Infant Scale Calibration: Load Points and Traceability

When a biomedical technician or clinical engineering manager receives an external calibration report or vendor service ticket for an infant scale, the document often displays a reassuring green label marked PASSED alongside a single recorded measurement, frequently at 10 kg or 20 kg. CMS itself uses the scale to illustrate setting-dependent risk: slight miscalibration may matter little in an adult outpatient clinic but could have very serious consequences in a NICU. A scale can read correctly with a 10 kg test weight and still be out of tolerance at 800 g or 1.5 kg, which is where newborns and preterm infants are actually weighed and where weight-based medication dosing depends on the reading. A single-point check does not test that part of the range at all.

To constitute an acceptable service record, an infant-scale calibration certificate must demonstrate metrological coverage across the operating profile of the specific model variant. The certificate should identify the exact model number, serial number, Instructions for Use (IfU) revision, and metrological variant (verified legal-metrology unit versus non-verified unit). It should give as-found and as-left readings at or near the specified minimum capacity (Min), at or near each point where the maximum permissible error (MPE) changes, at the weighing-range or interval switchover, and at or near maximum capacity (Max). It should document an eccentricity (off-centre) test, because an infant is rarely centred in a curved tray. It must state which acceptance criterion was applied and why that criterion fits the variant, and it must identify traceable test weights whose combined error and uncertainty stay below one-third of the tolerance being checked.

Auditing that evidence goes beyond administrative review. Service record requirements and generic calibration-certificate auditing establish whether the laboratory, dates, traceability statements and signatures hold up. This article covers what those reviews do not: whether the load points and acceptance criterion fit a multi-interval or multiple-range weighing instrument.

Electronic infant scales often combine fine resolution at low loads with a higher total capacity by using two scale intervals. The way they switch between intervals falls into two families defined in NIST Handbook 44 (2026), Appendix D:

• Multi-Interval Instruments: An instrument having one weighing range that is subdivided into partial weighing ranges, each with different scale intervals (graduation steps). The active weighing interval is determined automatically based on the applied load, transitioning seamlessly during both increasing and decreasing load sequences.

• Multiple-Range Instruments: An instrument having two or more weighing ranges with different maximum capacities and different scale intervals on the same load receptor. Each range extends independently from zero to its individual maximum capacity.

This architectural distinction directly governs how an instrument behaves during calibration and clinical use. Consider the verified seca 376 (a multiple-range class III instrument). In range 1 the scale weighs from a Min of 0.1 kg up to 7.5 kg with a verification interval (e) of 5 g. In range 2 it weighs up to 20 kg with e = 10 g. Range 1 is active when the scale is switched on, and the scale changes to range 2 automatically above a set weight. As documented in the seca instructions for use, the instrument returns to Range 1 only after the load is completely removed from the tray.

This return-to-range-1 rule matters for testing. If a technician loads 15 kg and then removes weights down to 5 kg without fully unloading the tray, the scale stays in range 2 and shows 10 g steps under range-2 limits instead of the 5 g range-1 steps used for newborns. Range-1 points should therefore be taken on a loading sequence that starts from an empty tray, and the certificate should show which sequence each reading came from. Whether to also exercise range 2 below 7.5 kg on a decreasing load is a documented choice: that state is rarely reached clinically, so it is lower value than testing range 1 thoroughly.

Non-verified variants of the same product family can switch graduation at a different load. While the verified seca 376 switches intervals at 7.5 kg, the non-verified seca 374 and non-verified seca 376 switch at 10 kg (5 g graduation below 10 kg; 10 g from 10 kg to 20 kg). A vendor that applies the verified 7.5 kg profile to a non-verified 374 tests the wrong boundary and expects the wrong graduation between 7.5 kg and 10 kg.

The worksheet below compiles ranges, Min, graduations and accuracy statements for five seca baby scales from the current seca IfU revisions (July 2026). In-service limits for the two verified models are derived, not printed in the IfU: they are twice the initial-verification values under Directive 2014/31/EU Annex I 4.2 and OIML R 76-1 3.5.2, and they apply only where that legal framework applies. Eccentricity loads are about one-third of Max, following OIML R 76-1 3.6.2.

