Tailored telephone counseling increases colorectal cancer screening

Abstract

To compare the efficacy of two interventions to promote colorectal cancer screening participation and forward stage movement of colorectal cancer screening adoption among first-degree relatives of individuals diagnosed with adenomatous polyps. One hundred fifty-eight first-degree relatives of individuals diagnosed with adenomatous polyps were randomly assigned to receive one of two interventions to promote colorectal cancer screening. Participants received either a tailored telephone counseling plus brochures intervention or a non-tailored print brochures intervention. Data were collected at baseline and 3 months post-baseline. Group differences and the effect of the interventions on adherence and stage movement for colorectal cancer screening were examined using t-tests, chi-square tests, and logistic regression. Individuals in the tailored telephone counseling plus brochures group were significantly more likely to complete colorectal cancer screening and to move forward on stage of change for fecal occult blood test, any colorectal cancer test stage and stage of the risk-appropriate test compared with individuals in the non-tailored brochure group at 3 months post-baseline. A tailored telephone counseling plus brochures intervention successfully promoted forward stage movement and colorectal cancer screening adherence among first-degree relatives of individuals diagnosed with adenomatous polyps.

Introduction

Colorectal cancer (CRC) is a common disease in the United States with 136 830 new cases estimated in 2014 [1]. Approximately half of the expected 50 310 CRC deaths could be prevented if appropriate CRC screening was widely implemented [1–3]. Removal of adenomatous polyps through endoscopic screening has been found to decrease CRC incidence by 75–90% [4–8]. Increasing CRC screening participation among persons at increased risk is especially critical. When compared with the general population, first-degree relatives (FDRs) of persons diagnosed with colorectal adenomatous polyps (CAP) have been shown to have a 2-fold relative risk of developing CRC [9–13]. Indeed, FDRs of individuals diagnosed with CAP at age 50 or younger have a >4-fold relative risk of developing CRC [9]. Unfortunately, many patients are unaware of their FDRs increased risk of CRC as a result of their own CAP diagnosis [14]. Furthermore, few CAP patients communicate their diagnosis with FDRs or recommend screening to their relatives, who are at increased risk for CRC due to their family history of polyps [14]. This study is innovative because it tested two CRC screening interventions aimed to promoting CRC screening adherence as well as forward stage movement among FDRs of individuals diagnosed with CAP, a group at increased risk of developing CRC.

Current guidelines for CRC screening are stratified by CRC risk [15]. Individuals are considered to be at ‘increased risk’ if they have: (i) a personal history of CRC or CAP; (ii) a personal history of inflammatory bowel disease; or (iii) a significant family history of CRC or CAP. Screening recommendations based upon a positive family history vary according to the number of relatives affected with CAP or CRC and their age(s) at time of diagnosis. People with a single FDR diagnosed with CRC or CAP at age 60 or older should be offered the screening options offered to those at average risk for CRC [i.e. annual fecal occult blood test (FOBT); fecal DNA testing; flexible sigmoidoscopy every 5 years; CT colonography every 5 years; double-contrast barium enema every 5 years; or colonoscopy every 10 years] starting at age 50 [16]. Persons who have an FDR diagnosed with CAP or CRC prior to age 60 or those with two or more affected FDRs should be screened with colonoscopy beginning at age 40 or 10 years earlier than the age at diagnosis of the youngest affected relative, whichever is earlier [15].

The efficacy of tailored telephone counseling (TTC) has been demonstrated in several studies designed to increase participation in cancer screening, particularly mammography [17–26]. However, to the authors’ knowledge, this is the first TTC trial aimed at individuals with a family history of CAP. The purposes of this trial were to: (i) compare the efficacy of a TTC plus brochures (TTC+) intervention versus non-tailored mailed brochures on CRC screening participation (primary outcome) and movement in stage of adoption for CRC screening among FDRs of persons diagnosed with CAP; and (ii) examine moderators of intervention efficacy. In addition, we assessed recall and satisfaction with the intervention for both groups. The specific research questions were:

1. Which intervention (TTC+ versus non-tailored mailed brochures) is more efficacious for promoting CRC screening adherence and forward stage movement at 3 months post-baseline in FDRs of persons diagnosed with CAP?

