How Efficacious Are Patient Education Interventions to Improve Bowel Preparation for Colonoscopy? A Systematic Review

Background Bowel preparation is inadequate in a large proportion of colonoscopies, leading to multiple clinical and economic harms. While most patients receive some form of education before colonoscopy, there is no consensus on the best approach. Aims This systematic review aimed to evaluate the efficacy of patient education interventions to improve bowel preparation. Methods We searched the Cochrane Database, CINAHL, EMBASE, Ovid, and Web of Science. Inclusion criteria were: (1) a patient education intervention; (2) a primary aim of improving bowel preparation; (3) a validated bowel preparation scale; (4) a prospective design; (5) a concurrent control group; and, (6) adult participants. Study validity was assessed using a modified Downs and Black scale. Results 1,080 abstracts were screened. Seven full text studies met inclusion criteria, including 2,660 patients. These studies evaluated multiple delivery platforms, including paper-based interventions (three studies), videos (two studies), re-education telephone calls the day before colonoscopy (one study), and in-person education by physicians (one study). Bowel preparation significantly improved with the intervention in all but one study. All but one study were done in a single center. Validity scores ranged from 13 to 24 (maximum 27). Four of five abstracts and research letters that met inclusion criteria also showed improvements in bowel preparation. Statistical and clinical heterogeneity precluded meta-analysis. Conclusion Compared to usual care, patient education interventions appear efficacious in improving the quality of bowel preparation. However, because of the small scale of the studies and individualized nature of the interventions, results of these studies may not be generalizable to other settings. Healthcare practices should consider systematically evaluating their current bowel preparation education methods before undertaking new interventions.


Introduction
Each year, patients in the United States undergo nearly 14 million colonoscopies [1]. These resource-limited, costly, and potentially harmful procedures are mostly performed in asymptomatic patients for prevention and early detection of colorectal cancer (CRC) [2]. The effectiveness of preventive colonoscopy depends on adequate preparation (i.e., cleansing) of the large bowel to allow for visualization of the mucosal surface and detection of adenomas, the pre-malignant lesions that lead to colon cancer [3]. Yet, a large fraction of patients have unsatisfactory or inadequate bowel preparation. A meta-analysis of trials using split-dose polyethylene glycol-based bowel preparations, the current standard of care, found 23% of patients had an unsatisfactory preparation [4] while observational studies yield varying estimates (17-33%) [5][6][7]. Inadequate bowel preparation leads not only to missed adenomas but also to incomplete procedures, longer procedure times, shorter intervals between procedures, increased cost, and patient inconvenience and harm [7][8][9][10].
With the complexity of bowel preparation, patient comprehension of preparation instructions is critical. Lack of comprehension may contribute to non-compliance, a strong patientlevel predictor of sub-optimal bowel preparation with split-dose preparations [11]. In one single-center observational study, nearly half of patients undergoing screening colonoscopy reported that they received no verbal instruction and nearly one-third reported that they received no written instruction for bowel preparation [11]. Others have reported that up to one in five patients may not follow recommendations for a split-dose bowel preparation [12].
Multiple studies have evaluated the efficacy of patient education interventions to improve the quality of bowel preparation. These studies have used various delivery platforms; yet, there is no consensus on how best to educate patients before colonoscopy. A recent recommendation from the US Multi-Society Taskforce on Colorectal Cancer recommended that all patients receive both oral and written instructions before colonoscopy but did not specify how or when these instructions should be delivered [13]. With this lack of standardization, endoscopy centers often develop educational materials and processes locally, resulting in significant variation in the instructional content of written bowel preparation instructions [14]. In this context, the aim of this systematic review is to evaluate the efficacy of patient education interventions on the quality of bowel preparation among adults undergoing colonoscopy in prospective, controlled studies.

