Abstract
Context:
The American Cancer Society's 80% by 2018 initiative aims to increase at-home colorectal cancer (CRC) screening.
Objective:
To assess the effectiveness of one-on-one conversations with patients about CRC screening on compliance with at-home fecal immunochemical tests (FITs).
Setting:
Federally funded health care center serving a largely minority, urban, underserved population in Vallejo, California.
Methods:
Patients aged 50 to 75 years were divided into 3 main intervention arms: (1) patients who received a one-on-one in-person consultation to discuss the need for CRC screening and the screening process, as well as a FIT; (2) patients who received a telephone call to discuss the need for CRC screening and the screening process before receiving a FIT in the mail; and (3) patients who were mailed a FIT with a letter explaining the importance of completing the FIT and instructions. The FITs were tracked to see if they were returned to the laboratory for processing, and the rates of return were compared in bivariate analyses using t and χ2 tests and in adjusted analyses using logistic regression with bootstrapping.
Results:
A total of 3415 patients were included in the study. One-on-one conversations either in person (OR, 24.63; 95% CI, 19.28-31.46) or via telephone (OR, 14.74; 95% CI, 10.96-19.82) were more effective at getting patients to complete the at-home CRC screening than not having one-on-one conversations before receiving the FIT.
Conclusion:
Patients may be more likely to complete at-home FITs to screen for CRC if they are first able to discuss the need for screening and the screening process with a health care professional.
Colorectal cancer (CRC) has the second highest mortality rate of all types of cancer, accounting for nearly 50,000 deaths in the United States in 2015.
1 Although the US Preventive Services Task Force recommends that all adults aged 50 to 75 years get screened,
2 28% of adults in this age category have never been screened, and nearly half of all men and women aged 50 to 75 years are overdue for a CRC screening.
3,4 Recognizing the need to increase the use of this life-saving preventive measure, the American Cancer Society and the Centers for Disease Control and Prevention created the National Colorectal Cancer Roundtable, which has the goal of increasing the CRC screening rate to 80% by the year 2018.
5 To reach this goal, the American Cancer Society distributed 80% by 2018 grants to health care centers and clinics to increase screening among their populations and to develop community-appropriate plans to do so. The 3 most common screening methods used to reach this goal are the annual fecal occult blood test, sigmoidoscopy every 5 years, and colonoscopy every 10 years.
2
To increase the CRC screening rate, health care professionals may offer the more cost-effective and convenient at-home fecal immunochemical test (FIT).
6 As new mass communication technologies become facilitated by electronic health records, and with the availability of at-home screening tests, it is necessary to evaluate which methods of outreach yield the best results in preventive health care maintenance. The purpose of this study was to assess the effectiveness of various CRC screening outreach methods to increase CRC screening in a racially and ethnically diverse patient population. We hypothesized that patients who received information on the importance of the test and the test process in a one-on-one fashion would be more likely to complete the screening compared with those who received the FIT without having any conversation.
Vallejo, CA, in Solano County, is one of the most racially/ethnically diverse cities in the state
7 and has undergone significant financial hardships over the years.
8 The city, which is home to 2 of La Clinica de la Raza's Bay Area clinics, has a population of 22.1% African American, 22.6% Hispanic, 24.9% Asian, and 32.8% white people.
9 In 2014, La Clinica de la Raza received a 1-year grant from the American Cancer Society's 80% by 2018 initiative to increase the rates of CRC screening among their client population in Solano County. In fall 2014, 2 Touro University California osteopathic medical students (T.J.F. and M.G.B.), under the auspices of the 2015-2016 Bay Area Albert Schweitzer Fellowship, worked in collaboration with La Clinica de la Raza's panel managers to develop a protocol for patient outreach. Previous studies have shown that CRC disproportionately affects minorities, particularly Hispanics and blacks,
10 as well as groups of low socioeconomic status. This setting, therefore, afforded an opportunity to engage a community that has a high number of people at increased risk of CRC.
Owing to the user-friendly nature of at-home fecal occult blood stool sample collection tests,
11 La Clinica de la Raza activities to reach the grant benchmarks included generating a list of patients due for a CRC screening according to an electronic medical record (EMR) search and subsequently mailing a FIT to 2133 patients from November 2014 through October 2015. Lists of patients potentially eligible for the CRC screening interventions were generated by panel managers using I2I and NextGen software. Initially, 2 inclusion criteria were used to identify patients who needed annual CRC screening: (1) age between 50 and 75 years and (2) not having completed a FIT or fecal occult blood test (FOBT) within the past year. Patients were identified as eligible for screening in November 2014. After a few months, it was discovered that using the initial 2 inclusion criteria was not sufficiently approximating the true target population; many patients were misclassified as needing screening when they did not. Before launching this CRC screening intervention, additional exclusion criteria were identified: history of colorectal cancer, adenomatous polyps, or inflammatory bowel disease; a family history of colorectal cancer; a normal colonoscopy result in the past 10 years; a normal flexible sigmoidoscopy result in the past 5 years; or a normal FIT result in the past year. (These patients were being monitored under other screening protocols completely separate from this intervention.)