OIML R 76-1 (A.4.4.1) calls for weighing tests from zero up to Max and back to zero, with at least five test loads that include Max, Min and values at or near the loads where the MPE changes. NIST Handbook 44 Section 2.20 uses a similar increasing- and decreasing-load structure for commercial scales. For multi-interval and multiple-range infant scales, those points should come from the scale's own boundaries rather than from whatever weights are convenient:

1. Minimum Capacity (Min): Min is the lowest load the scale is specified to weigh: 40 g on the seca 757, 20 g on the 727 and 100 g on the verified 376. A record with no point at or near Min, or at least at a low load representative of the smallest infants weighed, leaves the most clinically sensitive part of the range untested. If the provider does not carry suitable small weights, the certificate should say so rather than omit the point silently.

2. First Tolerance Transition Point (500e): For class III instruments, the initial-verification MPE steps up from ±0.5e (to 500e) to ±1e (above 500e to 2000e) and ±1.5e (above 2000e); in-service limits are twice those values. For seca 757 range 1 (e = 2 g), 500e is 1.0 kg, so a 1.0 kg point checks the scale where the derived in-service limit changes from ±2 g to ±4 g. For non-verified models, use the manufacturer's own break points instead: 3.6 kg on the non-verified 374/376 and 6 kg on the 727.

3. Range Switchover Boundary: The load where graduation changes: 5 kg on the 757, 6 kg on the 727, 7.5 kg on the verified 376, and 10 kg on the non-verified 374/376 and the 334. Readings close to each side of the boundary show that both graduations display correctly and stay within limits.

4. Second Tolerance Transition Point (2000e) / Upper Mid-Range: On seca 757 range 1, 2000e is 4.0 kg (derived in-service limit changes from ±4 g to ±6 g); in range 2 (e = 5 g), 2000e is 10 kg. On the verified 376, the range-2 change at 500e falls at 5.0 kg (±10 g to ±20 g in service), but because the 376 returns to range 1 only when unloaded, that point is reachable only on a decreasing load. Record which choice was made.

5. Maximum Capacity (Max): A point at or near Max (15 kg on the 757; 20 kg on the 376, 727, 374 and 334) confirms the upper end of the span. It is also the point that single-weight 'CAL' checks usually cover already, which is why it cannot stand alone.

Two further elements belong in the record: decreasing-load readings and an eccentricity (shift) test. OIML R 76-1 A.4.4.1 removes the test loads back to zero after reaching Max, and NIST Handbook 44 N.1.2 sets out a decreasing-load test for automatic-indicating scales. For eccentricity, OIML R 76-1 3.6.2 generally uses a load of about one-third of Max (plus any maximum additive tare effect) in different positions, while Handbook 44's shift test (N.1.3.7) uses 30 to 35% of capacity in each quadrant for scales of 500 kg or less. Infants are rarely centred on a curved tray, so an omitted or failing eccentricity test is a real gap in the evidence.

The table turns those rules into suggested points for each worksheet model. It is an editorial planning aid derived from the IfU data, not a seca procedure; the provider's own documented method still governs.

Audits often stall on which numerical limit applies. Biomedical teams reasonably ask whether a US hospital should use European class III limits, NIST Handbook 44 tolerances or the manufacturer's stated accuracy.

In the European Union, Directive 2014/31/EU Article 1(2)(d) names the determination of mass in the practice of medicine, for weighing patients for monitoring, diagnosis and medical treatment, as a regulated use. Instruments for that use carry CE marking with the supplementary metrology marking and must show Max, Min, e and d near the display. Under Annex I 4.2, the MPE in service is twice the initial-verification MPE (±1e to 500e, ±2e above 500e to 2000e, ±3e above 2000e). In-service verification itself is governed by member-state rules.

In the United States, the regulatory structure is distinct. The Food and Drug Administration (FDA) classifies medical patient scales under 21 CFR 880.2720 as Class I devices (general controls), exempt from 510(k) premarket notification. Separately, NIST publishes Handbook 44, which states adopt for commercial weights-and-measures purposes. NIST Handbook 44 Table 7a does not include medical patient weighing in its listing of Class III commercial applications. Its Table 6 class III maintenance tolerances (1e, 2e and 3e up to 4000e) match the EU in-service values numerically, so a provider using Handbook 44 may reach the same pass/fail numbers. The difference is legal status and documentation, not arithmetic: treat Handbook 44 as a reference method for hospital infant scales, not as an established mandate. This article did not survey individual state rules on clinical scales.

