Table 3.
Relative effectiveness of HPV testing versus cytology by income classifications
| Lower-middle-income areas | Upper-middle-income areas | |||||
|---|---|---|---|---|---|---|
| Cytology | Overall HPV testing | aOR (95%CI) | Cytology | Overall HPV testing | aOR (95%CI) | |
| Screen positivity (%, 95%CI) | 4879 (3.1) | 23 764 (9.6) | 3.40 (3.29–3.51) | 8345 (4.9) | 60 827 (10.4) | 2.27 (2.21–2.32) |
| Per protocol colposcopy referral (%, 95%CI)a | 4879 (3.1) | 9050 (3.7) | 1.21 (1.17–1.25) | 8345 (4.9) | 19 982 (3.4) | 0.69 (0.67–0.71) |
| Overall colposcopy referral (%, 95%CI) | 4206 (2.7) | 9721 (3.9) | 1.55 (1.50–1.61) | 8106 (4.8) | 23 346 (4.0) | 0.83 (0.80–0.85) |
| Per protocol detection rate of CIN or cancer (per 1000, 95%CI) | ||||||
| CIN2+ | 406 (3.0) | 993 (4.7) | 1.53 (1.37–1.70) | 816 (5.6) | 2867 (5.7) | 1.03 (0.96–1.11) |
| CIN2 or 3 | 368 (2.7) | 912 (4.3) | 1.55 (1.38–1.73) | 757 (5.2) | 2611 (5.2) | 1.01 (0.94–1.09) |
| Invasive cervical cancer | 38 (0.3) | 81 (0.4) | 1.32 (0.92–1.89) | 59 (0.4) | 256 (0.5) | 1.32 (1.02–1.71) |
| Overall detection rate of CIN or cancer b (per 1000, 95%CI) | ||||||
| CIN2+ | 417 (3.1) | 1039 (4.9) | 1.56 (1.40–1.73) | 885 (6.0) | 3060 (6.1) | 1.01 (0.95–1.09) |
| CIN2 or 3 | 373 (2.8) | 938 (4.4) | 1.57 (1.41–1.76) | 798 (5.5) | 2760 (5.5) | 1.01 (0.94–1.09) |
| Invasive cervical cancer | 44 (0.3) | 101 (0.5) | 1.44 (1.03–2.01) | 87 (0.6) | 300 (0.6) | 1.08 (0.86–1.35) |
| Positive predictive value c (%, 95%CI) | ||||||
| CIN2+ | 406 (9.9) | 1018 (12.4) | 1.23 (1.10–1.38) | 816 (11.4) | 3012 (14.3) | 1.31 (1.21–1.41) |
| CIN2 or 3 | 368 (9.0) | 936 (11.4) | 1.25 (1.11–1.40) | 757 (10.6) | 2732 (13.0) | 1.27 (1.17–1.38) |
| Invasive cervical cancer | 38 (0.9) | 82 (1.0) | 1.02 (0.71–1.46) | 59 (0.8) | 280 (1.3) | 1.65 (1.27–2.14) |
Note: HPV human papillomavirus, aOR adjusted odds ratio, CI confidential interval, CIN cervical intraepithelial neoplasia, ICC invasive cervical cancer, CI confidential interval. aPer protocol colposcopy referral was defined as the women who were screened positive and referred to colposcopy according to the protocol. b Overall detection rates included cases detected in per protocol colposcopy and others detected in opportunistic colposcopy. c Positive predictive value represented the detected cases from screened positivity. aOR was calculated by using multivariate logistic regression adjusted for age and ever screening