Efficacy of a Telehealth Intervention on Colonoscopy Uptake When Cost Is a Barrier: The Family CARE Cluster Randomized Controlled Trial

This article is featured in Highlights of This Issue, p. 1299

Screening is essential for the prevention and control of colorectal cancer (1–3), especially for people who are at increased familial risk of developing the disease (4). Despite an “A” rating from the United States Preventive Services Task Force (5), colonoscopies are underutilized among individuals who are considered at increased risk (6, 7). Approximately 40% of relatives of colorectal cancer patients adhere to screening guidelines (7, 8). Promoting colonoscopy use among family members of colorectal cancer patients is one efficient way to reduce colorectal cancer morbidity (9). However, increasing adherence to risk-based screening guidelines is a complex problem that requires addressing both structural (e.g., cost) and personal (e.g., motivation) barriers to screening (10, 11).

For many people, cost is a major factor in choosing whether or not to obtain a colonoscopy (12–15). According to the National Health Interview Survey (2010), only 21% of uninsured people in the United States are up to date with colorectal cancer screening—a stark contrast to the 59% of insured people who are up to date (8, 15, 16). Although the Patient Protection and Affordable Care Act eliminates screening colonoscopy copayments for those with private insurance or Medicare, individual states can choose whether to offer no-cost colorectal screening for patients with Medicaid (16). State laws that mandate colorectal cancer screening coverage through insurance vary in the amount of coverage required for different screening services (17). State laws cannot require coverage from insurance plans that are self-funded by the employer and the Affordable Care Act does not apply to health plans that were in place before it was passed (17). Thus, cost is likely to continue to be a barrier to colonoscopy use.

Even when cost is not a barrier, colonoscopy screening uptake remains suboptimal (18), underscoring the need for behavioral interventions. Although behavioral interventions vary in their delivery method and target populations, most are based on assumptions that patients should be educated about colorectal cancer screening, motivated to act, and assisted with overcoming barriers (16, 19–21). Most behavioral interventions have modest effects on colonoscopy uptake; however, few trials have examined how well intervention effects hold across population subgroups such as those with cost barriers (21, 22). The Family CARE trial took a pragmatic approach of targeting relatives of colorectal cancer patients to increase colonoscopy through evidenced-based communications strategies which addressed perception, barriers, motivation, and volition (23). Those in the intervention arm were nearly three times as likely to obtain a colonoscopy at 9 months postintervention compared with those in the comparison group, with no differences based on rural residency or income (24). Given that cost is a known barrier to colonoscopy use, we sought to determine the intervention effect by the 15-month follow-up when cost barriers to colonoscopy were considered. We also examined the impact of providing resource information for overcoming cost barriers to colonoscopy for those who remained nonadherent at 9 months and identified the time frame in which most participants obtained a colonoscopy after the intervention.

We describe the Family CARE trial in detail elsewhere (23–25). Family CARE was a cluster randomized, two-group trial conducted in the United States that tested the efficacy of a remote, tailored intervention, Tele-Cancer Risk Assessment and Evaluation (TeleCARE), to increase colonoscopy uptake among relatives of colorectal cancer patients who were considered to be at increased familial risk of colorectal cancer (ClinicalTrials.gov #: NCT01274143). Random assignment began in 2009 and ended two years later. Primary outcome assessments (colonoscopy within 9 months following the intervention) was completed in September 2012 and 15-month colonoscopy uptake assessment was completed in April 2013 (23–25).

The Institutional Review Boards of participating institutions approved the trial. All participants provided informed consent.

Primarily, population-based cancer registries (California, Colorado, Idaho, New Mexico, and Utah) identified colorectal cancer patients or their next of kin who were then contacted to request information about their relatives. Relatives were contacted about study participation. Relatives resided in 35 states.