2. Which demographic characteristics moderate intervention efficacy in FDRs of individuals with CAP?

Materials and methods

Theoretical framework

The Transtheoretical Model (TTM) and the Health Belief Model (HBM) provided a conceptual foundation for the study at two levels: (i) as a framework guiding the overall study and (ii) for the development of the TTC+ intervention. The TTM describes behavior change as a dynamic process in which individuals move through discrete stages (i.e. Pre-contemplation, Contemplation, Preparation, Action and Maintenance) [27]. The HBM proposes that individuals will take action to prevent, screen for, or control a health condition if they: (i) consider themselves to be at risk for the condition; (ii) believe the condition to have serious consequences; (iii) believe that action will reduce either their susceptibility to or the severity of the condition; (iv) believe that the anticipated barriers to taking action are outweighed by the benefits; and (v) are confident in their ability to take action [28, 29]. The theoretical framework for the study is outlined in Figure 1.

Fig. 1.

Theoretical framework. FOBT, fecal occult blood test.

Design

For this randomized controlled trial (RCT), data were collected via structured telephone interviews at baseline and 3 months post-baseline. Participants were randomized to either the TTC+ or non-tailored brochure arm following baseline data collection. Randomization was stratified by age (i.e. 50–64 or 65+) and the participant’s relationship to the CAP patient (i.e. mother, father, sister, brother, daughter or son). Both intervention groups received two non-tailored brochures—one produced by the American Cancer Society entitled ‘Colon Testing Can Save Your Life!’ and a second produced by the Centers for Disease Control entitled ‘Colorectal Cancer Screening Saves Lives.’ These brochures described CRC risk factors as well as the colon screening tests available. Individuals in the TTC+ intervention also received TTC. Figure 2 summarizes the flow of participants through the RCT.

Fig. 2.

Flow of FDR Participants through RCT. FDRs, first-degree relatives; FOBT, fecal occult blood test.

Setting, participants and eligibility criteria

A total of 158 FDRs of persons diagnosed within the past year with CAP at two large Midwestern hospitals were enrolled. Eligibility criteria included: (i) 50–80-years old; (ii) English-speaking; (iii) having no personal history of CRC; (iv) having not had an FOBT in the last 12 months, a flexible sigmoidoscopy in the last 5 years, or a colonoscopy in the last 10 years; and (v) having no medical conditions that prohibited CRC screening.

Recruitment

Physician approval was obtained to contact 1930 CAP patients to ask them to refer their potentially eligible FDRs to the study. We contacted 1365 patients, of whom 533 (39%) had no eligible FDRs to refer. Of the 832 who had potentially eligible relatives, 414 (50%) were willing to refer one or more relatives. Introductory letters that referenced the name and relationship of the relative who referred them were mailed to 559 FDRs (see Fig. 2). Letters included a toll-free telephone number to call to opt out for those who did not wish to be contacted. One week after letters were mailed, research staff telephoned all FDRs who did not decline. Trained recruiters reached 478 FDRs to explain the study, assess eligibility, obtain informed consent and answer questions.

Of the 478 FDRs reached, 225 (47%) did not meet eligibility criteria. The most common reason for ineligibility was that individuals were up-to-date with CRC screening (n = 213, 95%). Other reasons for ineligibility included: personal history of CRC (n = 5, 2%); <50 years or >80 years of age (n = 3, 1%); cognitively impaired (n = 1; <1%); speech or hearing impaired (n = 1, <1%); and other (n = 2; 1%). Of the 253 eligible FDRs, 95 (37.5%) refused participation and 158 (62.5 %) were consented and enrolled in the study. Only one FDR per CAP patient was enrolled.

Data collection, measures and randomization

Data were collected via structured telephone interviews at baseline and 3 months post-intervention. During the baseline interview, stage of adoption, knowledge, perceived risk, objective risk, perceived barriers to screening and self-efficacy were measured. Stage of adoption was measured with eight items addressing FOBT, seven items addressing sigmoidoscopy and seven items addressing colonoscopy. Knowledge of CRC and CRC screening tests were measured using 12 items. Perceived risk for CRC was measured with a single item assessing ‘compared with other (men/women) your same age, how would you rate your chance of getting colon cancer in the next 10 years? Would you say your chance of getting colon cancer is…’ with response options ranging from ‘much higher than average’ to ‘much lower than average’. Objective risk was measured using 10 items that assessed age, gender, race, family history of CRC and personal health history. One item assessed whether a physician had recommended each of the three screening tests using ‘yes’ or ‘no’ response options).

Perceived barriers to FOBT were assessed with eight items, sigmoidoscopy barriers were assessed with eight items, and colonoscopy barriers were assessed with nine items. Responses were measured using a 5-point Likert-type scale, ranging from ‘strongly disagree’ to ‘strongly agree’. In addition, participants were asked to describe in their own words the main reasons they had not received each of the three CRC screening tests. Examples of potential barriers include practical concerns such as transportation issues or cost emotional barriers such as embarrassment, and cognitive barriers such as lack of bowel symptoms.

