Background: Understanding patterns of chlamydia prevalence is important for addressing inequalities and planning cost-effective control programs. Population-based surveys are costly; the best data for England come from the Natsal national surveys, which are only available once per decade, and are nationally representative but not powered to compare prevalence in different localities. Prevalence estimates at finer spatial and temporal scales are required. Methods: We present a method for estimating local prevalence by modeling the infection, testing, and treatment processes. Prior probability distributions for parameters describing natural history and treatment-seeking behavior are informed by the literature or calibrated using national prevalence estimates. By combining them with surveillance data on numbers of chlamydia tests and diagnoses, we obtain estimates of local screening rates, incidence, and prevalence. We illustrate the method by application to data from England. Results: Our estimates of national prevalence by age group agree with the Natsal-3 survey. They could be improved by additional information on the number of diagnosed cases that were asymptomatic. There is substantial local-level variation in prevalence, with more infection in deprived areas. Incidence in each sex is strongly correlated with prevalence in the other. Importantly, we find that positivity (the proportion of tests which were positive) does not provide a reliable proxy for prevalence. Conclusion: This approach provides local chlamydia prevalence estimates from surveillance data, which could inform analyses to identify and understand local prevalence patterns and assess local programs. Estimates could be more accurate if surveillance systems recorded additional information, including on symptoms. See video abstract at, http://links.lww.com/EDE/B211.
Bibliographical noteFunding Information:
Financially supported by National Institute for Health Research: The NIHR Health Protection Research Unit in Modelling Methodology at Imperial College London in partnership with Public Health England (HPRU-2012-10080); Medical Research Council: MRC Centre for Outbreak Analysis and Modelling (MR/K010174/1). J.L. and P.J.W. are supported by the National Institute for Health Research Health Protection Research Unit in Modelling Methodology at Imperial College London in partnership with Public Health England (HPRU-2012-10080). P.J.W. is also supported by the MRC Centre for Outbreak Analysis and Modelling (MR/K010174/1). The views expressed are those of the authors and not necessarily those of the MRC, NHS, NIHR, Department of Health, or Public Health England.
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.