Enhanced Patient Education for Colonic Polyp and Adenoma Detection: Meta-Analysis of Randomized Controlled Trials

Introduction

Background

Colorectal cancer (CRC) is the third most common cancer and the second cause of cancer-related mortality among both sexes worldwide, with 1.8 million new cases and 0.88 million deaths in 2018 [1]. Colonoscopy is recommended as the principal approach for decreasing CRC incidence and associated mortality by detecting and then removing the precancerous lesions [2-5]. However, adequate bowel preparation is the prerequisite for a successful colonoscopy [6]. Evidence revealed that inadequate bowel preparation was associated with increased risk of missing colonic lesions, prolonged procedural time, and lower cecal intubation rate [7,8]. Issued data suggested approximately 18% to 30.5% of inadequate bowel preparation in patients undergoing colonoscopy [9-11]. Therefore, it is particularly urgent to apply an effective intervention to improve the quality of bowel preparation [12].

Previous studies have determined various factors that were linked to the quality of bowel preparation, such as the type of diet restriction, type of colon cleansing solutions ingested, methods of ingesting the solution, and patient’s adherence to the solution [10,13-15]. Adequate comprehension of details of instructions is a major contributor to the quality of colon cleansing because bowel preparation is very complex [16]. Patients usually receive written booklet and/or verbal instructions from professionals before colonoscopy for bowel preparation and dietary restriction, which are defined as standard patient education [2]. However, the effect of standard patient education on bowel preparation is not enough [10]. To improve the patient’s comprehension of bowel preparation instructions, enhanced patient education (EPE; such as cartoon pictures, phone calls, mobile apps, and social media apps) has been proposed and then tested [2]. So far, several meta-analyses have evaluated the efficacy of EPE in improving the quality of bowel preparation and demonstrated an improvement [2,16-19]. However, evidence revealed that adequate bowel preparation provides good colonoscopy vision and thus increases the detection rate of colonic polyps and adenomas [20-23]. The fact that 2 meta-analyses evaluated the efficacy of EPE interventions to detect polyps and both did not find significant differences is discouraging [2,17]. As a result, the magnitude of benefit of EPE interventions in detecting colonic polyps and adenomas remains uncertain. It is noteworthy that recently, several randomized controlled trials (RCTs) reporting conflicting results have been published. More importantly, as most CRCs transform from polyps and adenomas, early detection and then removal of premalignant colonic polyps and adenomas is crucial [19]. Thus, as one of the most important colonoscopy quality metrics, the detection of colonic polyps and adenomas should be primarily measured and evaluated [24].

Objective

The aim of this meta-analysis was to evaluate the efficacy of EPE interventions in detecting CRC precursor polyps and adenomas compared with standard patient education.

Methods

Methodological Standard

We conducted this meta-analysis according to the methods proposed by the Cochrane Collaboration [25] and reported the pooled estimates in accordance with the framework proposed by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement [26]. There was no formal protocol for this meta-analysis.

Search Strategy

A systematic search was performed in PubMed, EMBASE, and the Cochrane Central Register of Controlled Trials from their inception to June 2019 for the identification of relevant RCTs. All search strategies were built using Exploded Medical Subject Headings and the appropriate corresponding text words. Language and status of publication were not restricted. We have summarized the details of the full search strategy in Multimedia Appendix 1. We updated the search results on August 10, 2019. The bibliographies of previous meta-analyses and eligible studies were also manually checked to identify additional potentially eligible studies.

Study Selection

We used the following inclusion criteria to enroll any eligible studies in this study: (1) all adult participants aged more than 18 years who were instructed to receive elective outpatient colonoscopy, regardless of morning and afternoon colonoscopy; (2) the patients assigned in the study group were instructed with EPE regimes, and the ones enrolled in the control group were instructed with standard patient education regimes; (3) the eligible study design was RCTs; however, an abstract with sufficient information was also considered; and (4) studies published in English. Two investigators (XT and LX) independently searched citations; excluded duplicates; checked the titles and abstracts for eligibility; and then categorized the studies as included, excluded, or requiring further full-text assessment. We excluded duplicates with poor quality or relatively insufficient data. We also excluded conference abstracts without sufficient information. A third senior investigator (WC) was consulted for a final decision if there was any disagreement between the 2 investigators.