Self-efficacy for completing the three colon tests (e.g. FOBT, sigmoidoscopy and colonoscopy) was measured using three separate scales developed by the authors. Self-efficacy for FOBT was measured utilizing five items, whereas self-efficacy for sigmoidoscopy and colonoscopy were each measured utilizing eight items. Response options used a 5-point Likert-type scale, where 1 = ‘not at all confident’ and 5 = ‘very confident’. Examples of the self-efficacy items include confidence in one’s ability to find the time to perform the test, to follow instruction to clean one’s bowel (sigmoidoscopy and colonoscopy), and to collect stool samples (FOBT), among others.

Upon completion of the baseline interview, all 158 participants were randomized, using randomization lists that were stratified by age and relationship to the CAP patient. A total of 76 participants (48%) were assigned to the TTC+ group and 82 (52%) to the mailed non-tailored brochure group. The intervention nurse successfully counseled 75 (99%) of the participants assigned to the TTC+ intervention. One participant (1%) who had already scheduled a colonoscopy refused the counseling session. There were eight participants (5%) who already had an FOBT within 12 months prior to joining the study and they were excluded from further analyses. Follow-up telephone interviews were conducted at 3 months post-baseline with 153 participants (97%). Five participants (<3%) were unreachable for the follow-up interview. Attrition did not differ by group.

For the 3-month follow-up interview, all baseline measures were repeated and participant recall, relevance and satisfaction with the interventions were assessed using 12 items. Participants’ responded to eight items assessing satisfaction with the interventions using a 5-point Likert scale where 1 = strongly disagree and 5 = strongly agree. Participants were asked how much they agreed or disagreed with statements such as: the information received was new, interesting, easy to understand, prepared especially for them, they had discussed the information with their family, and with their doctor. Recall was assessed with one item about whether they remembered receiving the information. Relevance was assessed with three items asking whether they had received the right amount of information, how helpful it was, and how much of the information applied to them.

Development of the TTC intervention

The TTC+ intervention consisted of nine sections: (i) Introduction; (ii) Staging; (iii) Knowledge; (iv) Perceived Risk; (v) Objective Risk; (vi) Barriers to Screening; (vii) Self Efficacy; and (viii) Closing. Tailoring was featured in all sections except the closing. Tailored messages were generated from a computerized message library developed by Dr Rawl, Dr Champion and Dr Skinner. These messages were supplemented by generic messages designed to increase the participant’s knowledge of CRC and promote his/her active participation in the TTC+ dialogue. Tailored messages were derived from specific responses to items (variables) identified through the theoretical framework and measured on the baseline questionnaire. Responses to baseline questions were used to generate messages for the intervention (see Table I). Tailoring on the stage of adoption of a screening test was placed near the beginning of the intervention to either introduce or confirm the importance of taking control over one’s colon health and to guide selection of the screening test to be discussed.

Table I.

Intervention message tailoring

Intervention section	Tailoring items (variables) from baseline questionnaire
Stage of adoption for each screening test Pre-contemplation Contemplation Preparation Action	FOBT staging (eight items for algorithm computation) Sigmoidoscopy Staging (seven items for algorithm computation) Colonoscopy staging (seven items for algorithm computation) Sample item: ‘Are you planning to have a colonoscopy in the next 6 months?’
Perceived risk	Comparative risk (one item)
Objective risk	Age (one item) Gender, if female (one item) Race, if African American (one item) Personal Health History (three items) Colorectal polyps Ulcerative colitis Crohn’s disease Family History of CRC (four items) Number of FDRs affected Age of FDRs when affected (before age 60) Number of SDRs affected
Barriers to screening	Physician recommendation (one item) FOBT Barriers Scale (eight items) Sigmoidoscopy Barriers Scale (eight items) Colonoscopy Barriers scale (nine items) Main Reasons Open-Ended Query (one item) Sample item: ‘In your own words, please tell me what are the main reasons you have not had a colonoscopy in the last 10 years?’
Self-efficacy	FOBT Self-efficacy scale (five items) Sigmoidoscopy self-efficacy scale (eight items) Colonoscopy self-efficacy scale (eight items) Sample item: ‘How confident are you that you can go without eating solid food for a day before the colonoscopy?’

