Figure 3

One-way sensitivity analysis (SA) for aggregated costs of screening and diagnosis†. † (1) Base case assumptions included: a discount rate of 3% [3, 10], 11,475 women screened annually, a screening positivity rate of 14.6%, a cervical biopsy rate of 55% [27] and a ECC rate of 20% [28] for women with colposcopy. Various programmatic costs (including costs of advertising/ recruitment, training/quality control, and data management) ranged from $0.42 to $1.64. (2) SA for discount rate: 0%–6%; (3) SA for annual number of screened women: 2,000 and 6,000, based on previous experience, as explained in Methods. (4) SA for screening positivity rate: 7% [31] and 26% [32], two screening positivity rates observed in populations in China. (5) SA for multiplier of biopsy/ECC rate: because biopsy rate is usually clinician-dependent and subject to the underlying prevalence of disease in the population, we multiplied the rates at baseline for SA, as a function of multiplier of the biopsy/ECC rate at baseline (0.5, 1.5). (6) SA for multiplier of programmatic cost: the programmatic costs we used for baseline analysis were halved and doubled (0.5, 2.0), to account for the relatively large uncertainty associated with this parameter.