Data Extraction

Two independent investigators (XT and LX) were assigned to use a standardized Word (version 2013, Microsoft Office, Microsoft Corporation) table to extract essential data, and then, they completed the cross-checking of corresponding results. The following data were extracted: basic characteristics of eligible trials including leading author, publication year, country, and financial sources; risk of bias criteria based on the Cochrane Collaboration risk of bias tool [27]; and clinical characteristics including age, sex, sample size, indication for colonoscopy, details of diet restriction and colon cleansing solutions, details of education interventions, and outcomes of interest. A third senior investigator (WC) was consulted for a final decision if there was any disagreement between the 2 investigators.

Outcome Variables and Definitions

We defined the colonic polyp detection rate (PDR) and adenoma detection rate (ADR) as the primary outcomes, which were defined as the number of patients with at least one polyp or adenoma divided by the total number of patients undergoing selective colonoscopy. We considered the advanced adenoma (defined as adenoma ≥10 mm) detection rate (AADR), sessile serrated adenoma detection rate (SSADR), and cancer detection rate (CDR) as secondary outcomes, which were defined as the number of patients with at least one advanced adenoma, sessile serrated adenoma, or cancer divided, respectively, by the total number of patients undergoing selective colonoscopy. We also considered ADR-plus, which was defined as the number of additional adenomas found after the first adenoma per colonoscopy [28], and the right and left colon polyp and ADR, which was defined as the number of patients with at least one right and left colon polyp and adenoma divided by the total number of patients undergoing selective colonoscopy, as secondary outcomes.

Assessment of Risk of Bias

We assigned 2 independent investigators (XT and XL) to appraise the risk of bias with the Cochrane risk of bias tool (the Cochrane Collaboration) [27]. An individual trial would be labeled as low, unclear, or high risk of bias according to the following criteria: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. A third senior investigator (WC) was consulted for a final decision if there was any disagreement between the 2 investigators. Following the recommendations proposed by the Cochrane Collaboration, a trial was considered a high-level trial when all key domains are rated as having a low risk of bias, a trial was considered a low-level trial when any one or more key domains are rated as having a high risk of bias, and otherwise, a trial was considered a moderate-level trial.

Data Analysis

We expressed summary estimates as a risk ratio (RR) or odds ratio (OR) with 95% CI. Heterogeneity was measured by the I² statistic, which describes the percentage of total variation across studies that is due to heterogeneity rather than chance [29]. We performed all statistical analyses using random effects model regardless of heterogeneity. In addition, subgroup analyses for the primary outcomes were conducted according to geographical regions (Western vs Asian) and indications (screening vs diagnostic vs mixed). We also conducted subgroup analysis for primary outcomes according to the administration method of ingesting colon cleans solutions (single dose vs split dose) because Zawaly et al [30] demonstrated that split dose compared with single dose was associated with increased detection of adenomas and advanced adenomas. P<.05 was considered statistically significant, except where otherwise specified. All statistical analyses were performed using Cochrane Review Manager (RevMan, version 5.3.5, 2014; the Nordic Cochrane Centre, the Cochrane Collaboration) [31]. For pooled estimates of SSADR, we used inverse-variance statistic due to various data reported in analyzed trials, and for remaining pooled estimates, we used the Mantel-Haenszel model. An RR or OR value greater than 1 indicates that there was a higher detection rate of the specified colonic polyp and adenoma. We planned to assess publication bias if any pooled group consisted of 10 or more trials [32].

Trial Sequential Analysis

The magnitude of efficacy of summary estimates from cumulative meta-analyses and the risk of type I error are susceptible to repetitive hypothesis test of accumulating scarce information [33]. Thus, trial sequential analysis (TSA), which has the potential of constantly adjusting the significance level and then drawing monitoring boundaries and calculating adjusted information size, was proposed to address issues faced by the traditional meta-analysis [34-36]. The conclusion is conclusive if the accumulative sample size is more than the adjusted information size and the Z-curve is across the trial sequential monitoring boundary or futility boundary. We conducted TSA to test the robustness of summary estimates of primary outcomes according to an alpha error of .05, a beta error of .20 (a power of 80%), and an anticipated intervention effect of 20% relative risk reduction using TSA version 0.9 beta (Copenhagen Trial Unit, Center for Clinical Intervention Research) [37,38].

