Nutrition therapy in the critical care setting: What is best achievable practice? An international multicenter observational study*

Transcription

1 Continuing Medical Education Article Nutrition therapy in the critical care setting: What is best achievable practice? An international multicenter observational study* Naomi E. Cahill, RD, MSc; Rupinder Dhaliwal, RD; Andrew G. Day, MSc; Xuran Jiang, MSc; Daren K. Heyland, MD, FRCPC LEARNING OBJECTIVES After participating in this activity, the participant should be better able to: 1. Define best practices in nutritional support for critically ill patients. 2. Identify common areas of adherence to best practice and gaps in adherence to best practices. 3. Use this information in a clinical setting. Unless otherwise noted below, each faculty or staff s spouse/life partner (if any) has nothing to disclose. Dr. Heyland has disclosed that he received grant/research fees from Fresenius Kabi, Nestle, and Abbott; and was a consultant/advisor for GlaxoSmithKline and Baxter. The remaining authors have disclosed that they have no financial relationships with or interests in any commercial companies pertaining to this educational activity. All faculty and staff in a position to control the content of this CME activity have disclosed that they have no financial relationship with, or financial interests in, any commercial companies pertaining to this educational activity. Visit the Critical Care Medicine Web site ( for information on obtaining continuing medical education credit. Objective: To describe current nutrition practices in intensive care units and determine best achievable practice relative to evidence-based Critical Care Nutrition Clinical Practice Guidelines. Design: An international, prospective, observational, cohort study conducted January to June Setting: One hundred fifty-eight adult intensive care units from 20 countries. Patients: Two-thousand nine-hundred forty-six consecutively enrolled mechanically ventilated adult patients (mean, 18.6 per site) who stayed in the intensive care unit for at least 72 hrs. Interventions: Data on nutrition practices were collected from intensive care unit admission to intensive care unit discharge or a maximum of 12 days. Measurements and Main Results: Relative to recommendations of the Clinical Practice Guidelines, we report average, best, and worst site performance on key nutrition practices. Adherence to Clinical Practice Guideline recommendations was high for some recommendations: use of enteral nutrition in preference to parenteral nutrition, glycemic control, lack of utilization of arginine-enriched enteral formulas, delivery of hypocaloric parenteral nutrition, and the presence of a feeding protocol. However, significant practice gaps were identified for other recommendations. Average time to start of enteral nutrition was 46.5 hrs (site average range, hrs). The average use of motility agents and small bowel feeding in patients who had high gastric residual volumes was 58.7% (site average range, 0% 100%) and 14.7% (site average range, 0% 100%), respectively. There was poor adherence to recommendations for the use of enteral formulas enriched with fish oils, glutamine supplementation, timing of supplemental parenteral nutrition, and avoidance of soybean oil-based parenteral lipids. Average nutritional adequacy was 59% (site average range, 20.5% 94.4%) for energy and 60.3% (site average range, 18.6% 152.5%) for protein. Conclusions: Despite high adherence to some recommendations, large gaps exist between many recommendations and actual practice in intensive care units, and consequently nutrition therapy is suboptimal. We have identified best achievable practice that can serve as targets for future quality improvement initiatives. (Crit Care Med 2010; 38: ) KEY WORDS: nutrition support; enteral nutrition; parenteral nutrition; intensive care unit; critical illness; clinical practice guidelines *See also p Project Leader (NEC), Department of Community Health and Epidemiology, Queen s University, and Clinical Evaluation Research Unit, Kingston General Hospital, Kingston, ON, Canada; Project Leader (RD), Clinical Evaluation Research Unit, Kingston General Hospital, Kingston, ON, Canada; Senior Biostatistician (AGD), Kingston General Hospital, Kingston, ON, Canada; Research Analyst (XJ), Clinical Evaluation Research Unit, Kingston General Hospital, Kingston, ON, Canada; and Professor of Medicine (DKH), Queen s University, Kingston, ON, Canada. Ms. Cahill currently holds a Canadian Institutes for Health Research (CIHR) Fellowship in Knowledge Translation. For information regarding this article, dkh2@queensu.ca Copyright 2010 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: /CCM.0b013e3181c0263d 395

