David G. Regan

Background. The rate of diagnosis of Chlamydia trachomatis (chlamydia) infection has risen dramatically in Australia. In response, the Australian gov

Mathematical models have been used to investigate the dynamics of infectious disease transmission since Bernoulli's smallpox modelling in 1760. Their use has become widespread for exploring how epidemics can be prevented or contained. Here we discuss the importance of modelling the dynamics of sexually transmitted infections, the technology-driven dichotomy in methodology, and the need to 'keep it simple' to explore sensitivity, to link the models to reality and to provide understandable mechanistic explanations for real-world policy-makers. The aim of models, after all, is to influence or change public health policy by providing rational forecasting based on sound scientific principles.

Background: Vaccines are now available to prevent the development of cervical cancer from genital human papillomavirus (HPV) infection. The decision to vaccinate depends on a vaccine’s cost-effectiveness. A rigorous cost-effectiveness model for vaccinated individuals is presented in a companion paper; this paper investigates the additional benefits the community might receive from herd immunity. Methods: A mathematical model was developed to estimate the impact of a prophylactic vaccine on transmission of HPV type 16 in Australia. The model was used to estimate the expected reduction in HPV incidence and prevalence as a result of vaccination, the time required to achieve these reductions, and the coverage required for elimination. The modelled population was stratified according to age, gender, level of sexual activity and HPV infection status using a differential equation formulation. Clinical trials show that the vaccine is highly effective at preventing persistent infection and pre-cancerous lesions. These trials do not, however, provide conclusive evidence that infection is prevented altogether. The possible modes of vaccine action were investigated to see how vaccination might change the conclusions. Results: The model predicts that vaccination of 80% of 12-year-old girls will eventually reduce HPV 16 prevalence by 60–100% in vaccinated and 7–31% in unvaccinated females. If 80% of boys are also vaccinated, reductions will be 74–100% in vaccinated and 86–96% in unvaccinated females. A campaign covering only 12-year-old girls would require 5–7 years to achieve 50% of the eventual reduction. With a catch-up campaign covering 13–26-year-olds, this delay would be reduced to only 2 years. Unrealistically high coverage in both sexes would be required to eliminate HPV 16 from the population. Under pessimistic assumptions about the duration of vaccine-conferred immunity, HPV 16 incidence is predicted to rise in some older age groups. Conclusions: Mass vaccination with a highly effective vaccine against HPV 16 has the potential to substantially reduce the incidence and prevalence of infection. Catch-up vaccination offers the potential to substantially reduce the delay before the benefits of vaccination are observed. A booster vaccination might be required to prevent an increase in incidence of infection in women over 25 years of age.

Restricted diffusion in compartmentalized systems can lead to spatial coherence phenomena being observed in q-space plots from pulsed field gradient spin-echo (PGSE) nuclear magnetic resonance (NMR) experiments. The underlying features observed in these plots contain information on the geometry of the compartments that is otherwise difficult to obtain. A numerical procedure is proposed that accentuates these coherence features: the data are weighted with a bell-shaped window function, interpolated with a shifting cubic spline, and then the second derivative is taken prior to Fourier transformation. The window function provides apodization of the noisy data at high q values, while it and the second derivative are equivalent to applying a high-pass filter to remove the zero- or low-frequency components in the echo-signal attenuation. Using a combination of theory, Monte Carlo simulations, and data from PGSE NMR experiments on human red blood cells, we demonstrate this to be a valuable processing tool for delineating the underlying coherence features. It should prove particularly useful where the coherence features are poorly defined or where more than one pattern is present in a q-space plot. Magn Reson Med 52:907–912, 2004.