CD4 count testing is perceived to be an affordable strategy to diagnose treatment failure on first-line ART. We hypothesize that the superior accuracy of viral load (VL) testing will result in less patients being incorrectly switched to more expensive and toxic second-line regimens. Using data from a drug resistance cohort, we show that CD4 testing is approximately double the cost to make one correct regimen switch under certain diagnostic thresholds (CD4=US$499 vs. VL=US$186 or CD4=US$3,031 vs. VL=US$1,828). In line with World Health Organization guidelines, our findings show that VL testing can be both an accurate and cost-effective treatment monitoring strategy.