2 Quality healthcare is related to the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge (1). Achieving quality nutritional care is a daily challenge faced by critical care practitioners and dietitians around the world (2 4). This challenge is augmented by the fact that other complex and high technology therapies in the intensive care unit (ICU) have historically taken precedence over nutritional assessments and interventions. Recently, attention has shifted away from nutrition as support toward nutrition as therapy, with the emergence of high-grade evidence to support the concept that nutrients and nutritional interventions modulate the underlying disease process and impact on patient outcomes (5). Failure to translate this knowledge into practice may result in increased morbidity and mortality for thousands of critically ill patients. To determine best practice, we went through a rigorous process to develop practical, evidence-based Clinical Practice Guidelines (CPGs) on this topic. These CPGs were published in 2003 and are updated on a regular basis, most recently in January 2009 (6, 7). Before their publication, we validated these CPGs and demonstrated that ICU that were more adherent to these guidelines were more likely to successfully feed their patients (8). A natural next step for quality improvement efforts is to audit clinical practice to determine the level of adherence to best practice and identify opportunities for practice improvement. Building on our previous work in Canada (9), we were interested in describing clinical practice in ICUs across the world and, particularly, defining best achievable practice. We define best achievable practice as what is achievable by the top performing sites on key quality indicators. Ideally, 100% adherence with CPG recommendations is the gold standard for all sites, but often such performance may not be realistic and may be beyond the reach of some sites. By documenting actual performance and highlighting top performers, we can offer their performance as a secondary target in performance improvement initiatives. It is secondary in the sense that the primary goal is always 100% adherence with evidence-based CPGs. The purpose of this international, observational study was to describe current nutrition practices in the critical care setting. We sought to compare actual practices to the recommendations of the Critical Care Nutrition CPGs developed in Canada (7), and thus define the gaps or opportunities for improvement, and to identify the best achievable performance relative to key nutritional quality indicators. MATERIALS AND METHODS Study Design and Participants We conducted a prospective, observational study of nutrition practices in ICUs across the world. Participating ICU sites were recruited by disseminating study information to our existing network of Canadian dietitians who were involved in previous surveys (8 10), the membership registries of clinical nutrition societies and critical care societies across the world, and contacting individual health care providers known to have an interest in critical care or nutrition therapy. We also advertized the survey on our web site, (6). To be eligible, ICUs had to have at least eight beds and have a registered dietitian or an individual who had knowledge of clinical nutrition present to complete data collection. We excluded ICUs with less than eight beds because these units typically do not care for ventilated patients for 24 hrs and may not have adequate resources to complete data collection. However, if these smaller units wished to participate, and had the resources to do so, permission was granted on a case-by-case basis. Data Collection Participants were asked to enter information describing the characteristics of their hospital and ICU and general aspects of nutrition practice (e.g., use of feeding protocol or algorithms). Starting January 25, 2007, participants were asked to identify consecutive, eligible study patients who were being cared for in their units on that day. Critically ill adult (aged 18 yrs or older) patients who were mechanically ventilated within the first 48 hrs of admission to ICU and who remained in ICUs for 72 hrs were enrolled in this study. We excluded patients who were not intubated or who were receiving mask ventilation. Each participating ICUs was asked to enroll 20 patients, but only sites that enrolled at least eight patients were included in the final database. Data were collected retrospectively from the hospital records (with exception of head of the bed elevation, which was obtained through direct observation) and were entered online using a secure web-based data collection tool (6). Given the observational nature of this study, we made no attempt to standardize or influence care. We