# ISO 17025 Technical Records > Establish technical records and data management procedures per ISO/IEC 17025:2017 Clauses 7.5 and 7.11. Covers recording requirements (date, personnel, equipment, conditions, observations, calculations), correction procedures (single-line strike), original observation preservation, data integrity controls, LIMS/software validation, electronic record management with audit trails, and data retention policies. Tags: ISO 17025, Technical Records, Data Integrity, LIMS, Clause 7.5, Clause 7.11 ## Example Prompts - Create a technical records procedure for our testing laboratory - Define data integrity controls and LIMS validation requirements - Establish electronic record management procedures with audit trails - Design a data retention and archival policy for laboratory records URL: https://rakenne.app/skills/iso17025-technical-records/index.md Try this skill: https://rakenne.app/a/?skill=iso17025-technical-records ## Overview Technical records are the evidentiary backbone of every laboratory result. If an assessor or customer challenges a result, the laboratory must be able to reconstruct exactly what happened: who performed the test, on which instrument, under what environmental conditions, what raw data was observed, how calculations were performed, and who reviewed the work. ISO/IEC 17025:2017 Clause 7.5 mandates that records contain sufficient information to repeat the laboratory activity under conditions as close as possible to the original. This skill builds two complementary procedures. The technical records procedure (Clause 7.5) defines what must be recorded for every test or calibration activity, when it must be recorded (at the time of the activity, not retrospectively), how errors are corrected (single-line strike-through with initials and date -- never erasure or overwriting), and how records are traceable to the specific test item. The data management procedure (Clause 7.11) addresses the systems that store and protect those records: access controls, audit trails, LIMS/software validation, electronic record integrity, data transfer verification, protection against unauthorized modification, and retention policies. Together, these procedures ensure that your laboratory's data can withstand scrutiny from accreditation assessors, regulatory auditors, and legal challenges. ### What It Covers - **Recording elements** -- date, time, analyst identification, equipment ID and calibration status, environmental conditions, sample identification, raw observations, calculations, and derived results - **Recording timing** -- contemporaneous recording at the time of activity, not from memory after the fact - **Original observation preservation** -- raw data is never erased, overwritten, or made illegible; original observations are permanently retained - **Correction procedures** -- single-line strike-through showing the original entry, with correction, initials/signature, date, and reason - **Traceability to test item** -- every record links unambiguously to the sample or calibration item it describes - **Sufficient detail for repetition** -- records contain enough information that another competent analyst could reproduce the activity - **Access controls** -- role-based permissions defining who can view, enter, modify, and delete data - **Audit trails** -- automated tracking of who changed what and when, with timestamp and user identification - **LIMS/software validation** -- computer system validation approach per Clause 7.11.2, including functional testing, user acceptance, and revalidation triggers - **Electronic record controls** -- authentication, timestamps, change tracking, and protection against tampering - **Data transfer verification** -- accuracy checks when data moves between systems (instrument to LIMS, LIMS to report) - **Backup and disaster recovery** -- data backup procedures, offsite storage, recovery testing - **Data retention** -- defined retention periods per record type, archival procedures, and eventual destruction ## Example Conversation > **You:** We need technical records and data management procedures for LabTech Analytical. We use Element LIMS for most things, but our ICP-MS data starts in Agilent MassHunter software and gets transferred to LIMS. Some of our wet chemistry work (pH, conductivity, turbidity) is still recorded on paper bench sheets and then entered into LIMS manually. > > **Agent:** That hybrid setup -- instrument software plus LIMS plus paper -- is common but creates data integrity challenges at every handoff point. Let me map out your data flow: > > 1. **ICP-MS**: MassHunter generates raw spectra and calculates concentrations. How does data get into Element LIMS? Manual entry, CSV import, or direct interface? > 2. **IC (EPA 300.0)**: Same question -- what software, how does it reach LIMS? > 3. **Wet chemistry**: Paper bench sheets -- who enters data into LIMS, and when? > > **You:** MassHunter exports a CSV that gets imported into Element -- our IT guy set up an import routine. IC uses Chromeleon, same CSV import process. Wet chemistry: the analyst writes results on the bench sheet during the test, then enters them into LIMS at the end of the day. Sometimes the next morning if they're busy. > > **Agent:** The "next morning" part is a finding waiting to happen. Clause 7.5 requires records to be made at