Split- versus single-dose preparation tolerability in a multiethnic population: decreased side effects but greater social barriers

Introduction

Colorectal cancer (CRC) is the third leading cause of cancer death in both men and women in the United States, with an estimated 140,250 new cases and 50,630 deaths expected in 2018 [1]. Screening colonoscopy is known to decrease mortality from CRC [2,3] and the quality of the bowel preparation is critical to the success of the procedure [4]. However, concerns regarding bowel preparation have been one of the most commonly cited reasons for patients refusing to undergo screening colonoscopy. Bowel preparation has been consistently reported as one of the most unpleasant and burdensome parts of the procedure [5-8]. Poor quality bowel preparation has been shown to decrease the adenoma detection rate, slow the time to cecal intubation, prolong withdrawal time, and increase costs because of the need for repeat procedures [9-13]. Inadequate bowel preparation exposes patients to longer, less effective procedures and increases the financial burden on the patient and the hospital.

Efficacy and tolerability have been identified as the two essential “ingredients” of adequate bowel preparation [4]. To date, the majority of studies have focused primarily on efficacy rather than tolerability. Furthermore, most studies of bowel preparation have drawn on a patient population that tends to be racially and ethnically homogenous, native English-speaking, health literate, and privately insured. It is important to investigate potential barriers to colonoscopy in high-risk patient populations, including the uninsured, underinsured, those with limited English proficiency and patients with low health literacy served by safety-net hospitals. Thus, support for bowel preparation as a quality measure for colonoscopy is growing, suggesting it may soon play a role in reimbursement schemes in the emerging pay-for-performance climate and value-based care [14,15].

Safety-net hospitals, defined as those that predominantly treat the poor and underserved, have been shown to benefit less from quality performance measures than non-safety-net hospitals [16-18]. Bellevue Hospital Center (BHC) is a one of the largest safety-net public hospitals in the United States, as well as being the central referral center for the public hospital network of New York City. BHC is known for serving a racially and ethnically diverse high-risk immigrant population. Likewise, the Manhattan campus of the Veterans Affairs New York Harbor Healthcare system is a regional referral center serving a multiethnic veteran population. The issues recruiting underrepresented groups for research studies are well documented in the literature [17-19]. Studies of colonoscopy preparation quality measures including tolerability are similarly limited by a lack of inclusion of underrepresented groups.

Split-dose preparation has been shown to increase the quality of cleansing and has been associated with greater patient compliance in two recent meta-analyses, explaining why split-dosing has become an attractive quality measure [20,21]. The implementation of a split-dose preparation regimen and associated educational materials has even been shown to increase preparation quality in this safety-net hospital population [22]. However, there are only limited data concerning the impact of split-dose preparation on patient tolerability. The purpose of this study was to evaluate the determinants of tolerability between single- vs. split-dose bowel preparation in a socially diverse safety-net patient population.

Patients and methods

A prospective, dual-center study was conducted at BHC and the New York Harbor Healthcare Veterans Administration (VA) Manhattan Campus. All outpatients were asked to complete surveys upon presentation for colonoscopy. Surveys were available in English, Chinese and Spanish. Certified medical interpreter services were available to assist in administering the survey in other languages. Patients who declined to complete the survey were excluded from the study. The standard bowel preparation instruction was 4 L polyethylene glycol (PEG) administered as a split dose for patients at Bellevue and a single dose for patients at the VA, along with 20 mg of bisacodyl. At the time of study, the single-dose regimen was the standard practice at the VA. Per survey responses, a subset of the BHC population opted for a single-dose preparation. The study was classified as exempt after review by the Institutional Review Board across institutions.

Results

A total of 1291 patients returned the survey, of whom 1023 (79.2%) successfully specified whether the bowel preparation regimen was split or not and answered tolerability questions; these patients were included in the final analysis (Fig. 1). Patients were predominantly non-white, male, had Medicaid or were uninsured, reported a high-school education or less, and earned less than $25,000 per year (Table 1). More than one third of patients were not fluent in English, with 30.4% of respondents requiring the services of an interpreter to complete the survey.