Results

Literature Search

Figure 1 depicts the retrieval and selection of records. Initial search captured 588 records in 3 targeted databases. All records were imported to EndNote software (Thomson Reuters), and then, we deleted 110 duplicate records after running the finding duplicates function. We excluded additional 404 records after checking the titles and abstracts because of the following reasons: systematic review and meta-analysis and irrelevant to the analysis. We omitted 64 studies after carefully double-checking the full text in the remaining 74 studies because of the following reasons: 9 articles investigated a topic unrelated to this study, 1 article was a letter to the editor, 3 articles were editorials, 1 article was a duplicate publication, 1 article was a comment on published article, 36 were conference abstracts without sufficient information, 9 articles did not report essential outcomes or data that were considered in our study, and 4 articles used ineligible study design. We, thus, included 10 eligible RCTs in the final meta-analysis after checking the full text for eligibility [9,39-47].

‎View this figure

Study Characteristics

We document the details of basic characteristics of the 10 eligible RCTs in Table 1. All trials were reported between 2011 and 2018. The sample size in individual trial ranged from 94 to 969 (a total of 4560 participants). Of the 10 eligible RCTs, 3 [39,41,46] were from Western countries, including the United States [39,41] and Germany [46], and 7 [9,40,42-45,47] were from Asian countries, including China [9,43,47], Korea [40,45], and South Korea [42,44]. In total, 8 RCTs [9,39-41,43-46] were designed with two arms and remaining 2 RCTs [42,47] with three arms. The participants in 5 RCTs [39,40,42,44,45] received screening colonoscopy; in 4 RCTs [9,41,43,46] received screening, surveillance, or diagnostic colonoscopy; and in 1 RCT [47] received diagnostic colonoscopy. A total of 8 RCTs [9,40-44,46,47] described the details of diet restriction, and other 2 RCTs [39,45] did not report on the diet. Of the 10 included RCTs, 9 reported PDR as outcome [39-47], 5 reported ADR as outcome [9,41,42,46,47], 3 reported AADR as outcome [9,46,47], 3 reported SSADR as outcome [9,41,46], 2 reported CDR as outcome [9,43], and 1 reported ADR-plus as outcome [46].

Risk of Bias

Details of the risk of bias of individual trial are summarized in Table 2: 7 trials were rated as low level [39-42,44-46], 2 were moderate level [9,43], and 1 was high level [47]. A total of 8 trials appropriately generated randomization sequence [9,39,41-43,45-47], and 5 trials correctly conducted allocation concealment [9,42,43,46,47]. All trials [9,40-47] appropriately blinded endoscopist and reported anticipated outcomes.

Primary Outcomes

Figure 2 depicts the summary results of primary outcomes. Meta-analysis based on a random effects model suggested an increase in the detection rate of polyps (9 trials; 35.55% [715/2011] of participants vs 30.23% [535/1770] of participants; RR 1.19, 95% CI 1.05-1.35; P=.008; I²=42%) and adenomas (5 trials; 22.72% [271/1193] of participants vs 16.5% [155/940] of participants; RR 1.37, 95% CI 1.15-1.64; P<.001; I²=0%) in patients undergoing EPE. Subgroup analyses indicated that EPE increased the PDR in Asian patients; in patients ingested solutions with split dose and single dose; and in mixed patients undergoing screening, surveillance, and diagnostic colonoscopy, and that EPE increased the ADR in all patients regardless of geographical regions and in patients ingested solutions with split dose. The summary of subgroup analyses is shown in Multimedia Appendix 2.

TSA suggested that the accumulative Z-curve crossed the trial sequential monitoring boundary for benefit after the eighth trial of PDR (Figure 3) and after the fourth trial of ADR (Figure 4), showing that currently, the cumulative evidence for PDR and ADR is conclusive.

