Original Communication

0148-6071/05/2902-0074$03.00/0 JOURNAL OF PARENTERAL AND ENTERAL NUTRITION Copyright © 2005 by the American Society for Parenteral and Enteral Nutrition

Vol. 29, No. 2 Printed in U.S.A.

Original Communication Implementation of a Nutrition Support Protocol Increases the Proportion of Mechanically Ventilated Patients Reaching Enteral Nutrition Targets in the Adult Intensive Care Unit Shannon L. Mackenzie, BSc, RD*†; David A. Zygun, MD, FRCPC†‡¶; Bev L. Whitmore, MBA, RD*; Christopher J. Doig, MD, MSc, FRCPC†‡§; and Syed Morad Hameed, MD, MPH, FRCSC†! From the *Departments of Clinical Nutrition, †Critical Care Medicine, ‡Medicine, §Community Health Sciences, ! Surgery, and ¶Clinical Neuroscience, Calgary Health Region and University of Calgary, Calgary, Canada

ABSTRACT. Background: Despite the evidence that enteral feeding reduces morbidity in critically ill patients and is preferred to parenteral nutrition, the delivery of enteral nutrition (EN) is often inadequate. The purpose of this study was to determine whether implementation of an evidencebased nutrition support (NS) protocol could improve EN delivery. Methods: An NS protocol incorporating available scientific evidence; data from a retrospective survey of 30 intensive care unit (ICU) patients; and input from dietitians, intensive care physicians, surgeons, nurses, and pharmacists was developed. The impact of this protocol was evaluated prospectively in 123 consecutive adult patients admitted to a multisystem ICU who were eligible for EN. Results: The percentage of patients who received at least 80% of their estimated energy requirements during their ICU stay increased

from 20% before implementation of the NS protocol to 60% after implementation (p ! .001). After adjusting for confounders, those in the postimplementation group received significantly more kcal/kg/d than the preimplementation group (3.71 kcal/kg/d; 95% confidence interval, 1.64 to 5.78; p " .001). Parenteral nutrition kcal/kg/d use was reduced in the postimplementation group (1.6 vs 13%, p " .02). There was no difference in time to initiation of enteral nutrition between groups (1.76 days preprotocol vs 1.44 days postprotocol implementation, p " .9). Conclusions: The development and use of an evidence-based NS protocol improved the proportion of enterally fed ICU patients meeting their calculated nutrition requirements. ( Journal of Parenteral and Enteral Nutrition 29:74 –80, 2005)

Nutrition support (NS) has become an important priority in the comprehensive management of critically ill patients. Of the routes available, enteral nutrition (EN) has gained increasing favor both as an avenue for the provision of metabolic substrates and as a potential means to modify the role of the gastrointestinal tract in the systemic inflammatory response.1–7 EN may help in the maintenance of the nonspecific enteral immune barrier (local microflora, enterocytes, tight junctions), gut-associated lymphoid tissue, and systemic immune function mediated by extraintestinal mucosa-associated lymphoid tissue.8 Patients receiving EN in the intensive care unit (ICU) are often underfed, which may result in further impairment of immune function, increased risk of infections, and increased mortality.7,9 Although the term permissive underfeeding is seen in the ICU literature and implies hypocaloric NS may be safe, the favorable impact of NS on malnutrition has been demonstrated in the meta-analyses of both Braunschweig et al10 and Heyland et al11: malnourished patients

receiving NS may have lower risk of infection and mortality. However, despite the evidence promoting NS, a recent national survey of Canadian ICUs showed 40% of patients did not receive any form of NS.12 In this survey, on average only 55.6% of energy and 56.3% of protein goals were delivered by EN for those patients in the ICU #3 days.12 Work by McClave et al13 showed similar findings: only 51.6% of goal volume was infused over all days of infusion in the ICU. At a center within our region, a 1998 quality assurance project identified that patients met only 54% of their goals; these findings were once again duplicated in 2001 by a retrospective review of 30 patients at our study site. The medical literature provides little consensus on many components of our EN regimens (rate progression, gastric residual monitoring, reasons for cessation); however, there is evidence that the use of infusion protocols can improve the success of EN delivery as a result of standardized practice.13–15 Protocolized approaches to feeding have been shown to both increase delivery and minimize or address risks.15–17 Schwartz18 used a decision-tree format, for example, to communicate NS strategies to practitioners in their community hospital as part of an interdisciplinary performance improvement process; their findings included improved NS practice, which resulted in quality improvement and cost savings.