Eligibility criteria included: 30 to 74 years of age, considered at increased familial CRC risk (having a first-degree relative diagnosed with colorectal cancer before age 60 years or one first-degree relative diagnosed at age 60 years or older plus an additional first or second-degree biologic relative diagnosed at any age; refs. 2, 26, 27) for whom screening guidelines recommend colonoscopy, no colonoscopy within the last 5 years, a known family history of colorectal cancer, not a member of family with known hereditary cancer syndrome (28) or a candidate for germline mutation testing, no previous counseling about familial cancer risk or participated in a family cancer trial, and no history of in situ or invasive cancer other than non-melanoma skin cancer. We did not enroll individuals under the age of 30 years to limit the possibility of erroneously enrolling members of families with high penetrance hereditary conditions such as familial adenomatous polyposis or Lynch syndrome. We did not enroll individuals over the age of 74 years as the guidelines for screening at the time of the study did not apply to people 75 years of age or older.

After completing baseline measures, participants were randomized to study arm using a computer-generated allocation algorithm based on randomized block method (four to eight blocks). Family members were assigned to the same study arm to avoid study contamination. Staff collecting baseline assessments were unaware of the identity of a person's participating relatives to prevent them from predicting group assignment.

Participants assigned to the educational brochure group received a brochure that described colorectal cancer, the role of family history in determining risk, and the ability of colonoscopy to prevent and detect cancer early. It encouraged participants to discuss colorectal cancer and colonoscopy with health care providers. The brochure listed colonoscopy as the recommended screening test for their level of familial risk, but encouraged participants to seek some other form of screening such as stool blood testing if colonoscopy was not feasible. The brochure was designed specifically for the target population by the study investigators using investigator expertise and information gleaned from cognitive interviews conducted with colorectal cancer patients and close relatives representing the study's target population. The brochure was further reviewed by the study's community advisory board for acceptability. At the 9-month follow-up, a cost resource letter listing national and state-specific resources for free or reduced-cost colonoscopy was mailed to all participants who had not yet had a colonoscopy and had indicated that cost was a barrier at any of the assessment points. The educational brochure group received this cost resource letter at the 9-month follow-up as the brochure condition did not include communication with a study genetic counselor and therefore did not include a clinical encounter.

TeleCARE is a multifaceted, remote risk-communication intervention. Participants assigned to the TeleCARE group received the same educational brochure plus mailed visual aids tailored from their baseline assessment of likelihood to engage in colorectal cancer screening (i.e., risk perception of cancer, self-efficacy for obtaining colonoscopy). TeleCARE participants discussed this information over the phone with one of five genetic counselors who were trained in cancer risk assessment, behavior theory, and motivational interviewing techniques (29, 30). Participants were mailed a tailored letter within one week after the telephone call that summarized the session and the participant's action plan to obtain a colonoscopy. Participants' health care providers were mailed copies of the letter and family history of cancer if participants consented to it. Consistent with usual care, if participants indicated that cost was a barrier during the telephone call, counselors addressed cost barriers and the same type of cost resource letter that was sent to comparison group participants at 9 months was included in the post-phone call letter. Approximately 6 weeks after the intervention, participants received a reminder card with their personalized action plan for colonoscopy. If at 9 months participants remained nonadherent and indicated that cost was a barrier at a previous assessment point, a cost resource letter was sent again. As in the educational brochure group, participants were encouraged to seek some other form of screening if colonoscopy was not feasible.

Primary outcome (medically verified colonoscopy) was assessed at the individual level. Colonoscopies were verified through physician or clinic confirmation if participants provided a medical release. Participants who self-reported a colonoscopy but did not provide a release for medical verification were considered nonadherent to screening. Cancer screening, knowledge, cognitions, and screening barriers were assessed at baseline (prerandomization), 1 month, and 9 months postintervention. A 15-month follow-up questionnaire was mailed to participants who reported that they had not had a colonoscopy by the 9-month follow-up or if their screening status was unknown.

