This is a ready-to-use register. It is the list every other hybrid control depends on: you cannot declare a governing record, schedule a reconciliation, or show a migration plan for a hybrid nobody has written down, and the most common opening question on this topic is simply “how many hybrids do you have?” Replace every <<FILL: ...>> placeholder, keep one entry per hybrid, and route the register through your normal document control, review, and approval. Four filled specimen entries follow, covering deliberately different hybrid patterns. Maintaining this register does not by itself create control; it makes the population visible so the rest of the programme can act on it. Verify each cited regulation against the current source before you rely on it, and treat this as general guidance to adapt rather than legal or regulatory advice.
The procedure that governs how these hybrids are controlled and reconciled is the companion hybrid record control and reconciliation SOP. The background reasoning is in hybrid paper-and-electronic records.
Document control header
| Field | Entry |
|---|---|
| Document title | Hybrid System Inventory and Governing Record Declaration |
| Document number | <<FILL: LOG-ID, e.g. LOG-QA-DI-007>> |
| Version | <<FILL: version, e.g. 1.0>> |
| Effective date | <<FILL: effective date>> |
| Supersedes | <<FILL: prior version or "New">> |
| Register owner | <<FILL: role, e.g. Data Integrity Lead, Quality>> |
| Approvers | <<FILL: e.g. Head of Quality, Head of Validation>> |
| Applies to | <<FILL: sites / departments / GxP domains in scope>> |
| Review cadence | <<FILL: e.g. quarterly reconciliation, full review annually>> |
| Governing SOP | <<FILL: SOP-ID for hybrid record control and reconciliation>> |
How to use this register
- Set the document number, owner, and approvers in the header.
- Enter one row per hybrid. A record is hybrid when three questions all answer yes: does a system generate electronic data for this activity, is any part of the official record on paper, and are both halves needed to reconstruct the activity.
- Build the initial population from your GxP computerized system inventory and your instrument list, then confirm every candidate by watching a real execution. Desk review alone reliably misses the balance whose only output is a printed slip taped into a logbook, and the spreadsheet that re-keys results for a trend chart.
- Complete Part A (identification and declaration) and Part B (controls, gaps, and plan) for every entry. The two parts share the Hybrid ID as the key, so they can be maintained as one table in a controlled spreadsheet or as two linked tables in a document.
- Route every governing-record declaration through Quality Assurance approval before the entry is considered complete. An entry with an unapproved declaration is an open action, not a finished row.
- Keep the register under change control. Reconcile it on the cadence in the header and update it on any trigger listed in the review section.
Field definitions and completion rules
Part A: identification and declaration
| Field | Format | Required | Who completes | When |
|---|---|---|---|---|
| Hybrid ID | Short code, e.g. HYB-001 | Yes | Register owner | At entry creation, never reused |
| Activity or process | Free text, one line | Yes | Process owner | At entry creation |
| Site, area, department | Controlled site and area codes | Yes | Process owner | At entry creation |
| GxP domain | GMP / GLP / GCP / GDP / PV / Combination | Yes | Process owner | At entry creation |
| Electronic component | System name, version, hosting, physical location | Yes | System owner | At entry creation, updated on any version or hosting change |
| Paper component | Controlled form ID, logbook ID, or batch record section | Yes | Process owner | At entry creation |
| Hybrid pattern | P1 / P2 / P3 (see pattern codes below) | Yes | Process owner | At entry creation |
| Data type | Dynamic / Static / Mixed | Yes | System owner with SME | At entry creation |
| Governing record declared | Electronic / Paper / Split-field | Yes | Process owner, proposed | At entry creation, before first use |
| Declaration basis | Decision branch code (D1 to D5) plus one line of reasoning | Yes | Process owner | With the declaration |
| Field-level declaration (split-field only) | List of fields and which half governs each | Yes if pattern is P3 | Process owner with SME | With the declaration |
| QA approval of declaration | Name and date | Yes | Quality Assurance | Before the entry is closed as complete |
| Link mechanism | The machine-generated identifier and where it is recorded on each half | Yes | System owner | At entry creation |
| Data criticality | High / Medium / Low, with a one-line basis | Yes | Process owner, QA concurs | At entry creation, revisited at each review |
Part B: controls, gaps, and plan