Study selection
Before commencing the search, an outline of the search strategy was submitted to and approved by PROSPERO, an international prospective database of systematic reviews [15]. We searched CINAHL, EMBASE, Ovid, Web of Science, and the Cochrane Database of Systematic Reviews from database inception to August 2014. CINAHL was used to search the nursing and allied health literature. EMBASE was used to search European literature. Abstracts from meetings of the American Gastroenterology Association, American Society of Gastrointestinal Endoscopy, and American College of Gastroenterology were searched using EMBASE. The search strategy was not restricted by language of the article.
The search strategy for each database was designed in consultation with a health librarian to maximize search sensitivity and was iteratively refined to ensure that all known articles relevant to the topic were captured. The searches combined the terms colonoscopy, patient education/ counseling, and bowel preparation, or synonyms of those terms. For example, the Ovid search included the terms (endoscopy OR colonoscopy) AND (purgative$ OR clean$ OR cathartic$ OR cathartics) AND (educat$ OR instruct$ OR teach$ OR counsel$ OR patient education). Details on the search strategies are provided in S1 Table. We also screened all articles that cited, or were cited by, studies ultimately included in the review. Additionally, we searched clinicaltrials.gov for active clinical trials, and we contacted three content experts who have previously published on this topic for information about potential published or unpublished studies relevant to the review that were not captured by our search strategy.
Studies were evaluated for six inclusion criteria: (1) testing of a patient education intervention, defined as a structured or formal education or counseling intervention intended to improve adherence [16,17]; (2) a primary aim of improving bowel preparation quality; (3) assessment of bowel preparation using one of three validated scales [18][19][20]; (4) prospective design; (5) inclusion of a concurrent control group, including usual care; and, (6) study participants limited to adults at least 18 years of age. Studies that examined text messaging reminders without additional educational content were not considered educational interventions and were therefore not included. Studies of patient navigation interventions, which are typically multicomponent interventions intended to address individual patient barriers, were also excluded even if one aim was improvement of bowel preparation (since the impact of education could not be separated from the remaining components of the intervention). When the search strategy identified more than one abstract pertaining to the same intervention, a published article (if one existed) or the most recent abstract was used.
All abstracts were screened independently by two members of the study team (JK and AS) to determine whether they met inclusion criteria. If the abstract appeared potentially relevant, the full text article, if one existed, was reviewed to determine whether it met inclusion criteria. For studies published in abstract but not full-text form that appeared relevant, the authors were contacted for additional information. If the author of an abstract was not responsive to questions or did not provide enough information to establish that the abstract met inclusion criteria, it was excluded. Disagreements about inclusion criteria were adjudicated through a consensus process involving a third study team member (SS).

Data Extraction and Strength of Body of Evidence
For studies published in full text that met inclusion criteria, a structured form was used to extract data on patient sample (size and inclusion/exclusion criteria), study design, study setting (year, location), content and format of both intervention and control, bowel preparation used, and outcomes. When necessary, we contacted authors by email for additional study details. Since the majority of studies reported preparation scores as dichotomous or categorical variables, we contacted the authors of studies that reported only means or medians for additional data. Studies were rated for methodological quality using the Downs and Black scale, which is endorsed by the Cochrane handbook for the assessment of the validity of randomized and non-randomized studies [21,22]. The original instrument included 27 questions. For this review, the instrument was modified in three ways. First, a single question about methodological adjustments for length of follow up was excluded since outcomes in all cases were determined at the time of colonoscopy. Second, a question on whether principal confounders across groups were described not at all (0 points), partially (1 point), or completely (2 points) was defined more specifically (0 points if no confounders were described; 1 point if only age and gender were described; and 2 points if additional variables were described). Third, a question about reporting of adverse events was counted as positive (1 point) if non-attendance at the colonoscopy appointment and/or cancellations were reported since the educational interventions would not be expected to cause clinical adverse events but could conceivably affect appointment attendance. After these modifications, a maximum of 27 points were possible. For abstracts and research letters that met inclusion criteria, only basic study characteristics were extracted given the limited available information.

