Implementation of a Bowel Protocol to Improve Enteral Nutrition and Reduce Clostridium difficile Testing

Participants comprised a convenience sample of both medical and surgical adult patients (≥ 18 years old) admitted to the critical care unit from January 1, 2016, through January 31, 2017. We included eligible patients who we anticipated would require EN support and who remained in the ICU for 72 hours or longer.

This study retrospectively investigated a before-and-after intervention. Registered dietitian nutritionists and a dietetic intern collected data from inpatients before protocol implementation, from January 1 to March 1, 2016. The constipation protocol was implemented in April 2016 and the diarrhea protocol in July 2016. Data were also collected after the protocols had been implemented, from September through January 2017. Patients were identified from daily Department of Food and Nutrition ICU census sheets. All patients who started receiving EN were considered eligible and were recorded on a tracking form. Once identified, a patient’s information was obtained from the electronic health record. Patients were retrospectively included in the project if they had orders for EN for at least 4 days (≥ 96 hours). Patients were excluded if they were transferred out of the ICU or did not have an order for at least 3 days of EN.

Enteral nutrition order information and start time, and the type of formula, were obtained from the electronic health record with each patient’s body mass index, weight at admission, and initial nutritional assessment by and recommendations from the registered dietitian nutritionist. We used documentation of intake and output to determine each patient’s total formula intake, receipt of propofol (for sedation), additional protein supplements, and GRV. All data were recorded in a spreadsheet.

Since January 2013, all acute care hospitals must report C difficile LabID events to the Centers for Medicare and Medicaid Services via the National Healthcare Safety Network. Thus data are robust regarding the identification of C difficile at our facility. Our hospital’s SIR increased from 0.710 in 2015 to 0.952 in 2016—a 34% increase. In 2016 a multidisciplinary team performed a root cause analysis of each hospital-onset event. Team members reviewed time lines to ascertain deficiencies in the testing process. A report was built in the electronic health record in order to identify pending orders for C difficile testing, thus allowing infection prevention team members to collaborate with nurses and providers caring for patients at the bedside to address potentially inappropriate testing before collecting specimens. A best practice alert was also built into the electronic health record to alert providers ordering a test for C difficile to review the administration of cathartics within the preceding 48 hours.

To evaluate nursing staff knowledge regarding care of patients receiving EN, a dietetic intern surveyed ICU staff nurses about EN protocols, documentation in the electronic health record, and practices for checking GRV. In an effort to standardize practices, we created protocol algorithms, with clear steps to address either constipation or diarrhea, or potential C difficile infection (Figures 1–3 and Table 1). We developed algorithms by using a multidisciplinary approach that considered evidence-based practice guidelines and followed the format of current facility protocols to help guide providers and nurses in appropriately identifying diarrhea and testing for C difficile. In addition, we reviewed published algorithms addressing these topics.^15,16  Multipronged education on these new protocols was provided to nurses, nurse technicians, and providers. The provider and nursing algorithms were distributed throughout the inpatient units; infectious disease physicians presented during grand rounds and attended hospitalists’ meetings. Nurses received education on competency days, via electronic learning modules, and during daily safety huddles.

We used SPSS software version 24.0 to perform statistical analysis. Data were cleaned and screened for outliers, missing values, and invalid values before being analyzed. The a level was less than 0.05 for all analyses. Descriptive analyses included frequency distributions for patient sex, age, and ethnicity and service team (medical or surgical). We analyzed continuous variables (body mass index, time since admission that EN was ordered, and number of days receiving EN) with descriptive statistics (mean, median, and range). We used an independent sample t test for the continuous variables (changes in EN and tests for hospital-onset C difficile).

We included in the analysis a total of 43 patients: 23 in the preprotocol group and 20 in the postprotocol group. Baseline characteristics were similar between the groups. Male sex was the only significantly different variable: the preprotocol group contained significantly more men (78%; P = .046) than did the postprotocol group. In the postprotocol group, the mean patient age was 65.7 years; 65% were female and 85% were white. Approximately 80% of the patients were on the medical service, while the other 20% were cared for by the surgical service, which is consistent with our daily census breakdown.