Received for publication January 19, 2004. Accepted for publication December 2, 2004. Correspondence: Shannon Mackenzie, BSc, RD, Clinical Nutrition, Calgary Health Region, Foothills Medical Centre, 1403–29 St NW, Calgary, AB, T2N 2T9 Canada. Electronic mail may be sent to [email protected]. 74

Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on November 8, 2013

March–April 2005

NUTRITION SUPPORT PROTOCOL AND ENTERAL NUTRITION TARGETS

An evidence-based protocol has been similarly developed by a multidisciplinary working group at our site as part of a quality-improvement initiative to optimize delivery of EN to critically ill patients. Historically, the dietitian (RD) recommended NS route and goals on ICU morning rounds according to best practice; however, final decisions on route, goal rate, and starting and stopping was often determined by the attending physician, with no available standard or evidencebased guide to practice. The only written guidelines available were a regional gastric residual monitoring algorithm. The primary objective for instituting a set of feeding algorithms was to improve the safety and delivery of appropriate NS to our critically ill patients via consistent NS practice with regard to route, initiation, advancement and goal rate of EN delivery, monitoring of tolerance, use of prokinetic agents, and avoidance of parenteral nutrition (PN) if not clinically indicated. The purpose of this study was to evaluate the changes in EN delivery as a result of the implementation of this protocol as judged by the proportion of patients meeting on average !80% of their goals, the proportion of goal energy received by patients during their ICU admission, the timing of EN support, and the inappropriate use of PN. MATERIALS AND METHODS

Foothills Medical Centre ICU is a tertiary-care 22-bed medical-surgical unit serving a referral population of approximately 1.2 million. It is 1 of 3 adult ICUs in the Calgary Health Region. Before the development of our NS algorithm, there was no standardized approach to NS in the region, and our site lacked written guidelines. Individual patient feeding regimens are historically based on RD input on bedside rounds, but decisions were ultimately made by the attending physician. This represents the practice before protocol implementation. After implementation of the algorithm, the RDs continued to provide input on rounds but also referred to our guideline and used it as a teaching tool to standardize practice. We used the RD-determined goal as the standard to compare intake to both pre– and postimplementation because of the lack of standardized methods for determining goals historically. Indirect calorimetry was not used to determine energy goals. Enteral feeding formula is provided via orogastric tube as a 1 kcal/mL, high-protein, fibercontaining, polymeric formula, via closed system, continuously over 24 hours a day, unless feeding cessation is required for therapy or specific medications (phenytoin, levofloxacin, ciprofloxacin where EN is held 2 hours before and after each dose; hourly EN goals are usually adjusted to meet needs over the shorter period of time). Individualized formulas were used at the discretion of the RD or team as needed. Data were collected and the protocol was implemented at Foothills Medical Centre, with the plan that, if successful, the protocol would be considered for adoption at the other 2 sites. This was a quality-assurance project with institutional review and approval.