Results
Literature selection overview 1,080 unique abstracts were retrieved and reviewed by hand (Fig 1). 16 articles were reviewed in full-text form, five of which were excluded because they did not use a validated preparation scale [23][24][25][26][27], three because they lacked a primary aim to improve preparation quality [28-30], and one because it did not include a patient education intervention [31]. We attempted to contact the authors of 12 abstracts or letters, eight of whom supplied additional information. No additional studies were identified through clinicaltrials.gov or by content experts. Ultimately, 12 studies met our inclusion criteria, including seven full-text articles, four abstracts, and one research letter. No additional studies were identified after hand searching the references of the seven full-text articles, or after searching for articles that subsequently cited any of the seven articles.

Study characteristics
Characteristics of the seven full-text articles are shown in Table 1

Effect on bowel preparation
Six of the seven studies showed a positive effect of the intervention on the quality of bowel preparation ( Table 3). The magnitude of effects could not be compared or summarized across studies since there was significant statistical and clinical heterogeneity (e.g, diverse patient samples and intervention formats). This precluded meta-analysis. Studies also used three different bowel preparation scales and reported preparation scores with means, medians, and/or dichotomous categories with proportions. In the one study using the Aronchick scale, the rate of "good" or "excellent" preparations increased by 32% in the treatment group compared to usual care [35]. Among the three studies using the Ottawa Bowel Preparation Scale (in which a lower score is indicative of superior preparation), one study showed that the rate of scores

Non-full text studies
Among the four abstracts and one research letter that met study inclusion criteria (Appendices 3-5), three reported on educational videos [40][41][42], and two reported on printed instructions presented as a brochure and "photo aids" [43,44]. Four of the five non-full text studies reported