Phone conversations and messages took place between August 2015 and February 2016. There was a 2-month period of transition between the mass mailing and telephone conversation protocol because the clinic personnel had to be retrained on the new exclusion criteria and telephone call intervention. In-person clinic appointments spanned the entire study period from November 2014 through February 2016 (
Figure 1). The protocol was designed based on the hypothesis that having a one-on-one conversation with patients about the need for CRC screening, the screening process, as well as answering patient questions by certified medical assistants, can increase the rate of FIT completion among the populations at greatest risk for the disease over the mass-mailing strategy that was being used. Osteopathic principles were applied to promote preventive screening; our approach was to integrate patients as partners in their own care through personal conversations, education, and motivation that included eliciting commitment to complete the screen, thereby increasing the engagement between the clinic and its patients. Mailings, conversations, and outreach attempts were all recorded in patients’ EMRs.
The 2 medical students conducted a medical record review using patient EMRs to assess which La Clinica de la Raza patients received a FIT between November 1, 2014, and February 29, 2016. The data collected included how they were informed by La Clinica de la Raza of the need to complete a FIT (in person, by telephone, by mail, or some combination), whether they had another contact with the clinic via the aforementioned methods, and whether they returned the FIT to the laboratory for processing, thereby completing the screening protocol. The medical students were trained in La Clinica's standard EMR and Health Insurance Portability and Accountability Act regulations before accessing any patient data, and the study was approved by the Touro University California Institutional Review Board.
Patients were exposed to a major intervention arm determined by the mode through which the patient was first contacted: in-clinic appointment, telephone conversation or message, or mass mailing. Additionally, we recorded whether patients were exposed to additional interventions or a combination of the arms as seen in
Figure 1; thus, there were 3 main intervention arms and 10 sub-intervention arms. Due to the small sample size of some of the sub-intervention arms, statistical analyses were performed comparing the 3 major intervention arms in this study. Patients were exposed to a main intervention arm determined by highest level of contact they had with the clinic: in-clinic appointment, telephone conversation or message with either medical student, or mass mailing. There was no washout period between each intervention.
Patients who were eligible for preventive CRC screening and had an appointment with their primary care physician engaged in a one-on-one conversation and received a FIT in person.
Patients under this arm had their EMR audited by 1 of 2 medical students to ensure eligibility according to the February 2015 exclusion criteria (
Figure 2). Outreach via telephone in English or in Spanish was attempted for eligible patients. Results of the telephone calls included (a) personal contact was made with a thorough conversation about the need for screening, instructions for use of the FIT, answers to patient questions, commitment by patient to complete the screening, and subsequent mailing of a FIT to the patient; (b) personal contact could not be made and an educational message was left informing patients that they were due for a simple, noninvasive screening for CRC; or (c) no contact could be made, and a letter was mailed to the patient's address stating the purpose of the outreach along with contact information.
Patients received a FIT with low-literacy instructions in both English and Spanish, a letter explaining the importance of completing the FIT, and a postage-paid envelope to the laboratory for test processing.
In an effort to keep track of the varied contact between patients and the clinic, it was recorded what levels of contact the patient received between the time the first kit was provided to them and the time the kit was received by the lab. Both the highest level of contact with a patient and the mode of contact were tracked. Modes of contact tracked included if they received a letter, had phone counseling, a personalized phone message left by one of the medical students, or came into clinic. Not all sub-intervention categories were analyzed in the regression model as some of the combinations had small populations, but the highest level of care was analyzed in the regression.
Bivariate analysis with
t tests for age, a continuous variable, and χ
2 tests for categorical variables were used to assess variations between all patients and those who returned the FIT according to patient characteristics (age, sex, race/ethnicity) and intervention arm. Statistical significance was defined as
P<.05. Logistic regression on all 3415 study participants with 1000 bootstrapped replications, deemed adequate for bias-correction estimates of CIs,
12 was then used to model the association between FIT return rate and intervention arms, controlling for patient characteristics. Analyses were performed using STATA version 13.0 (Stata Corp).
In November 2014, 4280 patients were identified as eligible for screening. The added exclusion criteria resulted in the exclusion of 865 patients (20% of the initial patient list) who had been misclassified. The final study sample was 3415.
Table 1 shows the overall distribution of patient characteristics and intervention arms according to all participants and stratified by participants according to FIT return status. Of the 3415 study patients, 1456 patients (42.6%) returned a FIT. There were 1239 participants (36.2%) who were exposed to some combination of interventions compared with 2176 participants (63.7%) who experienced 1 intervention. Bivariate analyses, comparing patients according to FIT return status, revealed that more women (
P=.039) and more non-English speakers (
P=.001) returned their FITs. Additionally, variations in FIT return status were found according to intervention arm (
P<.001). A majority of FITs returned came from patients who had an in-person consultation in the clinic at some point (1167). Of that group, most only had the in-clinic consultation (848) and no other intervention. The remaining kits returned by patients who had been in the clinic also had additional interventions such as a mass mailed letter (179), both a letter and phone consultation (81), or a letter and phone message left by either medical student (13), and some patients received all the interventions before completing the test (46). Associations between the main intervention arm and FIT return status found in bivariate analysis persisted in the multivariate logistic regression analysis controlling for patient characteristics, as presented in
Table 2.