| Field | Format | Required | Who completes | When |
|---|---|---|---|---|
| Hybrid ID | Matches Part A | Yes | Register owner | Carried forward |
| Reconciliation control | Reference to the SOP, form, or step that performs reconciliation, plus its frequency | Yes | Process owner | At entry creation |
| Audit trail status | Present and reviewable / Present but not reviewable / Absent, plus review frequency | Yes | System owner | At entry creation, re-checked at review |
| Attribution status | Individual accounts / Shared account with compensating control / No account control | Yes | IT or instrument support | At entry creation, re-checked at review |
| Clock control | Synchronised and protected / Synchronised only / Uncontrolled | Yes | IT or instrument support | At entry creation, re-checked at review |
| Retention and backup | Where each half is retained, backup status, date of last tested restore | Yes | System owner with archive owner | At entry creation, restore date updated after each test |
| Open control gaps | Numbered list of specific gaps, each stated as an observable condition | Yes, or “None identified” | Process owner with QA | At entry creation and at each review |
| Risk class | Critical / High / Medium / Low, from the risk assessment | Yes | QA, from <<FILL: RA-ID for hybrid data integrity risk assessment>> | After the risk assessment is complete |
| Remediation action, owner, target date | One row per gap: action, named owner, target date, status | Yes for every open gap | Process owner, QA concurs | Within <<FILL: number>> working days of the gap being logged |
| Migration wave and target state | Wave number from the retirement plan, plus the intended end state | Yes | Register owner with process owner | After the retirement plan is approved |
| Last verified by observation | Date and observer name | Yes | Register owner or QA | At entry creation and at each periodic review |
| Status | Active hybrid / Remediation in progress / Migration in progress / Retired | Yes | Register owner | Updated on any change |
Hybrid pattern codes
| Code | Pattern | Typical example |
|---|---|---|
| P1 | Electronic generation, paper review and signature | A chromatography workstation generates dynamic data; the reviewed and signed record is a worksheet |
| P2 | Paper generation, electronic storage or summary | A paper batch record or logbook is the master and is later scanned, transcribed, or summarised into a system |
| P3 | Split-field record | Some fields exist only electronically and some only on paper, so neither half is a complete original |
An entry can carry a primary pattern and a secondary note. A paper batch record that also depends on a historian is P2 for the record as a whole and P3 for the specific parameters that exist continuously in the historian and hourly on paper. Record the primary code and describe the secondary condition in the declaration basis.
Declaration basis codes
Apply these tests in order and stop at the first that resolves. Record the code that resolved it.
| Code | Test | Resulting declaration |
|---|---|---|
| D1 | No system captures the observation electronically at the moment it occurs | Paper is the original; any later electronic entry is secondary |
| D2 | The electronic data is dynamic, meaning it can be reprocessed, re-integrated, re-queried, or re-plotted | Electronic data with its audit trail and metadata is the original; a printout cannot be |
| D3 | The electronic output is static and the system retains it as an attributable, time-stamped entry | The electronic value is the original; paper is a controlled extension |
| D4 | The instrument retains nothing after the output is produced | The paper capture is the only original; transcription requires second-person verification |
| D5 | One or more fields exist only on one half | Split-field: declare each field individually |
A declaration recorded as D2 with the printout named as the original is not approvable. That combination is the single most cited hybrid failure and the register should make it impossible to record silently.
Completing the open control gap and remediation columns
Write gaps as observable conditions, not as judgements. “Shared Windows login on the acquisition workstation, three named users share one credential” is a gap. “Weak access control” is not, because nobody can later prove it was closed. Each gap needs a corresponding remediation row with a named individual owner, not a department, and a target date that survives contact with a real project schedule. Where a gap cannot be closed before the system is replaced, record the compensating control and the residual risk explicitly rather than leaving the remediation column optimistic. A register full of overdue target dates is worse evidence than a register that states honestly which gaps are being carried until migration and why.