‎View this figure

‎View this figure

‎View this figure

Secondary Outcomes

Figure 5 delineated the summarized results of the secondary outcomes. Meta-analysis showed no significant difference in AADR (3 trials; 5.6% [50/892] of participants vs 3.8% [29/757] of participants; RR 1.54, 95% CI 0.67-3.55; P=.31; I²=57%) and CDR (2 trials; 1.59% vs 1.02%; RR 1.54, 95% CI 0.60-3.98; P=.37; I²=0%), and a significant difference in SSADR based on inverse-variance model (3 trials; OR 1.76, 95% CI 1.22-2.53; P<.05; I²=0%) between the EPE and standard patient education groups. Only 1 trial reported the ADR-plus, and adjusted estimate found a superior result in the EPE group (1 trial; RR 4.39, 95% CI 2.91-6.61; P<.001) [47]. One trial reported ADR according to segments of colon (right vs left) and did not find a significant difference between EPE and standard patient education interventions (right colon: RR 1.86, 95% CI 0.66-5.24 and left colon: RR 2.45, 95% CI 0.59-10.11) [41].

Publication Bias

Although the accumulated number of analyzed trials for all outcomes was less than 10, we also constructed the funnel plot for PDR because 9 trials were incorporated into this outcome. The funnel plot is not symmetrical, and the publication bias cannot be excluded (Figure 6).

‎View this figure

‎View this figure

Discussion

Principal Findings

Adequate bowel preparation is a critical contributor to successful colonoscopy; however, the performance of traditional instructions of bowel preparation before colonoscopy in improving the quality of bowel preparation is not enough [10]. So, several enhanced patient instructions such as WeChat and SMS were developed to cover the shortcomings of the traditional instructions [2]. In this meta-analysis evaluating the detection rate of colonic polyps and adenomas, we found that EPE relatively increased PDR by 19% (>57/1000) and ADR by 37% (>61/1000), when compared with standard patient education before colonoscopy. Moreover, we also found that EPE was associated with increased SSADR (OR 1.76, 95% CI 1.22-2.53) and ADR-plus (RR 4.39, 95% CI 2.91-6.61). We found no evidence of statistical differences in AADR and CDR between the 2 groups. After performing a subgroup analysis, we found that EPE increased PDR by 21% in Asian patients and by 44% in mixed outpatients, and ADR by 39% and 37% in Western and Asian patients undergoing colonoscopy, respectively. Although our study found an increased PDR, ADR, SSADR, and ADR-plus in the EPE group, the pure efficacy of EPE alone in the detection of colonic polyps and adenomas needs to be further investigated because all patients in the EPE group also received standard patient education.

Mechanism of Enhanced Patient Education

The EPE regimes were quite diverse in the 10 eligible trials, including visual aid, new visual aids, phone call, SMS, mobile apps, and multimedia education. We compared the essential characteristics of these EPE regimes with those of standard patient education (Multimedia Appendix 3). EPE has been shown to be effective in improving bowel preparation quality [2,19]. In this study, we further found that EPE increases the detection rate of colonic polyp and adenoma. EPE has several advantages including easy understanding of the education materials, easy access to the information of bowel preparation and dietary recommendations, more interactive approaches for seeking solutions to problems, or an additional approach to enhance correct memory compared with standard patient education [2]. After instructing patients with the EPE regime, the compliance of patients with instructions, including dietary recommendations and digestion of bowel preparation solutions, improves, which may be the possible reason for an increase in the colonic polyp and adenomas detection rate [2,9,42,43].

Comparison With Other Studies

To date, 5 systematic reviews and meta-analyses have been performed to comprehensively investigate the impact of EPE on the quality of bowel preparation compared with standard patient education [2,16-19]. However, only 2 of those 5 evaluated the PDR as a secondary outcome [2,17]. Chang et al [37] conducted a meta-analysis of 3 trials to evaluate the impact of EPE on PDR in patients undergoing colonoscopy and found no significant difference between the 2 groups (RR 1.14, 95% CI 0.87-1.51; I²=79.1%). However, there are some limitations that need to be considered in this meta-analysis. First, only 3 trials including 1779 patients were analyzed, which may cause summary estimates to be inflated and thereby limit the strength of the inference that can be drawn due to inadequate accumulated sample size (adjusted information size=6356) [48]. Therefore, the finding of this study may not be considered as definitive. Furthermore, this meta-analysis did not consider other important colonoscopy quality metrics such as ADR and adenomas per positive participant [49], which greatly reduced the relevance for clinical decision.