Risk for developing CRC was a central theme of the intervention. Participants received one of four-tailored counseling messages depending upon their perceived risk of CRC compared with other people their same age. This message was immediately followed by a tailored message regarding each user’s objective risk for CRC. This message was selected from one of 51 different objective risk profiles in the message library based upon participants’ individual risk factors (e.g. age, race, personal health history and familial CRC history). The objective risk message section included a recommendation for CRC screening based upon the most current American Cancer Society guidelines combined with individual risk factors. If participants were eligible for any of the three tests, their stage of test adoption was used to produce the final screening recommendation. For example, individuals with a family history of only one FDR diagnosed with CAP had a choice of screening tests. Participants were told that they could consider: (i) an annual FOBT; (ii) a sigmoidoscopy once every 5 years; or (iii) a colonoscopy once every 10 years. If participants were already in contemplation for colonoscopy, they received detailed information on colonoscopy only. When participants were eligible for any of the three tests and their stage of adoption was either pre-contemplation or contemplation for all three tests, participants were asked which testing option they were more likely to do. Per American Cancer Society guidelines, the combination of FOBT plus sigmoidoscopy or colonoscopy alone was emphasized as the best screening options. Based upon the participant’s preference, FOBT plus sigmoidoscopy or colonoscopy was selected for discussion.

Barriers to the participant-selected screening test were addressed with tailored messages that were written and programmed using computer algorithms to automatically populate into the TTC+ intervention. Messages to overcome common barriers were developed and pre-programmed into the computerized counseling script. These included messages related to: (i) participants’ feeling fine/not having any symptoms and seeing no need for the test; (ii) embarrassment; (iii) lack of time; (iv) fear of positive findings; (v) not knowing how to do a stool test; (vi) not knowing what will happen during a colonoscopy; (vii) cost; (viii) lack of transportation; and (ix) pain. Barriers reported by participants during the baseline interview were pre-populated into the telephone script and available on-screen for the nurse to deliver. Messages to overcome any additional barriers reported by participants during the counseling call were immediately retrievable from the computer program. Another feature of the barriers section was the ability to generate a customized message to counter the main reasons for not receiving screening as reported by participants in their own words. Examples of barrier messages delivered during the intervention are shown in Figure 3.

Fig. 3.

Sample barrier messages delivered during the intervention.

The self-efficacy section of the TTC+ intervention was designed similarly to the barriers section. Tailored messages for the test under discussion automatically appeared on the computer screen whenever a participant was not confident in some aspect of test performance. Tailored messages for self-efficacy items that arose during counseling were retrieved and delivered by the nurse. The computer-generated counseling script guided all messages to be delivered. During pilot testing, a frequently asked questions bank was developed which the nurse had immediate access to on the computer. In addition, the nurse interventionist was experienced with cancer screening tests.

Intervention delivery

The TTC+ intervention was delivered within 1 month of the baseline interview to each participant who was randomized to the TTC+ arm. To ensure accuracy and completeness prior to delivery, the nurse interventionist would compare a hard copy of the computer-tailored intervention to data collected on the baseline questionnaire. Approximately 1 week after the CRC screening brochures were mailed to the participant, the nurse interventionist began calling the participant to verify receipt of the brochures and schedule the counseling session. If the participant did not have the brochures, they were re-mailed and telephone follow-up was made approximately 1-week later. TTC+ was delivered at the time of the participant’s choosing after receipt of the brochures.

Data analyses

All variables were described using summary statistics, including proportions for binary variables, frequency tables for categorical variables, and mean, standard deviation, minimum and maximum for continuous variables. Comparisons of continuous variables between intervention groups were made by t-tests. If the distributional assumptions were not met, non-parametric Wilcoxon Rank Sum tests were used. The two groups were compared on ordinal variables using the 1 df Mantel chi-square test of trend. To compare categorical screening stage movement between the intervention groups, the contingency-table Pearson chi-square test of general association was used. When 20% or more of cells in contingency tables had expected counts <5, the two-sided Fisher’s exact test was used.

Screening participation was assessed in the following ways. First, each individual screening method was considered (i.e. FOBT, sigmoidoscopy, colonoscopy). Then, two other screening outcomes were assessed: (i) adherence to any test—defined as participating in ‘any’ of the three screening methods; and (ii) adherence to the risk-appropriate test—defined as participating in the screening method appropriate for, or more than appropriate for, an individual’s risk level. Adherence to the risk-appropriate test was defined as having had ‘either’ of the following: (i) FOBT and flexible sigmoidoscopy; or (ii) colonoscopy.

