Assessing the Accuracy of a Point of Care Analyzer for Hyperlipidemia in Western Kenya

The prevalence of hyperlipidemia, along with other non-communicable diseases, is on the rise in low- and middle-income countries. Given the resource-limited setting, a myriad of diagnostic challenges exist with traditional laboratory-based lipid tests, including mobility, timeliness, and laboratory infrastructure. Novel technology in the form of "point of care" devices seeks to overcome such barriers by providing immediate results without dependency on significant laboratory infrastructure. CardioChek PA (Polymer Technology Systems, Inc., Indianapolis, United States) is a point of care lipid measuring device and is readily available in Kenya. However, it has not been validated in this setting. In this study, I assess the accuracy of CardioChek PA with respect to standard laboratory-based testing, which is currently the gold standard.

In Webuye, Kenya, two blood samples were collected from 246 subjects to simultaneously measure the lipid levels via both CardioChek PA and the gold standard. All subjects were adults, and geographic stratified sampling methods were applied. Statistical analysis of the novel device's accuracy was based on percent bias, which is the standardized approach established by the National Cholesterol Education Program (NCEP) of the National Institute of Health (NIH). The NCEP suggests that percent bias be ≤±3% for low-density lipoprotein (LDL) cholesterol, ≤±5% for high-density lipoprotein (HDL) cholesterol, ≤±5% for total cholesterol (TC), and ≤±4% for triglycerides (TG). Misclassification rates and absolute percent bias were also analyzed.

This study found the CardioChek PA analyzer to be substantially inaccurate for LDL cholesterol (-25.9% bias), HDL cholesterol (-8.2% bias), and TC (-15.9% bias). For TG, the CardioChek PA performed well with a percent bias of 0.03%. However, the TG absolute percent bias (27.7%) and proportion of patients outside of the NCEP range (85%) reflected substantial inaccuracy of measurements. Moreover, those patients at higher risk of complications from hyperlipidemia were most likely to be misclassified into a lower risk category. Thus, we conclude that CardioChek PA is inaccurate and not suitable for our clinical setting. Furthermore, the findings highlight the need to validate new diagnostic tools in the appropriate setting prior to scale up regardless of its potential for novel utility.