Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.

BackgroundIn drug trials, adverse events (AEs) burden can induce treatment non-adherence or discontinuation. The non-adherence and discontinuation induce selection bias, affecting drug safety interpretation. Nested case-control (NCC) study can efficiently quantify the impact of the AEs, although choice of sampling approach is challenging. We investigated whether NCC study with incidence density sampling is more efficient than NCC with path sampling under conditional logistic or weighted Cox models in assessing the effect of AEs on treatment non-adherence and participation in preventive antimalarial drug during pregnancy trial.MethodsUsing data from a trial of medication to prevent malaria in pregnancy that randomized 600 women to receive chloroquine or sulfadoxine-pyrimethamine during pregnancy, we conducted a NCC study assessing the role of prospectively collected AEs, as exposure of interest, on treatment non-adherence and study non-completion. We compared estimates from NCC study with incidence density against those from NCC with path sampling under conditional logistic and weighted Cox models.ResultsOut of 599 women with the outcomes of interest, 474 (79%) experienced at least one AE before delivery. For conditional logistic model, the hazard ratio for the effect of AE occurrence on treatment non-adherence was 0.70 (95% CI: 0.42, 1.17; p = 0.175) under incidence density sampling and 0.68 (95% CI: 0.41, 1.13; p = 0.137) for path sampling. For study non-completion, the hazard ratio was 1.02 (95% CI: 0.56, 1.83; p = 0.955) under incidence density sampling and 0.85 (95% CI: 0.45, 1.60; p = 0.619) under path sampling. We obtained similar hazard ratios and standard errors under incidence density sampling and path sampling whether weighted Cox or conditional logistic models were used.ConclusionNCC with incidence density sampling and NCC with path sampling are practically similar in efficiency whether conditional logistic or weighted Cox analytical methods although path sampling uses more unique controls to achieve the similar estimates.Trial NCT01443130.

Original publication





PloS one

Publication Date





School of Public Health, University of the Witwatersrand, Johannesburg, South Africa.


Humans, Pregnancy Complications, Parasitic, Malaria, Sulfadoxine, Pyrimethamine, Chloroquine, Drug Combinations, Antimalarials, Chemoprevention, Incidence, Case-Control Studies, Pregnancy, Models, Theoretical, Female, Medication Adherence, Data Analysis