75

Creation of an ICU NS Protocol An interdisciplinary working group, including an intensivist/trauma surgeon, RD, and nurse (RN) was created to develop an evidence-based EN protocol. An extensive literature review was undertaken. Relevant articles including randomized controlled trials, metaanalyses, and well-designed cohort studies were identified from a MEDLINE search and from personal collections. The working group also considered data from a retrospective chart review of 30 consecutive patients from October 2001. The protocol developed is based on literature from the critical care, general surgery, and medical literature to meet the needs of patients in our mixed unit.3– 6,10,11,16,17,19 –26 We included specific guidelines for enteral NS in postoperative and pancreatitis patients because these were identified as areas for improvement due to inappropriate use of PN in the past; these are also used for education of medical and allied health professionals outside of the ICU environment. The protocol specifically addresses when to consider safe EN, identifying high-risk enteral feeding scenarios, identification of malnourished patients, progression of EN to a minimum goal, monitoring gastric residuals, ordering prokinetic agents or small bowel access when intolerance precludes gastric EN support, and the appropriate use of PN. We had limited historical data on frequency, duration, and reasons for interruptions of EN at our site; however, as the literature suggests this topic is often an issue, and as our experience was that EN is often held for “high gastric residuals,” we collected these data and included in our protocol common situations in which EN can safely be continued. Protocol development included feedback from the local ICU quality council, from multidisciplinary grand rounds (trauma, surgery, ICU, anesthesia, pharmacy, nutrition, nursing), and from other regional RDs, RNs, pharmacists, and surgeons. Our protocol was developed in PowerPoint 2002 (Microsoft Corporation) format for presentation purposes and for easy access on the internal Calgary Health Region Critical Care Medicine website at each bedside. The details of the literature review and development process of our protocol are beyond the scope of this paper and will be published elsewhere. Refer to Figure 1 for protocol content (or website www.calgaryhealthregion. ca/ccm/nutritional_support_in_the_icu.pdf). Protocol Evaluation We prospectively evaluated a cohort of 123 patients in 2002 at Foothills Medical Centre. Patients were evaluated on admission to unit and followed daily; patients who met the inclusion criteria were followed to study end point, defined as progression to oral diet (with discontinuation of EN), discharge from ICU, or death. Criteria for inclusion into the study were patients !16 years of age who were mechanically ventilated in ICU !72 hours and eligible to receive EN. Eligible patients who receive either zero EN or PN were included in the study; these patients would receive 0% for daily proportion of EN goal received for each day they received no EN. Exclusion criteria were


76

MACKENZIE ET AL

Vol. 29, No. 2

FIG. 1. Calgary Health Region ICU nutrition support protocol.

ICU length of stay (LOS) "72 hours, on NS "24 hours, receiving oral diet, diagnosis precluding use of EN, or PN use deemed appropriate according to predetermined criteria: mechanical bowel obstruction, GI hemorrhage, high output fistula (#500 mL/24 hours), new short-bowel syndrome with !60 cm remaining small bowel. The inclusion and exclusion criteria are designed to help us evaluate if the protocol improves adequacy and timing of EN delivery in all eligible patients, regardless of actual feeding mode. The study was conducted in 2 phases. The preprotocol-implementation cohort included 61 consecutive patients who met the criteria, admitted from January– March 2002. After development, refinement, and implementation of the protocol, the second prospective cohort of 62 consecutive patients (“postprotocol-implementation”) was collected October–December 2002. Delay in data collection between phases was planned in order to develop the protocol. Protocol development included initial development, communication with nursing and medical staff, processing of feedback, and final refinement. Daily data collected included diagnosis, surgical status, date and time of admission to unit, date and time of initiation of EN, energy and protein goals as determined by the RD (18 –30 kcal/kg/d and 1.5 g protein/kg/d on average), energy and protein delivered from EN or PN, hours receiving EN per day, reasons for holding feedings, and length of ICU stay. In both groups, RDs provided recommendations on morning rounds. A daily proportion of EN delivered to goal recommended was calculated for all ICU days from admission (day 1) until extubation and transition to oral diet, death, or discharge from unit so as to compute our primary outcomes (see below). If oral diet was limited after extubation and EN continued, the EN goal was not adjusted for oral intake. If the patient was

extubated, however, and transitioned to oral diet and EN was stopped, this indicated study end for that patient. The frequency, duration, and reasons for interruption of enteral feeding were documented daily. Statistical Analysis The primary end points evaluated were (1) the proportion of patients who met on average !80% of their goal energy requirements, as determined by the RD prescription, over all days of their ICU stay through to study end; and (2) the median proportion of delivered kilocalories over prescribed kilocalories. Secondary outcomes included kcal/kg/d delivered on average per patient over ICU stay, time to initiation of feeding, reasons EN was held, and PN use. These end points are comparable to those in the survey by Heyland et al12 of feeding practices in Canadian ICUs. Heyland et al refer to 80% as a target for “adequacy” of intake, where Canadian patients in the ICU !3 days met at least 80% of their EN goal by approximately day 10. We also used this figure for “adequacy” to determine the proportion of patients who meet !80% of their target over the course of their entire ICU stay. Means were compared using the Student t test and medians using the Mann-Whitney U test. Differences in proportions among categorical data were assessed using Fisher exact test. Individual multivariable linear regression models were created to analyze the differences in the kcal/kg/d delivered and prescribed. Variables significant in univariate analysis (p ! .2) were entered into a multivariable model. Backward stepwise elimination was used to create the most parsimonious model. Twosided p values !.05 were considered significant for all comparisons.


