Patient-centered approaches to targeting incomplete bowel preparations for inpatient colonoscopies

Introduction

Poor-quality inpatient bowel preparation for colonoscopy has been well described and affects up to 50% of procedures [1]. Incomplete bowel preparation may delay treatment decisions, decrease gastrointestinal (GI) pathology detection, and increase repeat procedures and cost [2,3]. Standardized orders for bowel preparation can help reduce poor preparation, as demonstrated in previous studies [1]. Patient education is a vital aspect of successful bowel preparation [4], as a recent study demonstrated improved bowel preparation scores after the implementation of an educational booklet [5]. Patients’ experience is affected by the instructions and education around colonoscopies [4,6].

The University of Alberta Hospital (UAH) has 2 GI inpatient units and the institutional standard colonoscopy preparation was 4 L of Golytely^®. However, recent data suggest that a 2-L split prep is superior to a 4-L one [7]. When we surveyed 20 GI staff physicians, we found that only 45% of physicians ordered the 2-L split prep. In a randomized chart audit conducted from November 2018 to March 2019, we found that 34% (44/128) of all inpatient bowel preparations were rated as poor and 14 colonoscopies were repeated (11%; 14/128). The aim of this quality improvement (QI) project was to decrease the number of incomplete inpatient colonoscopies as a result of poor bowel preparations, and to improve patient preparation experience through simple and easy-to-use interventions.

Materials and methods

The Model for Improvement [8] and Donabedian [9] conceptual evaluation framework guided this project (Supplementary Fig. 1). We developed an interdisciplinary QI team that included GI/endoscopy unit managers, physicians, medical students, and a quality consultant. To identify improvement opportunities we completed QI tools, including direct observation, force-field and fishbone analysis (Supplementary Fig. 1), reviewed colonoscopy data, surveyed physicians and nurses, and completed a randomized chart audit on 129 charts from 2 GI units. This study was conducted in accordance with local organizational policies and formal Ethical Committee approval was not required [10].

Three improvement opportunities were identified: bowel preparation ordering, nursing education, and patient education (Supplementary Fig. 2). A bowel preparation order label for a 2-L split prep of Golytely^® was created to promote order consistency to international standards [7]. A patient educational paper “placemat” was developed to facilitate standardized education about colonoscopy preparation, including a self-tracking section with the paper placemat placed on the patient’s bedside table during bowel preparation (Fig. 1). This tool was initially pilot tested with 6 patients and the feedback obtained was incorporated, resulting in a revised placemat. To support nursing education a “Nursing Tip” sheet was created to provide strategies for patients who were having difficulty consuming the preparation fluid (Fig. 2). The nursing tip sheet was designed with the help of 2 unit managers based on commonly reported issues. Plan-Do-Study-Act (PDSA) cycles were carried out to test and analyze the effect of the interventions. The outcome measure was a 30% reduction in incomplete inpatient colonoscopies. The process measures included the monthly number of colonoscopy order labels and placemats used. The balancing measure was the number of repeat colonoscopies. The SQUIRE 2.0 guideline framework was used to report findings for this project (Supplementary Table 1).

Figure 1 Patient education paper placemat, used to help educate patients and track bowel preparation

Figure 2 Educational tool for nurses to use when facing difficulties with patient bowel preparation

Data was collected for analysis through retrospective chart audits, nursing notes, and colonoscopy reports. In addition, surveys were developed on a 5-point Likert scale (Supplementary Fig. 3) and completed by patients, physicians and nurses to determine the feasibility and satisfaction of each intervention. Bowel preparation was determined by keywords in the procedure note and noted to be the overall bowel preparation quality as deemed by the endoscopist. Only procedures with clearly good or bad prep noted in the procedure note were included in the analysis.

Results

A total of 128 colonoscopies were audited prior to intervention and 66 colonoscopies post intervention. After removal of those with no comments on colonoscopy prep on the procedure notes, 99 colonoscopies prior to intervention and 47 colonoscopies post intervention were analyzed. Prior to the intervention, 44% (44/99) of inpatient colonoscopies reported poor prep, resulting in 10 repeat procedures (10%; 10/99). Of the repeated procedures, the most common indication for colonoscopy was lower GI bleeding (4/10) and the most common reason for repeated colonoscopy was poor prep (8/10 repeated procedures). The interventions did not have to be adjusted during the PDSA cycles. However, we did find that the standardized ordering label needed to be available in the Emergency Department for new GI admissions and we needed to provide education to rotating residents on the GI unit and new nursing staff during the study period of May 14-August 31, 2018.

Post intervention, approximately 60% (28/47) of the colonoscopies used the standardized order label and 66% (31/47) of physician orders used the 2-L split prep. A total of 80% of patients were provided with the placemat at their bedside. Of the 47 colonoscopy reports audited post intervention, 27.7% (13/47) had poor bowel prep and 72.3% (34/47) had good prep (Table 1). There was a significant reduction in poor bowel preparation compared to pre intervention (P=0.038) (Table 1).