There were three possible levels of stage of adoption at baseline and four at follow-up (1 = Pre-contemplation, 2 = Contemplation, 3 = Preparation, 4 = Action). By definition, none of the participants were in the action stage at baseline. Stage of adoption was assessed for FOBT, colonoscopy and sigmoidoscopy. In addition, stage was assessed for ‘any’ CRC test and risk-appropriate CRC test by assigning the highest stage category observed among the stages of the respective tests for each participant. Participants could move forward three stages, have no stage movement or move backward three stages between baseline and intervention. Statistical analysis of stage change was based on categorization into two groups (forward movement versus no movement or backward movement) because of sparseness in the categories of those individuals moving either backward any number of stages or forward more than one stage. The effect of the interventions on adherence and forward stage movement adjusting for covariates was assessed by using logistic regression models. Multivariate relationships were examined and adjusted for potentially confounding demographic variables and patient characteristics.

Group differences in satisfaction with the intervention material also were examined. Differences in the mean scores of individual items measuring satisfaction and the total satisfaction score were examined using independent t-tests. Responses to the eight items assessing satisfaction with the intervention were summed to create a total satisfaction score and the mean was calculated. Recall and relevance items were examined using Pearson chi-square tests in the case of dichotomous variables and using1 df Mantel chi-square test of trend in the case of ordinal variables.

Results

Participants in both groups had similar distributions on demographic characteristics (see Table II). The groups showed an average age near 60 and an average number of years of education equivalent to 2 years of college. Most participants were female, Caucasian, currently employed with medium to high incomes, had insurance and were married or living with a partner (see Table II). Greater proportions of the TTC+ group had lower incomes than the non-tailored brochure group.

Table II.

Participant characteristics

Variable	Tailored phone counseling		Non-tailored print brochure		P value
	n = 71		n = 79
	Mean	SD	Mean	SD
Age	60.4	7.4	60.0	9.2	0.80
Education, highest grade completed	13.9	2.7	14	3.4	0.82
	n	%	n	%
Gender
Female	48	67.6	52	65.8	0.82
Male	23	32.4	27	34.2
Race
Caucasian	53	74.7	65	82.3	0.46
African American	14	19.7	12	15.2
More than one race	4	5.6	2	2.5
Education
Less than high school degree	7	9.9	12	15.2	0.26
High school graduate	29	40.9	22	27.9
Vocational school or some college	19	26.8	19	24.1
College graduate	6	8.5	14	17.7
Some graduate work or graduate degree	10	14.1	12	15.2
Employed full-time/part-time
Full-time	25	35.2	36	45.6	0.35
Part-time	11	15.5	13	16.5
Not employed	35	49.3	30	38.0
Income
<$30 000	33	47.8	23	31.1	0.04
$30 000–$75 000	27	39.1	30	40.5
>$75 000	9	13.0	21	28.4
Marital status
Not partnered	29	40.9	36	45.6	0.54
Married or living with a partner	42	59.2	43	54.4
Health insurance
No	6	8.5	7	8.9	0.93
Yes	65	91.6	72	91.1
FOBT stage at baseline
Pre-contemplation	58	81.7	69	87.3	0.056
Contemplation	8	11.3	10	12.7
Preparation	5	7	0	0
Sigmoidoscopy stage at baseline					1.00
Pre-contemplation	69	97.2	76	96.2
Contemplation	2	2.8	3	3.8
Colonoscopy stage at baseline
Pre-contemplation	56	78.9	62	78.5	0.95
Contemplation	15	21.1	17	21.5
Stage of any CRC test, baseline
Pre-contemplation	48	67.6	53	67.1	0.042
Contemplation	18	25.4	26	32.9
Preparation	5	7.0	0	0.0
Stage of risk appropriate CRC test at baseline
Pre-contemplation	49	69.0	53	67.1	0.036
Contemplation	17	23.9	26	32.9
Preparation	5	7.0	0	0.0
Study site
Safety net hospital	22	31.4	19	24.1	0.31
University-affiliated hospital	48	68.6	60	76.0

Note: For the three-level and dichotomous (<$30 000 versus ≥$30 000) versions of income, P = 0.04.

FOBT pre-contemplation versus contemplation or preparation, P = 0.34.

There were no differences between groups on staging for sigmoidoscopy and colonoscopy. For FOBT, all five participants in preparation were in the TTC+ group, which led to the marginally significant difference in FOBT staging at baseline, and a significant difference for any and risk-appropriate baseline stage. There were no significant differences between the two groups on the remaining demographic variables.

Analyses were performed to determine whether significantly greater proportions of FDRs in the TTC+ group reported that: (i) they recalled receiving the intervention; (ii) they learned new information; (iii) the information provided was helpful; and (iv) the information was relevant to them compared with the non-tailored group 3 months after baseline (see Table III). The TTC+ group displayed significantly greater satisfaction on the total score compared with the non-tailored group. The TTC+ group had significantly higher mean scores, indicating greater agreement with the statements that: (i) the information was prepared especially for them and; (ii) they discussed the information with their family. A significantly greater percentage of the TTC+ group (66%) agreed that the information was very helpful to them compared with 39% in the non-tailored brochure group. A significantly greater percentage of the participants in the TTC+ group agreed that all or most of the information applied to them (76%) compared with 53% in the non-tailored group.

