# Cost Effectiveness of Interventions to Promote Screening for Colorectal Cancer: A Randomized Trial

## Article information

J Prev Med Public Health. 2011;44(3):101-110
Publication date (electronic) : 2010 May 17
doi : https://doi.org/10.3961/jpmph.2011.44.3.101
1School of Public Health, University of Texas Health Science Center at Houston, Houston, USA.
2Kelsey Research Foundation, Houston, USA.
3Washington University School of Medicine, St. Louis, USA.
Corresponding author: David R. Lairson, PhD. 1200 Herman Pressler Dr., Houston, TX 77030 Suite RAS-E307, USA. Tel: +1-713-500-9176, Fax: +1-713-500-9171, David.R.Lairson@uth.tmc.edu

## Abstract

### Objectives

Screening for colorectal cancer is considered cost effective, but is underutilized in the U.S. Information on the efficiency of "tailored interventions" to promote colorectal cancer screening in primary care settings is limited. The paper reports the results of a cost effectiveness analysis that compared a survey-only control group to a Centers for Disease Control (CDC) web-based intervention (screen for life) and to a tailored interactive computer-based intervention.

### Methods

A randomized controlled trial of people 50 and over, was conducted to test the interventions. The sample was 1224 partcipants 50-70 years of age, recruited from Kelsey-Seybold Clinic, a large multi-specialty clinic in Houston, Texas. Screening status was obtained by medical chart review after a 12-month follow-up period. An "intention to treat" analysis and micro costing from the patient and provider perspectives were used to estimate the costs and effects. Analysis of statistical uncertainty was conducted using nonparametric bootstrapping.

### I. Costs and Cost-Effectiveness

Table 1 presents a breakdown of the cost per person for administering the web-based educational intervention and the tailored, interactive intervention.

Cost of intervention by activity

The average cost was $39.82 and the standard deviation was$6.45 per person for implementing the web-based intervention. The average cost per person for implementing the tailored intervention was $44.87 and the standard deviation was$6.69 per person. Patients were not randomized to intervention or control groups until after they completed the baseline survey. Thus any differences in costs to recruit people were not due to any known/systematic differences between groups. The variability in cost was primarily due to the time cost of both patients and staff during the intervention delivery; more time was required to complete the tailored intervention. For the tailored intervention, about 56% of the direct cost was for intervention delivery, compared to 51% for the web-based education intervention. The remainder of direct cost was primarily for recruitment in each intervention. Estimates of the base case incremental cost-effectiveness analysis are presented in Table 2.

Incremental cost effectiveness (cost per additional individual screened)

The screening compliance in the control group was 33.9% compared with 35.4% in the web-based group and 32.0% in the tailored group. The randomized trial did not yield statistically significant differences in screening rates [13]. For the economic evaluation, the point estimates represent the "best" available estimate of program effects and costs. The mean ICER was $2602 moving from no intervention to the web-based intervention, whereas the tailored intervention had a negative mean ICER compared to both the control group and the website intervention, due to negative effect. The tailored intervention was dominated by the website intervention as the tailored intervention was more costly and less effective. Tables A1, A2 and A3 present the results of the sensitivity analyses. The average cost and the incremental cost effectiveness ratios showed relatively small changes to alternative assumptions regarding the inclusion of surveys and the overhead rate. The ICER comparing the web-based intervention to the control group declined by 30% when the percent time the HIC personnel devoted to the intervention was decreased from 100% to 25%. However, the cost per additional person screened remained above$1800.

### II. Results of Uncertainty Analysis

The joint density of ΔC (difference in cost) and ΔE (difference in effect) when comparing the survey-only control group to the web-based intervention along with the 95% Confidence Interval (CI) for ICER is summarized in Figure 2. The joint density of ΔC and ΔE when comparing the web-based intervention to the tailored intervention along with the 95% CI for ICER is summarized in Figure 3.

ICER confidence intervals with joint density of ΔC and ΔE comparing no intervention control group to web-based intervention group in promoting CRCS.

ICER: incremental cost-effectiveness ratios, CRCS: cases of colorectal cancer screening, CI: confidence interval.

ICER confidence intervals with joint density of ΔC and ΔE comparing web-based intervention group to tailored intervention in promoting CRCS.

ICER : incremental cost-effectiveness ratios, CRCS : cases of colorectal cancer screening, CI : confidence interval.