March–April 2005

NUTRITION SUPPORT PROTOCOL AND ENTERAL NUTRITION TARGETS RESULTS

The demographics of each group are presented in Table I. The preimplementation group was significantly older than the postimplementation group. Severity of illness, as measured by the admission APACHE II score, was similar. There were trends to a lower proportion of men and a higher proportion of postsurgical patients in the postimplementation group, but these did not reach statistical significance. The ICU LOS in the postimplementation group was marginally shorter than preimplementation (p " ns). The distribution of diagnostic categories is presented in Table II. ICU mortality was not significantly different between the groups. Not all patients received EN (5 in the preimplementation group vs 1 in the postimplementation group). Sixty percent of the postimplementation group compared with 20% of the preimplementation group received at least 80% of RD-determined goal kilocalories on average during their ICU stay. This difference was highly significant (p ! .001). The median proportion of RD prescribed kilocalories actually received by the patient was significantly higher in the postimplementation group compared with the preimplementation group (0.83 vs 0.56, p ! .001). Although surgical patients were fed less well than nonsurgical patients, the surgical patients had improvement in the proportion of RD-prescribed kilocalories delivered from 43% preimplementation to 74% postimplementation (p ! .001). The overall improvements in the proportion of goal kilocalories received by the postimplementation group may have been affected by a decrease in the goal kilocalories prescribed in the postimplementation group. After adjusting for age, gender, postsurgical status, a diagnosis of trauma or burn, admission APACHE II score, and ICU mortality in a multivariable linear regression model, the postimplementation group was prescribed a significantly lower number of kcal/kg/d on a daily basis (coefficient: $4.9 kcal/kg/d; 95% confidence interval: $6.7 to $3.1; p ! .001). The mean (% SD) number of kcal/kg/d delivered to patients was 13.8 (% 6.8). There was a trend to a greater kcal/kg/d delivered to the patients in the postimplementation compared with the preimplementation group (preimplementation group mean % SD: 12.7 % 7.0; postimplementation group mean % SD: 14.8 % 6.5; p " .09). Univariate analysis of a priori identified factors related to the number of kcal/kg/d delivered is presented in Table III. Although the pre-

77

protocol group was older, this did not predict kcal/kg/d delivered. The most parsimonious model is presented in Table IV (n " 123, F " 20.2, p ! .0001, R2 " .34). Implementation group, postsurgical status, and ICU LOS were found to be independent predictors of the kcal/kg/d delivered. The model predicts that postsurgical patients receive fewer kcal/kg/d than nonsurgical patients and that patients with longer ICU LOS receive more kcal/kg/d than those with shorter LOS. After adjusting for postsurgical status and ICU LOS, those in the postimplementation group actually received significantly more kcal/kg/d than the preimplementation group (kcal/kg/d: 3.71; 95% confidence interval: 1.64 to 5.78; p " .001), despite being prescribed less. Changes in protein intake followed the same trends as energy intake. The median (IQR) time to initiation of enteral feedings for all patients was 1.18 days (0.56, 2.14). There was no significant difference in time to initiation between groups (median 1.17 days pre- vs 1.23 days postprotocol implementation, p " .7). Use of the protocol had little impact on cessation of EN. The median (IQR) number of times enteral feedings were held per day of ICU stay was 0.43 (0.20, 0.67). The number of times held were similar between the pre- and postimplementation groups (p " .8). The median (IQR) hours enteral feeding was held per day was 3.29 hours (1.37, 5.15). Again, this was not different between groups (p " .8). The most common reasons for holding feedings are listed in Table V. Eight patients (13%) preimplementation received PN (not clinically indicated), whereas 1 patient (1.6%) postimplementation received PN. This difference was statistically significant (p " .02). Prokinetic agent usage and small-bowel feeding-tube placement related to elevated residual volumes during the study period were low and therefore were not statistically analyzed between study groups. DISCUSSION

The implementation of our evidence-based NS algorithm improved the proportion of patients meeting !80% of their goals in the ICU, improved overall delivery of EN, and reduced the inappropriate use of PN. It provides a standard of care where previous guidelines did not exist and serves as a guide to safe EN practice while minimizing the use of inappropriate PN. The benefit of a program that emphasizes EN may lead to improvement in patient outcomes, lower costs, and

