Modelling the impact of prevention strategies on cervical cancer incidence in South Africa

Abstract

In 2020, the World Health Organisation (WHO) published a strategy to eliminate cervical cancer as a public health concern. In South Africa, despite having a national screening policy in place since 2000, diagnosed cervical cancer incidence has shown no signs of decline. We extend a previously developed individual‐based model for human immunodeficiency virus (HIV) and human papillomavirus (HPV) infection to include progression to cervical cancer. The model accounts for future reductions in HIV incidence and prevalence and includes a detailed cervical cancer screening algorithm, based on individual‐level data from the public health sector. We estimate the impact of the current prevention programme and alternative screening scenarios on cervical cancer incidence. The South African screening programme prevented 8600 (95%CI 4700‐12 300) cervical cancer cases between 2000 and 2019. At current levels of prevention (status quo vaccination, screening, and treatment), age‐standardised cervical cancer incidence will reduce from 49.4 per 100 000 women (95%CI 36.6‐67.2) in 2020, to 12.0 per 100 000 women (95%CI 8.0‐17.2) in 2120. Reaching WHO's prevention targets by 2030 could help South Africa reach elimination (at the 10/100 000 threshold) by 2077 (94% probability of elimination by 2120). Using new screening technologies could reduce incidence to 4.7 per 100 000 women (95%CI 2.8‐6.7) in 2120 (44% probability of elimination at the 4/100 000 threshold). HPV vaccination and decreasing HIV prevalence will substantially reduce cervical cancer incidence in the long term, but improvements to South Africa's current screening strategy will be required to prevent cases in the short term. Switching to new screening technologies will have the greatest impact.

Mathematical Models for Evaluating Effectiveness and Cost-Effectiveness of Cervical Cancer Control Policies in Populations Including Women Living With Human Immunodeficiency Virus: A Scoping Review

“…15,17,18,26,32,34 Strategies combining HPV vaccination and primary HPV testing are predicted to be particularly effective in preventing cervical cancer. 24,33 A South African study compared cervical cell collection devices for cytology-based screening and found that the more expensive plastic brush would yield lower costs per woman screened than the wooden spatula because fewer smears would have to be repeated. 29 In a study on women living with HIV in Kenya, conventional VIA and HPV testing similarly reduced the number of cervical cancer cases and deaths because the effect of better diagnostic performance of HPV testing was balanced by minimized losses to follow-up through same-day treatment for VIA-positive women.…”

“…Only 5 studies modeled HIV transmission-1 because it focused on the effectiveness of HIV control strategies on cervical cancer incidence and mortality, 23 and the other studies extended HIV and HPV transmission models to assess the impact of cervical prevention strategies. 24,27,31,33 If HIV disease progression was incorporated, this was mostly done by including CD4 cell count stages in the model, yet CD4 cell categories varied across studies.…”

“…Thirteen studies (59%) conducted cost-effectiveness analyses (CEAs), and 9 (41%) examined the effectiveness of different prevention strategies on cervical cancer incidence or mortality. Fourteen studies (64%) were based on adaptations of previous modeling work: of these, 9 studies 15,16,18,25,28,[30][31][32][33] built on published HIV, HPV, or cervical cancer models and 5 studies 14,22,24,27,35 were adaptations of already included models. 19,20,23,31 Most studies compared cervical cancer prevention strategies in 1 country; 18 (82%) focused on African countries (Fig.…”

“…17,26,29 Five studies modeled HIV transmission. 23,24,27,31,33 HIV disease progression was typically characterized by CD4 cell counts (n = 9) or WHO clinical stage of HIV/AIDS (n = 1). Five studies did not model HIV disease progression but included HIV-related mortality in the model.…”

“…Six studies modeled HPV transmission between men and women. 16,23,24,27,31,33 All but 2 study 17,26 incorporated precancerous progression in their models, using the cervical intraepithelial neoplasia classification (n = 13) or Bethesda system (n = 6) of low-grade intraepithelial lesions or high-grade lesions. One study incorporated progression, no change, and regression as arms in a decision tree model.…”