Preferences for surveillance strategies for women treated for high-grade precancerous cervical lesions

OBJECTIVES: Data are lacking on how women view alternative approaches to surveillance for cervical cancer after treatment of high-grade cervical intraepithelial neoplasia. We measured and compared patient preferences (utilities) for scenarios with varying surveillance strategies and outcomes to inform guidelines and cost-effectiveness analyses of post-treatment surveillance options. METHODS: English- or Spanish-speaking women who had received an abnormal Pap test result within the past 2 years were recruited from general gynecology and colposcopy clinics and newspaper and online advertisements in 2007 and 2008. Participation consisted of one face-to-face interview, during which utilities for 11 different surveillance scenarios and their associated outcomes were elicited using the time tradeoff metric. A sociodemographic questionnaire also was administered. RESULTS: 65 women agreed to participate and successfully completed the preference elicitation exercises. Mean utilities ranged from 0.989 (undergoing only a Pap test, receiving normal results) to 0.666 (invasive cervical cancer treated with radical hysterectomy or radiation and chemotherapy). Undergoing both Pap and HPV tests and receiving normal/negative results had a lower mean utility (0.953) then undergoing only a Pap test and receiving normal results (0.989). Having both tests and receiving normal Pap but positive HPV results was assigned an even lower mean utility (0.909). 15.9% of the respondents gave higher utility scores to the Pap plus HPV testing scenario (with normal/negative results) than to the "Pap test alone" scenario (with normal results), while 17.5% gave the Pap test alone scenario a higher utility score. CONCLUSIONS: Preferences for outcomes ending with normal results but involving alternative surveillance processes differ substantially. The observed differences in utilities have important implications for clinical guidelines and cost-effectiveness analyses.