For cervical cancer screening to be feasible in developing countries, it must be accurate, inexpensive, and easy to administer. We conducted a pilot study in rural Shanxi Province, People's Republic of China, to determine disease prevalence and study feasibility in preparation for a large-scale comparative trial of 6 screening tests. One-hundred and thirty-six nonpregnant women with no history of hysterectomy, pelvic radiation, or Papanicolaou tests were screened in a rural clinic. Ten percent of the women enrolled reported abnormal vaginal bleeding and 45% reported abnormal vaginal discharge. The tests were the Papanicolaou test (both conventional and ThinPrep), a self-administered swab test by Hybrid Capture II for high-risk human papillomavirus (HPV), a test for high-risk HPV from residual PreservCyt medium, fluorescence spectroscopy, and visual inspection of the cervix by a clinician. All women also underwent colposcopy and biopsies as the reference standard. Biopsies showed 12 of 136 women had >/= high-grade squamous intraepithelial lesions (HGSIL). Screening was completed in 5 half-day sessions, the procedures went smoothly, and local cooperation was enthusiastic. Disease prevalence in Xiangyuan and Yangcheng Counties, Shanxi Province, can be estimated at 8.8% (95% CI, 4.5% to 15.0%). Screening 1000-2000 patients would be sufficient to detect a 10% difference in accuracy between diagnostic tests. The proposed large-scale trial is feasible.