Review cycle and triggers
- Reconcile the register
<<FILL: cadence, e.g. quarterly>>: confirm every entry is still accurate, every remediation date is current, and no new hybrid has appeared. - Perform a full review at least
<<FILL: cadence, e.g. annually>>, including at least<<FILL: number>>entries verified by direct observation of a real execution rather than by document review. - Update immediately, outside the cycle, on any of these triggers: a new system or instrument is introduced; an existing system is upgraded, networked, or re-hosted; a method, form, or process step changes how the record is created; a deviation or audit finding reveals a hybrid that was not listed; a hybrid is retired.
- Route every change through change control per
<<FILL: SOP-ID for change control>>. - Feed the register into the quality management review per
<<FILL: SOP-ID for management review>>, reporting at minimum the total count, the count by risk class, the number of open gaps, and progress against migration waves.
Retention
Retain the register and each superseded version for <<FILL: retention period, e.g. the life of the quality system plus the longest applicable record retention period>>. Superseded versions matter more here than in most registers, because they are the evidence of what was declared as the governing record at the time a historical batch or study record was created. Do not overwrite entries in place without preserving the prior version.
Acceptance criteria for the register
- Every GxP instrument, system, and process has been classified as pure-paper, pure-electronic, or hybrid, with a recorded rationale, and the register accounts for the hybrid population.
- Every entry names one declared governing record, with a basis code and Quality Assurance approval.
- No entry declares a printout as the original for dynamic data.
- Every entry names a machine-generated identifier that appears on both halves, or logs the absence of one as a control gap with remediation.
- Every open gap has a named owner and a target date, and overdue items are visible rather than silently rolled forward.
- Every entry carries a risk class traceable to the hybrid data integrity risk assessment, and a migration wave traceable to the retirement plan.
- At least
<<FILL: number>>entries per review cycle were verified by direct observation, and the observation date and observer are recorded.
References
21 CFR 211.68 (automatic, mechanical, and electronic equipment), 211.180 (general records requirements), 211.194 (laboratory records). 21 CFR Part 11 (electronic records and electronic signatures), in particular 11.10, 11.50 (signature manifestations), and 11.70 (signature and record linking, including handwritten signatures executed to electronic records). EU GMP Annex 11 (Computerised Systems) and EU GMP Chapter 4 (Documentation). FDA guidance, Data Integrity and Compliance With Drug CGMP, Questions and Answers (December 2018). MHRA GXP Data Integrity Guidance and Definitions (March 2018). PIC/S PI 041, Good Practices for Data Management and Integrity in Regulated GMP/GDP Environments. ICH E6 Good Clinical Practice, for clinical source and certified copy expectations. ICH Q9, Quality Risk Management, for the risk basis of criticality and prioritisation.
Confirm the current version and clause numbers of each reference before issue.
Blank register
Part A
| Hybrid ID | Activity | Site / area | GxP domain | Electronic component | Paper component | Pattern | Data type | Governing record | Basis | QA approval | Link mechanism | Criticality |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | P1/P2/P3 | Dynamic/Static/Mixed | <<FILL>> | D1-D5 | <<FILL: name, date>> | <<FILL>> | High/Med/Low |
Part B
| Hybrid ID | Reconciliation control | Audit trail | Attribution | Clock | Retention and backup | Open control gaps | Risk class | Remediation, owner, target | Migration wave | Last observed | Status |
|---|---|---|---|---|---|---|---|---|---|---|---|
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL: date, observer>> | <<FILL>> |
Revision history
| Version | Date | Author | Summary of change |
|---|---|---|---|
<<FILL: 1.0>> | <<FILL: date>> | <<FILL: author>> | Initial issue. |
Approvals
| Role | Name | Signature | Date |
|---|---|---|---|
| Author / register owner | <<FILL>> | ||
| Reviewer (QA) | <<FILL>> | ||
| Approver (Quality Head) | <<FILL>> |
Filled specimen
Four entries follow, chosen because they fail in different ways. Company, systems, names, and dates are illustrative; replace them with your own. Each entry is shown as Part A followed by Part B so the full row is readable.