Differences in demographic variables between study arms were tested with χ² tests. Generalized mixed logistic regression models were used to account for the cluster (i.e., nuclear family) effect when evaluating the impact of intervention on colonoscopy uptake by the 15-month assessment. Data were analyzed using participants with known outcomes, negative outcome imputation, and multiple imputation. Family was considered a random effect variable. Negative outcome imputation and multiple imputation were used for intent-to-treat analysis and included all eligible participants who were randomly assigned. Negative outcome imputation assumed that colonoscopy did not occur if there was no documented verification of colonoscopy. Multiple imputation was based on data from age at baseline, sex, household income, and health insurance coverage. Five imputed datasets were used to provide a combined estimate for missing values. ORs with 95% confidence intervals (CI) compared colonoscopy uptake at 15 months.

To estimate the cumulative incidence of medically verified colonoscopy following the intervention to 15-month follow-up, curves based on Kaplan–Meier estimates were created along with associated 95% Hall–Werner confidence bands. Cumulative incidences were stratified by study arm. For outcome analysis involving cost as a barrier, stratified logistic regression estimated the intervention effect within those who indicated that cost was a barrier and those who indicated that cost was not a barrier at baseline. Those with missing data on cost as a barrier or colonoscopy use were excluded for this analysis. Family was not included as a random effects variable as clustering could no longer apply when stratifying based on cost as a barrier. The effect of receiving a resource letter describing low-cost options for colonoscopies was examined within the subgroup of people who had not yet received a colonoscopy by 9 months postintervention. The model controlled for the intervention effect and a potential interaction between intervention and resource letter. Two biostatisticians (K.M. Boucher and L.M. Pappas) conducted the analyses using SAS (version 9.3; SAS Institute).

Sample size and power calculations were based on the trial's primary outcome analysis at 9-month follow-up (24).

Study enrollment, randomization, and retention data are shown in Fig. 1. The TeleCARE group had 167 eligible clusters and the educational brochure group had 170 clusters. The number of family members enrolled per cluster did not differ significantly by treatment arm (24). Demographic characteristics did not differ between the study groups (see Table 1) or when further stratified by participant reported cost barriers to colonoscopy. The participants who received the 15-month assessment did not differ in demographics specified in Table 1 from those who received the 9-month assessment. Twenty-one participants (TeleCARE = 12; educational brochure = 9) reported another type of colorectal cancer screening.

Overall, 42.7% of those in the TeleCARE group obtained a colonoscopy by 15 months postintervention compared with 24.1% of those in the educational brochure group (OR, 2.37; 95% CI, 1.59–3.52; Table 2). The intervention effect was similar across known outcome and imputed models.

As shown in the cumulative incidence curves (Fig. 2), the TeleCARE group had higher cumulative incidence of colonoscopy throughout the 15-month follow-up period. The divergence of cumulative incidence slopes between the TeleCARE and educational brochure group appears primarily within the first nine months of the intervention (i.e., within 276 days after the intervention), suggesting that the intervention effect occurred primarily within those 9 months. Thirty-six percent of the TeleCARE participants had a colonoscopy by 9 months and 44.3% by 15 months. Within the educational brochure group, 16% had a colonoscopy by 9 months and 24.2% by 15 months.

At 15 months, TeleCARE participants were more likely than those in the educational brochure group to obtain a colonoscopy even when cost had been indicated as a barrier at baseline. Among those who did not identify cost as a barrier [TeleCARE = 66 (37.5%), educational brochure = 89 (43.0%)], 51.5% of the TeleCARE group had a colonoscopy compared with 24.7% of the educational brochure group (OR, 3.66; 95% CI, 1.85–7.24; Table 3). Of the 228 people who identified cost as barrier [TeleCARE = 110 (62.5%), educational brochure = 118 (57.0%)], 38.2% of those in the TeleCARE group had a colonoscopy compared with 23.7% of those in the educational brochure group (OR, 1.99; 95% CI, 1.12–3.52).

