Table 1. Parameters used in the model and ranges explored in probabilistic sensitivity analysis*.
| Baseline values | Ranges explored in sensitivity analysis | Data informing assumptions | |
|---|---|---|---|
| Diagnostic test accuracy | |||
| Primary HPV test sensitivity (specificity) toCIN2+ | 96% (90%) | 95% - 98% (87% - 93%)**for one-way sensitivity analysis, we assumed that in the worst case, the HPV test would be unable to detect the presence of 5.8% of cervical cancers (which are assumed to be truly HPV negative). [52] | Positivity rates informed by international meta-analyses.[53] |
| HPV test-of-cure sensitivity (specificity) toCIN2+ | 93% (81%) | 86% - 97% (74% - 86%) | |
| Unsatisfactory rate for the HPV test | 0% | 0%-1% | |
| LBC test sensitivity (specificity) toCIN2+ | ASC-US threshold: 77% (94%) | ASC-US threshold: 72% - 81% (92% - 95%) | Informed by a systematic review and meta-analysis comparing liquid-based and conventional cytology[54] |
| Unsatisfactory rate for LBC test | 1.80% | 0.3%-2.6% | Informed by data from a prospective Australia-based study[55] |
| Colposcopy sensitivity (specificity) to CIN2+ | 88% (52%) | 80–91% (49–74%) | Informed by a large dataset of >20,000 colposcopies from the Royal Women’s Hospital in Victoria.[41,42] |
| Percent of positive HPV test results where HPV 16/18 is misclassified as other high-risk types | 0% | 0–5% | Assumption |
| Percent of positive HPV test results where OHR HPV types are misclassified as HPV 16/18 | 0% | 0–5% | Assumption |
| Screening attendance rates | |||
| Probability of attending routine screening | Cumulative attendance in under 5 years is 9%; by 5 years is 80%; by 6 years in 85% | Attendance delayed by a year: cumulative attendance in under 6 years is 9%, and by 6 years is 80%. | Early and on-time attendance was informed by VCCR data and data from England (which has a call-recall system). Screening behaviour in under-screened women unchanged from current rates (based on VCCR data). |
| Probability of attendance for 12-month follow-up visits within a year | 65–85% (depends on age) | Rate increased/decreased by 10%** | Informed by 12 month attendance as derived from VCCR data. |
| Probability of screening uptake at first invitation | 82% at age 25, 91% attend before the age of 30 | 74% at age 25, 85% attend before the age of 30 | Uptake at 25 years was assumed to be equivalent to update at or by 25 years as observed under current practice. |
| Probability of colposcopy attendance within a year after referral | 85–95% (depends on age) | Rate increased/decreased by 10%** | Informed by data from the VCS and the Royal Women’s Hospital.[41,42] |
| Cost assumptions | |||
| cytology test cost | $30.50 | $19.45-$42.00 | Expert advice from the Renewal Steering committee |
| HPV test cost | $30.00 | $20.00-$45.00** | |
| Natural history assumptions | |||
| Natural history aggressiveness | 5% reduced progression rates; 5% increased progression rates^ | This variation produces cancer rates that fit within uncertainty ranges derived from AIHW.[56] |
* The uniform distribution is used for probabilistic sensitivity analysis when parameters from the ranges are selected for all parameters except for the choice of test sensitivity and specificity for the HPV and LBC test–in these cases, one of the three combinations of sensitivity and specificity are selected each time (i.e. there is no continuous distribution, just three discrete choices).
** Variations in these parameters were also explored in one-way sensitivity analysis.