Use of Point-Of-Care Cholesterol Testing in Population Based Non-Communicable Disease Surveillance: Caveats and Challenges

Abstract

Introduction: Point of care testing (POCT) for total cholesterol (TC) is invaluable in non-communicable disease (NCD) surveillance programs, as it may permit rapid risk stratification for efficient channeling of limited finances in resource constrained settings. Nevertheless, one needs to be aware of some caveats to the dependability of POCT results for TC in high load situations.

Aims & Objectives: To evaluate the analytical performance of POCT for TC in a population-based NCD surveillance study, by comparing its results with a laboratory assay, and to identify sources of error.

Place and duration of study: Mangamandi, Lahore (sampling); Services Institute of Medical Sciences, Lahore (laboratory), from December 2019 to March 2020.

Material & Methods: POCT for TC was done as part of CVD risk stratification in a large NCD surveillance project. Lower than expected readings of TC on POCT were flagged during routine data quality checking, and this prospective study was designed to determine accuracy of POCT readings by testing the same sample in a laboratory. Mean ±SD of two methods were compared in overall sample and in subgroups. Linear regression analysis was done to determine correlation between the two methods. After a significant disparity was confirmed, POCT process was scrutinized to identify its cause, and re -testing after its correction confirmed the source of interference.

Results: Mean TC level in overall sample (n= 699) by POCT was significantly lower than that of laboratory method: 2.80 (±0.30 SD) mmol/l vs. 5.28 (±1.27 SD) mmol/l (p <0.0001) R² 0.085. This trend persisted in subgroup analysis. A significant difference between the two methods was seen in a Bland Altman plot. POCT process evaluation identified optical window interference as a possible cause of the discrepancy, and after this was corrected, POCT results started showing a higher trend and became comparable with laboratory: 4.67 (±1.50 SD) vs. 5.45 (±1.89 SD) mmol/l, R² 0.9157.

Conclusion: Even though the utility of POCT for CVD risk stratification in NCD surveillance programmes is undeniable, some caveats and challenges remain. Non-compliance with device maintenance protocols in high throughput situations encountered in field testing may contribute to inaccurate results. Cholesterol POCT requires careful operator training, technical support and strong quality assurance backup.