Of the 1,812 women completing study visits in 2006, 1,451 (80%) had completed questionnaires in English and are the focus of this study. Of these, 1,214 completed questionnaires again in 2008, but 93 were missing responses to at least one question and were excluded. Of the remaining 1,121 women included in this analysis, 773 (69%) were HIV seropositive and 348 (31%) were seronegative. The 93 women missing data on the follow-up questionnaire were slightly older (45.7 vs 43.1 years, P = 0.01 vs included patients), with lower education levels (P = 0.004) and lower WRAT scores (median 26 vs 29, P = 0.0009), were more likely to be a current or former intravenous drug users than those with complete questionnaire data (48.4% vs 23.2%, P< 0.0001), and were more likely to have CD4 counts <200/cmm (20% vs 13%, P = 0.04). Demographic and medical characteristics of included women are presented in . Of the women with HIV, 500 (65%) were using highly active antiretroviral therapy at the time of the first survey.
Baseline demographic and medical characteristics of women who completed questionnaires at baseline and follow-up (n=1121). N (%)
Summary knowledge scores increased modestly between baseline and follow-up (12.7 +/− 5.8 vs 13.8 +/− 5.3, P < 0.0001). HIV-seropositive women had higher scores than seronegative women. This was true both at baseline (13.2 +/− 5.7 vs 11.8 +/− 6.0, P = 0.0002) and follow-up (14.1 +/− 5.3 vs 13.2 +/− 5.5, P = 0.01). However, the change in scores was statistically similar for the two groups (0.94 +/− 5.3 vs 1.47 +/− 5.5, P = 0.13).
Knowledge of cervical cancer prevention improved across a range of questions during follow-up. Nevertheless, substantial knowledge deficits persisted. For example, while significantly more women at follow-up understood what part of the body a Pap test evaluated (P = 0.0002), the proportion rose only to 52%. Knowledge of risk factors, indicating understanding of the causal factors underlying cervical cancer, remained marginal, with substantial proportions of women not understanding the link between cervical cancer and sexual activity, screening compliance, and smoking. Knowledge that cervical cancer is caused by a virus rose significantly (P = 0.005), but only to 24%. Belief that cervical cancer is preventable only rose from 52% to 55% (P = 0.04). Despite this, more than 90% of women continued to believe that regular Pap testing was important for both HIV infected and uninfected women. The proportion of women who knew HPV is a sexually transmitted virus causing warts and cervical cancer rose from 66% to 71% (P < 0.0002). About a third believed incorrectly that HPV could be cured with medication and roughly half believed that individuals can tell when they are HPV infected, proportions that did not change significantly across time. Awareness of the availability of HPV vaccination rose between 2006 and 2008 among the 1,121 women completing follow-up questionnaires, from 505 (45%) to 739 (66%), (P < 0.0001). This increased awareness appeared to arise from multiple sources, as women’s recognition of most sources of information increased over baseline (doctors 18% in 2006 vs 20% in 2008, nurses 10% vs 19%, WIHS staff 16% vs 33%, and advertising 69% vs 79%, P <0.0001). The proportion of women citing news reports or not recalling their information source did not change significantly (63% vs 55% for news, 7% vs 6% for unknown, P >0.2).
Knowledge about HPV vaccination also improved dramatically between surveys (), although many women did not understand the utility of vaccination in preventing perianal lesions and many believed it prevented herpes infections. Women also failed to appreciate the importance of targeting young girls for vaccination and incorrectly considered older women as good vaccine candidates. The proportion of women who believed HPV vaccination was extremely or very important for cervical cancer prevention, as compared to those who considered it somewhat/not important or were unsure, rose from 70% to 78% (P < 0.0001). Compared to 2006, more women at follow-up in 2008 believed recommending HPV vaccination to female relatives and friends to be extremely or very important (61% vs 66%, P = 0.001).
Percentage of correct responses to questions about vaccination against the human papillomavirus (HPV) at baseline and follow-up. (n=1,121)
Factors associated with an improvement in knowledge score are shown in . The first model demonstrated that higher baseline score, younger age, higher education level, higher income, and former as opposed to never drug use were associated with greater improvement in knowledge score. R2 for this model was 0.32, indicating that these factors explained about a third of the magnitude of change. HIV status was not significant after controlling for these factors. When added to the final model (model 3) WRAT reading level replaced education as a significant correlate (0.09, 95% C.I. 0.04-0.13, P < 0.001) and improved R2 to 0.35, suggesting that educational quality was a more important predictor of change in knowledge than the number of years in school.
Regression coefficients for ANCOVA models among participants completing questionnaires assessing cervical cancer prevention knowledge, assessing association between follow-up knowledge score and other factors.