Diagnostics and Lab Equipment Editing and Proofreading Services

A test is 99% sensitive and 99% specific. In a population where one person in a thousand has the disease, most of its positive results will be wrong. Both statements are true, neither is a criticism of the test, and the gap between them is where diagnostics marketing routinely misleads people who are not thinking about base rates at nine in the morning.

We edit what diagnostics developers and laboratory equipment manufacturers produce — performance claims and their supporting studies, instructions for use and limitations sections, package inserts and product labelling, analytical and clinical validation reports, regulatory submissions and technical files, method comparison and precision study reports, customer-facing performance summaries and brochures, laboratory implementation and verification guidance, interpretation guidance for clinicians, instrument manuals and maintenance documentation, and troubleshooting and error resolution material. Our editors work on the claim a laboratory will verify.

The performance claim is where a diagnostic is either honestly represented or set up for a complaint, and its failure is presenting sensitivity and specificity as properties of the test alone. They are properties of the test in a population, measured against a comparator, at a threshold. We work through these so every performance figure carries the population it was measured in — including the prevalence, the setting and the spectrum of disease, since a test validated in symptomatic hospital patients does not carry that performance into screening; so the reference standard is named, because a test compared against another imperfect test cannot exceed it and "compared with the leading assay" is not a truth claim; so confidence intervals are given, given that a sensitivity of 97% from 34 positives has an interval most readers would find sobering; so the predictive values are shown at two or three realistic prevalences rather than at the study's own, as this is the number the clinician actually needs and it is almost never published; so the cut-off is stated with what it was optimised for, and the effect of moving it shown; so the limitations section names the actual interfering substances, cross-reacting conditions and specimen problems rather than referring to them generally. Claims written this way survive a laboratory's own verification, which is where they are otherwise found out.

Everything you send is treated in confidence, including validation data, submissions and product documentation. We are editors rather than clinical scientists, statisticians or regulatory advisers, and we offer no view on performance, validation or claims. What we can do is keep the population attached to the number.