For Medicare-participating hospitals, the binding expectation comes from CMS. Under State Operations Manual Appendix A, as revised in CMS Memorandum QSO-25-24-Hospitals, equipment should be inspected and tested for performance and safety before initial use and after major repairs or upgrades. Hospitals comply by following manufacturer-recommended maintenance activities and schedules, or through a documented Alternate Equipment Maintenance (AEM) program based on accepted standards such as ANSI/AAMI EQ56:2024, with decisions made by qualified personnel.

CMS uses the scale as its example of setting-dependent risk: a slightly miscalibrated scale in an adult outpatient clinic might not present significant risk, while a miscalibrated scale in a NICU could have very serious consequences. In practice that means the criterion should be explicit and matched to the variant. For a non-verified seca scale, the natural criterion is the manufacturer's stated accuracy, with any gap (such as 10 to 20 kg on the non-verified 374/376) closed by a documented decision. For a verified scale in the EU, the class III in-service limits apply under national rules. For a verified-model unit in a US hospital, the derived class III or Handbook 44 values are a reasonable, clearly labelled reference, not a legal requirement.

A scale reading is only as trustworthy as the weights used to check it. NIST Handbook 44 Appendix A, Section 3.2, states that a standard used without correction must have combined error and uncertainty less than one-third of the applicable device tolerance. OIML R 111-1 similarly says the error of a weight used to verify a weighing instrument shall not exceed one-third of the instrument's MPE. This is the one-third rule.

OIML R 111-1 sorts weights into classes from E1 (finest) through F1, F2, M1, M2 and M3. It describes class M1 weights as intended for use with medium-accuracy class III instruments, and lists class M2 for class III instruments as well. Class alone therefore does not settle the question; the one-third arithmetic at each test point does.

A worked example at the 1 kg point on seca 757 range 1 shows how much margin a class M1 weight leaves:

• Instrument Under Test: seca 757, Range 1 (e = 2 g).

• Derived In-Service Tolerance at 1 kg: 1 kg is exactly 500e for e = 2 g, so the derived in-service MPE is ±1e = ±2 g (±2,000 mg), twice the ±1 g initial-verification value in the IfU.

• Maximum Allowable Standard Error Budget (1/3 Rule): 2,000 mg / 3 ≈ 667 mg.

• OIML Class M1 Standard Specification: A 1 kg class M1 weight has an MPE of ±50 mg. OIML R 111-1 (5.2) limits the expanded uncertainty of its conventional mass to one-third of that MPE, about 16.7 mg. Worst-case error plus uncertainty is therefore about 66.7 mg.

• Compliance Ratio: About 66.7 mg uses roughly 10% of the 667 mg budget, comfortably inside the one-third rule.

The same arithmetic shows that weight class is rarely the limiting factor at infant-scale loads. A 1 kg class M2 weight (±160 mg, up to about 213 mg with its permitted uncertainty) fits easily. A class M3 weight (±500 mg) fits on MPE alone but reaches the 667 mg budget once its permitted uncertainty is added, so it leaves no margin. The table applies the same check at other loads, using derived in-service limits from the worksheet. What usually decides whether the weight evidence is acceptable is a current, traceable calibration certificate, not whether the weights are M1 rather than M2.

Beyond weight class, the certificate must establish unbroken traceability to national standards (for example NIST in the US, PTB in Germany or NPL in the UK). The report should list the weight set's identification or serial numbers, the weight calibration certificate number and date, the calibrating laboratory (and its accreditation, where claimed) and the recalibration due date. Readings taken with weights that lack a current certificate cannot support the one-third check, so they do not demonstrate accuracy. When qualifying external contractors, use service provider qualification standards to confirm that field technicians carry certified weight kits suited to infant-scale loads.

