Early Impact and Performance Characteristics of an Established Anal Dysplasia Screening Program: Program Evaluation Considerations

Christopher Mathews, Joseph Caperna, Edward R. Cachay and Bard Cosman Published Date: (29 November, 2007) Background: Screening for invasive anal cancer and its precursors is being increasingly advocated as a response to increasing incidence among HIV-infected persons. We implemented a comprehensive screening program in 2001 and report our early experience to inform monitoring and evaluation of such programs. Our research aims were: (1) to estimate incidence of and mortality from invasive anal cancer (IAC) before (1995-2000) and after (2001-2005) screening program implementation and (2) to examine potential screening program quality indicators. Methods: The study cohort included all patients under care for HIV infection at UCSD Owen Clinic between 1995-2005. Person-time incidence rates (IR) and case survival of IAC were estimated for the pre-screening (1995-2000) and postscreening (2001-2005) periods. High resolution anoscopy (HRA) operator accuracy was estimated by kappa agreement between cyto-histologic comparisons. Program quality indicators included: (1) screening coverage; (2) percent technically unsatisfactory cytology smears; (3) time between 1st abnormal cytology and 1st HRA; and (4) time between last clinic visit and last cytology. Results: 28 cases of IAC and 13,411 person-years were observed between 1995-2005. IRs (95% CI) pre-screening and post-screening were 199 and 216 per 100,000 person-years, respectively. There was no routine treatment of high grade squamous intraepithelial lesions (HSIL) during the study period. The percent of patients with IAC requiring chemoradiation decreased from 90.9% to 70.6% (p=0.36). There was a significant improvement in cyto-histologic agreement at HRA with increasing operator experience (r=0.92, p=0.025). Screening coverage was 73% of the target population. Among 14 providers, the percent unsatisfactory cytology smears averaged 27% but varied from 0 - 62%. The median time from 1st abnormal cytology to 1st HRA was 258 days. The median interval between the last cytology and the last clinic visit was 207 days. Conclusions: (1) The overall IR of IAC did not decline in the screening era and was higher than previous estimates for HIV cohorts; (2) stage shift to IAC of more favorable prognosis is a reasonable screening goal; (3) HRA accuracy varied by provider experience; (4) because of delay in access to HRA, digital rectal exam should be combined with cytology screening to detect palpable disease.