We found that a multi-modal intervention, which included mailing a patient reminder with a colon cancer decision aid to patients of attending physicians and system changes allowing patients direct access to schedule screening tests, increased cancer screening test completion by 11% (95%; CI 3%;18% p = 0.01) when compared to a control group. We estimated the cost of the intervention to be $94 per additional patient screened. The response to the mailings was modest (42%), and about half who sent a mailed response did not complete the written materials.
Our multi-modal intervention included a reminder letter from their physician, a decision aid to educate and facilitate choice, and system changes to decrease barriers to colon cancer screening test ordering. In a recent review Stone et al.
identified several studies that combined interventions in clinical practices to target colon cancer screening barriers at multiple levels [7
]. However, none included a decision aid as a part of the intervention. We identified a randomized controlled trial performed by Zapka and colleagues that mailed a colon cancer screening educational video to 450 primary care patients before their scheduled appointment for a physical exam. Their study found no significant difference in screening rates between the intervention and control groups [13
]. A potential reason for the discrepancy in findings with our study could be the intent of the video. The video used in our study was aimed at facilitating choice between screening options, whereas the video used in the Zapka et al.
study was focused on increasing screening by flexible sigmoidoscopy. Additionally, our study provided direct access to screening tests, while the Zapka and colleagues' study encouraged discussion with providers and depended on the practice to arrange the screening appointment. Our previous work has shown that screening tests are often not ordered, despite patient interest [12
Recent studies have used a similar mass mailing to patients to attempt to increase colon cancer screening rates with mixed results A study using self-reported screening completion as an outcome showed an increase in colon cancer screening of 23% over one year when fecal occult blood test cards and reminders were sent directly to patients [14
]. Using claims data instead of self-reporting, Walsh and colleagues found no significant increase in colon cancer screening rates for colonoscopy and fecal occult blood tests after mailing patient educational materials and a letter encouraging screening; however, a small (3%) increase in flexible sigmoidoscopy was observed [15
]. The 11% increase we found in screening rates could be attributable to the decision aid that differed from the educational materials sent by Walsh and colleagues, the system changes we implemented, or a combination of both.
Evaluating the costs of cancer screening programs is important in deciding which programs are most effective and efficient at promoting the targeted screening activity [16
]. Most assessments of screening program costs have been conducted for breast cancer screening programs directed to patients [17
]. The effectiveness and efficiency of mammography screening programs varies widely depending on the baseline screening rates, the target population, the intensity of the intervention, and the method for calculating costs. A low-intensity intervention similar to our study that included a tailored letter and telephone call to patients in a large HMO increased screening rates by 20%. This intervention had a relatively high estimated cost of additional patient screened of $818 [18
]. Others have estimated costs for comparable programs to be much lower. Fishman et al.
found that a telephone reminder increased screening by 16%, with an estimated cost of $92 for previously adherent women and $100 for those previously non-adherent [19
]. Saywell targeted non-adherent women at a large HMO using telephone counseling and a physician reminder letter, which was shown to increase screening by 17% at an estimated a cost of $14 per additional patient screened [20
]. Our estimated costs of $95 per additional patient screened compares favorably with the breast cancer screening programs because colon cancer screening may be less acceptable and is complicated by multiple testing options.
We were able to find only one other study that estimated the cost per additional patient screened for colon cancer screening. This study found comparable results to ours, but was aimed specifically at physicians. The intervention provided VA physicians with quarterly feedback on screening rates of their patients. This intervention increased colon cancer screening rates by 9%, with an associated cost per additional patient screened of $196 [17
Our study has several important limitations. First, our sample size was small, making it possible that we found a difference in screening rates by chance. The relatively tight confidence intervals, however, suggest that this is unlikely. Second, the study included only one academic practice and a sub-sample of attending physicians' patients within the practice, limiting the generalizability of our findings. Third, the study was non-randomized, which could introduce bias if there were unmeasured differences between the intervention and control group patients. Because we had a limited number of patient characteristics available in our database we are not able to exclude differences between the groups as a possible cause for our findings. Although age, gender, and race were similar between the groups, insurance status was not available and could bias the control group to no screening if there were fewer insured patients in the control group. Although we found a difference in screening rates between intervention and control groups, we were unable to identify which part of the intervention was responsible for the difference in screening. Finally, our response rate to the initial mailings and calls was modest, despite five calls per patient. The views of the patients who refused or who we were unable to reach may differ significantly from those who we were able to interview.
A multi-modal intervention, which included mailing a patient reminder with a colon cancer decision aid to patients and system changes allowing patients direct access to schedule screening tests, was effective in increasing cancer screening test completion in a subset of patients within a single academic practice. Although the uptake of the decision aid was low, the cost was also modest, suggesting that this method could be a viable cancer screening program.