Using Patients’ Social Network to Improve Compliance to Outpatient Screening Colonoscopy Appointments among Blacks: A Randomized Clinical Trial

Abstract

Objectives:

Patient navigation improves colorectal cancer (CRC) screening among underserved populations, but limited resources preclude widespread adoption in minority serving institutions.

We evaluated whether a patient’s self-selected social contact person can effectively facilitate outpatient screening colonoscopy.

Methods:

From September 2014 to March 2017 in an urban tertiary center, 399 black participants scheduled for outpatient screening colonoscopy self-selected a social contact person to be a facilitator and provided the person’s phone number. Of these, 201 (50.4%) participants were randomly assigned to the intervention arm for their social contact persons to be engaged by phone. The study was explained to the social contact person with details about colonoscopy screening and bowel preparation process. The social contacts were asked to assist the participants, provide support and encourage compliance with the procedures. The social contact person was not contacted in the usual care arm, n=198 (49.6%). We evaluated attendance to the scheduled outpatient colonoscopy and adequacy of bowel preparation. Analysis was performed by intention to treat.

Results:

The social contact person was reached and agreed to be involved for 130/201 (64.7%) participants. No differences were found in the proportion who underwent screening colonoscopy (77.3% versus 77.2%; RR=1.01; 95% CI: 0.91–1.12) but there was a modest increase in the proportion with adequate bowel preparation with social contact involvement (89.1% versus 80.9%; RR=1.10; 95%CI: 1.00-1.21).

Conclusions:

Engaging a patient’s social network to serve in the role of a patient navigator did not improve compliance to outpatient screening colonoscopy, but modestly improved adequacy of bowel preparation.

INTRODUCTION

Blacks have the highest mortality from colorectal cancer (CRC) in the United States. ^1–2 Studies have suggested that lower healthcare access and poorer utilization of healthcare resources among blacks contribute to this disparity.^3,4 Although CRC screening reduced CRC mortality in randomized trials, ^5–7 screening rates are lower among minority populations even when CRC screening is a covered benefit.^8–10 Furthermore, a high rate of non-compliance to outpatient screening colonoscopy, has been reported among underserved populations especially among blacks in safety net hospitals.^11–14

An effective intervention to improve cancer screening and adherence to diagnostic follow-up care after detection of an abnormality among underserved populations is through the use of patient navigators.^14–16 Patient navigators are paid personnel who assist patients in overcoming logistic challenges in order to facilitate receipt of medical care. Unfortunately, in many minority serving institutions where financial resources may be more limited, the use of patient navigators is reserved for patients with actual cancer diagnosis and these services are not routinely available for preventive services such CRC screening.

We have previously reported that patients who were married (as a measure of social support) were more compliant with attendance to out-patient colonoscopy. ¹⁷ Therefore, in this study, we evaluated the effect of directly involving a social contact person chosen by the patient to be a facilitator (akin to a patient navigator) on attendance to and quality of out-patient screening colonoscopy in a randomized controlled trial.

METHODS

Study population

We recruited competent, non-institutionalized black men and women who were 45 years of age or older and were scheduled for screening colonoscopy from gastroenterology clinics of Howard University Hospital. The referral base included Howard University primary care clinics and non-affiliated primary care clinics serving low income populations in the District of Columbia. Most of these non-affiliated clinics are run by the local government and charity organizations. Howard University Hospital does not have open access endoscopy, in part, because of vast outside network referrals. Therefore, every patient undergoes face to face consultation with an endoscopist prior to scheduling. For this trial, we recruited patients referred for screening colonoscopies. The patients provided information on their demographic characteristics and lifestyle factors. All participants gave a written informed consent and signed medical records release forms giving permission to obtain their colonoscopy reports. Every recruited participant received a $10 gift certificate immediately after enrollment, prior to randomization. The project was approved by Howard University Institutional Review Board (14-IRB-Med-11).

The interval between consultation and procedure date was generally between 2 to 4 weeks. An information booklet to educate patients about colonoscopy with detailed information about the required bowel preparation as well as information about their endoscopists was given to each subject during consultation. It also provided instructions for patients to call the facility if they were unable to keep their appointments. All participants were instructed to take only clear liquids the day before the procedure. We used split dose 4-liter polyethylene glycol solution for bowel preparation in which 2-liters were consumed the evening before the procedure and the remaining 2-liters were consumed on the morning of the procedure.