TABLE I Patient demographics Patient characteristic

Preimplementation

Postimplementation

p

Number of patients Age, mean (% SD) Gender (% male) Weight Admission APACHE mean (% SD) Postsurgical ICU LOS median days (IQR) ICU mortality

61 58 (% 17) 72% 70.7 kg 17 (% 8) 39% 8.8 (6.2, 15.0) 16.4%

62 51 (% 21) 55% 66.8 kg 16 (% 7) 56% 7.5 (4.9, 11.5) 21.0%

.05 .06 .06 .35 .07 .09 .66

ICU, intensive care unit; LOS, length of stay. Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on November 8, 2013

78

Vol. 29, No. 2

MACKENZIE ET AL TABLE II Diagnostic categories Diagnostic category

Preimplementation group (%)

Postimplementation group (%)

All patients (%)

Cardiovascular Gastrointestinal Metabolic, endocrine or renal Poisoning Neurological Peripartum Shock Trauma Monitoring Respiratory Burn

2 (3.3) 3 (4.9) 0 (0) 1 (1.6) 10 (16.4) 0 (0) 8 (13.1) 16 (26.2) 0 (0) 17 (27.9) 4 (6.6)

2 (3.2) 0 (0) 2 (3.2) 2 (3.2) 15 (24.2) 1 (1.6) 4 (6.5) 19 (30.7) 2 (3.2) 14 (22.2) 1 (1.6)

4 (3.3) 3 (2.4) 2 (1.6) 3 (2.4) 25 (20.3) 1 (0.8) 12 (9.8) 35 (28.5) 2 (1.6) 31 (25.2) 5 (4.1)

improved patient safety, and a recent Canadian study showed that the introduction of evidence-based guidelines increases EN days, reduced hospital stay, and mortality.27 Unfortunately, despite evidence in support of early enteral feeding, adequate delivery of EN is extremely challenging.7,13–15,28 Our data confirm that delivery remains well below goals despite introducing written evidence-based guidelines. Spain found that EN delivery was improved by use of a protocol, primarily because of better MD ordering of EN and more rapid advancement of EN15; however, MDs in their study used the protocol for only 58% of patients in their treatment group. At our site, attending bias was a factor even after grand rounds and in-servicing; despite this, we increased the median proportion of delivered to prescribed kilocalories from 56% preprotocol to 83% postprotocol implementation. The improvements we have seen in the proportion of goal kilocalories received may have been due in part to a reduction in the goal kilocalories prescribed in the postimplementation group; however, the impact was minimal as we simultaneously improved delivery by almost 4 kcal/kg/d. Interrater variability between groups may be the primary limitation to this evaluation; 2 different RDs assessed the patients’ goals in the pre- vs the postprotocol groups, and delivery was simultaneously affected by RN and MD bias. Although RDs used their discretion in determining daily goals, the protocol did not intend to modify the prescription. The mean prescribed kilocalories preprotocol implementation was 23.5 % 7.5 kcal/kg/d compared with 20 % 10 kcal/kg/d postprotocol implementation, which is a small difference in prescribing patterns between TABLE III Univariate analysis: kcal/kg/d delivered Variable

Coefficient

SEM

p Value

Age Gender APACHE Burn Trauma ICU LOS ICU mortality Postsurgical status Implementation group

$0.009 $1.89 0.019 4.57 0.197 0.31 1.62 $3.11 2.06

0.032 1.27 0.085 3.11 1.37 0.05 1.58 1.21 1.22

.78 .14 .82 .14 .89 !.001 .31 .01 .09

ICU LOS, intensive care unit length of stay.