HYB-001: standalone chromatography workstation, release assay
Part A
| Field | Entry |
|---|---|
| Activity | Assay and related substances by HPLC, finished product release testing |
| Site / area | Site 1, QC Laboratory, Chemistry |
| GxP domain | GMP |
| Electronic component | Chromatography data system, standalone acquisition workstation HPLC-07, application v7.2, local disk, no network share |
| Paper component | Controlled analyst worksheet CF-QC-114, sequentially numbered |
| Pattern | P1 |
| Data type | Dynamic |
| Governing record declared | Electronic: raw signal, integration parameters, sequence file, and audit trail |
| Basis | D2. Chromatographic data can be re-integrated and reprocessed, so the printed report is a static summary and cannot be the original. |
| QA approval | L. Fernandes, 03 March 2026 |
| Link mechanism | Result ID and sequence name printed on every report; analyst records both on the worksheet; worksheet number recorded in the CDS sample comment field |
| Criticality | High. Results support batch disposition. |
Part B
| Field | Entry |
|---|---|
| Reconciliation control | Per-run reconciliation at second-person review, SOP-QA-031 section 5.4; every run before the result is used |
| Audit trail | Present and reviewable; reviewed each run |
| Attribution | Shared Windows account at operating system level; individual application accounts inside the CDS. Compensating control: workstation in a card-access room, access log CF-QC-009. |
| Clock | Synchronised to the site time source but the local administrator can change it; change is captured in the application audit trail |
| Retention and backup | Electronic on local disk with nightly copy to the validated file store; paper to QC archive. Last tested restore 12 June 2026, chromatogram re-integrated successfully after restore. |
| Open control gaps | 1. Shared operating system login permits file-level actions outside the application audit trail. 2. Local disk is the primary store, so a disk failure between nightly copies loses up to one working day. |
| Risk class | High |
| Remediation, owner, target | Gap 1: migrate acquisition to the networked CDS server with individual domain accounts. Owner: R. Okonkwo (Lab Systems). Target 31 December 2026, status In progress. Gap 2: redirect project path to the validated share. Owner: R. Okonkwo. Target 30 September 2026, status Open. |
| Migration wave | Wave 1, target state fully electronic review and approval on the networked CDS |
| Last observed | 08 July 2026, observer M. Haddad (QA) |
| Status | Remediation in progress |
HYB-002: analytical balance with printout counter, dispensing
Part A
| Field | Entry |
|---|---|
| Activity | Weighing of reference standards and sample preparations |
| Site / area | Site 1, QC Laboratory, Sample Prep |
| GxP domain | GMP |
| Electronic component | Analytical balance BAL-12 with serial thermal printer. No memory, no user accounts, no audit trail. The balance retains nothing after the slip prints. |
| Paper component | Dispensing worksheet CF-QC-088 with a printout counter column; thermal slips affixed to the worksheet |
| Pattern | P1 in appearance, resolved as D4 |
| Data type | Static |
| Governing record declared | Paper: the affixed printed slip together with the worksheet entry |
| Basis | D4. The instrument retains nothing, so no electronic original exists. Second-person verification of every transcribed weight is required because there is no electronic source to check against later. |
| QA approval | L. Fernandes, 03 March 2026 |
| Link mechanism | Printer sequential counter value recorded in the worksheet column, so a gap in the counter reveals a discarded weighing |
| Criticality | High. Weights feed assay calculations used for release. |
Part B
| Field | Entry |
|---|---|
| Reconciliation control | Counter continuity check and weight verification at second-person review, every worksheet |
| Audit trail | Absent. Compensating control is the printer counter plus the controlled sequentially numbered worksheet. |
| Attribution | No account control on the instrument; attribution rests entirely on the paper signature and the controlled-access room |
| Clock | Balance prints a date and time that a user can set. Recorded as uncontrolled. Compensating control: analyst records the wall-clock time on the worksheet and the reviewer confirms the two agree. |
| Retention and backup | Paper to QC archive. Thermal slips photocopied as verified true copies on the day of use and the copy retained alongside the original, because thermal print fades well inside the retention period. |
| Open control gaps | 1. Instrument clock is user-settable and unlogged. 2. No electronic source exists, so an undetected transcription error is only caught by the second-person check. |
| Risk class | High |
| Remediation, owner, target | Gap 1 and 2: replace with a balance interfaced to the laboratory system so weights transfer without transcription. Owner: P. Sandoval (Metrology). Target Q3 2027, status Planned. Interim: second-person verification enforced 100 percent, monitored in the quarterly QA sample. |