The bootstrapped mean differences for cost and effectiveness per participant presented in Figure 2 imply that the web-based intervention has an added cost with a positive effect compared with no intervention as 68.0% of the bootstrapped replicates fall in the NE quadrant of the cost effectiveness plane. Decision makers may consider implementing this intervention depending on how much they are willing to pay per additional person screened for CRC and their assessment of the chance that the intervention could result in lower screening compliance than the no intervention alternative. Implementing a tailored intervention compared to the web-based education has a negative effect and an added cost as most of the bootstrapped joint density for cost and effect differences per participant fall in the NW quadrant of the CE plane. The results therefore do not provide support for disseminating this type of tailored intervention.

## DISCUSSION

The results of this large, carefully executed trial of a tailored intervention in a well-organized multi-specialty group practice raise concerns about the efficiency of this resource-intensive approach to CRCS promotion. The tailored computer based intervention was dominated (more costly and less effective) than a simpler web-based education intervention. The web-based education intervention was relatively costly compared to other mail and system reminder based interventions in primary care settings.

The results are consistent with the general findings from a systematic review of patient directed tailored intervention studies [23]. For example, Myers et al. [24] found that implementation of targeted and tailored interventions for CRCS promotion improved screening compliance compared to regular care or no intervention. However, the value of tailoring interventions was questioned when compliance was shown to be slightly lower than for the interventions involving mail and telephone reminders. Ling et al. [25] presented results of a randomized trial comparing tailored vs. non-tailored physician recommendation letters and enhanced vs. non-enhanced physician office and patient management systems to improve screening rates for CRC in a population of eligible patients aged 50-79 years in 10 primary care physician office practices in Philadelphia. At one year follow-up, a flexible sigmoidoscopy or colonoscopy was obtained by 54.2% of participants in the non-tailored letter, enhanced management group, 53.3% in the tailored-letter, enhanced management group, 43.6% in the tailored-letter, non-enhanced management group and 37.6% in the non-tailored letter, non-enhanced management group. Thus, in the enhanced management group the tailored interventions showed a smaller impact on CRCS compared to providing non-tailored information on CRCS, similar to our findings.

Our web-based intervention was less cost-effective compared to previous studies of CRCS promotion. In a study by Chirikos et al. [10] the authors presented results of a low-cost inreach intervention to improve mammography rates, FOBT compliance and pap testing compliance in a low-income population of eligible patients aged 50-75 years in a clinical setting. Cancer screening office systems, a reminder system for physicians to check the current screening status of eligible patients and ensure completion of timely screening, was implemented in primary care clinics in Florida. The ICER results for an FOBT was $12 and the compliance rate was 40% for the intervention group, suggesting that the method of a non-computerized reminder system utilizing the available clinic personnel was cost effective for improving cancer screening compliance in a primary care setting. However, participants were low income and the control group screening rate was 12%. Our participants were middle class, insured, with a 30 percent screening rate in the control group. A more recent study evaluated the cost-effectiveness of an intervention among patients of a multi-specialty group practice in eastern Massachusetts [26]. The intervention included a tailored letter, educational brochure, a fecal occult blood test kit with a stamped return envelope, and a phone number to a dedicated line for scheduling a sigmoidosopy or colonoscopy. The baseline screening rate was 63%, the increase in the screening rate was 5.8% and the incremental cost per additional person screened was$94. Costs may be low due to a fully operational electronic medical record system, no mention of overhead costs, and low cost for the FOBT kits. While letters were tailored, there was no discussion of the degree or cost of the tailoring. With multiple program elements, it was not possible to determine the effectivness of the tailoring component.

The literature on cancer screening promotion supports implementing some form of intervention, whether a targeted intervention with mail or phone reminders or a tailored intervention or generic print based information about cancer screening to patients, or reminder systems to physicians to positively affect screening compliance among the target population of 50-70 years when compared to no intervention [9-12,19,24,25]. However, our findings and those of studies described above, raise questions about the information provided to participants about CRCS. One possible explanation of the limited effect of the interventions is that the baseline survey about colorectal cancer and screening may have cued the control group to action [27]. However, this would not explain why the tailored intervention performed worse than the web-based intervention. Participants spent an average of 23 minutes on the tailored intervention compared to 17 minutes on the web-based program [13]. Perhaps, the more extensive individualized information about cancer and screening raises apprehension about screening, which could result in lower compliance. Also, the role of physicians is very important for screening compliance [28]. Although the physicians in the KSC group practice were informed about the study and the importance of CRCS and patients were encouraged to discuss CRCS with their physician, including doctors more directly in the intervention design may improve screening compliance. A qualitative analysis of a subset of patients' discussions with providers about CRCS in the current study found that providers focused on colonoscopy, which reduced discussion of patient's test preferences and conflicted with our interventions' focus on choice between multiple test options [29].

