A mobile point-of-care analyser agreed reasonably with the laboratory for metabolic syndrome factors

Design

Pilot reliability and validity study against a reference laboratory, Bland-Altman analysis

Population

50 adults, mean age 52, in a mobile examination unit; 16 retested

Primary outcome

Agreement of point-of-care glucose, triglyceride, HDL and HbA1c with laboratory values and metabolic syndrome classification

Effect

Glucose bias -5 mg/dl (limits -25 to 6); triglycerides +8 (-58 to 30); HDL -2 (-14 to 9)

A 50-person pilot (mean age 52) compared several point-of-care analysers in a mobile unit with a reference laboratory for glucose, triglycerides, HDL cholesterol and HbA1c. The Alere device was chosen for the smallest bias and few missing values: glucose bias -5 mg/dl (limits of agreement -25 to 6), triglycerides 8 mg/dl (-58 to 30) and HDL -2 mg/dl (-14 to 9).

Metabolic syndrome classification matched the laboratory best with this device, and in 16 participants retested on another day, readings ranged from poor to excellent reliability but classification did not change between days.

The sample is small and the participants were not selected for risk, so this is a feasibility signal for camp-based or community screening rather than a basis to replace laboratory testing. Capillary readings near a diagnostic threshold still need confirmation.

• Use point-of-care readings for screening, and confirm results close to diagnostic thresholds in a laboratory
• Pair them with blood pressure and waist circumference to apply the metabolic syndrome criteria
• Check that any analyser you rely on has its own agreement data
• Record the device used so results can be compared over time

Why it matters

Opportunities for risk-factor screening in the community depend on devices that agree with laboratory results.

Don't overread it

A 50-person pilot cannot establish diagnostic accuracy, and one device's performance does not extend to others.

The statistics, in plain English

Bland-Altman limits of agreement show the range within which most individual readings differ from the laboratory. A bias of -5 mg/dl for glucose is small on average, but individual readings could differ by about -25 to +6, which matters near thresholds. Only 16 people were retested, so reliability estimates are imprecise.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.