Inclusion and exclusion criteria

All study participants gave written informed consent, had access to a telephone (cell phone or land phone) and gave information about at least one social contact person who could serve as facilitators for them. The social contact person could be any adult identified by the participant such as spouse, family members, friends and neighbors. We excluded patients who could not provide information about at least one social contact and those referred for diagnostic procedures such as iron deficiency anemia, unexplained weight loss, and gastrointestinal bleeding. We also excluded patients at high risk for CRC such as those with personal history of familial adenomatous polyposis syndrome (FAP), those with family history of hereditary non-polyposis colorectal cancer syndrome (HNPCC), those with inflammatory bowel disease (Crohn’s disease or ulcerative colitis) and those with personal histories of CRC. Furthermore, we also excluded participants who had undergone bowel resection regardless of the indication.

Sample size and power calculation

Our previous study, ¹⁷ which involved screening and diagnostic colonoscopy, revealed an attendance rate of 78% and our baseline adequate bowel preparation rate of 80%. Therefore, using a baseline range of 70-80% for our outcomes, we planned to recruit 400 patients giving our study 85%-99% power to detect an absolute 15% increase in colonoscopy attendance and adequacy of bowel preparation in the social contact group using a two sided test and alpha error at 0.05.

Randomization assignment

A total of 425 subjects were recruited. Of these, 26 subjects were excluded for ineligibility. The eligible subjects were randomly assigned into usual care (control) arm (n = 198) and the intervention arm (n = 201) using a computer-generated, simple randomization sequence.

For participants in the intervention arm (social contact group), we called the participant’s social contact person on the phone and obtained a verbal consent to participate from him/her. In the event that we could not reach the primary social contact after 3 attempts within 1 week of participant’s recruitment or if he/she refused to participate, we contacted the alternate social contact person when the information is available. If the social contact agreed to participate, we informed him/her about the study using the study transcript which explained the process involved in undergoing colonoscopy including bowel preparation, conscious sedation, need for an escort, and the potential benefits and risks of colonoscopy screening. We did not discuss the patient’s medical condition with the social contact. We requested the social contact to perform the typical tasks of a patient navigator which includes encouraging the patient to comply with instructions for the bowel preparation, assisting with filling healthcare forms as needed, serving as an escort and arranging transportation as needed. We also mailed information about CRC screening and the process involved with colonoscopy to the social contact.

Participants in the usual care arm also received detailed information about colonoscopy, completed the intake form and provided information about social contacts similar to those in intervention arm, but their social contacts were not approached about the study.

Outcome assessment

We identified study participants who attended their scheduled procedures. We also determined the quality of bowel preparation from their endoscopy reports. The bowel preparation quality was recorded using the Aronchick scale which was selected in a drop down menu format in our Endoworks ® software system. This was rated as excellent, very good, good, fair, and poor based on how clearly the colonic mucosa was seen and the percentage examined. The designation of fair and poor were considered to be inadequate bowel preparation while good, very good and excellent designations were considered to be adequate bowel preparation. The endoscopists were not aware of the participants’ group assignments.

In addition, we conducted a brief telephone interview among 83 participants in the intervention arm to assess their opinions regarding their interactions with their social contact persons. We conducted similar interview with 72 social contacts inquiring about their interactions with the study participants also.

Statistical analyses

We compared the baseline demographic and lifestyle characteristics of the participants in the intervention and control arms. We determined the percentage, characteristics and the relationship of social contacts who agreed to participate in the study. We calculated the percentage of study participants who attended their colonoscopy appointments. We determined the percentage of participants with adequate bowel preparation as graded by the endoscopist in the colonoscopy reports. We also examined the opinion of the participants in intervention arm and their social contacts regarding their interaction. We used log-binomial models to the compare the attendance to colonoscopy appointments in the intervention group as compared to the control group (usual care). Among subjects who underwent colonoscopy, we also compared the percentage of participants with adequate bowel preparation by randomization assignment. Adequate bowel preparation was defined as a description of good, very good and excellent on the Aronchick scale and inadequate bowel preparation was defined as fair and poor designations. Our primary analysis was by intention to treat.

We performed an exploratory per-protocol analysis in which we restricted the subjects in the intervention arm to those whose social contacts agreed to participate 130/201 (64.7%) and repeated our analysis. We used Stata statistical software version 14.2 for all our analyses. We used log-binomial modeling for our outcomes and calculated the relative risks (RR) and 95% confidence intervals (95% CI). A p value less than 0.05 was considered to be statistically significant.