Promoting CRCS via website interventions may be a suitable strategy to improve screening rates compared to similar print materials. However, the impact of the website intervention on screening compliance was minimal and the cost per additional person screened was high compared to other studies. The recruitment cost may be reduced with an automated system to determine eligibility for screening. The study relied on staff making multiple phone contacts to determine eligibility from a list of potentially eligible patients. The research context of the intervention contributed to the cost and therefore the cost estimates represent an upper bound of feasible cost in a practice setting. However, KSC was well organized and had partially automated patient records and appointment schedules that facilitated the study.

The interventions were tested in a single multispecialty clinic thereby limiting the generalizability of the results. Time estimates for the project personnel were self-reported, but time logs were completed on a weekly or monthly basis throughout the trial. Overhead cost was calculated as a percent of the direct cost instead of exact measurement. However, analysis of uncertainty showed that a reasonable range of overhead costs had no substantive effect on the overall findings. While the randomized trial did not yield statistically significant results, the point estimates along with an analysis of uncertainty, provide valuable information for decision-making, which is not about testing hypotheses but about using the best available estimates of program effect and cost to inform decisions about resource allocation [30].

More research is required to understand the cost and the effectiveness of tailoring and other methods to motivate patients to obtain recommended CRCS tests. This requires an assessment of system, physician and patient-centered efforts to educate patients and providers and to address any perceived barriers to screening [28]. Costs should be fully assessed, including the cost of planning and recruitment of eligible participants, overhead costs, and the cost of developing information for tailoring.

## Notes

The authors have no conflicts of interest with the material presented in this paper.

