Using a discrete choice experiment to inform the design of programs to promote colon cancer screening for vulnerable populations in North Carolina