Skip to main content

ECG and Point-of-Care Records: Source, Quality and Clinical Review

Record an in-clinic ECG, glucose measurement or urine dipstick result with its patient, actual measurement time, source, operator, relevant device or kit details, original output and clinical review status. An uploaded strip or entered number is not automatically a clinician’s interpretation. Clinics need a way to distinguish patient measurements from quality-control checks, explain unusable or repeated measurements, and connect the finding to the doctor’s documented decision without delaying care to complete software fields.

By Bractis · Updated · 15 min read

Indian clinics recording ECGs and point-of-care measurements during outpatient visits.

Separate performing the test from interpreting it

A staff member may acquire an ECG or enter a point-of-care reading, while a doctor reviews its clinical meaning. Keep those roles visible. The person who uploads a file is not necessarily the operator, and the operator is not necessarily the clinician who interprets the finding.

The record should answer three questions: what was actually obtained, whether there were quality or context limitations, and what the responsible clinician decided. A single “test done” checkbox cannot distinguish a complete reviewed measurement from an unusable attempt.

This guide concerns documentation and handover. Staff performing investigations need appropriate training and the applicable device, laboratory and clinical procedures. A software article should not replace instructions for acquisition, infection prevention, calibration or patient management.

Keep the evidence appropriate to each type of investigation

An ECG is not adequately represented by a text label alone when the tracing is relevant to review. A glucose measurement needs its reported value and unit. A dipstick may produce multiple categorical observations, and these should not be forced into one numerical field.

Preserve automated device comments separately from the clinician’s interpretation. An imported phrase such as an algorithm-generated interpretation must remain attributed to the device; copying it into a signed assessment without review changes its apparent authority.

On a narrow screen, scroll the table horizontally to read all columns.

Evidence and context for common in-clinic tests
InvestigationRecord the source evidenceKeep distinct from it
ECGIdentified tracing, acquisition time and relevant acquisition settings or limitationsDevice interpretation, clinician assessment and later amended opinion
Glucose measurementDisplayed value, units, specimen context, time and source deviceHome-reported values, control solution checks and later laboratory results
Urine dipstickActual component results, specimen context, time and kit/device information as applicableUnperformed components, missing readings and inferred diagnoses
Unusable attemptWhat was attempted and why the result cannot be relied onA completed valid measurement or a fabricated normal result
Repeat measurementNew measurement with its own time and link to the earlier attemptAn edit that silently overwrites the first observation

Use measurement time rather than upload time in the clinical sequence

HL7 FHIR R4 distinguishes the physiologically relevant observation time from the time information is made available. For direct patient measurements, this helps explain why a reading taken during the visit should not become a later measurement simply because staff uploaded it after clinic hours.

If the precise acquisition time is unavailable, record the available evidence and uncertainty. Do not manufacture minute-level accuracy from a file name. A scanned paper strip may show a device clock that needs verification; the upload timestamp only establishes when the file entered the system.

Origin also matters. A patient’s photograph of a home meter reading and a clinic measurement can both be useful, but they are different sources. Keep them labelled so a later reader does not assume that clinic staff performed a test merely because they entered it.

Source: HL7 FHIR R4: Observation time and source context

Prevent quality-control material from becoming a patient result

A device check using control material belongs to the quality process, not to a patient chart. The clinic should be able to identify the device or kit involved in patient testing and locate the relevant quality records without displaying control values as clinical trends.

WHO’s laboratory quality management handbook treats equipment, process control and occurrence management as parts of laboratory quality. It provides background principles; the exact checks and acceptable limits for a particular device come from the applicable controlled procedures and manufacturer information.

If a quality concern is identified, follow the designated procedure for stopping or restricting affected testing and assessing earlier measurements. Do not let an attractive result screen imply that a measurement is reliable when its validity is unresolved. The responsible clinical and technical staff determine the next action.

Source: WHO: Laboratory quality management system handbook, 2011

Record repeat attempts without choosing the most convenient number

A repeat can follow an acquisition problem, a clinician’s request or a concern about the first result. Retain the reason and distinguish an invalid attempt from a valid observation that the doctor still needs to consider. Deleting the less convenient value makes the record look more certain than the actual sequence.

Where the device shows a non-numeric message, preserve its meaning rather than entering the nearest display limit as though it were a measured number. Ask the responsible clinician and technical staff how the result should be represented and handled under the applicable procedure.

The same principle applies to a blurred ECG scan. Re-uploading a clearer copy of the same tracing is a document replacement, while acquiring a new ECG is another measurement. The record should make that difference visible.

Worked visit: eight recorded events do not mean eight patient tests

Fictional E-73 extends the ECG example below into a mixed paper-and-digital clinic visit. All times are on 28 October 2026 in IST. Assume the operator records the patient acquisitions contemporaneously against the clinic clock and identifies the patient and source equipment. The clinical indication, values, tracing findings and treatment are deliberately not supplied. This is an evidence-classification exercise, not an acquisition protocol or a demonstration of actual Bractis behaviour.