Two-hundred eighty-seven participants (113 in TeleCARE group, 174 in educational brochure group) reported no colonoscopy by 9 months postintervention. Colonoscopy uptake by 15 months was not increased by the addition of the cost-resource letter (OR, 0.80; 95% CI, 0.52–1.23), with no evidence of the effect of the letter differing based on group assignment (P = 0.50).

TeleCARE, a remote, tailored intervention targeted to people at increased familial risk for colorectal cancer, was effective in increasing colonoscopy screening even among those who reported cost barriers. However, the intervention effect was considerably stronger among those who did not report cost as a barrier, a particularly relevant finding in the context of the Affordable Care Act.

It remains to be seen what the full impact of the Affordable Care Act will be on risk-based colorectal cancer screening. Evidence to date suggests that eliminating copayments has a modest effect on colorectal cancer screening (18, 31, 32). In addition to reducing colorectal cancer incidence and mortality, improving colonoscopy rates among those at increased risk may demonstrate to insurers the value of providing access to cancer screening and prevention services, including remote behavioral interventions. Insurer decisions to uniformly waive colonoscopy copayments or extend waivers to situations in which a colonoscopy becomes diagnostic (i.e., removal of polyps) will likely be influenced by whether colonoscopies provide cost savings (33). If colonoscopy uptake among those with increased colorectal cancer risk is improved, it may become financially appealing to insurers to waive copayments for colonoscopy because of the potential to avoid the costs of expensive cancer treatments (18, 33). Interventions such as TeleCARE are one way to promote colonoscopy use among people with higher risk.

The increase in screening rates between 9 months and 15 months was not attributed to the cost-resource letter, which raises questions as to other contributing factors. It is possible that some people needed the additional time to plan and overcome individual barriers to screening such as scheduling and making arrangements for work, childcare, or transportation. It is also possible that participants had a health maintenance visit with their provider that prompted a discussion about and completion of colonoscopy.

We are unable to identify which components of TeleCARE influenced the decision to undergo colonoscopy. A stepped-care approach should be evaluated to determine whether the intervention can be streamlined and remain effective (e.g., waiting to send reminders until a person has not obtained screening by 3 months or providing telephone counseling to those who do not have a colonoscopy within 6 months of the intervention). It may be possible to increase the potency of TeleCARE by incorporating patient navigation for those who need it. Navigation services are likely to become more widely available due to the Centers for Medicare and Medicaid Services' 2014 rule offering state Medicaid agencies the option to reimburse for more community-based preventive services, including those of community health workers (34). Other third party payers are expected to adopt this reimbursement policy. Thus, there is potential for community health workers to provide health promotion services, including navigation, to help patients overcome cost and other barriers to cancer prevention (6).

Our study was not designed to determine the effect of individual components of TeleCARE's intervention (i.e., telephone call, reminder card). Reminder cards may have some benefit (35), although their stand-alone effectiveness is unknown for people who are ambivalent about having a colonoscopy. TeleCARE is multifaceted: all participants assigned to TeleCARE received more intervention-related contact than those in the educational brochure (three vs. one, respectively). Therefore, we cannot tease apart the effect of attention (i.e., dose) versus that of the intervention content and delivery (36).

The majority of our study population was non-Latino White, which precluded evaluation of TeleCARE's efficacy among ethnically diverse subgroups. Although motivational interviewing accommodates a wide variety of cultural beliefs, the intervention did not explicitly address cultural beliefs that may conflict with having colonoscopy (37, 38). People with additional sociocultural barriers (e.g., competing life concerns or cultural beliefs such as fatalism) may require patient navigation interventions beyond the scope of TeleCARE (14, 37, 39). Our study was also not powered to examine differences in uptake by state, which is important as states and communities vary in their capacity for colonoscopy (10, 40).

Finally, it is important to note that although our intervention increased adherence to guideline-concordant colorectal cancer screening from 0% to 42.7%, which is higher than other intervention trials among individuals at increased familial risk (41–43)—the national goal for screening adherence in the general population is 70% to 80% (44, 45). Given the increased risk in relatives of colorectal cancer patients, screening adherence needs to be higher than what our intervention achieved. As highlighted previously, adherence to colonoscopy is a multifaceted problem. The efficacy of TeleCARE may be improved by complementing it with an intervention that intervenes on system-level barriers such as provider communication about screening, increasing access to screening patient navigation, direct assistance with coverage for screening, and patient follow up in primary care (12, 16).