Table III.

Comparison of groups on recall, relevance and satisfaction with interventions

Variable	Tailored phone counseling			Non-tailored print brochure			p-value
	n	Mean	SD	n	Mean	SD
I received a lot of new information	58	3.8	1.0	59	3.5	1.1	0.18
The information was easy to follow	59	4.2	0.5	59	4.2	0.4	0.87
The information was interesting	58	4.0	0.6	59	4.0	0.8	0.87
The information was easy to understand	59	4.2	0.4	59	4.2	0.5	0.97
The information was prepared especially for me	53	3.4	1.1	57	2.8	1.1	0.005
Discussed information w/family	61	3.6	1.0	59	3.0	1.1	0.008
Discussed info w/ doctor	61	2.4	1.0	59	2.3	0.8	0.33
Planning discuss info w/doctor	60	2.5	1.7	58	2.7	1.4	0.63
Satisfaction total score (sum of eight items above)	59	29.0	3.7	59	27.0	4.0	0.002
	n	%		n	%		p-value
Do you recall receiving the info?
No	6	8.8		12	15.6		0.22
Yes	62	91.2		65	84.4
The information you received was:
More than you liked	2	3.4		5	8.6		0.46
Just the right amount	52	89.7		48	82.8
Less than you liked	4	6.9		5	8.6
Was information helpful?
Very Helpful	41	66.1		23	39.0		0.007
Somewhat helpful	17	27.4		32	54.2
Not very helpful	2	3.2		4	6.8
How much info applied to you?
All of it	12	20.3		12	21.4		0.035
Most of it	33	55.9		18	32.1
Some of it	13	22.0		23	41.1
None of it	1	1.7		3	5.4

Note: Pearson chi-square was used for dichotomous variables, and the 1 df Mantel chi-square of trend was used for ordinal variables.

‘Research Question 1: Which intervention (TTC+ versus non-tailored mailed brochures) is more efficacious for promoting CRC screening adherence and forward stage movement at 3 months post-baseline in FDRs of persons diagnosed with CAP?’

The bivariate relationships between intervention group and adherence and change in stage are reported in Table IV. When compared with the non-tailored brochure group, the TTC+ group had significantly higher percentages of participants who were adherent to colonoscopy, any CRC test, and the risk-appropriate test at follow-up. Those who were in the TTC+ group were approximately five times more likely to be adherent to any CRC test than those in the non-tailored group. Furthermore, the TTC+ group displayed significantly higher percentages in moving forward on stage for FOBT, any CRC test stage, and stage of the risk-appropriate test, and the greater movement for the TTC+ group was marginally significant (P = 0.07) for colonoscopy stage. For sigmoidoscopy, no participant moved forward or showed adherence at follow-up.

Table IV.

Adherence and stage change by group at follow-up

Variable	Tailored phone counseling		Non-tailored print brochure
	n	%	n	%	Odds ratio	95% CI	p-value*
FOBT adherence
Not adherent	63	92.7	74	96.1	2.0	(0.5, 8.5)	0.475
Adherent	5	7.4	3	3.9
Colonoscopy adherence
Not adherent	57	83.8	76	98.7	14.7	(1.8, 116.9)	0.001
Adherent	11	16.2	1	1.3
Sigmoidoscopy dherence
Not adherent	68	100.0	77	100.0
Adherent to any test
Not adherent	53	77.9	73	94.8	5.2	(1.6, 16.5)	0.003
Adherent	15	22.1	4	5.2
Adherent to the risk-appropriate test
Not adherent	53	77.9	74	96.1	7.0	(1.9, 25.3)	0.002
Adherent	15	22.1	3	3.9
FOBT stage changes (two categories)
Backward and no movement	54	76.1	72	94.1	3.2	(1.3, 8.4)	0.014
Forward movement	17	24.0	7	8.9
Colonoscopy stage changes (two categories)
Backward and no movement	46	64.8	62	78.5	1.98	(1.0, 4.1)	0.071
Forward movement	25	35.2	17	21.5
Sigmoidoscopy stage changes (two categories)
Backward and no movement	70	98.6	79	100.0			0.473
Forward movement	1	1.4	0	0.0
Any stage changes (two categories)
Backward and no movement	35	51.5	58	75.3	2.9	(1.4, 5.8)	0.003
Forward movement	33	48.5	19	24.7
Risk appropriate stage changes (two categories)
Backward and no movement	36	52.9	59	76.6	2.9	(1.4, 5.9)	0.003
Forward movement	32	47.1	18	23.4

Note: *p-value is from two-sided Fisher exact test.