Guo et al [2] performed an updated systematic review and meta-analysis after including recent trials to evaluate the efficacy of EPE for bowel preparation before colonoscopy. In this study, PDR was also evaluated as a secondary outcome, and a pooled estimate based on 5 trials did not detect a significant difference in PDR between the 2 groups (OR 1.25, 95% CI 0.93-1.68; P=.14). It must be noted, however, that there are also some flaws that need to be acknowledged in this study. First, although this study included more trials for final analysis [39,42,43,50,51], 2 studies that did not report appropriate data were inappropriately considered to be eligible [50,51]. As a result, the findings of this study must be interpreted cautiously. Similarly, this study did not also consider other important colonoscopy quality metrics [49].

Differences between our study and the 2 previous meta-analyses should be emphasized. In the 2 previous meta-analyses, limited number of eligible trials were accumulated for PDR, and since then, additional eligible trials with a high quality and large sample size have been published [9,40,41,44,46,47]. ADR is widely accepted as one of the objective colonoscopy quality metrics [52], and published evidence suggested that increased ADR was associated with decreased risk of post colonoscopy CRCs [53]. Moreover, ADR-plus was also suggested as one of the colonoscopy quality metrics with significant clinical relevance [28]. It is noted, however, that the previous meta-analyses only evaluated PDR, and other important quality metrics were not considered because of limited data, which reduced the clinical relevance of summary estimates. Our meta-analysis of 10 RCTs involving 4560 patients suggests that EPE is associated with an increase in PDR, ADR, ADR-plus, and SSADR. There were no significant differences in ADR and CDR. To test the robustness of summary results of primary outcomes, we performed a trial sequential meta-analysis to adjust the significance level and calculate the adjusted information size. Trial sequential analyses of both outcomes indicated that currently, cumulative evidences are conclusive. Moreover, subgroup meta-analyses found that EPE is only associated with increased PDR in Asian patients and ADR in all patients.

Strengths and Limitations of This Study

Strengths of this review include a comprehensive literature search and the inclusion of multiple types of precursor polyps as outcomes. Moreover, we also performed a trial sequential meta-analysis for PDR and ADR. However, limitations of this meta-analysis should be acknowledged. First, 2 trials with a three-arm design were included, and we simply combined the data reported in the two positive groups according to the methodology recommended by the Cochrane Collaboration [25]. However, it may be rational to use the network meta-analysis to compare all interventions if more eligible trials can be captured. Second, most trials analyzed in our meta-analysis assigned patients to follow different diet restrictions such as clear liquid and low-residue diets, but we did not perform subgroup analysis according to this condition because 2 trials did not introduce the details of diet restriction [39,45]. It is worth mentioning, however, that previous meta-analyses have demonstrated comparable efficacy between clear liquid and low-residue diets for bowel preparation before colonoscopy [54,55]. Third, our previous meta-analysis detected no significant difference between the low- and traditional-volume polyethylene glycol regimens in bowel preparation [56]. Thus, bowel preparation regimen was not considered to be a factor for performing a subgroup analysis. However, prokinetic agents were used in some trials [39,40,42,43,46]. So, additional analyses should be performed when sufficient data on bowel preparation regimen can be obtained. Fourth, the start time of initiating education was different among all included trials. This factor and potential influence should be considered with caution. If possible, a further study should be designed to investigate the impact of starting time of education on the detection of colonic polyps or adenomas. Fifth, the EPE methods were quite diverse in the eligible RCTs; however, additional analysis cannot be performed to investigate which EPE method or which combination is the best for bowel preparation instruction because of insufficient data.

Conclusions

In summary, current evidence indicates that there was a significant difference between EPE and standard patient education in PDR (RR 1.19, 95% CI 1.05-1.35; P<.05), ADR (RR 1.37, 95% CI 1.15-1.64; P<.001), ADR-plus (RR 4.39, 95% CI 2.91-6.61; P<.001), and SSADR (OR 1.76, 95% CI 1.22-2.53; P<.05). However, results for AADR (RR 1.54, 95% CI 0.67-3.55; P=.31), ADR-plus, and CDR (RR 1.54, 95% CI 0.60-3.98; P=.37) should be interpreted cautiously as data are still limited. Large-scale trials addressing this question may provide data better applicable to clinical practice.