The morning folder contains a meter control check, a training tracing, two patient ECG attempts and a point-of-care glucose record. They do not all belong in the patient’s trend. C1 uses control material; T1 comes from an ECG training simulator with no patient attached. E1 is the unsuccessful patient attempt. E2 is the separate repeat tracing obtained on the responsible clinician’s direction. G1 is the patient measurement, with its actual value, unit and device context retained in the fictional source record rather than invented in this article.

The clinician reviews the original E2 paper tracing and G1 record at 10:25 and records the scope as R1. No diagnostic conclusion is supplied here. Reception later scans E2 twice because the first scan is cropped. S2 improves the representation of E2; it does not record another ECG or a second clinical interpretation. Eight selected ledger events therefore include three patient acquisition attempts, two nonpatient activities, one review event and two scan events. The two outputs covered by R1 are E2 and G1, not every item in the folder.

On a narrow screen, scroll the table horizontally to read all columns.

Fictional E-73: classify eight events before building the patient record
Time IST / itemWhat occurredHow it belongs in the evidence trail
08:45 · C1Meter check using control materialDevice quality record; never a patient glucose observation
09:00 · T1Tracing from a training simulatorNonpatient training material; not evidence about this patient
10:05 · E1Patient ECG attempt with a recorded acquisition limitationPreserve the unsuccessful attempt and reason; do not label it normal
10:12 · E2Separate patient ECG acquisition following the clinician’s directionNew acquisition linked to E1; retain its original tracing
10:18 · G1Patient point-of-care glucose measurementRetain the actual source value, unit, operator and device context
10:25 · R1Doctor reviews original E2 and G1One review event with two named targets; no diagnosis invented here
13:00 · S1Reception scans E2 but crops part of the outputIncomplete representation of E2; not another acquisition
13:08 · S2Reception makes a complete scan of the same E2 paperReplacement copy linked to E2 and S1; clinical review remains R1

Check the copy against the source without rewriting either clock

For E-73, suppose the E2 header prints 09:57 while the operator’s contemporaneous clinic record says 10:12. The times differ by fifteen minutes. Preserve both and flag their sources; do not edit the printed header or silently substitute the 13:08 scanning time. One disagreement does not prove that the device clock was exactly fifteen minutes slow throughout the day or that every older tracing needs the same correction.

The relevant technical staff should investigate the clock discrepancy using the applicable procedure, while the clinician considers whether uncertainty about timing affects the patient’s assessment. If a correction is established, record its basis, author and actual entry time alongside the original evidence. The scan cannot resolve the discrepancy simply by being sharper. FHIR R4 distinguishes the clinically relevant observation time from the availability time of a version; it provides a modelling distinction, not a rule for choosing which unverified clock is correct.

At the 13:15 preparation cutoff, use the copy checks below to explain what a referral packet can claim. A complete S2 file can represent the tracing reviewed earlier if its relationship to E2 has been established. It does not imply that the doctor reviewed the scan at 13:08, endorsed every machine-generated phrase, or retrospectively assessed an item absent from R1. A source-copy check and a new clinical interpretation should have different labels and accountable people.

On a narrow screen, scroll the table horizontally to read all columns.

E-73 source-copy checks for a proposed referral packet
Evidence itemRequired distinctionState or limit in this example
S1 versus S2Incomplete scan versus complete representation of the same tracingUse the reconciled S2 copy; retain the earlier scan’s limitation in the record
E2 time fieldsPrinted 09:57 versus operator-recorded 10:12Fifteen-minute discrepancy remains attributed and unresolved
Device comment on E2Machine text versus clinician interpretationKeep its device attribution; R1 is the separate clinical review reference
G1 result recordPatient value and unit versus C1 control materialOnly G1 is the patient glucose measurement in this case
Review reference R1Review of E2 and G1 versus review of all folder contentsNo review or diagnostic conclusion is inferred for T1, C1 or a later addition

Source: HL7 FHIR R4: Observation time and source context

Link the measurement to the clinical review that followed

A clinician reviewing a result should be able to locate its source evidence and record the assessment and plan in the encounter. The handover needs a responsible recipient, not just a general notification. If the doctor is unavailable, use the clinic’s accepted coverage arrangement.

An urgent clinical concern should follow the established care process immediately; completing an upload or filling every metadata field must not become a prerequisite for obtaining help. Any necessary retrospective documentation should preserve the actual sequence of events.

Make exported evidence readable to the next clinician

Before sending an ECG with a referral, check that the tracing is legible, correctly associated with the patient and not cropped at the edges. If the source includes settings needed to interpret the tracing, those should survive export. A thumbnail that looks acceptable in a list may be useless when printed.

For numerical or categorical results, retain units, observation time and source context in the exported summary. Avoid a bare number detached from its method or origin. If clinical review remains outstanding, do not present the export as a reviewed report.

Where a patient shares a family phone, confirm how the output will reach the intended recipient. Sending a file is a release event; it does not establish that another doctor received it, interpreted it or accepted responsibility for follow-up.

