Estimating adjusted prevalence ratio in clustered cross-sectional epidemiological data