Next, exact-probability logistic regression models were used to test intervention effects on adherence and forward stage movement while adjusting for potentially confounding covariates. Table V has eight models. Each model has a different dependent variable, either adherence or forward stage movement with respect to four different outcomes: FOBT, colonoscopy, any test and risk-appropriate test. The models were adjusted for the following baseline characteristics: income (<$30 000 versus ≥$30 000) and stage of adoption at baseline (the only characteristics in Table I with P < 0.25), and the theoretically-important covariates of age (50–64 versus 65+), gender, and race (Caucasian versus other). In each model, the relevant baseline stage variable was used. Specifically, models of FOBT adherence and stage movement were adjusted for baseline FOBT stage; models of adherence to any CRC test and stage movement were adjusted for baseline stage of any test; and models of adherence to risk-appropriate test and stage movement were adjusted for baseline stage of risk-appropriate test. The outcome of adherence to sigmoidoscopy could not be modeled because none of the participants reported receipt of sigmoidoscopy at follow-up.

Table V.

Multivariable exact-probability logistic regression models of follow-up adherence and follow-up forward stage movement

Dependent variable	Odds ratio	95% CI	p-value
Adherence
Model 1: Adherence to FOBT	1.5	(0.2, 11.5)	0.89
Model 2: Adherence to colonoscopy	13.6	(1.8, 611.9)	0.004
Model 3: Adherence to any test	6.1	(1.6, 30.8)	0.005
Model 4: Adherence to risk-appropriate test	9.8	(2.1, 67.5)	0.001
Stage movement
Model 5: FOBT stage movement	2.9	(1.0, 9.1)	0.047
Model 6: Colonoscopy stage movement	2.4	(1.0, 5.6)	0.044
Model 7: Any test stage movement	3.9	(1.7, 9.6)	0.001
Model 8: Risk-appropriate test stage movement	4.1	(1.7, 10.3)	0.001

Note: The odds ratio reported is for the tailored group versus non-tailored group. Independent variables: group (tailored versus non-tailored), age (50–64 versus 65+), race (Caucasian versus other), gender, income (<$30 000 versus ≥$30 000) and baseline stage. For each model, the relevant baseline stage variable was used, as described in the ‘Materials and methods’ section.

For each logistic model, the adjusted odds ratios, its 95% Wald confidence interval (CI), and Wald P-values are reported in Table V for the group effect (tailored versus non-tailored) adjusted for baseline covariates.

The multivariate regression results were consistent with the bivariate analysis results. After adjusting for potentially confounding covariates forward colonoscopy stage movement was significantly greater for the TTC+ group (P = 0.044), whereas the difference was only marginally significant when confounding covariates were not considered.

Table V shows that after adjusting for potentially confounding covariates, those who were in the TTC+ group were approximately six times more likely to be adherent to any CRC test than those in the non-tailored group.

‘Research Question 2: Which demographic characteristics moderate intervention efficacy in FDRs of colorectal CAP?’

Moderators of intervention efficacy were investigated using logistic regression to test the interaction terms between randomized group and baseline characteristics. There were no significant interactions with screening adherence. However, there were several interactions with forward stage movement. Specifically, significant interactions were found with education, marital status and employment, which were further explored by performing separate logistic regression models for each of two levels of the baseline characteristic (see Table VI). Forward stage movement on colonoscopy, any test and the risk-appropriate test was significantly greater for TTC+ compared with the non-tailored group among those with less education, not among those with more education. This does not imply that those with more education did not move forward in stage of adoption, but rather that superior efficacy of TTC+ over non-tailored was observed in the low education group, whereas the forward movement rates among those more educated was similar between intervention groups.

Table VI.