## Appendices

Supplementary material

Sensitivity analysis by varying inclusion of surveys

Sensitivity analysis by varying inclusion of follow-up survey and overhead rate

Sensitivity analysis by varying the percentage of HIC personnel cost

## References

1. Colon and rectal cancer. National Cancer Institute cited 2011 Feb 17. Available from: http://www.cancer.gov/cancertopics/types/colon-and-rectal.
2. Seifeldin R, Hantsch JJ. The economic burden associated with colon cancer in the United States. Clin Ther 1999;21(8):1370–1379. 10485508.
3. Smith RA, von Eschenbach AC, Wender R, Levin B, Byers T, Rothenberger D, et al. American Cancer Society guidelines for the early detection of cancer: update of early detection guidelines for prostate, colorectal, and endometrial cancers. Also: update 2001-testing for early lung cancer detection. CA Cancer J Clin 2001;51(1):38–75. 11577479.
4. Levin B, Lieberman DA, McFarland B, Smith RA, Brooks D, Andrews KS, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. CA Cancer J Clin 2008;58(3):130–160. 18322143.
5. Screening for colorectal cancer. US Preventive Services Task Force cited 2011 Feb 17. Available from: http://www.uspreventiveservicestaskforce.org/uspstf/uspscolo.htm.
6. Colorectal cancer screening tests. Centers for Disease Control and Prevention (CDC) cited 2011 Feb 17. Available from: http://www.cdc.gov/cancer/colorectal/basic_info/screening/tests.htm.
7. Healthy people 2010 targets. Centers for Disease Control and Prevention (CDC) cited 2011 Feb 17. Available from: http://apps.nccd.cdc.gov/SAHA/Default/HealthyTargets.aspx.
8. Transtheoretical model: detailed overview of the transtheoretical model. Cancer Prevention Research Center cited 2011 Feb 17. Available from: http://www.uri.edu/research/cprc/TTM/detailedoverview.htm.
9. Sequist TD, Zaslavsky AM, Marshall R, Fletcher RH, Ayanian JZ. Patient and physician reminders to promote colorectal cancer screening: a randomized controlled trial. Arch Intern Med 2009;169(4):364–371. 19237720.
10. Chirikos TN, Christman LK, Hunter S, Roetzheim RG. Cost-effectiveness of an intervention to increase cancer screening in primary care settings. Prev Med 2004;39(2):230–238. 15226030.
11. Lairson DR, DiCarlo M, Myers RE, Wolf T, Cocroft J, Sifri R, et al. Cost-effectiveness of targeted and tailored interventions on colorectal cancer screening use. Cancer 2008;112(4):779–788. 18098272.
12. Andersen MR, Urban N, Ramsey S, Briss PA. Examining the cost-effectiveness of cancer screening promotion. Cancer 2004;101(5 Suppl):1229–1238. 15316909.
13. Vernon SW, Bartholomew LK, McQueen A, Bettencourt JL, Greisinger A, Coan SP, et al. A randomized controlled trial of a tailored interactive computer intervention to promote colorectal cancer screening: sometimes more is just the same. Ann Behav Med DOI: 10.1007/s12160-010-9258-5.
14. Bartholomew LK, Parcel GS, Kok G. Intervention mapping: a process for developing theory and evidence-based health education programs. Health Educ Behav 1998;25(5):545–563. 9768376.
15. Prochaska JO, DiClemente CC, Velicer WF, Rossi JS. Standardized, individualized, interactive, and personalized self-help programs for smoking cessation. Health Psychol 1993;12(5):399–405. 8223364.
16. Velicer WF, Prochaska JO, Bellis JM, DiClemente CC, Rossi JS, Fava JL, et al. An expert system intervention for smoking cessation. Addict Behav 1993;18(3):269–290. 8342440.
17. Lairson DR, Chang YC, Bettencourt JL, Vernon SW, Greisinger A. Estimating development cost for a tailored interactive computer program to enhance colorectal cancer screening compliance. J Am Med Inform Assoc 2006;13(5):476–484. 16799126.
18. Colorectal (colon) cancer: screen for life print materials. Centers for Disease Control and Prevention (CDC) cited 2011 Feb 17. Available from: http://www.cdc.gov/cancer/colorectal/sfl/print_materials.htm#factsheets.
19. Andersen MR, Hager M, Su C, Urban N. Analysis of the cost-effectiveness of mammography promotion by volunteers in rural communities. Health Educ Behav 2002;29(6):755–770. 12456132.
20. Drummond MF, Sculpher MJ, Torrance GW, O'brien BJ, Stoddart GL. Methods for the economic evaluation of health care programmes 2005. New York: Oxford University Press.
21. Briggs AH, Wonderling DE, Mooney CZ. Pulling cost-effectiveness analysis up by its bootstraps: a non-parametric approach to confidence interval estimation. Health Econ 1997;6(4):327–340. 9285227.
22. Briggs A, Clark T, Wolstenholme J, Clarke P. Missing... presumed at random: cost-analysis of incomplete data. Health Econ 2003;12(5):377–392. 12720255.
23. Albada A, Ausems MG, Bensing JM, van Dulmen S. Tailored information about cancer risk and screening: a systematic review. Patient Educ Couns 2009;77(2):155–171. 19376676.
24. Myers RE, Sifri R, Hyslop T, Rosenthal M, Vernon SW, Cocroft J, et al. A randomized controlled trial of the impact of targeted and tailored interventions on colorectal cancer screening. Cancer 2007;110(9):2083–2091. 17893869.
25. Ling BS, Schoen RE, Trauth JM, Wahed AS, Eury T, Simak DM, et al. Physicians encouraging colorectal screening: a randomized controlled trial of enhanced office and patient management on compliance with colorectal cancer screening. Arch Intern Med 2009;169(1):47–55. 19139323.
26. Sequist TD, Franz C, Ayanian JZ. Cost-effectiveness of patient mailings to promote colorectal cancer screening. Med Care 2010;48(6):553–557. 20473196.
27. Godin G, Sheeran P, Conner M, Germain M. Asking questions changes behavior: mere measurement effects on frequency of blood donation. Health Psychol 2008;27(2):179–184. 18377136.
28. Sarfaty M, Wender R. How to increase colorectal cancer screening rates in practice. CA Cancer J Clin 2007;57(6):354–366. 17989130.
29. McQueen A, Bartholomew LK, Greisinger AJ, Medina GG, Hawley ST, Haidet P, et al. Behind closed doors: physician-patient discussions about colorectal cancer screening. J Gen Intern Med 2009;24(11):1228–1235. 19763699.
30. Briggs AH, O'Brien BJ. The death of cost-minimization analysis? Health Econ 2001;10(2):179–184. 11252048.

## Article information Continued

### Appendix.

Supplementary material

 Bootstrapping using STATA was done in the following manner: 1. A “set seed” command was used for random number generation, which assured reproduction of results. 2. All 413 observations in the tailored computer intervention group and 398 observations in the web based intervention group were chosen to run the bootstrapping and the number of replications was set to 1000. 3. Then bootstrapping was done using the following command: 4. Bs (location: mean=r(mean)), reps (1000) saving (Group 1, replace): summarize total cost, detail 5. Then, the difference in cost (ΔC) and difference in screening compliance (ΔE) were calculated by comparing control group bootstrapped data with the web based intervention group and the tailored computer based intervention group. 6. Incremental cost effectiveness ratios were calculated using the expression ΔC/ΔE. 7. Scatter plot was produced by plotting the cost and effect difference pairs for tailored and control groups to show the distribution of incremental cost- effectiveness ratios. The same were done for web and control groups, and tailored and web groups. 8. To calculate the 95% confidence limits, we identified 25 cases from the worst and best cases of ICERs. To draw 95% confidence limit lines on the scatter plot, we identified the 26th case from both the worst and best cases of ICERs, and drew the lines by connecting 0 and (ΔC, ΔE). The same calculation was done for the base case (point estimate of ICER). This procedure was followed for the different intervention groups.