groups; both RDs prescribed goals in the range of 18 –30 kcal/kg/d. Our methods and results are comparable to other work in the field, where prescribed goals were left to the RD’s discretion and were not blinded.12,13,27 Notably, the case mix was different between our groups; there were more burns and fewer surgical patients in the before group, which correlates with higher overall prescribed kcal/kg/d in that group. We cannot rule out that the regression model used to control for confounding factors did not fully compensate for postsurgical status and case mix. In summary, we fed more to the postprotocol group despite prescribing less, which highlights the strength of the result. And of interest, it should be noted that preprotocol, it took 10 days for patients still in the unit by day 10 to receive !80% of their goals. This time to achieve goals decreased postimplementation of the protocol: it took only 5 days to receive !80% of the goal for all patients in the unit. This suggests that a combination of factors, including time to initiation and daily adequacy of EN, is leading to quicker achievement of goals. Although we increased delivery overall, only 15 kcal/kg/d were delivered over ICU stay, on average. Our delivery data appear lower than other authors’. For example, 1 author looked only at NS days,29 rather than entire ICU stay, which gives the impression of better feeding success. De Jonghe et al28 began data collection when NS was prescribed, rather than when NS should have been initiated (ie, early EN); as a result, their patients received a higher intake of 20 kcal/kg/d. Because there is no direct evidence that method of goal determination or that more closely meeting that feeding goal has an impact on outcomes,1 we cannot predict that this rate of underfeeding will be detrimental to our patients. The reduction in inappropriate PN use observed may be a better indicator of the success of our protocol and potential for improved outcomes, and this factor alone may have affected the adequacy of EN delivery. We only included the patients TABLE IV Multivariable analysis: kcal/kg/d delivered Variable

Coefficient

95% Confidence interval

p Value

Implementation group Postsurgical status ICU LOS

3.71 $3.66 0.34

1.64–5.78 $5.71, $1.61 0.23–0.44

.001 .001 !.001


March–April 2005

NUTRITION SUPPORT PROTOCOL AND ENTERAL NUTRITION TARGETS

TABLE V Reasons for holding feedings Reason

Percentage

Medications Physiotherapy Surgery Extubation Transport for CT scan or procedures

21.9 18.4 13.1 11.3 6.5

receiving PN in the study who, according to the literature, should have been receiving some EN. The reduction in PN use might be expected to affect daily goals determined by the RD; whereas in the past patients received their goal kilocalories from PN (and likely no EN), in the future the RD may compromise on the short-term “goal” from EN kilocalories to support the use of the GI tract via low-dose EN (for example, 25 mL/h of a 1 kcal/mL product is only 500 kcal/d). More specifically, before protocol implementation 13% of patients received PN later deemed not clinically indicated. Across Canadian ICUs, 14.6% of patients received PN alone or in combination with EN12; whether or not the use of PN was appropriate was not indicated, and therefore it is difficult to compare results. Patients were excluded from our study if PN use was deemed appropriate according to our predetermined criteria. We saw a significant reduction in the inappropriate use of PN after implementation of the protocol, which over time may equate to cost savings and benefit to overall patient outcome in the ICU.11,24 Education to medical and nursing staff within and outside the ICU (eg, surgeons) on the appropriate use of PN and early EN may have facilitated our success and continues to be a priority to ensure we are using the GI tract safely yet maximally throughout our region. For example, we noted that surgical patients are fed later and less adequately than medical patients; however, despite this, delivery of EN to surgical patients improved from 43% of goal preimplementation to 74% of goal postprotocol implementation (p ! .001); surgical patients also started EN 2.78 days (median) after admission to the unit preimplementation, which was reduced to 1.6 days postprotocol implementation. Although not statistically significant, a 50% reduction in days to feeding is clinically significant and suggests a change in practice. Use of the protocol had little impact on route of EN or cessation of EN. Our data indicated we were feeding patients early at 1.18 days. In comparison, the median time to initiation of EN among patients in Canadian ICUs was 1.7 days; the subgroup of patients in the ICU #72 hours was fed at day 2 (median).12 We did not statistically change the timing of initiation of EN; however, according to papers including the recently published Canadian Clinical Practice Guidelines for NS in Mechanically Ventilated Critically Ill Adults, our timing is within appropriate limits.1,19 As already mentioned, we improved time to feeding in surgical patients by implementing the protocol, and those patients are now fed within the guidelines.12 This earlier use of the GI tract may reduce septic complications and costs.3–5,23–26,29