simply collected the following information related to the enrolled patients: admission category (surgical vs. medical), admission diagnosis, gender, age, body mass index, Acute Physiology and Chronic Health Evaluation II score (11), and baseline nutrition assessment (i.e., prescribed energy and protein intake). The presence or absence of acute respiratory distress syndrome was based on chart abstraction using standard definitions (12); this determination was not verified centrally. The type and amount of nutrition received, morning blood glucose levels, insulin dose, and use of motility agents were recorded daily until death or ICU discharge, or for a maximum of 12 days. ICU and hospital outcomes were followed-up for 60 days. Data Management Range, data completeness, and logic checks were incorporated into the web-based data collection tool and database. The entered data were checked by the Methods Centre to identify errors, inconsistencies, and omissions. Data queries were sent back to the participating ICU for verification. Statistical Analysis Categorical variables are described as counts and percentages. Length of stay variables are reported as medians with quartiles because of their positive skew. Other continuous variables are reported as means with standard deviations. Variables related to the CPGs are reported as overall averages (or percentages) with the range of site averages. The site ranges are interpreted as the best and worst site performances relative to the Nutrition CPGs. Adequacy of nutrition (an indicator of overall performance) was calculated as the amount of calories or protein received (from either enteral nutrition [EN] or parenteral nutrition [PN] but not oral) plus propofol, divided by the amount prescribed as per the baseline assessment and expressed as a percentage. Days without EN or PN were included and counted as 0% adequacy. Days after permanent progression to exclusive oral intake were excluded from the calculation of nutritional adequacy. We arbitrarily selected 80% nutritional adequacy as an indicator of high performance. Institutional ethics approval was obtained from the Health Sciences Research Ethics Board at Queen s University, Kingston, Ontario, and additional centers if required for their participation. The need for informed patient consent was waived given the nature of this study. 396

3 Table 1. Characteristics of participating intensive care units RESULTS A total of 208 ICUs registered to participate in the survey, 50 ICUs were excluded because they had entered data on less than eight patients. The remaining 158 ICUs from 20 countries enrolled an average of 18.6 patients (range, 9 26) for a total of 2946 patients and 27,944 patient days. Characteristics describing the participating ICUs and patient demographics are found in Tables 1 and 2, respectively. Type of Nutrition Total, n 158 Geographic region Canada 48 (30.4%) USA 39 (24.7%) UK 22 (13.9%) Europe (other) 14 (8.9%) Australia and New Zealand 22 (13.9%) Asia 8 (5.1%) Latin America 5 (3.2%) Hospital type Teaching 122 (77.7%) Nonteaching 35 (22.3%) Size of hospital (beds) Mean (SD) 638 (493) Multiple ICUs in hospital Yes 80 (50.6%) ICU structure Open 30 (19%) Closed 125 (79.1%) Other 3 (1.9%) a Case-mix groups admitted to ICU Medical 136 (86.1%) Surgical 140 (88.6%) Trauma 98 (62.0%) Pediatrics 26 (16.5%) Neurologic 103 (65.2%) Neurosurgical 78 (49.4%) Cardiac surgery 51 (32.3%) Burns 28 (17.7%) Other 18 (11.4%) Presence of medical director Yes 148 (93.7%) Size of ICU (beds) Mean (SD) 17.4 (11.3) Presence of dietitian(s) Yes 134 (84.8%) Full-time equivalent dietitians (per 10 beds) Mean (SD) 0.4 (0.2) a One intensive care unit (ICU) may admit multiple case-mix groups. The CPGs strongly recommend that EN be used rather than PN, and in critically ill patients with an intact gastrointestinal tract, it is strongly recommend Table 2. Patient characteristics Total, n 2946 Number of patients, per site Mean (SD) 18.6 (3.3) Age, yrs Mean (SD) 59.7 (17.6) Gender Female 1217 (41.3%) Male 1729 (58.7%) Admission category Medical 1785 (60.6%) Surgical: elective 419 (14.2%) Surgical: emergency 741 (25.2%) Admission diagnosis Cardiovascular/vascular 565 (19.2%) Respiratory 780 (26.5%) Gastrointestinal 411 (14.0%) Neurologic 359 (12.2%) Sepsis 253 (8.6%) Trauma 284 (9.6%) Other 293 (9.9%) APACHE II score Mean (SD) 21.5 (7.9) Presence of ARDS Yes 379 (12.9%) BMI, kg/m 2 Mean (SD) 27.5 (8.2) Length of mechanical ventilation, days Median (IQR) 7.7 ( ) Length of ICU stay, days Median (IQR) 10.8 ( ) Length of hospital stay, days Median (IQR) 22.0 ( ) Patient died within 60 days Yes 802 (29.3%) APACHE, Acute Physiology and Chronic Health Evaluation; ARDS, acute respiratory distress syndrome; BMI, body mass index; ICU, intensive