Interactions between demographic characteristics and intervention efficacy

Dependent variable	Tailored phone versus non-tailored brochure			Tailored phone versus non-tailored brochure
	OR	95% CI	P	OR	95% CI	P
	Less Education			More Education
Forward stage movement on colonoscopy	5.2	(1.2, 28.9)	0.023	1.3	(0.4, 4.0)	0.84
Forward stage movement on any test	13.3	(2.8, 98.4)	<0.0001	1.6	(0.5, 5.0)	0.50
Forward stage movement on risk-appropriate test	22.7	(3.5, 311.8)	<0.0001	1.6	(0.5, 5.1)	0.48
	Not married			Married
Forward stage movement on colonoscopy	9.4	(1.6, 108.5)	0.008	1.2	(0.4, 3.6)	0.84
	Not employed			Employed
Forward stage movement on any test	13.1	(1.9, 184.1)	0.004	2.5	(0.9, 7.4)	0.09
Forward stage movement on risk-appropriate test	13.1	(1.9, 184.1)	0.004	2.5	(0.9, 7.8)	0.10

Note: OR, odds ratio. For each logistic model, the adjusted odds ratios, 95% CI, and P-values are for the group effect (tailored versus non-tailored) adjusted for these covariates: age (50–64 versus 65+), race (Caucasian versus other), gender, income (<$30 000 versus ≥$30 000) and baseline stage. For each model, the relevant baseline stage variable was used as described in the ‘Materials and methods’ section.

In addition, forward stage movement on colonoscopy was significantly greater for the TTC+ compared with the non-tailored group among those not married (see Table VI). Among those who were married, rates of forward movement were similar between intervention groups. Likewise, forward stage movement on any test and the risk-appropriate test was significantly greater for TTC+ compared with the non-tailored group among those not employed. Among those who were employed, the rates of forward movement were not significantly different, although the TTC+ group did show marginally greater movement than the non-tailored group (P < 0.10).

Discussion and conclusion

Individuals in the TTC+ intervention group were significantly more likely to complete CRC screening and to move forward on stage for FOBT, any CRC test stage, and stage of the risk-appropriate test compared with individuals in the non-tailored brochure group 3 months post-baseline. After adjusting for potentially confounding covariates, individuals who received TTC+ were six times more likely to complete CRC screening to any test, and ten times more likely to complete risk-appropriate screening, when compared with individuals in the non-tailored brochure group.

Our study is unique in that it focused on a population at greater than average risk for CRC due to a family history of CAP. To the authors’ knowledge, this is the first tailored intervention to target the CRC screening behavior of FDRs of patients with CAP. Prior CRC screening intervention studies have tested telephone counseling, but most have focused upon those at average risk for CRC or those with a significant family history of CRC [24, 26, 30–34]. Prior telephone counseling interventions in other populations have demonstrated mixed results [24, 26, 30–34]. Two interventions which featured TTC with FDRs of CRC survivors resulted in significantly higher rates of post-intervention screening of participants in tailored intervention groups [26, 32]. However, in a study comparing tailored print brochure, tailored brochure plus TTC, and a generic brochure, significant differences in screening rates were not observed between the two tailored intervention groups [32]. Additional trials to test telephone-based CRC screening interventions are currently ongoing [35–37].

Strengths of the study include the uniqueness of the study in focusing on promoting CRC screening and forward stage movement in FDRs of individuals diagnosed with CAP. Further strengths of the present study include a RCT design and low attrition. However, the majority of participants were White and well-educated and had health insurance, a limitation for generalization. Another limitation includes the use of self-report measures. Also, in the current study, the nurse who delivered the tailored intervention was able to assist participants with problem-solving to overcome barriers they reported during the intervention as well as those reported during the baseline interview.

The success of the TTC+ intervention to promote forward stage movement and CRC screening completion may reflect the ability of the nurse to speak to the specific barriers faced by the participant [38]. Future interventions may use a nurse in a patient navigator role to answer questions and to assist participants in overcoming barriers to CRC screening. Furthermore, technology-based intervention utilizing DVD or computers also may be considered to promote CRC screening and forward stage movement among FDRs of those diagnosed with CAP [39–40]. Such technology-based interventions also might include counseling from either a lay advisor or patient navigator to promote CRC screening and forward stage movement [41]. In addition, future studies might explore the efficacy of TTC interventions with a more diverse sample, especially among underserved populations at increased risk for CRC. More broadly, similar interventions may be designed and tested among FDRs of individuals with other diseases that place the FDRs at increased risk for also developing the disease.

The results of this study suggest that TTC+ among FDRs of individuals diagnosed with CAP can successfully promote forward stage movement and CRC screening adherence among this population at increased risk for CRC. The TTC+ intervention efficacy was strong in the overall sample, and was even stronger in women with less education, those not employed, and those not married. Future research is needed to evaluate the efficacy of technology-based interventions to promote CRC screening and forward stage movement in this population at increased risk for CRC.

Funding

This work was supported by the National Institute of Nursing Research at the National Institutes of Health (R15NR07999 to S.M.R.). The work of the second author was supported by the National Cancer Institute at the National Institutes of Health (R25CA117865 to V.L.C.).

Conflict of interest statement

None declared.