On verified scales (such as verified seca 757 and 376 units), quality seals and a verification counter show whether verification-relevant settings have changed since the last legal verification. Check them during routine service and post-repair return-to-service verification.

The seca IfUs describe two complementary safeguards:

• Quality seals: seca states that verification is required whenever one or more quality seals are damaged, and that seca Service should be contacted directly in that case. The person who performs verification secures it with an additional seal.

• Verification counter: The scale records each change in verification-related data in a counter. After verification, a counter sticker identifies the counter reading at that time.

At each service visit, compare the counter shown on the display with the number on the verification counter sticker. On the seca 757, the IfU describes pressing and holding any key while switching the scale on; the counter value then flashes on the display for a few seconds. Follow the procedure in the IfU for the specific model. Both numbers must match for the verification to be valid; if they do not, the scale must be re-verified.

Do not attempt adjustments, seal removal or service-mode changes on a verified scale. seca states that verification may be performed only by authorized persons in line with national legal regulations, recommends servicing before verification, and directs servicing and repairs to seca Service or an authorized service partner. Escalate drift, broken seals or a counter mismatch on that path, and record the escalation in the service history.

Weighing accuracy is only part of return-to-service acceptance. Three configuration checks also affect how readings are used:

• Metric resolution versus imperial display: Imperial display resolution is coarser than metric on these models. On the seca 727, the graduation is 1 g below 6 kg but 0.1 oz (about 2.8 g) in imperial units; on the seca 374 it is 5 g versus 0.2 oz (about 5.7 g). To reduce weight-based dosing errors, the National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) recommends measuring and documenting patient weights in metric units only, locking out pounds where a scale can measure in both, and buying scales that measure in, or can be locked to, metric units. Where the model allows a lock, confirm metric-only configuration before release. Where unit switching remains a user function (as the seca 334 IfU describes, while also telling users to rely only on results in SI units), record the setting and rely on the facility's metric-only policy.

• Gravity zone (GAL marking): A load cell responds to force, so a scale's reading depends on local gravitational acceleration. seca type plates carry a GAL marking for the location-dependent gravity value. When a scale is relocated to a distant site or a replacement unit arrives, intake should confirm that the marking matches the installed location and escalate any mismatch to seca Service or an authorized partner rather than adjusting in-house.

• Level and pre-use function check: seca's function check before each use covers visual inspection, alignment, display, controls and accessories, and the seca 376 has a spirit level. Confirm the scale is level and stable before testing and before release, and record any fault that the IfU troubleshooting does not resolve as a reason to keep the scale out of service.

A consistent review sequence keeps decisions defensible. The flow below summarizes the gates before an infant scale returns to clinical use:

graph TD
    A["Receive calibration certificate"] --> B{"Model variant and IfU revision match the unit?"}
    B -- "No" --> R1["REJECT: Re-test against the correct variant"]
    B -- "Yes" --> K{"Acceptance criterion stated and fits the variant?"}
    K -- "No" --> R0["REJECT: Re-issue with stated criterion"]
    K -- "Yes" --> C{"Five or more loads incl. Min or low load, error-change points, switchover and Max?"}
    C -- "No" --> R2["REJECT: Inadequate load-point coverage"]
    C -- "Yes" --> D{"Eccentricity test recorded and within limits?"}
    D -- "No" --> R3["DO NOT RELEASE: Repeat or refer to authorized service"]
    D -- "Yes" --> E{"Traceable weights meeting the one-third rule?"}
    E -- "No" --> R4["REJECT: Weight evidence missing or insufficient"]
    E -- "Yes" --> F{"Verified unit: seals intact and counter matches sticker?"}
    F -- "No" --> R5["ESCALATE: Authorized re-verification under national rules"]
    F -- "Yes or not applicable" --> G{"As-found readings within criterion?"}
    G -- "No" --> H["Impact review, then confirm as-left within criterion"]
    H --> I["ACCEPT: File certificate in CMMS"]
    G -- "Yes" --> I

The checklist below gives the acceptance condition, the typical deficiency and the corrective action for each gate:

A certificate that passes these checks shows the scale was tested where it is actually used, against a criterion that fits its variant, with weights that can be trusted. One that does not should go back to the provider rather than into the CMMS as proof of calibration.