Discussion
Colonoscopy is an important preventive, diagnostic, and therapeutic modality, but its effectiveness hinges on successful pre-procedure bowel preparation [13]. When bowel preparation fails, the harms are numerous, including shorter interval repeat procedures, increased expense, missed neoplastic lesions, patient inconvenience, and increased risk of procedural adverse events [7-10,45]. However, the most efficacious manner in which to educate patients about bowel preparation is not known. Our review indicates that multiple different patient education interventions have been demonstrated to improve bowel preparation quality. All but one of the seven published studies resulted in a significant improvement in bowel preparation scores compared to usual care. The single study that failed to show a positive effect tested one of the least intensive educational interventions (a visual aid depicting adequate and inadequate bowel preparations), and had a relatively high rate of treatment contamination [32]. However, these data do not demonstrate the superiority of any one of these strategies over the others. Rather, we can conclude that multiple strategies have the potential to improve bowel preparation. Local factors, such as current bowel preparation education practices, patient characteristics, and staff availability, are likely to affect which intervention is most effective in a given healthcare setting. Before deciding on a specific intervention, healthcare practices might consider rigorously analyzing their current bowel preparation approach, and evaluating the impact of any interventions as part of a Plan-Do-Study-Act (PDSA) cycle [46]. While our review suggests that many educational interventions are efficacious in improving bowel preparation prior to colonoscopy, it also highlights important methodological concerns. Under the Affordable Care Act, Accountable Care Organizations will face greater pressure to provide cost-effective colorectal cancer screening for their patient populations [49]. This pressure will likely come in the form of increasingly prevalent bundled payments for screening [50] and colonoscopy performance measures [51]. Rates of adequate bowel preparation, adenoma detection and cecal intubation will likely benefit from improved bowel preparation education [3]. In this context, practices considering educational interventions to improve bowel preparation should consider several factors. First, the cost-effectiveness of physician-staffed interventions is likely to be lower than medical assistant-staffed interventions, videos, or brochures. Second, the medium (e.g., online versus printed materials) should ensure accessibility across a full range of age and socioeconomic status. Third, interventions that can be delivered while a patient is present in clinic may be more reliable than interventions that depend on the patient to take additional steps afterward.
The preparation for colonoscopy is more complex, and more important than the preparation for almost any other diagnostic test in medicine. It is therefore incumbent upon gastroenterologists to ensure that all patients undergoing colonoscopy, whether self-referred or open access, receive effective teaching beforehand. While the studies reviewed do not provide definitive guidance on how best to educate patients for colonoscopy, in the opinion of the authors, the first step should be clear and concise written instructions, since every patient likely receives some form of written instructions. Indeed, in six of the seven studies reviewed, the treatment group received standard written instructions along with the patient education intervention.
Healthcare practices may consider adoption of new instructions or review and revision of existing instructions using accepted evaluation tools such as those produced by the Agency for Healthcare Research and Quality (AHRQ) [52]. Ideally high quality written instructions would be combined with an opportunity for verbal in-person or telephone education by an ancillary medical provider, at least for patients at increased risk for preparation failure. When it comes to education about bowel preparation, healthcare practices should recall general principles of health education, including the importance of teach-back, encouraging questions, and hands on demonstration, which are all topics appropriate for review with clinic staff. As noted by AHRQ, "Health materials are effective only when used as part of an overall patient education strategy. . . If the information is critical, make sure you or someone in your office reviews the information with your patient and/or the patient's caregiver" [52].
Our review highlights several areas for future research. First, few of the studies that met criteria for inclusion in the review comprehensively assessed known risk factors for colonoscopy preparation failure (e.g., comorbidities, medications, socio-economic status, insurance type, literacy). Future studies should systematically assess these risk factors, which could potentially allow for identification of heterogeneity of treatment effects and lead to targeted approaches that match intervention to selected patient characteristics. Second, only two of the studies included in the review used intention-to-treat analysis, limiting the ability to assess adherence to the interventions and to make conclusions about real-world effectiveness. Third, along the same lines, future studies should systematically assess patient adherence to the bowel preparation, in addition to the education intervention, which will provide further support for efficacy. Fourth, while all three of the studies that used split-dose bowel preparation showed significant effects, others have suggested that educational interventions may not be as efficacious with split dosing [53] (NB: This study used the same brochure as Spiegel, et al. and was published only in abstract form, not meeting criteria for inclusion in this review). Split-dose bowel preparation, which has become standard of care, should be used in future studies of educational interventions. Fifth, future studies would benefit from larger sample sizes with adequate power to detect changes in clinically meaningful outcomes, such as adenoma detection and need for repeat colonoscopy, as well as economic outcomes, such as time and cost [34]. Finally, future studies should carefully consider the advantages and disadvantages of the available bowel preparation scales [54]. For evaluation of patient education interventions to improve bowel preparation, the scale used to measure preparation quality should ideally assess preparation quality before lavage by the endoscopist rather than after, so that the effect of the intervention can be evaluated independently. The scale should also have an established cutoff to distinguish between adequate and inadequate preparation (a dichotomous characterization that is endorsed by GI professional societies but is rarely standardized).
Others have published meta-analyses of patient education interventions to improve bowel preparation [55,56]. Rather than focusing on efforts to address this topic quantitatively with meta-analysis (an approach that we believe is inappropriate in the context of substantial clinical heterogeneity, as highlighted above), our review adds to this existing literature by calling attention to gaps in current knowledge through the lens of health education, discussing important issues of study design for future studies, and helping to contextualize how existing studies, with their limitations, might still be used to improve clinical practice. We would refer readers to the meta-analysis by Guo, et al [56], for a listing of several more recent studies published after the conclusion of our search, which all involved smartphone applications or mobile message reminders, which were intentionally excluded from this systematic review, and the findings of which do not substantially alter our conclusions.

Limitations
The findings of this review should be interpreted with some caveats. First, because of statistical and clinical heterogeneity, meta-analysis could not be performed. Second, the studies included in the review may be subject to publication bias. We attempted to minimize this risk by searching relevant conference proceedings and including unpublished studies. Third, our results are subject to the same threats to validity that exist in the constituent studies.

Clinical implications and future directions
This review suggests that several different patient education interventions are efficacious in improving the quality of colonoscopy preparation compared to usual care conditions. These include illustrated brochures, videos, education groups, and re-education phone calls. While none of these interventions has sufficient evidence to recommend broadly, healthcare practices should be aware of these education options as they consider strategies for improving bowel preparation quality. This topic is likely to grow in importance as colonoscopy preparation becomes a greater focus of quality measures and payer reimbursement.