The most effective mode of outreach for CRC screening was an in-clinic appointment, during which the patient had an in-person conversation with a clinician about the FIT and received the FIT during the appointment (OR, 24.63; 95% CI, 19.28-31.46), as shown in
Table 2. The second most effective mode of outreach for CRC screening, also involving a person-to-person interaction, was the telephone contact with one-on-one conversation or with a recorded message (OR, 14.74; 95% CI, 10.96-19.82).
We found that the FIT return rate was highest among patients who received clinician-to-patient counseling and those who engaged in an educational telephone conversation with or had an educational message left for them. These findings are consistent with the literature, suggesting that an in-person one-on-one conversation about CRC screening is the best way to ensure that FITs are returned to the laboratory.
6,11
The current study provides evidence for an alternative strategy—namely, providing at-home screening, particularly to those who come to the clinic infrequently or with difficulty or who do not have CRC screening conversations during their appointments, can be effectively encouraged to complete the test by one-on-one telephone outreach from a live person. These results suggest that having personal contact and education prior to sending out a FIT is a key component in increasing their rate of return. If patients do not come to the clinic, however, a telephone conversation with a health care professional regarding the value of the screening FIT may be the next best way to get a FIT returned.
The current study's findings are consistent with those of previous studies. Green et al
13 compared usual care to 3 levels of additional intervention. The “automated group” received automatic mailings generated by the EMR. The “automated-assisted group” got additional attention from a medical assistant to encourage screening and identify patient barriers; and the “automated-assisted-navigated group” received guidance from a registered nurse to identify the various methods of screening and the individual's risk for CRC and to develop a patient-specific treatment plan. Their work showed that all intervention groups had a significantly higher return rate than simple mass mailings, but the automated-assisted and automated-assisted-navigated group had the most returns.
13 Mosen et al
14 found that in patients who were overdue for CRC screening, the most effective means to patients completing CRC screening was in the setting of a comprehensive conversation with their provider. This finding reflects the impact that one-on-one conversations can make in establishing patient commitment to CRC screening.
Work by Liles et al
11 determined that patients’ concerns about at-home testing, including perceived unpleasantness of the test, ease of completion, and convenience were major barriers to CRC screening completion. When FIT and FOBT were directly compared, patients reported less perceived unpleasantness and greater ease and convenience with the FIT testing, so much so that switching from FOBT to the FIT alone would lead to an increase in screening completion. Thus, our study design of using the FIT took advantage of the innate user-friendly aspects of the FIT over FOBT. Our design also took advantage of findings by Mosen et al,
14 who found that in the setting of patient-centered conversations, more questions can be answered and fears can be allayed, thus increasing patient compliance. The protocol of conversation and motivation enabled the medical students to answer patient questions about the FIT, its use, and its significance, thereby combining the findings of the 2 studies.
Additionally, De Jesus et al
15 highlighted the importance of social fluency and competency in health care professionals’ discussion of CRC screening with patients in low-income urban settings. An understanding of a patient's social context was significantly associated with completion of CRC screening, independent of recommendation for screening. Patients who were provided one-on-one conversations, either in the clinic or over the telephone, during which personal beliefs could be addressed, individual questions answered, and reassurance given, had higher rates of return than mass mailing. Furthermore, our medical students were able to address patients in either English or Spanish, and one of the panel managers spoke Tagalog, which made culturally sensitive and appropriate communication possible for the majority of patients.
Given the small sample size of some of the sub-intervention arms, the nuances of these interventions were too difficult to assess statistically in the regression model, especially after we controlled for covariates, which further stratified the sample. With larger samples in the various sub-intervention arms, we would be able to further assess the nuances of how the 3 main methods of outreach and follow-up interacted with each other to influence FIT return and is an area of future study.
Our study had limitations. Patients were called on Tuesday, Wednesday, Thursday, and Friday mornings, missing patients who did not respond to their telephones during these early hours. Telephone calls were made in either English or Spanish; therefore patients who spoke other languages did not receive a telephone call in their preferred language. Additionally, contact with homeless patients was primarily limited to in-clinic visits. As a result, this high-risk group may have been underrepresented in the overall analyses. Limitations to generalizability include the resources required. Nonphysician staff thoroughly reviewed EMRs and telephoned patients, which amounted to nearly 300 hours during the study.
In-person interactions between patients and health care professionals were the most effective way of guaranteeing FIT returns, followed by one-on-one contact by telephone. This study suggests that if clinics that serve populations comprised of people at high risk for CRC in urban, resource-poor communities engage their patients in one-on-one conversation about screening, they will be more successful in getting their patients screened and ultimately preventing CRC. Future areas of study may include cost comparison of automated telephone calls vs personal telephone calls in CRC and other preventive screening tests. Such studies could help community clinics to understand how best practices from large health maintenance organizations, which rely on mass mailing and automated telephone calls, can be scaled in a low-resource setting.
All authors provided substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; all authors drafted the article or revised it critically for important intellectual content; all authors gave final approval of the version of the article to be published; and all authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.