In: Statistics and Probability
The Hemoccult test and the colonoscopy test for colon cancer screening are being conducted in 5000 people in a community. The Hemoccult test has a sensitivity of 78% and a specificity of 75%. The colonoscopy test has a sensitivity of 80% and a specificity of 90%. If the community has a prevalence of 12/1,000 for colon cancer, what is the positive predictive value (PPV) of the hemoccult test and the colonoscopy test, respectively? Which one would be good to recommend for screening test in a large population?
For Hemoccult test,
Disease | No Disease | Total | |
Positive | 60 * 0.78 = 46.8 | 4940 - 3705 = 1235 | 46.8 + 1235 = 1281.8 |
Negative | 60 - 46.8 = 13.2 | 4940 * 0.75 = 3705 | 13.2 + 3705 = 3718.2 |
Total | 5000 * 12/1000 = 60 | 5000 - 60 = 4940 | 5000 |
Positive predictive value (PPV) of the Hemoccult test = Total number of Positive test results with Disease / Total Positive results
= 46.8/ 1281.8 = 0.0365 = 3.65%
For Colonoscopy test,
Disease | No Disease | Total | |
Positive | 60 * 0.80 = 48 | 4940 - 4446 = 494 | 494 + 48 = 542 |
Negative | 60 - 48 = 12 | 4940 * 0.90 = 4446 | 12 + 4446 = 4458 |
Total | 5000 * 12/1000 = 60 | 5000 - 60 = 4940 | 5000 |
Positive predictive value (PPV) of the Colonoscopy test = Total number of Positive test results with Disease / Total Positive results
= 48/ 542 = 0.0886 = 8.86%
Since the Positive predictive value (PPV) of the Colonoscopy test is greater than that for Hemoccult test , we would recommend Colonoscopy test for screening test in a large population.