Cost-effectivess of Colonoscopy as a Screening Tool for Colorectal Cancer

Cost-effectivess of Colonoscopy as a Screening Tool for Colorectal Cancer

Karen Harper

National University

March 27, 2016

Cost-effectiveness of Colonocopy as a Screening Tool for Colorectal Cancer

Colonoscopy is widely regarded as the “gold standard” for colorectal cancer screening in the United States. During this procedure, the clinician can remove polyps in the colon which may be precancerous (adenomas). Although many clinicians continue to recommend colonoscopy for colorectal cancer screening, questions arise as to the procedure’s efficacy and cost-effectiveness. Colonoscopy is a highly invasive procedure, requiring the patient to be either sedated or placed under general anesthesia while the colon is examined. Preparation for the procedure is uncomfortable and inconvenient for many patients as a large amount of laxative must be consumed to flush out the bowel. In addition, colonoscopy is a very expensive procedure. Although many adenomas have been identified and removed via colonoscopy, most adenomas do not eventually become cancerous, again bringing into question whether this invasive procedure is really cost-effective. This paper will compare two studies in which the efficacy and cost-effectiveness of colonoscopy were examined.

Study by Sonnenberg, Delco, and Inadomi

Amnon Sonnenberg, Fabiola Delco, and John Inadomi published a study in 2000 with the objective of examining the cost-effectiveness of colonoscopy in comparison to other screening tests (fecal occult blood testing and flexible sigmoidoscopy). This study used a hypothetical sample of 100,000 adults age 50 and older. This hypothetical cohort was followed until death. Computer models used a Markov process to randomly determine whether each of the study participants received screening via annual fecal occult blood testing, flexible sigmoidoscopy every five years, or colonoscopy every ten years starting at age 50 until death. For those who had positive fecal occult blood tests or adenomatous polyps discovered during sigmoidoscopy, a colonoscopy was then performed every three years until all polyps were gone. Data for the computer model came from published U.S. vital statistics and cancer statistics on specificity and sensitivity of the screening methods, as well as efficacy of each method. Costs were estimated based on Medicare reimbursement rates for the screenings and cancer care.

Results of this study showed that although fecal occult blood testing was the least expensive of the three methods, it saved fewer life-years than colonoscopy. Flexible sigmoidoscopy was also found to be less cost-effective than fecal occult blood testing or colonoscopy. Low patient compliance with annual fecal occult blood testing also contributed to colonoscopy every ten years being a more cost-effective screening tool.

Statistical analyses of potential risks involved with each method were included as part of the study. Sensitivity of each method was also taken into consideration. Although colonoscopy was associated with higher rates of complications (perforation and bleeding) than the other two methods, it was far superior at detecting and removing polyps. Even when the cost of colonoscopy was taken into account, the study’s authors still concluded that colonoscopy is more cost-effective than flexible sigmoidoscopy or fecal occult blood testing.

This study does have some limitations. It is based on a computer model and uses a Markov process, which seems appropriate for this scenario in determining efficacy in a controlled environment, but the results may not translate to effectiveness within the larger population. The study also did not take into account patients who opted to receive no screening at all. It also used Medicare reimbursement rates to estimate costs, which may be misleading as the authors started following the cohort at age 50, when most participants would have been ineligible for Medicare. Other sources such as Medicaid or private insurance may have had lower or higher reimbursement rates which should have been taken into account when estimating costs. Finally, this study was done in 2000 and is now sixteen years old. It is possible that factors such as cost, sensitivity, and risks of each procedure may have changed since then as technologies evolved. Therefore, I would hesitate to apply this study’s results to the population at large.

Study by Hassan, Rex, Zullo, and Kaminski

A 2015 study by Cesare Hassan, Douglas Rex, Angelo Zullo, and Michal Kaminski used a similar Markov computer model with a hypothetical sample size of 100,000 participants. The objective was to estimate the impact of adenoma detection rate on the long-term colorectal cancer prevention rate. The authors used the computer model to estimate the adenoma detection rates of endoscopists who had a history of low adenoma detection rates versus those who had average or high detection rates. The theoretical cohort consisted of 100,000 American men and women ages 50-80. They received colonoscopies every ten years. As in the previous study, costs were estimated using available Medicare reimbursement data.

