Computer-aided System of Evaluation for Population-based All-in-One Service Screening (CASE-PASS): From Study Design to Outcome Analysis with Bias Adjustment

Chen, Li-Sheng; Yen, Amy Ming-Fang; Duffy, Stephen W.; Tabar, Laszlo; Lin, Wen-Chou; Chen, Hsiu-Hsi

doi:10.1016/j.annepidem.2010.06.003

« Previous Next »Annals of Epidemiology
Volume 20, Issue 10 , Pages 786-796, October 2010

Computer-aided System of Evaluation for Population-based All-in-One Service Screening (CASE-PASS): From Study Design to Outcome Analysis with Bias Adjustment

Li-Sheng Chen, PhD
- Division of Biostatistics, Graduate Institute of Epidemiology, College of Public Health, National Taiwan University, Taipei, Taiwan
- School of Oral Hygiene, College of Oral Medicine, Taipei Medical University
Amy Ming-Fang Yen, PhD
- Division of Biostatistics, Graduate Institute of Epidemiology, College of Public Health, National Taiwan University, Taipei, Taiwan
- Centre of Biostatistics Consultation, College of Public Health, National Taiwan University, Taipei, Taiwan
- School of Oral Hygiene, College of Oral Medicine, Taipei Medical University
Stephen W. Duffy, MS
- Department of Epidemiology, Mathematics and Statistics, Cancer Research UK Centre, London
Laszlo Tabar, MD, PhD
- Department of Mammography, Falun Central Hospital, Falun, Sweden
Wen-Chou Lin, MS
- Center for Disease Control, Department of Health, Taipei, Taiwan
Hsiu-Hsi Chen, DDS, PhD
- Division of Biostatistics, Graduate Institute of Epidemiology, College of Public Health, National Taiwan University, Taipei, Taiwan
- Centre of Biostatistics Consultation, College of Public Health, National Taiwan University, Taipei, Taiwan
- Address correspondence and reprint requests to: Professor Hsiu-Hsi Chen, Division of Biostatistics, Graduate Institute of Epidemiology, College of Public Health, National Taiwan University, Room 533, No. 17, Hsuchow Road, Taipei, Taiwan. Tel: +886-2-33228033. Fax: +886-2-23587707.

Received 6 January 2010; accepted 23 June 2010.

Purpose

Population-based routine service screening has gained popularity following an era of randomized controlled trials. The evaluation of these service screening programs is subject to study design, data availability, and the precise data analysis for adjusting bias. We developed a computer-aided system that allows the evaluation of population-based service screening to unify these aspects and facilitate and guide the program assessor to efficiently perform an evaluation.

Methods

This system underpins two experimental designs: the posttest-only non-equivalent design and the one-group pretest–posttest design and demonstrates the type of data required at both the population and individual levels. Three major analyses were developed that included a cumulative mortality analysis, survival analysis with lead-time adjustment, and self-selection bias adjustment. We used SAS AF software to develop a graphic interface system with a pull-down menu style.

Results

We demonstrate the application of this system with data obtained from a Swedish population-based service screen and a population-based randomized controlled trial for the screening of breast, colorectal, and prostate cancer, and one service screening program for cervical cancer with Pap smears. The system provided automated descriptive results based on the various sources of available data and cumulative mortality curves corresponding to the study designs. The comparison of cumulative survival between clinically and screen-detected cases without a lead-time adjustment are also demonstrated. The intention-to-treat and noncompliance analysis with self-selection bias adjustments are also shown to assess the effectiveness of the population-based service screening program. Model validation was composed of a comparison between our adjusted self-selection bias estimates and the empirical results on effectiveness reported in the literature.

Conclusions

We demonstrate a computer-aided system allowing the evaluation of population-based service screening programs with an adjustment for self-selection and lead-time bias. This is achieved by providing a tutorial guide from the study design to the data analysis, with bias adjustment.

Key Words: Computer-aided System, Lead Time, Population-based Screening, Quasi-Experimental Design, Self-Selection Bias

Selected Abbreviations and Acronyms: CASE-PASS, Computer-aided System of Evaluation for Population-based All-in-One Screening, RR, relative risk, CI, confidence interval, IACCS, International Asian Conference on Cancer Screening