Hand over the evidence relationships and the unresolved clock question

A referral-preparation task should say which patient evidence is being assembled, which version is readable and what uncertainty remains. In E-73, the clinical record retains E1, E2 and G1 with their distinct roles, while the proposed referral material uses the reconciled S2 representation of E2 and the actual G1 result record with their review reference. Excluding C1 and T1 from a patient packet does not mean deleting quality or training records from the appropriate record system.

A family contact may be helping an older patient transfer the packet to another doctor. Establish the authorised recipient and patient identity before sharing, and explain in the preferred language that the two ECG scans represent the same tracing. Otherwise a family member might understandably describe them as two new tests. The person preparing the packet should not give a diagnostic explanation of the tracing merely to answer a question about duplicate files.

The filled record below is deliberately a handover at a preparation cutoff, not a claim that a referral was sent or accepted. If the responsible clinician considers a timestamp uncertainty or unreadable source clinically significant, use the appropriate clinical route without waiting for routine file administration. A later recipient acknowledgement would be a further event; it would not retrospectively change the acquisition or review times.

Evaluate whether the record can reconstruct a complicated visit

A useful demonstration includes one unsuccessful acquisition, one successful repeat and a later upload. The record should allow a reviewer to reconstruct the actual visit without relying on staff memory. Check that the trend view and exported summary use the right event times.

Also test a device-generated comment and a clinician-authored interpretation. Their attribution should remain distinct in the interface, audit history and PDF. These checks assess the documentation workflow; they do not certify a device or validate an interpretation algorithm.

Worked example: two acquisitions and one later scan

Fictional case: staff attempt an ECG at 10:05, record an acquisition limitation, and obtain another tracing at 10:12 following the responsible clinician’s direction. The doctor reviews the second tracing during the visit. At 13:00, reception scans the paper output. The chart records the unsuccessful attempt, the separate 10:12 acquisition, the doctor’s review and the 13:00 file-entry event. It does not describe the scan as a new afternoon ECG or silently delete the earlier attempt. No diagnostic conclusion is supplied by this example.

In-Clinic Investigation Evidence Worksheet

Use these prompts in your clinic’s own notes. Record the result, responsible person and next action for each item.

Source and role
Who acquired this measurement, who entered it and who interpreted it?
Validity evidence
Which device or kit records and acquisition limitations are relevant to this result?
Chronology
Can a reader distinguish measurement, repeat, review and upload times?
Portable output
Does the referral copy preserve the tracing or result, units, source and review status?

Questions doctors ask

Does uploading an ECG mean it has been clinically reviewed?

No. File receipt, acquisition and clinician review are separate events. Record the responsible clinician’s assessment and preserve the tracing that was actually reviewed.

Can an ECG machine’s interpretation become the doctor’s note automatically?

Keep automated comments attributed to the device. The clinician should review the evidence and record their own assessment rather than allowing copied text to imply an independent professional interpretation.

Should a glucometer control check appear in the patient trend?

Control material is part of the device quality process, not a patient measurement. Keep the quality record connected to the device without treating its value as a patient result.

Should the first reading be deleted after a repeat?

Preserve the sequence and explain whether the first attempt was unusable or remained a valid observation. A repeat has its own time and reason; it should not silently replace the earlier evidence.

Should clinical action wait until every result field is filled?

No administrative form should delay necessary clinical assessment or the clinic’s escalation process. Record the care sequence accurately, including later documentation when needed.

Is a clearer scan of the same ECG another test or another clinical review?

Neither follows automatically. Link the clearer copy to the original acquisition and record why it replaces the earlier scan. Preserve the actual review event and the source it covered. If a clinician performs a new interpretation, document that as its own event rather than inferring it from the upload.

What should staff do when an ECG header time differs from the clinic record?

Keep both times with their sources and investigate the discrepancy through the applicable device and clinical process. Do not change the printed source or use upload time as a substitute. A single mismatch does not justify applying one clock adjustment to every historical tracing; any established correction needs its evidence and actual entry time.

Sources and scope

Updated on . This article combines the cited references with Bractis’s original workflow analysis and illustrative examples. Product capabilities depend on the configured scope. No independent clinical or legal review is claimed.

  • HL7 FHIR R4: Observation time and source context

    Version 4.0.1 directly checked 28 October 2026 for effective and issued times, source device and observation context. A version-specific modelling reference, not a clinical acquisition procedure, current-release claim or evidence of Bractis conformance.

  • WHO: Laboratory quality management system handbook, 2011

    Publication overview directly checked 28 October 2026. The linked PDF returned 403; official indexed handbook text for equipment management and occurrence management supplied the limited supporting context. These 2011 background principles do not establish current device limits, certification or clinical decisions.

Try the workflow
with Bractis.

Inspect a fictional ECG with an unsuccessful first attempt, a reviewed repeat and a later scan, then check the exported referral evidence.

Request a clinic demo