ICER : incremental cost-effectiveness ratios.

### Table A1

Sensitivity analysis by varying inclusion of surveys

Scenarios Web-based group average cost ($) Web vs. control ICER ($)
Base case (include 6 month survey cost) 39.82 2602
Worst case (include 6 month and baseline survey costs) 42.89 2803
Best case (exclude 6 month and baseline survey costs) 37.33 2439

ICER : incremental cost-effectiveness ratios.

### Table A2

Sensitivity analysis by varying inclusion of follow-up survey and overhead rate

Scenarios Web-based group average cost ($) Web vs. control ICER ($)
Base case (include 6 month survey cost, 35% overhead) 39.82 2602
Worst case (include 6 month survey cost, 40% overhead) 41.29 2698
Best case (exclude 6 month survey cost, 30% overhead) 35.11 2294

ICER : incremental cost-effectiveness ratios.

### Table A3

Sensitivity analysis by varying the percentage of HIC personnel cost

Intervention Cost ($) Incremental cost ($) Effect (% screened) Incremental effect (%) CE ($) ICER ($)
25% HIC cost
Control 0 - 0.3390 - - -
Web-based 28.25 28.25 0.3543 0.0153 79.73 1846.40
Tailored 30.74 2.49 0.3204 - 0.0339 95.94 Dominated
50% HIC cost
Control 0 - 0.3390 - - -
Web-based 32.11 32.11 0.3543 0.0153 90.63 2098.69
Tailored 35.45 3.34 0.3204 - 0.0339 110.64 Dominated
75% HIC cost
Control 0 - 0.3390 - - -
Web-based 35.96 35.96 0.3543 0.0153 101.49 2350.33
Tailored 40.16 4.20 0.3204 - 0.0339 125.34 Dominated
100% HIC cost
Control 0 - 0.3390 - - -
Web-based 39.82 39.82 0.3543 0.0153 112.39 2602.6
Tailored 44.87 5.05 0.3204 - 0.0339 140.04 Dominated

HIC : health information center, ICER : incremental cost-effectiveness ratios, CE : cost-effectiveness.

### Figure 1

Project PCCaSO intervention process flowchart.

HIC: health information center, CRC: colorectal cancer, HIC: health information center,HIPAA: The Health Insurance Portability and Accountability Act, KSC: Kelsey- Seybold Clinic, PCP: primary care physician, CRCS: cases of colorectal cancer screening.

### Figure 2

ICER confidence intervals with joint density of ΔC and ΔE comparing no intervention control group to web-based intervention group in promoting CRCS.

ICER: incremental cost-effectiveness ratios, CRCS: cases of colorectal cancer screening, CI: confidence interval.

### Figure 3

ICER confidence intervals with joint density of ΔC and ΔE comparing web-based intervention group to tailored intervention in promoting CRCS.

ICER : incremental cost-effectiveness ratios, CRCS : cases of colorectal cancer screening, CI : confidence interval.

### Table 1.

Cost of intervention by activity

Program activity Tailored intervention cost ($) Webbased intervention cost ($)
Step 1: Recruitment
(Identification of eligible participants from KRF database, recruitment letter development time cost- personnel, Recruitment letter mailing time cost- personnel, enrollment phone call)
Total recruitment time cost 4728.85 4557.10
Recruitment material cost for letter supplies 235.41 226.86
Total Step 1 4964.26 4783.96
Step 2: Intervention delivery
HIC visit time cost- personnel 5764.85 4548.32
Tailored program viewing time costparticipant 1946.35 0.0
Web-based program viewing time costparticipant 0.0 1416.61
Total step 2 7711.2 5964.93
Step 3: Follow-up
6 month phone survey time cost-personnel 598.85 577.10
6 month phone survey time cost- participant 432.64 413.58
Total step 3 1031.49 990.68
Total direct cost 13 726.28 11 739.57
Overhead: utilities, equipment, computers, database management, facility, office space. 4804.20 4108.85