Self-report and smoking tests: what does each measure?

01

Start with the question, not the test label

‘Have you smoked at all in the last seven days?’ and ‘Have you smoked since your quit date?’ ask about different periods. Self-report can describe a long interval, but relies on understanding the question and recalling and reporting the behaviour. A sample collected today does not observe every day in that interval.

The SRNT 2019 update discusses biochemical verification as a research method, including its value and its practical limits. A study can combine a behavioural definition with a specified marker; ‘verified’ still needs a method and a time window.

[1]

02

CO and cotinine are not interchangeable

Exhaled carbon monoxide (CO) reflects recent exposure to combustion. Its window is relatively short, and sources other than tobacco smoking, including environmental combustion, can affect it. A low reading cannot establish that every earlier day was smoke-free.

Cotinine is a metabolite of nicotine. Nicotine medicines and nicotine-containing products other than cigarettes can contribute to it. Cotinine on its own therefore does not identify cigarettes as the source. The specimen, assay and individual metabolism also affect interpretation.

[1]

03

Two fictional mismatches

Fictional example A: a research participant reports no cigarettes in the last week but is using a nicotine patch; the study finds cotinine. Before treating this as cigarette use, the researchers need the product history and their prespecified verification rule. No personal threshold is being supplied here.

Fictional example B: someone reports smoking earlier in the study but not recently, and the final breath sample is below the study’s CO cut-point. The result does not erase the earlier report; the sample and the report cover different periods. These examples explain questions to ask, not how to interpret your own test.

04

A missing sample is not a positive result

Sample collection can fail or be incomplete. A protocol may count an unverified person as not meeting the study endpoint, exclude an observation or use another stated analysis. These choices change the reported result, but failure to obtain a sample is not the same fact as detecting a marker.

Read how many people reported abstinence, how many supplied usable specimens and how many met the verification rule. If only a subset was sampled, check how it was selected before treating that subset as the whole study.

[1]

05

What to ask about a published verified rate

Look for the behaviour window, marker, specimen, collection timing, assay or device, cut-point rationale, other product exposure and treatment of missing specimens. You do not need to publish your own smoking history or upload a laboratory report to ask how a public study defined its outcome.

For a personal result, ask the clinician or laboratory which exposure and period the test can address, and what other products matter. In England, the NHS Stop Smoking Service directory is a route to qualified support; check with the service before visiting. Research guidance alone should not decide a clinical, employment or insurance classification.

[1][2]

06

What to keep in mind

• A behaviour window and a detection window are different.
• Cotinine measures nicotine exposure, not a cigarette-specific history.
• A missing specimen and a detected marker must not be conflated.
• Discordance calls for context, not an accusation.

07

Common questions