Research Laboratories: Using POCT Without Breaking a Dataset

Quick Answer — Point-of-care assays are excellent for screening, stratification, welfare monitoring and time-course experiments where the question is relative change. They are weak for primary endpoints that require traceable absolute values. The dividing line is not assay quality — it is whether the study needs a number that can be compared with literature produced on another platform. Get that decision right and the rest is documentation.

Quick Answer

Three questions decide the fit: does the endpoint need absolute traceability, will the study span a lot change, and will a reviewer be able to reproduce the method from what you write?

Where POCT fits well

Use	Why it fits
Screening animals for inclusion/exclusion	Fast, low volume, result available before randomisation
Stratification by severity	Quantitative value supports grouping better than a clinical score alone
Welfare monitoring in longitudinal studies	Same-day results allow intervention thresholds
Time-course experiments	Frequent sampling is impractical with sent-out testing
Pilot work before committing to a reference-lab study	Cheap way to size the effect

Where it does not

Use	Why not
Primary endpoint requiring absolute comparison with published values	Platform-specific calibration and bias
Multi-year study across lots without bridging	Lot shift confounds the time series
Measurements where the unit is defined by a reference method	Assay-defined units are not comparable
Regulatory submission data	Traceability chain must terminate at a recognised reference

Data pending. This section reports measured values from internal validation studies. They are published once the supporting study is complete; the current report is available from technical support on request.

The methods paragraph a reviewer expects

A reproducible POCT methods section contains, at minimum:

1. Platform and model, with firmware/software version where relevant.
2. Analyte, reagent catalogue number and lot number.
3. Calibration approach — lot-specific code, on-board curve, or shared curve.
4. Sample type and handling, including time to analysis and storage.
5. Precision at the concentrations used, from the platform’s own data or an in-study repeatability check.
6. Reference interval source, and a statement that it is platform-specific.
7. Statistical treatment of values below LoQ or above the measuring interval.

Omitting items 2, 3 and 5 is the most common reason a methods section is challenged.

Lot bridging, in practice

When a long study crosses a reagent lot change:

1. Retain 20–30 samples spanning the measuring interval.
2. Measure them on the old lot before depletion.
3. Measure the same aliquots on the new lot within the new lot’s stability window.
4. Compute the bias and decide: accept, correct, or report as a limitation.
5. Document the decision in the study file, not only in the paper.

A lot change without bridging is not a small documentation gap; it is a plausible alternative explanation for any effect the study reports.

Data integrity

• Export raw values, with units and timestamps, rather than transcribing into a spreadsheet by hand.
• Keep the instrument’s own records for audit.
• Record exclusions — clotted samples, insufficient volume, repeat runs — because they affect the denominator.
• Store the lot numbers with the dataset, not only in the lab notebook.

Six traps

1. Using POCT for a traceable primary endpoint. The most common design error.
2. Crossing a lot change without bridging. Confound that reviewers find.
3. Missing lot numbers in the methods. Not reproducible.
4. Reporting below-LoQ values as numbers. Manufactures precision.
5. Ignoring hemolysis in the sample set. Interference becomes a study effect.
6. Assuming reference intervals transfer. They do not, and the paper will be criticised for it.

The Author’s Take

Position. The research segment is small in revenue but disproportionate in credibility: a vendor whose platform appears in a well-documented published study gains a reference that marketing cannot buy.

Reasoning. Researchers are the harshest technical reviewers a diagnostic product will ever face, and they publish their verdicts. A platform that supports reproducible methods — lot documentation, stated traceability, precision at the concentrations used — accumulates citations that function as third-party validation, which is precisely what this category lacks. The commercial mistake is to treat research as a low-margin channel and under-support it; the correct move is to publish a methods template and a lot-bridging protocol, which cost nothing and turn every study into a reference.

Disclosure. This page is methodological guidance, not a performance claim. Figures marked indicate where platform documentation must be inserted.

Sources

1. Reference to be added.
2. Reference to be added.
3. CLSI EP05 / EP09 / EP28 — precision, method comparison, reference intervals.