Our results suggest that TeleCARE increases colonoscopy screening among relatives of colorectal cancer patients, especially, but not exclusively, when cost is not a barrier. Interventions such as TeleCARE may help maximize the impact of programs and policies that increase access to preventive services by bolstering cancer screening.

R. Burt has received speakers bureau honoraria from Myriad Genetics. No potential conflicts of interest were disclosed by the other authors.

This content is solely the responsibility of the authors and does not necessarily reflect the opinions or views of the funding and supporting agencies.

Conception and design: K.M. Boucher, S.T. Walters, W. Boonyasiriwat, M.D. Schwartz, W.K. Kohlmann, R. Burt, R.G. Simmons, A.Y. Kinney

Development of methodology: S.T. Walters, W. Boonyasiriwat, A.M. Stroup, M.D. Schwartz, W.K. Kohlmann, A.Y. Kinney

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.M. Stroup, W.K. Kohlmann, J.T. Lowery, C.L. Wiggins, D.A. Hill, A.Y. Kinney

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L.E. Steffen, K.M. Boucher, L.M. Pappas, S.W. Vernon, S.L. Edwards, D.A. Hill, A.Y. Kinney

Writing, review, and/or revision of the manuscript: L.E. Steffen, K.M. Boucher, B.H. Damron, S.T. Walters, K.G. Flores, S.W. Vernon, M.D. Schwartz, S.L. Edwards, W.K. Kohlmann, J.T. Lowery, C.L. Wiggins, D.A. Hill, J.C. Higginbotham, R. Burt, R.G. Simmons, A.Y. Kinney

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K.G. Flores, A.M. Stroup, S.L. Edwards, R.G. Simmons, A.Y. Kinney

Study supervision: S.T. Walters, A.M. Stroup, A.Y. Kinney

Other (developed some materials used in the intervention—reminder postcard): W. Boonyasiriwat

The authors thank the study's genetic counselors, Kory Jasperson, Amanda Gammon, Anne Naumer, and Lisa Wadge, for providing the intervention.

This work was supported by grants from the NCI (1R01CA125194-0305; to A.Y. Kinney, Principal Investigator) and the Huntsman Cancer Foundation. The work was also supported by the Shared Resources (P30 CA042014) at Huntsman Cancer Institute [Biostatistics and Research Design, Genetic Counseling, Research Informatics, the Tissue Resource and Applications Core (TRAC), and the Utah Population Database (UPDB)]; the Utah Cancer Registry, which is funded by Contract No. HHSN261201000026C from the NCI's Surveillance, Epidemiology, and End Results (SEER) Program with additional support from the Utah State Department of Health and the University of Utah; the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885, the National Cancer Institute's SEER Program under contract N01PC-2010-00034C awarded to the Northern California Cancer Center, contract N01-PC-35139 awarded to the University of Southern California, and contract N01-PC-54404 awarded to the Public Health Institute, and the Centers for Disease Control and Prevention's National Program of Cancer Registries, under agreement U58CCU000807-05 awarded to the Public Health Institute; the Colorado Central Cancer Registry program in the Colorado Department of Public Health and Environment funded by the National Program of Cancer Registries of the Centers for Disease Control and Prevention; the Cancer Data Registry of Idaho supported in part by the National Program of Cancer Registries of the Centers for Disease Control and Prevention; the New Mexico Tumor Registry which is funded by NCI Contract No. HHSN261201300010I; the Rocky Mountain Cancer Genetics Network (HHSN261200744000C); the Huntsman Cancer Registry; and the Intermountain Healthcare Oncology Clinical Program and Intermountain Clinical Genetics Institute.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.