care unit; IQR, interquartile range. that PN not be used routinely. Figure 1 shows the distribution of the type of nutrition provided by day. EN alone was provided on most of the patient days (61.7%). The proportion of patient days receiving EN alone ranged from 97.3% at one site to 1.0% at the worst performing site. PN alone was received on 11.8% of patient days with sites ranging from 0% to 72.3%. PN was used in combination with EN on 6.7% of patient days (site range, 0% 71.8%). For those patients who received PN, there was no contraindication to EN on half of the patient days on which PN was received (357 days [50.3%]). Reasons EN was not provided to patients on PN are found in Table 3. No nutrition (i.e., via the enteral, parenteral, or oral route) was provided on 19.7% of patient days. However, one site was successful in providing nutrition on 100% of days; however, at the other extreme, one site did not provide nutrition on 57.6% of patient days. Timing of Initiation of Nutrition The CPGs recommend that EN be initiated within 24 to 48 hrs of admission to the ICU. Of the patients who received EN after ICU admission but within the 12 days of observation (n 2221), overall, the average time from admission to start of EN was 46.5 hrs across all sites. The best performing site was successful in initiating EN an average of 8.2 hrs after ICU admission, with the worst site delaying initiation of EN for 6 days (149.1 hrs), on average. The mean time to initiation of EN was 48 hrs in 96 (60.8%) of participating ICUs, and 24 hrs in 21 (13.3%) of participating ICUs. The CPGs recommend that for critically ill patients started on EN, PN should not be started at the same time, and should not be started until all strategies to maximize EN delivery have been attempted. For those patients (486 of %) who received a combination of EN and PN, PN was started after EN in 153 (153 of %) patients. Among these patients, PN was started on average 81.6 hrs after EN (range, hrs), and 49 of 153 (32.0%) received either motility agents or small bowel feeding to maximize EN delivery before initiation of PN. There were 14 of 127 (11.0%) ICUs that complied with the recommendation and started PN after EN in all their supplemental PN patients. Of these sites, there was a wide variation in the timing of initiation of PN after the initiation of EN, with ICUs on average starting PN at hrs (range, hrs) after EN. Of these sites that started PN after EN, 4 of 14 (28.6%) ICUs provided either motility agents or small bowl feeding before initiating PN to all their patients who received supplemental PN (100% compliance). Strategies to Maximize Delivery of EN The CPGs recommend the use of motility agents and small bowel feedings, especially in patients who demonstrate high gastric residual volumes and who are not tolerating adequate amounts of EN. Of all the patients who received EN or EN in combination with PN, (n 2908), a total of 782 (27%) had high gastric residual volumes. Overall, the average use of motility agents and small 397

4 compliance with the use of other recommended pharmaconutrients (Table 4). Strategies to Optimize Benefits and Minimize Risks of Parenteral Nutrition Figure 1. Type of nutrition provided by day. Hatched bars indicate overall percentages and the error bars represent the range of site-specific percentages. The best achievable site percentages are circled. ICU, intensive care unit; EN, enternal nutrition; PN, parenteral nutrition. Table 3. Reasons enteral nutrition was not provided in patients administered parenteral nutrition Total, n 1039 Number of patients-days Contraindication to EN Mechanical bowel 28 (2.7%) obstruction Bowel ischemia 56 (5.4%) Small bowel ileus 176 (16.9%) Small bowel fistulas 36 (3.5%) GI perforation 12 (1.2%) Short gut syndrome 2 (0.2%) No contraindication/ 357 (34.4%) clinical reason Hemodynamic instability 108 (10.4%) Proximal bowel anastomosis 95 (9.1%) Not tolerating EN (high 34 (3.3%) gastric residuals/ emesis) GI surgery 23 (2.2%) GI bleed 20 (1.9%) No access to small bowel/ 17 (1.6%) feeding tube blockage Pancreatitis 15 (1.4%) Other 60 (5.8%) EN, enteral nutrition; GI, gastrointestinal. bowel feeding in patients with high gastric residual volumes was 58.7% (461 of 782) and 14.7% (115 of 782), respectively (Fig. 2). At the site level, only 34 (21.5%) and four (2.5%) ICUs utilized motility agents and small bowel tubes, respectively, in all patients with high gastric residual volumes (100% compliance). Whereas a total of 24 (15.2%) and 92 (58.2%) sites did not utilize motility agents and small bowel tubes in any of their patients who had high gastric residuals (100% noncompliance). The CPGs also recommend that patients who receive EN have the head of bed elevated to 45 degrees where