the time of the activity. If an analyst records a pH reading at 2 PM but doesn't enter it into LIMS until 9 AM the next day, the LIMS timestamp misrepresents when the observation was made. Two options: > > 1. Keep the paper bench sheet as the primary record (timestamped when the measurement happens) and treat the LIMS entry as a transcription -- but then you need a verification step > 2. Move to direct LIMS entry at the bench using a tablet or workstation > > Which is more feasible in the short term? > > **You:** We're planning to get tablets eventually, but for now we'll stick with paper bench sheets. Our assessor last year at a different lab said paper is fine as long as the process is controlled. > > **Agent:** Agreed -- paper is perfectly acceptable under 17025. I'll structure it this way: > > - **Primary record**: Paper bench sheet completed at time of analysis. Must include: date, time, analyst initials, equipment ID, environmental conditions (from the wall-mounted thermometer/hygrometer), sample IDs, raw readings, and any calculations > - **Transcription**: LIMS entry within the same business day. A second person (or the analyst using a different session) verifies the transcription against the original bench sheet > - **Data transfer verification**: Spot-check 10% of transcribed entries monthly; document in the QA log > > For your CSV imports from MassHunter and Chromeleon -- those are electronic data transfers. The procedure will require: > - Import validation: check record count (number of samples in export matches LIMS import) > - Spot-check: compare 3 random results between source software and LIMS per import batch > - Audit trail: LIMS must log who imported, when, and the source file name > > Now for corrections. How do your analysts currently correct errors on paper bench sheets? > > **You:** They cross it out and write the new value. Some of them use white-out -- I know that's bad. > > **Agent:** White-out is a critical finding -- it destroys the original observation, which violates Clause 7.5. The procedure will mandate: > > - **Single-line strike-through only** -- the original entry must remain legible > - **Correction written adjacent** to the struck-through entry > - **Initials and date** of the person making the correction > - **Reason for correction** noted (e.g., "transcription error," "recalculated per supervisor review") > - **No white-out, erasers, or overwriting** under any circumstances > > I'll include a training requirement: all analysts must be trained on the correction procedure, and it should be part of your new-hire orientation checklist. > > I've drafted both procedures. Running validators... > > `technical_record_completeness_checker` -- **PASS**. All recording elements, timing, correction procedures, traceability, and reviewer identification addressed. > > `data_integrity_validator` -- **REVIEW**: > - MISSING: Timestamp requirements for data entries and modifications not defined > > I'll add explicit timestamp requirements: LIMS entries must carry system-generated timestamps (not user-editable), instrument exports must retain original acquisition timestamps, and paper records must include time of observation. Running again -- **PASS**. > > Both procedures saved to `output/`. ## Sample Output Excerpt The skill produces a technical records procedure and a data management procedure. Here is a representative excerpt from the technical records procedure. --- ## 4. Recording Requirements ### 4.1 Mandatory Recording Elements Every technical record at LabTech Analytical shall include, at minimum: | Element | Requirement | Example | |---------|-------------|---------| | **Date and time** | Date and time of the activity, recorded contemporaneously | 2024-10-14, 14:30 CDT | | **Analyst identification** | Name or unique identifier of the person performing the activity | J. Park (JP) or Employee ID E-007 | | **Equipment identification** | Unique equipment ID and current calibration status | Agilent 7850 ICP-MS, Asset ID ICP-001, Cal due 2025-03-15 | | **Environmental conditions** | Temperature and humidity from calibrated monitoring device | 22.1 C, 45% RH (Thermometer TH-003) | | **Sample identification** | Laboratory sample ID linking to the chain of custody | WO-2024-0847-001 | | **Raw observations** | Original measurement readings as observed, without rounding or adjustment | Instrument response: 14,847 cps; Calibration slope: 2,341 cps/(ug/L) | | **Calculations** | Formulas applied, intermediate values, final results | C = (14847 - 312) / 2341 * 1.0 = 6.21 ug/L | | **QC results** | Associated quality control results for the batch | CCV recovery: 98.2%; Method blank: < 0.5 ug/L | | **Reviewer identification** | Person who checked or reviewed the work | Reviewed by: M. Rodriguez, 2024-10-15 | ### 4.2 Contemporaneous Recording All observations and data shall be recorded at the time of the laboratory activity, not from memory at a later time. Specifically: - **Instrument-generated data**: Retained in the instrument software with original acquisition timestamps. CSV exports for LIMS import shall occur within the same work shift as the analysis - **Paper bench sheets**: Completed during the activity. Entries timestamped by the analyst. LIMS transcription shall occur within the same business day - **Field records**: Field sampling data recorded on-site