and Section 2.4/2.4.1 (investigator oversight of remote and home-based activity). ICH E6(R2) Section 5.18.4 (monitoring, including remote and centralized approaches). FDA guidance on decentralized clinical trials for drugs, biological products, and devices.
Confirm current versions before issue.
Filled specimen
A vendor-level pattern surfaced through pooled trending, illustrative.
Register (excerpt)
| Event ID | Source | Subject ID | What happened | Discovered by | Root owner | Safety/endpoint affected? | Escalated to | CAPA | Close |
|---|---|---|---|---|---|---|---|---|---|
| DCT-041 | Home-health vendor | S-014-007 | PK sample held at room temperature 40 minutes past the protocol’s 15-minute cold-chain window before courier pickup | Site review of courier log | Vendor training | N | Vendor quality contact | DCT-CAPA-009 | 14 Jul |
| DCT-044 | Home-health vendor | S-009-011 | PK sample held 35 minutes past the cold-chain window, same vendor, different nurse | Monitor SDV | Vendor training | N | Vendor quality contact | DCT-CAPA-009 | 21 Jul |
| DCT-047 | Home-health vendor | S-021-003 | PK sample held 50 minutes past the cold-chain window, same vendor, third site | Platform alert (courier tracking) | Vendor training | Y, PK concentration data potentially affected | Sponsor clinical quality | DCT-CAPA-009 | 28 Jul |
Metric and decision
| Metric | Value | Decision |
|---|---|---|
| Repeated identical event, same vendor | 3 cold-chain window misses across 3 sites in six weeks, same home-health vendor | Not a site issue, not a single-nurse issue; the vendor’s phlebotomy training does not cover this protocol’s specific cold-chain window |
| Decision | Vendor-level CAPA | Require the vendor to add protocol-specific cold-chain training before any nurse is assigned to this study, and add a courier-pickup time alert; verify the miss rate drops to zero over the next two months across all vendor-served sites |
Pooling the three events by vendor, rather than leaving them inside three separate site-level deviation logs, is what turned an apparently isolated miss at each site into one vendor training gap with one fix that protects every site the vendor serves.
Common inspection findings this log prevents
- A vendor-caused deviation logged and closed at the site level with a site retraining action that cannot fix a failure the site never controlled.
- A pattern across sites invisible because each occurrence sits in a different site’s own deviation log.
- A safety or endpoint-affecting decentralized-trial event escalated late because no defined channel routes it to sponsor clinical quality outside the normal site-monitoring cadence.
- A wearable or device transmission gap treated as a data query rather than logged and trended as a deviation source in its own right.
How to adapt this log
- Set your controlled source, root-owner, and “what happened” lists so the fields stay queryable across every vendor and platform used in the program.
- Point the escalation targets at your real vendor quality contacts and sponsor clinical operations/clinical quality functions.
- Calibrate thresholds to program scale; a single-vendor, single-region trial needs a lower repeat-count trigger than a multi-vendor global one.
- Feed vendor-level outliers into the sponsor’s vendor oversight plan and, where the pattern is severe or repeated, into vendor contract or requalification decisions.