Recalibrate assuming increase HPV/CIN/CC in KZN

[carajbro]

To approximate cervical cancer incidence in KZN rather than South Africa as a whole, we adjusted the Globocan 2018 rates by age taking into account increase HIV prevalence in KZN. To make the model match our estimated rates for KZN, we increased CIN3 to cervical cancer progression. However, it would be more accurate to adjust our HPV and CIN observed data and assume that more HPV is what leads to more cancer. (Our current HPV and CIN prevalence data is for Cape Town rather than KZN, Issue #28 ).

[carajbro]

To estimate cervical cancer incidence in KZN, adjusted the 2018 Globocan rates by age. Assumed WLWHIV have 4x increased risk of cervical cancer and accounted for increased HIV prevalence in KZN compared to SA nationally. HIV prevalence data in older ages are sparse, so assume KZN incidence is increased by the same proportion in women aged 50+. The figure below shows estimated cancer incidence rates by age for KZN (orange line) compared to Globocan 2018 (blue line).

Limitations:

• 4x multiplier for CC incidence probably doesn’t apply uniformly across all ages groups due to a) different peak ages of HPV acquisition and rates of progression to cancer by HIV status and b) competing risks.
• This change increased the credibility interval of cervical cancer incidence by age (range of 25 simulations)

Increased CIN3->CC progression multipliers:

[carajbro]

Elvira Singh heads the SA NCR. They don’t have the rates for KZN already calculated. What they would be able to do is give us numbers of cases for KZN for the years that we are interested in. Then we would have to calculate the rates. However, this is still only diagnosed cases. We would need to send a concept sheet and have an ethics waiver or ethics approval to use the raw data since it is not in the public domain. Deciding not to follow up on this as it is still not exactly what we want.