possible. Of all the patients who received EN (or EN and PN combined), the mean head of the bed angle was degrees. At a site level, a total of 8 (5.1%) ICUs were successful in achieving an average head of bed elevation angle of 45 degrees or greater in all their patients who received EN (100% compliance), whereas at one site all patients who received EN remained supine (100% noncompliance). The CPGs recommend the use of a feeding protocol that incorporates motility agents (e.g., metoclopramide) at initiation of feeding and tolerates a higher gastric residual volume of 250 ml. Of the 158 participating ICUs, 126 (79.7%) reported using a feeding protocol or bedside algorithm. The majority of these protocols included motility agents (88 ICUs [69.8%]). The mean threshold gastric residual volume in these protocols was 217mL ( ml), with 24 ICUs (15.2%) using the recommended threshold volume of 250 ml. Composition of Enteral Nutrition and Use of Pharmaconutrition Whereas the majority of sites were compliant with the recommendation to use polymeric enteral formulas, and to avoid using arginine-supplemented enteral formulas, there was suboptimal The guidelines recommend that patients who receive PN should receive parenteral glutamine supplementation, when available. Of the patients in the study who received PN (n 697 from 141 sites), the average utilization of parenteral glutamine supplementation in these patients was 9.3%. Only five (3.5%) ICUs provided parenteral glutamine to all their patients who received PN, and 118 (83.7%) did not provide parenteral glutamine supplementation to any of their PN patients. The CPGs recommend that low-dose PN and withholding lipids high in soybean oil should be considered in critically ill patients who are not malnourished, are tolerating some EN, or when PN is indicated for short-term use ( 10 days). The mean daily calories received per kilogram from PN across all ICUs was 16.8 kcal/kg per day, the lowest average dose of PN provided at one ICU was 2.7 kcal/kg per day, and the highest average dose at a single ICU was 35.5 kcal/kg per day. A total of 108 (68.4%) ICUs adhered to the recommendation and provided 20 kcal/kg per day to all their patients who received PN alone or EN and PN combined. Of the 4863 patient days on which PN was received, the average number of days during which lipids were withheld was 1331 (27.4%). Only 18 (11.4%) ICUs adhered to the recommendation and withheld soybean oil-based lipids or provided non-soybean lipid formulations to their patients who were receiving PN, whereas 64 (40.5%) ICUs were noncompliant with the recommendation and provided soybean oil-based lipids on all patient days during which PN was administered. Intensive Insulin Therapy The CPGs recommend that avoiding hyperglycemia (blood glucose 10 mmol/l) is recommended in all critically ill patients. Across sites, the mean daily glucose level was 7.5 mmol/l (site average range, mmol/l). Overall, one ICU successfully avoided hyperglycemia on all 241 patient days, whereas three (1.9%) ICUs had average daily blood glucose levels 10 mmol/l. 398

5 3 illustrates the adequacy of calories in the best, average, and worst performing site across the 12 days of observation. At a site level of analysis, only 14 (8.9%) and 18 (11.4%) ICUs successfully achieved overall average adequacy of calories and protein of 80%, respectively. Because the CPGs focus on the use of EN in preference to PN, when we only considered calories or protein from EN and from PN in the presence of a contraindication to EN (excluding calories from PN when used with an intact gastrointestinal tract), the average adequacy of calories and protein was 51.5% (site range, 4.1% 90.5%) and 52.3% (3.5% 117.7%), respectively. At the site level, a total of five (3.2%) and eight (5.1%) ICUs successfully achieved average adequacy of calories and protein from EN and appropriate PN of 80%, respectively. The average adequacy of calories and protein from EN in patients who received EN was 45.0% (4.1% 85.4%) and 45.5% (3.5% 93.7%), respectively. At the site level, a total of one (0.6%) and four (2.5%) ICUs successfully achieved overall average adequacy of calories and protein from EN of 80%, respectively. Figure 2. Use of motility agents and small bowel feeding in patients with high gastric residual volumes (GRV). Hatched bars indicate overall percentages and the error bars represent the range of site-specific percentages. The best achievable site percentages are circled. Table 4. Strategies to maximize benefits of enteral nutrition: Percent of patients at each site receiving the recommendation Guideline Recommendation Overall Best Site Worst Site Use polymeric solutions 94% 100% 10.5% Do not use arginine-supplemented diets 3.8% 0% 92.3% Use diet enriched