RESULTS

Baseline characteristics of participants

A total of 425 subjects were recruited, 26 subjects were excluded for ineligibility and 399 were included in the final analysis. CONSORT diagram (figure 1) shows the flow of participants through the study. The subjects in both arms were comparable except that there was a slightly higher percentage with annual household income less than $25,000 in the control group (P = 0.04) (Table 1).

Figure.

CONSORT diagram of flow of participants through the study

Table 1.

Characteristics of the study participants

Characteristics	Total (n = 399)	Control group (n = 198)	Social contact group (n = 201)	P value
Mean age in years	58.1	57.8	58.4	0.37
Sex, n (%)				0.80
Male	188 (47.1)	92 (46.4)	96 (47.8)
Female	211 (52.9)	106 (53.6)	105 (52.2)
Married, n (%)				0.81
No	306 (77.3)	154 (77.8)	152 (76.8)
Yes	90 (22.7)	44 (22.2)	46 (23.2)
Highest education, n (%)				0.64
High school or less	253 (63.6)	123 (62.4)	130 (64.7)
More than high school	145 (36.4)	74 (37.6)	71 (35.3)
Yearly household income, n (%)				0.04
More than $25,000	104 (26.3)	42 (21.5)	62 (30.9)
$25,000 or less	292 (73.7)	153 (78.5)	139 (69.1)
Smoking status, n (%)				0.23
Never	152 (39.0)	77 (39.5)	75 (38.5)
Former	114 (29.2)	50 (25.6)	64 (32.8)
Current	124 (31.8)	68 (34.9)	56 (28.7)
Has family history of CRC, n (%)				0.87
No	349 (89.0)	175 (89.3)	174 (88.8)
Yes	43 (11.0)	21 (10.7)	22 (11.2)
Had previous CRC screening, n (%)				0.74
No	239 (60.2)	117 (59.4)	122 (61.0)
Yes	158 (39.8)	80 (40.6)	78 (39.0)

Characteristics of the social contacts

In the entire study, suggested social contacts were 98.2% blacks and 68.5% females. In terms of relationship with the study participants, 68.2% were spouses or had first degree blood relationship such as siblings, parents and children (immediate family), 9% had second degree relationships such nieces, nephews, aunts, uncles and current in-laws (other relatives), while 22.9% were friends and other acquaintances such as co-workers, neighbors i.e. non blood, non-marital relationships (friends). In the intervention arm, the social contacts were 98.5% blacks and 66.8% female. In terms of relationship with the study participants, 63.5% were immediate family members, 10.9% were other relatives while 25.5% were friends.

A total of 130/201 (64.7%) agreed to participate in the trial to act as facilitators for the study participants. Of these 130 social contacts, 117 answered the question about their personal history of colonoscopy experience and 70/117 (59.8%) social contacts have had colonoscopy previously themselves. Of the 7½01 (35.3%), who did not participate in the study, we were unable to reach 63/71 (88.7%) due to non-response to phone calls, no returned calls when messages were left, disconnected / inactive phones and wrong phone numbers. Only 8/71 (11.3%) actually refused to participate in the study.

Social contact and study subject interactions

The perceptions of the interaction between the participants and their social contacts were comparable. From our brief telephone interviews of 83 participants in the intervention arm, 72 (86.7%) recalled talking with their social contacts about their procedures, 68/75 (90.7%) felt the social contact was interested in their interaction, 61/71 (85.9%) opined that the social contact was helpful, 57/71 (80.3%) felt the social contact helped them keep their appointments, 41/65 (63.1 %) felt the social contact encouraged them to be adherent with bowel preparation process and 40/69 (58%) related that the social contact also served as the escort. In our similar telephone interview of 72 social contacts regarding their interactions with the study participants, 64 (88.9%) recalled talking with the participant about the procedure, 61/66 (92.4%) felt the participant was interested in their interaction, 58/66 (85.9%) felt they helped the participants keep their appointments, 34/58 (58.6 %) felt they encouraged the participant to be adherent with bowel preparation process and 35/60 (58.3%) of the social contacts acknowledged serving as escorts for the study participants.

Colonoscopy attendance and bowel preparation results

Overall, in the intention to treat analysis, there was no difference in attendance to scheduled colonoscopy by randomization assignment. In the intervention group, 156/201 (77.6%) subjects underwent colonoscopy as compared to 152/198 (76.8%) in control group (RR = 1.01; 95% CI: 0.91 – 1.12, P value = 0.84) (Table 2). However, we noted a modest 8.2% absolute increase in adequacy of bowel preparation among subjects in the intervention arm (RR = 1.10; 95% CI: 1.00 – 1.21, P value = 0.046) (Table 3).

