The objective of the study was to investigate the protective effect of Papanicolaou test screening and contacts with the medical care system when considered in conjunction with established risk factors for cervical cancer. The design was a statewide (Pennsylvania) case-control study of invasive cervical cancer. Cases were identified from the Pennsylvania Cancer Registry. Controls were identified from the neighborhoods of the cases and were matched to the cases on age at diagnosis and race. The average age of the cases was 45.3 years and the controls, 44.6 years. The independent variables were: checkup and physician's visits within the 3 years before diagnosis, Papanicolaou test within 3 years of diagnosis, marital status, income, smoking, number of pregnancies, number of long-term relationships, age at first pregnancy, education, ever use of birth control, and, for 1975 and later, ever use of oral contraceptives, condoms, or a diaphragm. Multiple logistic regression models including those variables that were significant on univariate analysis identified four significant risk factors for invasive cervical cancer: no Papanicolaou test within 3 years of diagnosis (odds ratio [OR] = 3.10, 95% confidence interval [CI] = 1.45-6.64), two or more long-term relationships (OR = 2.78, 95% CI = 1.26-6.13), use of protective condoms (OR = 0.31, 95% CI = 0.12-0.79), and income less than $15,000 (OR = 2.59, 95% CI = 1.17-5.72). Each of these risk factors appears to exert its own independent effect. This study demonstrates that future prevention strategies should include educating women about the need to incorporate Papanicolaou test screening into their ongoing medical care, the protection afforded by condoms, and the increased risk associated with multiple sexual partners.