79

Despite ongoing barriers to feeding maximally, the improvement we have seen in meeting EN targets was related to several factors, including the reduction in inappropriate PN usage, medical director support and endorsement of the protocol, efforts of local opinion leaders (MD and RD), and nursing compliance with the guidelines. Effective use of the protocol requires support from administration, physicians and surgeons, and nurse practitioners. Compliance or adherence to the protocol was not formally evaluated; however, it was noted that physician and nurse bias played a role in compliance with protocol use and may have affected initiation of EN or attainment of goals. Protocol use was not enforced but rather used as a guide to best practice. As previously noted, compliance with or adherence to a protocol is a challenge and therefore remains an ongoing quality-improvement initiative at our site. Sixmonth poststudy quality-improvement data indicated that since our postprotocol data collection, and without ongoing rigorous education and monitoring, achievement of goals had dropped. We managed to increase the rate to #80% again with reinstitution of nursing in-services, implementing a “contest” among the nursing teams, creating pocket cards containing protocol concepts for residents, among other tools to encourage use of the protocol. As part of these initiatives, we conduct regular audits and share results at the ICU quality council and other department meetings and on our internal website, which Landry and Sibbald30 suggest is one of several useful educational techniques aimed at changing behavior. It is important to note that the sample size is not large enough to correlate the improvement in EN delivery or reduction in PN use with clinically important outcomes, and therefore we cannot comment on improvement in mortality rates or LOS in relation to the use of a feeding protocol. CONCLUSIONS

Our study has resulted in the regional adoption of a protocol that standardizes feeding practices, optimizes enteral intake, provides guidelines regarding intolerance management, and is consistent with the subsequently published national guidelines. Our NS protocol improved the proportion of patients meeting their goals, EN delivery, and led to reduction in inappropriate PN usage. Further research will be needed to confirm the specific elements of the protocol and to determine if it affects clinically important outcomes and to look more specifically at the reasons we are not approaching our minimum goals. ACKNOWLEDGMENTS

Carmen Christman, Deonne Dersch, Anar Jamal, Elaine Lemieux, Joyce Mammel, Rob Mulloy, Kerri Pain (study design), and Karen Plett provided input into protocol development or data collection. The study was supported by a grant from the Canadian Foundation for Dietetic Research and the Department of Clinical Nutrition, Calgary Health Region.


80

MACKENZIE ET AL REFERENCES

1. Heyland DK, Dhaliwal R, Drover JW, Gramlich L, Dodek P. Canadian clinical practice guidelines for nutrition support in mechanically ventilated, critically ill adult patients. JPEN J Parenter Enteral Nutr. 2003;27:355–373. 2. Lewis SJ, Egger M, Sylvester PA, Thomas S. Early enteral feeding versus “nil by mouth” after gastrointestinal surgery: systematic review and meta-analysis of controlled trials. BMJ. 2001;323:773–776. 3. Moore EE, Jones TN. Benefits of immediate jejunostomy feeding after major abdominal trauma: a prospective, randomized study. J Trauma. 1986;26:874 – 881. 4. Moore FA, Moore EE, Jones TN, McCroskey BL, Peterson VM. TEN versus TPN following major abdominal trauma: reduced septic morbidity. J Trauma. 1989;29:916 –923. 5. Kudsk KA, Croce MA, Fabian TC, et al. Enteral versus parenteral feeding: effects on septic morbidity after blunt and penetrating abdominal trauma. Ann Surg. 1992;215:503–513. 6. Khalili TM, Navarro RA, Middleton Y, Margulies DR. Early postoperative enteral feeding increases anastomotic strength in a peritonitis model. Am J Surg. 2001;182:621– 624. 7. Adam S, Batson S. A study of problems associated with the delivery of enteral feed in critically ill patients in five ICUs in the UK. Intensive Care Med. 1997;23:261–266. 8. Sigalet DL, Mackenzie SL, Hameed SM. Enteral nutrition and mucosal immunity: implications for feeding strategies in surgery and trauma. Can J Surg. 2004;47:109 –116. 9. Dempsey DT, Mullen JL, Buzby GP. The link between nutritional status and clinical outcome: can nutritional intervention modify it? Am J Clin Nutr. 1988;47:352–356. 10. Braunschweig CL, Levy P, Sheean PM, Wang X. Enteral compared with parenteral nutrition: a meta-analysis. Am J Clin Nutr. 2001;74:534 –542. 11. Heyland DK, MacDonald S, Keefe L, Drover JW. Total parenteral nutrition in the critically ill patient: a meta-analysis. JAMA. 1998;280:2013–2019. 12. Heyland DK, Schroter-Noppe D, Drover JW, et al. Nutrition support in the critical care setting: current practice in canadian ICUs: opportunities for improvement? JPEN J Parenter Enteral Nutr. 2003;27:74 – 83. 13. McClave SA, Sexton LK, Spain DA, et al. Enteral tube feeding in the intensive care unit: factors impeding adequate delivery. Crit Care Med. 1999;27:1252–1256. 14. Chapman G, Curtas S, Meguid MM. Standardized enteral orders attain caloric goals sooner: a prospective study. JPEN J Parenter Enteral Nutr. 1992;16:149 –151. 15. Spain DA, McClave SA, Sexton LK, et al. Infusion protocol improves delivery of enteral tube feeding in the critical care unit. JPEN J Parenter Enteral Nutr. 1999;23:288 –292. 16. Mentec H, Dupont H, Bocchetti M, Cani P, Ponche F, Bleichner G. Upper digestive intolerance during enteral nutrition in criti-