Figure 1 Inclusion and cohort design

Of the entire study population, the most commonly reported social barriers to bowel preparation tolerability were concerns about medications (18.8%), availability of appropriate foods/drinks (18.6%), and patient responsibilities at home (18.6%). The most frequently reported moderate and severe gastrointestinal symptoms relating to colon preparation completion were stomach fullness, a bad taste in the mouth, and lack of sleep (Fig. 2A). Of the entire study population, 18.2% and 8.6% of patients reported that the colon preparation was “somewhat” or “very” intolerable, while 40.4% and 32.8% of patients reported that the preparation was “somewhat” or “very” tolerable, respectively.

Figure 2 Adverse effects. (A) Adverse effects for entire study cohort. (B) Gastrointestinal symptoms compared between single- vs. split-dose colon prep group

*Significantly different at P<0.05

The study population was stratified into two cohorts, single-dose (N=342, 33.4%) and split-dose (N=681, 66.6%) preparation, individually assessed for tolerability. The two cohorts were demographically similar, except for a slight predominance of older patients in the single-dose cohort (Table 1). Comparing tolerability between cohorts, patients reported single-dose preparation as “very intolerable” approximately twice as often as patients who completed split-dose preparation (12.6% vs. 6.6%; P=0.005, Table 2), with a significantly greater proportion of “severe” nausea/vomiting and stomach fullness in the single- vs. split-dose group (Fig. 2B).

Table 1 Patient demographics (%) compared between single- vs. split-dose colon preparation groups (N=1023)

Table 2 Tolerability of single- vs. split-dose preparation

Univariate analysis of the single-dose preparation cohort identified that tolerability was significantly associated with male sex (P<0.001), English fluency (P=0.005), instructions in the patient’s preferred language (P=0.016), and having an interpreter present for preparation instructions (P=0.007, Table 3). Medical comorbidities and social barriers did not correlate with tolerability in the single-dose preparation cohort. The four factors significant on univariate analysis (male sex, English fluency, nursing instructions in preferred language, and interpreter present) were included in a multiple logistic regression, of which male sex (OR 3.056, 95%CI 1.77-5.27; P<0.05) and nursing instructions in preferred language (OR 2.2, 95%CI 1.03-4.82; P<0.05) were independently associated with tolerability in patients who completed single-dose colon preparation (Table 3). The Hosmer-Lemeshow statistic was not significant (5.736; P=0.68), indicating that the model fit the data. In a sensitivity analysis, including race (P=0.28) as an additional factor in the regression equation did not change factor significance (P<0.05). Q. Please mention the exact P values; if P is less than 0.001, then quote P<0.001)

Table 3 Univariate analyses of patient demographics in single- and split-dose preparation cohorts, comparing patients who tolerated bowel preparation vs. those who could not

Male sex (P<0.001) and race (P<0.001) were significantly associated with patient-reported tolerability in the split-dose preparation cohort, with a greater proportion of white (19.7% vs. 10.0%) and black (25.5% vs. 18.8%) patients tolerating a split dose than Hispanic patients (35.1% vs. 55.0%; Table 3). Older age (P=0.051) and English fluency (P=0.077) also showed a trend toward tolerability, with more patients in the 60-69 year-old age group (35.4% vs. 24.2%) than in the 50-59 year-old age group (40.4% vs. 50.9%) tolerating split-dose preparation. Medical comorbidities, interpreter services (P=0.106) or respondents reporting nursing instructions in preferred language (P=0.878) did not differ between patients who could or could not tolerate split-dose colon preparation. However, patient-reported responsibilities at home (P=0.007) and concerns about medications (P<0.001) were significantly more common in patients who could not tolerate split-dose colon preparation (Table 4). Factors significant on univariate analysis (male sex, race, responsibilities at home, concerns about medications) were included in a multiple logistic regression, which revealed that male sex (OR 2.40, 95%CI 1.64-3.52; P<0.05) and concerns about medications (OR 0.55, 95%CI 0.35-0.86; P<0.05) were independently associated with tolerability in patients who completed split-dose colon preparation (Table 5). The Hosmer-Lemeshow statistic was not significant (8.25; P=0.41), indicating that the model fit the data. In a sensitivity analysis, including either race (P=0.33) or race (P=0.422) and English fluency (P=0.423) as additional factors into the regression equation did not change factor significance (P<0.05).

Table 4 Social barriers in single-dose vs. split-dose preparation by tolerability

Table 5 Multiple regression analysis

Discussion

In a large safety-net patient population, split-dose preparation was associated with significantly greater tolerability with fewer side effects than single-dose preparation. Males were more likely to tolerate either bowel preparation regimen. Language-related instructional factors more commonly affected patient tolerability of single-dose preparation, while social barriers, such as medication concerns, more commonly affected the tolerability of split-dose preparation.