at the time of collection Retrospective recording (e.g., writing up results the following day from memory) is prohibited. If circumstances prevent contemporaneous recording, the delay and reason must be documented on the record. ## 5. Correction of Records ### 5.1 Paper Records Corrections to paper records shall follow this procedure without exception: 1. Draw a **single line** through the incorrect entry so that the original remains legible 2. Write the **corrected value** adjacent to the strike-through 3. **Initial and date** the correction 4. Note the **reason** for the correction (e.g., "calculation error," "misread instrument," "wrong sample ID") **Prohibited practices:** - White-out, correction fluid, or correction tape - Erasing (pencil or otherwise) - Overwriting that renders the original illegible - Removal or replacement of pages from bound notebooks ### 5.2 Electronic Records (LIMS) Corrections in Element LIMS shall: 1. Be made through the system's **amendment function**, which preserves the original entry and creates a new version 2. Require the user to enter a **reason for change** 3. Generate an automatic **audit trail entry** recording: original value, new value, user ID, timestamp, and reason 4. Be restricted to **authorized personnel** only (Department Lead or above for result changes) ## Extension Tools ### `technical_record_completeness_checker` Validates the technical records procedure for all required elements per Clause 7.5. | Check | What It Validates | Severity | |-------|-------------------|----------| | **Date of activity** | Date listed as required recording element | MISSING if absent | | **Personnel identification** | Analyst/operator ID required in records | MISSING if absent | | **Equipment identification** | Equipment/instrument ID required | MISSING if absent | | **Environmental conditions** | Temperature, humidity, or relevant conditions recorded | MISSING if absent | | **Original observations** | Raw data or primary observations addressed | MISSING if absent | | **Calculations** | Formulas and derived results documented | MISSING if absent | | **Contemporaneous recording** | Requirement to record at time of activity | MISSING if absent | | **Original observation preservation** | No erasure or overwriting of original data | MISSING if absent | | **Single-line strike-through** | Correction procedure using single-line method | MISSING if absent | | **Correction initials** | Person making correction identified with initials/signature | MISSING if absent | | **Traceability to item** | Records linked to sample or calibration item | MISSING if absent | | **Sufficient for repetition** | Records contain enough detail to reproduce the activity | MISSING if absent | | **Reviewer identification** | Review or checking process with personnel identified | MISSING if absent | ### `data_integrity_validator` Validates the data management procedure for all required elements per Clause 7.11. | Check | What It Validates | Severity | |-------|-------------------|----------| | **Access controls** | Role-based permissions for data viewing, entry, modification, deletion | MISSING if absent | | **Backup procedures** | Data backup and disaster recovery defined | MISSING if absent | | **Audit trails** | Change logging with who, what, when tracking | MISSING if absent | | **LIMS/software validation** | Computer system validation requirements (Cl.7.11.2) | MISSING if absent | | **Electronic records** | Digital record and e-signature requirements | MISSING if absent | | **Timestamps** | Date-time stamps for data entries and modifications | MISSING if absent | | **Change tracking** | Version control or amendment tracking for electronic records | MISSING if absent | | **Data transfer verification** | Accuracy checks when data moves between systems | MISSING if absent | | **Unauthorized access protection** | Measures against tampering and unauthorized modification | MISSING if absent | | **Data retention** | Retention periods and archival policy defined | MISSING if absent | ## Getting Started Start by activating the *ISO 17025 Technical Records* skill. The agent will ask about your laboratory's data systems, record-keeping practices, and how data flows from instruments to final reports. Have this information ready: - Your LIMS system name and version (e.g., Element, STARLIMS, LabWare) - How data flows from instruments to LIMS (direct interface, CSV import, manual entry) - Whether you use paper bench sheets for any activities, and how they are managed - Your current correction practices (how analysts fix errors on paper and in LIMS) - Who has access to modify data in your LIMS, and how permissions are managed - Whether your LIMS has audit trail functionality enabled - Any software validation documentation you already have - Current data backup procedures and frequency - Record retention requirements from regulations or accreditation bodies - Whether you use electronic signatures for report authorization If you have already completed other ISO 17025 skills, the agent will ensure the technical records procedure is consistent with equipment identification from the equipment calibration skill, personnel roles from the personnel competence skill, and reporting requirements from the reporting of results skill. --- Back to [Skill Library](https://rakenne.app/skills/index.md)