Background: Chlamydia screening is recommended to prevent pelvic inflammatory disease (PID). A systematic review was conducted to determine how the natural history of Chlamydia trachomatis or Neisseria gonorrhoeae infection and progression to PID have been described in mathematical modeling studies.
Methods: Four databases, from their earliest dates to October 2009, and reference lists of included studies were searched. Models were defined as dynamic if progression from infection to PID was time dependent and static otherwise. Descriptions of the natural history of infection and parameter values used for progression to PID were extracted from all studies. Details of how disease progression was implemented were extracted from reports of dynamic models.
Results: Forty-five publications from 40 unique models were included. Nine models were classed as dynamic, including 4 Markov, 3 compartmental, and 2 individual-based models. There were 28 static decision analysis models. For 3 publications, the model type could not be determined. Among the dynamic models, there were explicit statements that C. trachomatis could progress to PID uniformly throughout the infection, in the first 6 months of infection, in the second half of infection, or that there is a most likely interval from the initial infection for the development of PID, which varies from 1 to 12 months. In static models, the average fraction of cases of chlamydia developing PID was 22%.
Conclusion: The reporting of key items in mathematical modeling studies about PID could be improved. The potential timings of progression to PID identified in this review can be investigated further to advance our understanding about how chlamydia screening interventions work to prevent PID.