with fish oil, borage oil, and 4.5% 100% 0% antioxidants in ARDS Use of enteral glutamine in burn patients 27.8% 100% 0% Use of enteral glutamine in trauma patients 7.9% 100% 0% ARDS, acute respiratory distress syndrome. Figure 3. Overall performance. The amount of calories received by enteral nutrition and parenteral nutrition as a percentage of the maximum calories prescribed at baseline assessment in all patients across the 12 days of observation. ICU, intensive care unit. Overall Performance Sixty-eight (2.3%) of study patients did not have a nutritional prescription recorded and were excluded from the analysis of adequacy of nutrition. The average adequacy of total calories and protein over the 12 days of observation was 59.0% (site range, 20.5% 94.4%) and 60.3% (site range, 18.6% 152.5%), respectively. Figure DISCUSSION We conducted a prospective, observational study in 158 ICUs across five continents to compare nutrition practice to the Critical Care Nutrition CPGs and determine the best achievable nutrition practices across participating sites. We observed that there are large gaps between guideline recommendations and current practice. For example, for more than one-third of their time in the ICU, patients did not receive any EN; more than one-third of ICUs failed to initiate EN within 48 hrs of admission to the ICU; and the utilization of motility agents and small bowel feeding as strategies to optimize EN in patients with high gastric residuals was very low. The consequence of poor adherence to these guideline recommendations is the low nutritional adequacy observed across sites, with ICUs delivering only approximately 60% of prescribed calories and protein. When we juxtapose these performance data with the results of other studies that suggest that ICU patients who receive inadequate amounts of nutrition have increased complications, prolonged mechanical ventilation, longer time in ICUs, and greater risk of death (13 15), we conclude that there is an urgent need to address these performance issues. There is a similar performance gap with respect to pharmaconutrition. Several randomized trials and meta-analyses of randomized trials have shown that supplementation of specialized nutrients clearly modulates the underlying inflammatory state and improve clinical outcomes (5). National and international nutrition CPGs endorse their use (7, 16, 17). 399

6 Yet, the results of our study show that there is limited use of enteral formulas enriched with fish oils for patients with ARDS, and the use of glutamine supplementation is rare, particularly in patients who are receiving PN, in whom it has been shown to significantly reduce infectious complications and hospital mortality (6). The identification of these gaps between evidence-based recommendations and actual nutrition practices highlight opportunities for improving the provision of nutrition therapy in critical care practice. However, the observation that these knowledge practice gaps do not exist across all ICUs, and that some sites are consistently able to adhere to the recommendations in daily practice or achieve at least a high degree of adherence, provide compelling evidence that the translation of evidence, in the form of CPGs, into practice is an achievable and worthwhile pursuit. This is the value of defining the best achievable level of performance, which sets the standard for all other ICUs. For some recommendations of the Critical Care Nutrition CPGs, namely, use of EN in preference to PN, glycemic control 10 mmol/l, lack of utilization of arginine-enriched enteral formulas, delivery of hypocaloric PN, and the presence of a feeding protocol, the evidence practice gap was observed to be narrow, with the majority of participating ICUs adhering to the recommendations. Whereas we should strive for 100% adherence to the CPGs in all patients, we must also be cognizant that these recommendations are derived from randomized, controlled trial data, and as such may be difficult to replicate in real-life practice. Thus, setting 100% adherence as our goal may not be an achievable level of performance in some situations. Hence, establishing a realistic, achievable performance goal is important to effectively improve quality of care. In the business sector, benchmarking has long been recognized as an effective quality improvement strategy and is beginning to be applied to healthcare (18). Benchmarking refers to a process of measuring performance, comparing performance to other organizations and against established standards of best practice, followed by modification of practice when indicated. Benchmarking is a form of clinical audit. A Cochrane review of randomized, controlled trials of audit and feedback concluded that this strategy can be