Table 2.

Completion of out-patient colonoscopy by randomization assignment

Randomization assignment	Did not undergo colonoscopy, n (%)	Underwent colonoscopy, n (%)	Risk ratio (95% CI)
Control (n = 198)	46 (23.2)	152 (76.8)	Reference
Social contact group (n =201)	45 (22.4)	156 (77.6)	1.01 (0.91 – 1.12)

Table 3.

Adequacy of bowel preparation among those who underwent colonoscopy by randomization assignment

Randomization assignment	Inadequate bowel preparation, n (%)	Adequate bowel preparation, n (%)	Risk ratio (95% CI)
Control (n = 152)	29 (19.1)	123 (80.9)	Reference
Social contact group (n = 156)	17 (10.9)	139 (89.1)	1.10 (1.00-1.21)

In our evaluation of sex as a biological variable, there was no difference in colonoscopy attendance and adequacy of bowel preparation by sex. Females had comparable attendance to colonoscopy (RR = 1.01; 95% CI: 0.91 – 1.13) and similar adequacy of bowel preparation (RR = 1.04; 95% CI: 0.95 – 1.14) as males.

Per protocol analyses

When we restricted our analysis only to those subjects whose social contacts were reached and agreed to participate (n = 130), the results were similar to that obtained in the intention to treat analysis. There was no difference in the attendance to colonoscopy (80% versus 76.8%, RR = 1.04; 95% CI: 0.93 – 1.17, P value = 0.48) and there was a modest 8.5% increase in the adequacy of bowel preparation (89.4% versus 80.9%, RR = 1.11; 95% CI: 1.00 – 1.22, P value = 0.054)

Other exploratory analysis

In the intervention arm, when we evaluated attendance to scheduled colonoscopy appointment as a function of the relationship of the subjects with their social contacts, there was no difference in the attendance to colonoscopy (82% versus 71.4%, RR = 1.15; 95% CI: 0.94 – 1.40, P value = 0.17) or adequacy of bowel preparation (89% versus 94.3%, RR = 0.94; 95% CI: 0.85 – 1.05, P value = 0.29) when the social contact was an immediate family member as compared to when the social contact was a friend or had non-marital, non-blood relationship. Similarly, there was no difference in colonoscopy attendance and bowel preparation when the social contact was a second degree relative (71.4% versus 71.4%, RR = 1.00; 95% CI: 0.72 – 1.38, P value = 1.00) and (80% versus 94.3%, RR = 0.85; 95% CI: 0.65 – 1.11, P value = 0.23), respectively, as compared to when the social contact was a friend or had non-marital, non-blood relationship.

We also found that previous colonoscopy experience by the social contact was borderline negatively associated with colonoscopy attendance (75.7% versus 89.4%, RR = 0.85; 95% CI: 0.72 – 1.00, P value = 0.049) as compared to when the social contact never had colonoscopy, but there was no association with adequacy of bowel preparation (86.8% versus 90.5%, RR = 0.96; 95% CI: 0.83 – 1.11, P value = 0.57).

DISCUSSION

Although it is well established that the use of patient navigators hired by healthcare institutions has been associated with an increase in compliance to CRC screening with colonoscopy among underserved populations, ¹⁴ the use of patient navigators in these settings typically has been through externally funded grants and demonstration projects in many minority serving institutions. The implication of this is that when such grants or external funds end, the patient navigation program is often terminated due to lack of resources committed to such endeavors in many of these institutions. Therefore, in this randomized controlled study, we evaluated the efficacy of using a patient’s self-selected social contact to serve as a facilitator, akin to a patient navigator, for completion of scheduled out-patient screening colonoscopy. Our premise was that if direct involvement of a patient’s social contact person as a facilitator is acceptable to patients and can also increase compliance to scheduled out-patient screening colonoscopy among an underserved population, it will provide a relatively inexpensive and readily available intervention to improve endoscopy delivery to the underserved. The use of social contact facilitators did not increase compliance to the scheduled screening colonoscopy among enrolled underserved blacks in our study, but we noted a modest improvement in the adequacy of bowel preparation, a quality indicator for colonoscopy.