A relative risk of 1.5 was assumed between endoscopists with low and average adenoma detection rates, and a relative risk of 11 was assumed between endoscopists with average and high adenoma detection rates. The impacts and costs of providing additional training to the endoscopists who had lower detection rates were also taken into account.

Unlike the previous study, this one did take into account the costs associated with receiving no screening at all (and costs and mortality associated with colorectal cancer). Those who received no screening had no reduction in rate of colorectal cancer detection, while overall rate of colorectal cancer detection in those who received colonoscopies was 75%. This percentage was lower for endoscopists with low adenoma detection rates (70%) and higher for those with average detection rates (71-77%). The best outcome occurred when endoscopists had high adenoma detection rates (92%).

The authors concluded that colonoscopy is a cost-effective method of screening for colorectal cancer, but that its usefulness at preventing colon cancer is significantly impacted by the clinician’s success rate at detecting adenomas. The efficacy of the colonoscopy can vary by 7-21% depending on the endoscopist’s skill level. The authors conclude, therefore, that it is also cost-effective to invest in adequately training endoscopists so that there will be a higher adenoma detection rate overall.

The authors acknowledge that a limitation of the study is that they do not have sufficient data to determine the long-term effect of regular screening colonoscopies (every ten years) by a clinician with a high adenoma detection rate. As in the previous study, I would also add that another limitation could be that this is a computer simulation which may not be applicable to the general population. The research may also be biased as the four authors of the study used their own adenoma detection rates as models for the computer simulation. It is not clear from the study what criteria were used to rate each clinician as “low,” “average,” or “high.” Also, as in the previous study, Medicare reimbursement rates were used to estimate costs. At the time the study took place, most people under age 65 were not eligible for Medicare unless disabled, so this leaves an age gap of ages 50-64 in the study who may have had other sources of payment (self-pay, Medicaid, or private insurance).

Comparison of the Two Studies

Both studies conclude that colonoscopy is a cost-effective method of screening for colorectal cancer. The second study by Hassan, et. al., also concludes that the endoscopist’s rate of adenoma detection is also crucial to the effectiveness of colonoscopy. Both studies used a theoretical sample size of 100,000 using a Markov process via computer simulation. It is difficult to know whether this simulation is representative of the population as a whole. Neither study took into account payer sources other than Medicare, which may have skewed the data on costs. Because the study by Hassan, et. al is more recent than the previous study I would hesitantly be more comfortable in applying the results to the general population.

If I were to design a study examining the efficacy and cost-effectiveness of colonoscopy, I would do a cohort study of participants ages 50 and older. I would continue to follow them until death and compare rates of colon cancer among the groups who had no screening versus those who had colonoscopy, flexible sigmoidoscopy, or fecal occult blood testing. I would consider stratifying the data by payer source such as Medicare, Medicaid, private insurance, or self-pay in order to get a more accurate measure of cost. I would collect the data via medical records and hospital bills, including bills for cancer care. I would also adjust for possible risk factors which may contribute to development of colorectal cancer. The disadvantage to this type of cohort study is that it might take a long time to get results since we are using actual patients in the study instead of a computer simulation. Another way to do this might be a retrospective cohort study, although it may be more difficult in this case to ascertain risk factors for colorectal cancer, especially among those patients who are no longer alive. Regardless, I feel that using actual patient data, including reimbursement rates from the patients’ various payer sources, would give more accurate results of cost effectiveness than a computer model.

References
Hassan, C., Rex, D., Zullo, A., & Kaminski, M. (n.d.). Efficacy and cost-effectiveness of screening colonoscopy according to the adenoma detection rate. United European Gastroenterology Journal, 3(2), 200-207.

Sonnenberg, Amnon, Delco, Fabiola, & Inadomi, John M. (2000). Cost-Effectiveness of Colonoscopy in Screening for Colorectal Cancer. Annals of Internal Medicine,133(8), 573.