| Migration wave | Wave 2, target state balance interfaced to the laboratory system, no manual transcription |
| Last observed | 08 July 2026, observer M. Haddad (QA) |
| Status | Active hybrid |
HYB-003: paper batch record dependent on the historian and DCS
Part A
| Field | Entry |
|---|---|
| Activity | Bulk drug substance fermentation and harvest, paper batch manufacturing record |
| Site / area | Site 2, Manufacturing, Upstream |
| GxP domain | GMP |
| Electronic component | Distributed control system with process historian; continuous trending of temperature, pH, dissolved oxygen, and agitation; alarm and event history |
| Paper component | Paper batch manufacturing record BMR-114 rev 6, including hourly parameter entries and step signatures |
| Pattern | P2 overall, P3 for the continuously monitored parameters |
| Data type | Mixed. Step signatures and manual observations are static and on paper; process parameters and alarms are dynamic and continuous in the historian. |
| Governing record declared | Split-field. Process parameters, alarms, and events: the historian is the original. Manual observations, material additions, line clearance, and signatures: the paper batch record is the original. |
| Basis | D5. The historian records continuously, including the intervals between hourly manual readings, so the manual entry is a sample of a record that already exists electronically. Fields captured only by an operator exist nowhere else. |
| QA approval | K. Aduba, 17 March 2026 |
| Link mechanism | Batch identifier and DCS batch or campaign ID recorded on the batch record cover page and on every historian extract; step numbers common to both |
| Criticality | High. Supports batch disposition. |
Part B
| Field | Entry |
|---|---|
| Reconciliation control | Batch record review includes retrieval of the historian trend and alarm and event list for the full batch window, per SOP-QA-031 section 5.4 and the batch record review procedure |
| Audit trail | Present and reviewable in the DCS and historian; reviewed at batch record review and at periodic system review |
| Attribution | Individual named accounts on the DCS; operator signatures on paper resolve to the same individuals per the specimen signature register |
| Clock | Synchronised and protected; DCS and historian take time from the site time source, users cannot change it, changes are logged |
| Retention and backup | Historian on the validated server with daily backup; batch record to the manufacturing archive. Last tested restore 04 May 2026, alarm history retrieved and readable. |
| Open control gaps | 1. Hourly manual entries can read as compliant while an excursion occurred between readings; the review step that catches this depends on the reviewer actually pulling the trend. 2. No automated alert links a historian alarm to the corresponding batch record section. |
| Risk class | High |
| Remediation, owner, target | Gap 1: batch record review checklist amended to require the trend and alarm extract to be attached, not merely consulted. Owner: J. Bertrand (Manufacturing QA). Target 30 September 2026, status In progress. Gap 2: scope an alarm-to-batch-section report with the automation group. Owner: J. Bertrand. Target 31 March 2027, status Open. |
| Migration wave | Wave 3, target state electronic batch record with historian integration |
| Last observed | 22 June 2026, observer K. Aduba (QA) |
| Status | Remediation in progress |
HYB-004: clinical site paper source document feeding the EDC system
Part A
| Field | Entry |
|---|---|
| Activity | Collection of subject vital signs and adverse event data at investigator sites, study <<illustrative study code>> |
| Site / area | Investigator sites, sponsor Clinical Operations |
| GxP domain | GCP |
| Electronic component | Electronic data capture system holding the case report form data and its audit trail, hosted by <<FILL: vendor>> |
| Paper component | Site source documents: medical chart entries, visit worksheets, subject diaries |
| Pattern | P2 |
| Data type | Mixed. Source observations are static paper; the EDC record is electronic with an audit trail. |
| Governing record declared | Paper source at the site is the original for the observations. The EDC entry is derived data. Where the site records an observation directly into the EDC with no prior paper capture, the EDC entry is the original and is identified as direct-entry source in the study source data agreement. |
| Basis | D1 for observations first written in the chart. D3 for identified direct-entry fields. Recorded per field group in the study source data agreement. |
| QA approval | S. Vainio, 11 May 2026 |
| Link mechanism | Subject identifier, visit identifier, and page or form identifier common to the source worksheet and the EDC form; the source data agreement lists which fields are direct-entry |