17. 18.

19. 20.

21.

22. 23. 24. 25.

26.

27.

28.

29.

30.

Vol. 29, No. 2

cally ill patients: frequency, risk factors, and complications. Crit Care Med. 2001;29:1955–1961. American Gastroenterological Association Medical Position Statement: guidelines for the use of enteral nutrition. Gastroenterology. 1995;108:1280 –1281. Schwartz DB. Enhanced enteral and parenteral nutrition practice and outcomes in an intensive care unit with a hospital-wide performance improvement process. J Am Diet Assoc. 1996;96: 484 – 489. Heyland DK, Konopod E, Alberda C. How well do critically ill patients tolerate early, intragastric enteral feeding? Results of a prospective, multicenter trial. Nutr Clin Pract. 1999;14:23–28. Adams S, Dellinger EP, Wertz MJ, Oreskovich MR, Simonowitz D, Johansen K. Enteral versus parenteral nutritional support following laparotomy for trauma: a randomized prospective trial. J Trauma. 1986;26:882– 891. Pinilla JC, Samphire J, Arnold C, Liu L, Thiessen B. Comparison of gastrointestinal tolerance to two enteral feeding protocols in critically ill patients: a prospective, randomized controlled trial. JPEN J Parenter Enteral Nutr. 2001;25:81– 86. McClave SA, Greene LM, Snider HL, et al. Comparison of the safety of early enteral vs parenteral nutrition in mild acute pancreatitis. JPEN J Parenter Enteral Nutr. 1997;21:14 –20. Singh G, Ram RP, Khanna SK. Early postoperative enteral feeding in patients with nontraumatic intestinal perforation and peritonitis. J Am Coll Surg. 1998;187:142–146. Hedberg AM, Lairson DR, Aday LA, et al. Economic implications of an early postoperative enteral feeding protocol. J Am Diet Assoc. 1999;99:802– 807. Windsor AC, Kanwar S, Li AG, et al. Compared with parenteral nutrition, enteral feeding attenuates the acute phase response and improves disease severity in acute pancreatitis. Gut. 1998; 42:431– 435. Kalfarentzos F, Kehagias J, Mead N, Kokkinis K, Gogos CA. Enteral nutrition is superior to parenteral nutrition in severe acute pancreatitis: results of a randomized prospective trial. Br J Surg. 1997;84:1665–1669. Martin CM, Doig GS, Heyland DK, Morrison T, Sibbald WJ. Multi-centre, cluster-randomized clinical trial of algorithms for critical care enteral and parenteral therapy (ACCEPT). CMAJ. 2004;170:197–204. De Jonghe B, Appere-De-Vechi C, Fournier M, et al. A prospective survey of nutritional support practices in intensive care unit patients: what is prescribed? What is delivered? Crit Care Med. 2001;29:8 –12. Taylor SJ, Fettes SB, Jewkes C, Nelson RJ. Prospective, randomized, controlled trial to determine the effect of early enhanced enteral nutrition on clinical outcome in mechanically ventilated patients suffering head injury. Crit Care Med. 1999; 27:2525–2531. Landry MD, Sibbald WJ. Changing physician behavior: a review of patient safety in critical care medicine. J Crit Care. 2002;17: 138 –145.