Table 1 Comparative variables between pre and post PDSA (Plan-Do-Study-Act) cycles. All continuous variables are reported as median with interquartile range (IQR)

As regards the balancing measure, there was a reduction in the number of repeat colonoscopies from 10% (10/99) to 4% (2/47). Notably, only 1/2 repeated colonoscopies post intervention was due to poor prep, compared to 8/10 of the repeated procedures prior to the intervention (P=0.27). Lower GI bleeding was the most common indication for colonoscopy in the patients who had poor preparation but did not have a repeat colonoscopy.

There was no significant difference in Charleston comorbidity index scores, age or Schmid Fall Risk scores between patients with good or poor bowel preparation or pre/post intervention (Tables 1,2). However, on subgroup analysis comparing good and poor bowel preparation either pre or post PDSA intervention, there were significantly more men with poor preparation in the pre-PDSA group (P=0.029) and significantly higher Schmid Fall Risk Score for patients with poor preparation in the post-PDSA group (P=0.048) (Table 3). A total of 59% of the surveyed patients found the placemat helpful, 100% of surveyed physicians indicated that they would use the standardized label, and 100% of surveyed nurses perceived the nursing tips sheet to be helpful.

Table 2 Comparison of variables between total patients with good prep and poor prep both pre- and post-intervention (total is 146 colonoscopies). All continuous variables are reported as median with interquartile range (IQR)

Table 3 Comparison of variables between good and bad prep colonoscopies within pre and post PDSA (Plan-Do-Study-Act) cycles. All continuous variables are reported as median with interquartile range (IQR)

Discussion

The patient-centered interventions described above significantly reduced the number of poor-quality colonoscopies and the number of repeated colonoscopies. Interestingly, other factors such as comorbidities, sex or age did not differ between colonoscopies with good or poor prep (Table 2). Empowering patients with easy-to-follow instructions improved the patient’s experience during the bowel preparation process and many patients suggested that the use of the placemat would be helpful for future colonoscopies. Using a standardized prep order label with an option for a split preparation reduced physician order variability and promoted the use of superior bowel preparation [7]. This also increased the ordering of the international standard 2-L split prep by GI physicians (66% vs. 45%). Providing nurses with strategies to navigate preparation challenges and complications offered consistency in approaching education and management for bowel preparation.

Prior literature on optimizing colonoscopy preparation have highlighted patient education as a key and beneficial factor to support bowel preparation [5]. However, this study used a large booklet and an educational session, which was time-consuming and required greater effort to implement [5]. The point-of-care placemat used in our study was an easy-to-read single page paper document, placed on the patient’s bedside table, offering simplified information about the bowel preparation process. The advantage of this approach is that it allowed for patient self-tracking during bowel prep and it standardized the explanation/teaching between nurses and patients. The novel nursing tip sheet provided nurses with management plans to help patients during the bowel prep process, as prior literature identified that poor and variable nursing education was independently associated with poor bowel preparation [13]. The split bowel preparation of 2 L is superior to the 4-L prep, as outlined in a prior meta-analysis [7], and the increased use of the split prep with the standardized order sticker for physicians allowed for the optimized prep to be ordered more consistently.

We anticipate that these interventions can be easily adopted by other hospital units with less experience regarding colonoscopy bowel preparation. The placemat is easy to read for both patients and nursing staff, and is inexpensive to distribute. In addition, all prior literature on optimizing colonoscopy preparation has highlighted education as a key and beneficial factor to support bowel preparation.

Strengths of this project include the extensive pre-project information gathered to contextualize the problem of poor bowel preparation through surveys and chart audits. In addition, using a multidisciplinary team of physicians, nursing staff, managers and a QI specialist allowed for various perspectives and experiential knowledge to be shared regarding inpatient bowel preparation. The patient educational placemat was easy to introduce and the financial cost of implementing this project was low. Finally, our target unit was the GI units, which have more patients requiring colonoscopy compared to other units, and where staff are already more knowledgeable about this procedure compared to other units, such as internal medicine or cardiology.

Study limitations included a non-validated survey, the absence of bowel preparation rating in some colonoscopy reports (28.7% of the original cohort did not have comments on the prep post intervention and were excluded from analysis), and the lack of a pre-intervention patient satisfaction survey. In addition, while we noted whether the placemat was given or not, we did not evaluate any teaching the nurses provided around the placemat. Future studies will focus on the sustainability of the placemat for hospitalized patients, the nursing tip sheet as well as cost analysis. In addition, given the small sample size of this project, future studies with larger sample sizes would be worthwhile to validate these results.

Overall, by implementing these interventions targeting patient education, physician ordering and nursing management, we can improve both patient experience and health outcomes for the hospitalized patient population undergoing colonoscopy.