Although the number of social contacts who refused to participate as a facilitator was low at 4% (8/201), less than two-thirds of the selected social contacts were eventually engaged in the trial due to not being able to reach them by telephone. Our finding that the participants and their social contacts opined at a high level that the parties were interested in their interactions to achieve a positive outcome was important and 58% of the chosen social contacts also served as an escort on the day of the procedure. Despite the fact that the main objective of our study was not achieved in terms of significant increase in attendance to the colonoscopy screening appointment, the improved bowel preparation quality suggests that there may be some limited roles for patients own social network to act as natural helpers in improving healthcare utilization among underserved blacks. Nonetheless, given the lack of efficacy of social contact facilitators, minority serving institutions should prioritize the use of sustainable, evidence-based methods to improve uptake of screening colonoscopy in underserved populations such as through the use of paid navigators.

Social support and social networks have played, and continue to play a substantial role in health as natural helpers.^{18, 19} Studies have suggested that social support from family members was associated with reduced smoking among youth ²⁰ and improved quality of life among those with depression.²¹ Although survival may not be impacted for patients with advanced cancers, ^{22, 23} robust social support has been associated with improved quality of life (QOL) among patients with cancer. ^24–26 For preventive services utilization specifically, it has been suggested that discussions between adolescent daughters and their mothers improved cervical cancer screening and may be explored as a health promotion initiative.²⁷ It is noteworthy that social support is mainly derived from family members,²⁸ but important health promotions do occur among African Americans through non-family social contacts such as cosmetologists,²⁹ and barbers.^{30, 31} Social networking and personalized contact methods (including word of mouth) have been reported to improve participation in fitness promotion research as well.³² This underscores the importance of the social environment, including interactions with family, friends, and the community as a whole to eliminate disparities in health outcomes.³³

We are not aware of any similar study in which a patient’s social contact was engaged to improve outpatient screening colonoscopy for a direct comparison with our study. However, a previous study in which a study employee acted as a peer coach demonstrated 11% increase in attendance when compared to a mailed colonoscopy brochure alone, ³⁴ and the use of patient navigators have been reported to increase colonoscopy attendance volume by 56% among Medicaid patients. ³⁵ It is noteworthy that studies of interventions to increase CRC screening were performed in different healthcare settings such as primary care setting ^{34, 36} and direct endoscopy access settings ³⁵ and combination of screening options were also used to determine study outcome ³⁶ whereas our study was conducted among patients who were seen in clinic by the gastrointestinal endoscopists and the outcome was attendance to colonoscopy screening.

An important strength of our study is that we studied an underserved, low income, urban blacks in a community with high incidence of CRC. It is noteworthy that the District of Columbia has among the highest incidence of CRC in the country. ³⁷ Furthermore, we achieved a high rate of participation among eligible patients and their social contacts.

Our study has some limitations. The study was conducted in a tertiary referral center and we focused only on non-Hispanic blacks since this group has the highest burden of the disease. There is also a possibility that the degree of social support influence may vary by race ethnicity and income. We have limited open access endoscopy and participants in our study underwent face to face consultations with the endoscopists which may increase adherence among the control group and reduce the effect of the intervention. However, the colonoscopy attendance rate (76.8%) and bowel preparation adequacy (80.9%) in the control group closely matched the baseline rate of 78% colonoscopy attendance and 80% adequate bowel preparation rate in our system. ¹⁷ Furthermore, we did not have information on the actual depth of interaction between the social contact and the patient detailing nature, extent, content and specific duration of their engagement with the patient.

In conclusion, we did not find an increased attendance with patients self selected social contact engagement as facilitators. However, there was a modest improvement in adequacy of bowel preparation at screening colonoscopy performed among compliant participants. Rather than investing in a social contact facilitators program, minority serving institutions should invest in sustainable models with a strong evidence base such as supporting paid patient navigators to improve healthcare delivery to underserved blacks. Future studies may evaluate comparison of efficacy of paid patient navigators versus patients’ social contact.

STUDY HIGHLIGHTS.

WHAT IS KNOWN

• Paid patient navigators improved colon cancer screening among underserved populations using externally funded grants.

• Many minority serving institutions do not hire paid patient navigators for preventive services delivery due to cost.

• Information about a contact person is usually collected from patients receiving healthcare services.

WHAT IS NEW HERE

• Patient’s self selected contact person was not an effective facilitator for improving attendance to screening colonoscopy among underserved blacks

• Engagement of patient’s self selected contact person modestly increased adequacy of bowel preparation for screening colonoscopy