Split-dose preparation has been previously associated with greater tolerability in relatively homogenous populations [20,21,25], and this study extends those findings to a multiethnic and socio-economically diverse cohort. In contrast to the current data, previous studies primarily focusing on efficacy endpoints did not report an association between split-dose preparation and tolerability [26,27]. Tellez-Avila et al reported that in-patients undergoing split-dose preparation tended to report fewer symptoms than those undergoing single-dose [28]. This study further found that split-dose preparation was associated with less severe symptoms (Fig. 2B).

Male patients were more likely to report a tolerable bowel preparation experience than female patients, consistently with previously published literature [29,30]. These earlier studies are unlikely to represent the diverse patient population included in this study, further suggesting that sex is an independent risk factor for bowel preparation tolerability. Earlier studies have speculated that men may simply have a higher tolerance threshold than women. However, we argue that men may simply be more likely to endorse tolerability and may in fact underreport the challenges of bowel preparation, presenting a complicated challenge for providers.

To date, there has been little investigation into the impact of language and race on bowel preparation tolerability; thus, our findings associating single-dose preparation with language barriers and split-dose preparation with race are novel. It is somewhat surprising that language barriers were significantly associated with single-dose and not the more complicated split-dose preparation. This highlights one limitation of our study, that patients facing language barriers may have been more likely to self-select the simpler single-dose preparation, introducing a potential patient selection bias that would have been best addressed by a randomized intention-to-treat analysis. Further studies of multilingual patient populations would help further elucidate these connections.

This study also sheds light on social barriers to tolerability that appear to be unique to split-dose bowel preparation. Taking into account the increased time and required morning of commitment, our study suggests that split-dose preparation would be more likely to interfere with domestic responsibilities and existing morning medication regimens. Another recent study assessing split-dose regimen uptake also found that social factors, including morning appointments, travel time to endoscopy, and low education level, were associated with limited uptake of the split-dose regimen [31]. It may be important for health systems and providers to specifically address these social barriers to ensure bowel preparation tolerability, given that split-dose is currently the standard of care. Furthermore, it will be critical to consider the variable impact of quality measures, as they may impact safety-net hospital populations differently compared to the homogenous and insured populations, more represented in the literature. Our study focused on patient-reported outcomes, intending to amplify the needs and concerns of this underrepresented and understudied population.

Major strengths of our study are the large sample size, the significant diversity of the population, and results based solely on patient-reported outcomes. In addition, our results are generalizable to the real-world setting, as our study population encompassed two of the largest healthcare organizations in America today: the VA and the public hospital system. The focus of this study on patient-reported determinants of tolerability is an additional strength that diversifies the existing knowledge on bowel preparation. However, while our questionnaire was constructed using a previously validated survey, the study format excluded patients without proficient reading and writing abilities in one of the few languages in which our survey was available. Furthermore, the design of our recruitment strategy did not allow for the number of patients who decline to participate to be counted, which may have introduced a response bias. Lastly, the discrepancy in regimen instructions between the VA and public hospitals, as well as issues relating to patients’ self-selection of a preparation regimen that could be age-related, may limit the generalizability of our findings. The split- and single-dose cohorts were well-matched (Table 1), but the possibility of sampling bias cannot be excluded. Despite these limitations, this study represents a critical early step towards a growing understanding of the barriers to tolerable bowel preparation in a broader and more diverse population than previously considered.

In conclusion, there are multiple factors affecting split-dose preparation tolerability in a racially diverse, safety-net population. Given its poor health literacy, there is concern that this population may be unable or unwilling to perform split-dose preparation. We have previously shown that a split-dose regimen increases the adequacy of bowel preparation [22]. A recent randomized controlled trial also found that a split-dose regimen increases the adenoma detection rate, making uptake of this regimen essential [32]. In this study, we demonstrated that split-dose preparation was more tolerable than single-dose preparation and reduced symptom severity. However, this more complicated regimen also introduces additional social barriers of which providers need to be aware. Our findings are in agreement with the current recommendations for the use of split-dose preparation from the United States Multi-Society Task Force on optimizing bowel cleansing for CRC screening [33]. Further studies of emerging quality measures in colonoscopy preparation, including this patient population, are warranted.