Can we screen for pelvic organ prolapse without a physical examination in epidemiologic studies?

Objective: Large population-based epidemiologic studies of pelvic organ prolapse are rare. One barrier is the need for physical examination in order to confirm disease status. The objectives of this study were to develop a simple screening question for pelvic organ prolapse (POP) and to evaluate its test characteristics in high and low prevalence populations.
Study Design: Data from 100 women enrolled in the validation study of the Pelvic Floor Distress Inventory (PFDI) were used to identify the question or questions that most accurately identified women with advanced pelvic organ prolapse. After identifying an accurate and reliable screening question from this original group, its test characteristics were evaluated prospectively in 2 additional distinct populations: a group of 120 women presenting to a tertiary care urogynecology clinic (High prior probability of POP) and 448 women presenting to a nurse practitioner for annual gynecologic examination (Low prior probability of POP). Subjects in these 2 groups each completed the screening question and underwent a POPQ examination by a blinded examiner.
Results: A single question was identified from the original study population that most accurately and reliably identified those women with POP "Do you usually have a bulge or something falling out that you can see or feel in your vaginal area?" An affirmative answer to this question was 96% sensitive (95%CI 92-100) and 79% specific (95%CI 77-92) for prolapse beyond the hymen. The 1-week test-retest reliability was good (kappa .84). The prevalence of POP in this group was 29%. No other single question or group of questions had better test characteristics. When prospectively evaluated in the second High probability population (prevalence 39%), similar test characteristics were noted: sensitivity 85% (95%CI 71-93), specificity 86% (95%CI 75-92). However, when evaluated in the Low prior probability group (POP prevalence 3.8%) the specificity improved to 99% (95%CI 98-99), while the sensitivity decreased dramatically to 35% (95%CI 15-61).
Conclusion: Screening for POP without a physical examination is subject to spectrum bias. Spectrum bias occurs when a diagnostic test performs differently in different groups of patients. In groups with a high prior probability of POP, a simple screening question can accurately screen for advanced POP without a physical exam. However, in groups with a low prior probability of POP such as might be seen in a population-based epidemiologic study, this question has poor sensitivity.