| Criticality | High. Data supports safety reporting and the primary analysis. |
Part B
| Field | Entry |
|---|---|
| Reconciliation control | Source data verification and source data review at the frequency defined in the monitoring plan, risk-based per subject and per field group |
| Audit trail | Present and reviewable in the EDC; reviewed by the data manager and at monitoring visits |
| Attribution | Individual named accounts in the EDC for every site user; delegation log maintained at each site |
| Clock | EDC server time is controlled by the vendor and stated in the vendor qualification; site paper entries are dated and timed by hand and cannot be independently verified |
| Retention and backup | EDC data retained and archived per the sponsor archive procedure; site source retained by the site per its own obligations. Certified copies of source taken during monitoring are indexed in the trial master file. |
| Open control gaps | 1. Direct-entry field list is defined per study but not always reflected in site training, so some sites double-record and create an avoidable transcription step. 2. Certified copies of source collected remotely are not consistently verified before the paper leaves the reviewer’s hands. |
| Risk class | High |
| Remediation, owner, target | Gap 1: add the direct-entry field list to site initiation training and the site file. Owner: A. Reyes (Clinical Operations). Target 31 August 2026, status In progress. Gap 2: apply the certified copy work instruction to all remotely collected source. Owner: A. Reyes. Target 30 September 2026, status Open. |
| Migration wave | Wave 4, target state direct data capture for the field groups where it is feasible; paper source is expected to persist for chart-based observations |
| Last observed | 19 June 2026, observer S. Vainio (Clinical QA) |
| Status | Active hybrid |
What makes these four entries useful as a set is that they resolve to four different governing records for four different reasons. Only the chromatography entry lands on the electronic half. The balance lands on paper because the instrument keeps nothing. The batch record splits, because the historian genuinely holds a record the paper only samples. The clinical entry splits along a field list defined per study. A register that records “electronic” for every row has almost certainly not applied the decision tree.
Common inspection findings this register prevents
- The firm cannot state how many hybrid records it has, and the count offered during the inspection differs from the count in the quality system.
- A hybrid is found in use on the floor that appears nowhere on any inventory, most often a standalone instrument or a spreadsheet that re-keys results.
- No document states which half of a record is the original, so different reviewers treat different halves as authoritative.
- A printout is recorded as the original for dynamic electronic data, and the electronic data behind it is uncontrolled or was deleted.
- The paper batch record is not recognised as a hybrid at all, so nobody reviews the historian data it depends on.
- Split-field records are forced into a single declaration, leaving the fields that exist on only one half without any declared owner or control.
- Control gaps are described in general terms that cannot be verified as closed, and remediation dates roll forward with no visible owner.
- The register exists but was last updated before two systems were upgraded and one was networked, so its risk classes and gaps no longer describe reality.
How to adapt this register
- Set the document number, owner, approvers, and review cadence in the header, and point the governing SOP field at your real hybrid control procedure.
- Build the first population from your system inventory and instrument list, then budget time to walk the floor and the laboratory. Expect the observed count to exceed the desk count, and record the difference as a finding about the inventory process rather than quietly correcting it.
- Name the actual machine-generated identifier each system produces in the link mechanism field. This is the highest-value field in the register and it is entirely system-specific.
- Add or remove GxP domain values to match your operations. A cell and gene therapy site will want chain of identity and chain of custody records in scope; a distribution site will want temperature records and shipment documentation.
- Bind the risk class column to your hybrid data integrity risk assessment and the migration wave column to your retirement plan, so the three documents cannot drift apart.
- Decide where the register physically lives. A controlled spreadsheet is acceptable if it is itself under validation and access control appropriate to its use; a register maintained in an uncontrolled file with no version history undermines the point of the exercise.
- Confirm every regulation in the references section against the current published version before issue.