Abstract
Dear Editor,
We read with interest the recently published article, “Correlates of HIV pre-exposure prophylaxis (PrEP) uptake among female sex workers in northern Nigeria”. 1 The authors address a critical evidence gap in an underserved region and deserve recognition for conducting fieldwork in a context marked by legal restrictions and substantial stigma. At the same time, several methodological aspects warrant clarification to support accurate interpretation of the findings and to strengthen future investigations in similar settings.
The use of venue-based time-location sampling (TLS) is appropriate for recruiting mobile and hard-to-reach populations. However, TLS requires transparent reporting of venue mapping procedures, sampling probabilities, and the use of sampling weights. These components are essential to reduce selection bias and improve external validity, as emphasized in the methodological guidelines by Muhib and colleagues. 2 In the present study, the absence of information on weighting and the reported 100% response rate raises concerns about potential overrepresentation of individuals already connected to One Stop Shop services. Such participants are more likely to have greater exposure to HIV prevention programming and may not reflect the broader population of female sex workers in northern Nigeria.
The definition of PrEP uptake as “ever use” is another area where methodological clarity is needed. Although self-reported measures are common in behavioral research, broad definitions that include any lifetime use can inflate estimates of meaningful PrEP engagement. Evidence from adherence research indicates that prevention benefit depends on consistent and sustained use rather than experimentation. 3 This distinction is especially important given the finding that only 41.7% of users reported daily adherence, while a majority reported irregular use. Without additional validation measures such as pharmacy refill verification or drug level biomarkers, self-reported adherence may also be susceptible to social desirability bias.
In the multivariate analysis, participation in group HIV prevention sessions emerged as the sole independent predictor of PrEP uptake, with an adjusted odds ratio of 14.22. Effect sizes of this magnitude in logistic regression models with modest sample sizes require careful scrutiny. Sparse data bias, model overfitting, and unmeasured confounding are known challenges in such scenarios, as outlined by Vittinghoff and McCulloch. 4 Although the authors note that variance inflation factors were examined, reporting these values would allow readers to assess potential multicollinearity and overall model stability.
Finally, the study is framed using the Social Ecological Model, yet structural variables such as policing practices, mobility constraints, and stigma were not incorporated into the regression analysis. Including multilevel determinants would strengthen coherence between the conceptual framework and the analytical approach.
Despite these limitations, the study provides valuable insight into PrEP awareness and service engagement in a region where data are scarce. Addressing the methodological considerations outlined above will enhance the robustness and generalizability of future research in similar contexts.
