Scenario 1 – treatment regime based on treating one index case and all first degree contacts every seven days (Q = 10−1.5).

Results of simulating our model at equilibrium as a function of varying network size and structure. All results are given as the means and three standard deviations of 24 individual one-year simulations at equilibrium. Each simulation is based on a randomly generated network of the given architecture. For prevalence rates only, results denoted by an asterix indicate mean values significantly different from the mean values given in the baseline column. (Student's t-test with significance accepted if pBL: Baseline network (Boldface, N = 200, clustered, broad degree distribution).A: N = 200; clustered, broad degree distribution, low degree.B: N = 200; clustered, regular.C: N = 200; Non-clustered, approximately scale free.D: N = 200; Non-clustered, regular.Expected CC: The clustering coefficient of a randomly generated graph with the same connectivity.**Incidence rates per 100 person years.***Effective treatments per 100 person years.