Patients in the preprotocol group received EN support for a mean of 4.2 days and received 69% of their ordered volume. Patients in the postprotocol group received EN support for a mean of 5.6 days and received 78% of their ordered volume. Although the postprotocol group did not meet the goal of 80%, the EN volumes those patients did receive increased significantly (t₄₁ = 2.1836; P = .048). Table 2 presents baseline demographic and clinical characteristics of ICU patients before and after the bowel protocol was implemented.

Fifteen nurses completed the EN survey. Survey results are presented in Table 3. One nurse reported that they hold EN if the GRV is greater than 200 mL, but all other nurses reported that they hold EN only if the GRV is greater than 300 mL, which is consistent with the protocol in place. Five nurses reported that they were unaware that the policy stated to check GRV only with large-bore decompression tubes and not with small-bore feeding tubes. Two nurses reported that they knew not to check residuals with small-bore feeding tubes but they sometimes still do it out of habit (Table 3).

Root cause analyses of hospital-onset C difficile events indicated that 49% of events were due to late or inappropriate testing. By disseminating the C difficile testing algorithm to nurses and providers and proactively reviewing pending orders for C difficile testing, inappropriate testing was reduced by 54%. The SIR decreased from 1.290 in the third quarter of 2016 to 0.594 in the fourth quarter—a 46% reduction.

Optimizing the delivery of EN while minimizing complications (constipation, C difficile infection, and diarrhea) is a fundamental issue when managing the care of critically ill adults. The implementation of constipation and diarrhea protocols was associated with significant improvements in goal EN caloric volumes to 78% (baseline 68%). Patients in the postprotocol group, however, did not meet the goal of receiving at least 80% of their prescribed target nutritional recommendations. These findings demonstrate a need for further education of the critical care unit staff and an updated protocol to improve EN therapy in ICU patients. The wide range of recommended calorie targets, with EN volumes at 70% to 100% of the target associated with improved survival in critically ill patients, warrants future exploration.

Although the primary quality improvement project focused on increasing the delivery of EN volumes by implementing bowel protocols, we identified a secondary opportunity to clarify procedures for C difficile testing. We implemented an algorithm and best practice alert and we educated staff during the course of this project. We achieved significant results in decreasing inappropriate C difficile testing. This decrease in inappropriate testing is important, as fewer inappropriately diagnosed C difficile infections provide substantial benefits to the institution. Providers did not stop ordering C difficile tests, but the orders were more appropriate (eg, based on timing and symptoms) after they received the education. Resulting benefits to the institution included improvements in quality and safety, and cost savings. These improvements, in turn, improve patient satisfaction and safety. Currently, enteric precautions are used for patients identified as being infected with C difficile, necessitating a private room and the use of personal protective equipment by both staff and visitors. These requirements can cause patients to feel isolated and involve more facility resources, including employee time and supplies.

Limitations to this quality improvement initiative include the small sample size, which limited our ability to draw from the statistical analysis robust conclusions regarding the effect of the protocols. The small numbers of patients were due to several factors, including the short period of data collection (6 months), patient transfers out of the ICU before completing the 4-day EN period, or orders for fewer than 4 days of active EN.

Documentation of complete enteral information was inconsistent in the electronic health record. Patients were excluded from the analysis if EN volume data were missing from a shift. Because data collection was retrospective, it is unclear whether the data were simply not documented or whether EN had been stopped for part of a day for medical reasons. Because we initially focused on improving enteral volume delivery, data were not collected on timing and the number of bowel movements or the specific bowel medications used.

The survey of ICU registered nurses, which assessed their knowledge of existing EN policy, was limited to those nurses working the day shift. Future surveys should also include staff working night and weekend shifts.

Potential next steps include revising the facility EN protocol, as new guidelines now recommend no longer checking GRV,¹  and collecting additional data that include specifics on patient bowel movements and concurrent medication administration. Studies that use bowel protocols among larger sample sizes are warranted.

At one institution, implementation of constipation and diarrhea protocols significantly improved EN delivery, which approached the target level. Associated education and implementation of C difficile testing protocols decreased the number of hospital-onset C difficile infections. Bowel protocols and flowcharts can improve patient safety and facilitate positive patient outcomes. CCN

The authors thank all staff in the Hospital of Central Connecticut Critical Care Unit; Kelly Williams, a dietetic intern at the University of Connecticut; and Virginia Bieluch, MD, Director of Infectious Disease at the Hospital of Central Connecticut.