effective in improving professional practice, and that the relative effectiveness of audit and feedback is likely to be greater when baseline adherence to recommended practice is low and when feedback is delivered more intensively (19). Historically, feedback compared an individual s performance to the mean performance of a peer group. There is emerging evidence that the effectiveness of feedback in motivating change in practice may be enhanced if an achievable standard is used based on the practices of top performers (20). Thus, benchmarking provides an opportunity to learn from others, set goals based on real, achievable levels of performance, while also stimulating interest among the nutrition community to collaborate in efforts to change practice. Our findings are consistent with other observational studies of nutrition practices in critically ill patients who also report considerable variation in practices across ICUs or document that provision of nutritions therapy, particularly via the enteral route, is suboptimal (2 4, 9). Through our previous work in Canadian ICUs, we observed that patients only received 58% of their prescribed calories and protein (9, 10). In contrast, in a recent single center observational study in a medical ICU in Germany, the daily administered amount of EN was much higher than in our study at 86.2% (SD 30.4%) (13). However, this analysis was based on only 61 patients who had received EN for a minimum of 7 days. Because it is common practice to gradually increase nutrition over the first few days of ICU stay, with little or none given on the first few days, the average amount of calories and protein received is strongly related to the total number of days of nutrition received. Hence, the lower nutritional adequacy observed in our study may be attributable to the inclusion of patients who received EN for 7 days. Nevertheless, the successful performance of this single German site confirms our assertion that some ICUs are able to achieve excellent performance, with respect to adequacy of nutrition via the enteral route, whereas others struggle to provide acceptable amounts of energy and protein. It was not the intent of the current study to explain why variation exists or to elucidate the mechanisms by which sites achieve best practice. However, such an analysis may offer further insights in how to improve performance. Achieving adherence to guidelines is profoundly complex, being determined by the interplay of a number of factors that act as barriers or enablers to practice. We have conducted a qualitative study to better understand the factors associated with adherence of the Critical Care Nutrition CPGs (21). We discovered that the characteristics of the specific guideline (e.g., topic, format, developer), the strategies used to implement the guidelines, institutional factors, the critical care provider s knowledge and attitude, and the clinical condition of the patient all influence the readiness with which guideline recommendations are adopted. To the extent that specific products are recommended (such as parenteral glutamine or alternatives to soy bean lipid emulsions), their availability in hospital formularies further impacts on compliance with adoption of the guidelines. Designing and conducting more effective quality improvement interventions by tailoring educational strategies to overcome these site-specific barriers and enablers may aid in facilitating the realization of attaining the goal of best achievable nutrition practices across ICUs. The major strength of this study is the observation of actual nutrition practices compared to evidence-based recommendations in a large sample of ICUs from across 20 countries. The high response to our invitation to participate illustrates that critical care practitioners are interested in improving the provision of nutrition therapy in their ICU. To audit practice within a single site is time-consuming and the inferences permitted from the results are limited by the omission of an external comparison. The benchmarking aspect of this study, highlighting what is best achievable, provides powerful information for individual critical care practitioners and researchers alike. A limitation of the study is that the participating ICUs are not a representative sample of all ICUs across the world. Some geographic regions did not participate or had low representation. Because participation was voluntary, we are aware of the potential for response bias. Consequently, it is likely that the evidence practice gap that we identified in our study is in reality larger. Another limitation of this study may be that we used the Critical Care Nutrition CPGs developed in Canada, and these recommendations may not represent the standard of care for all ICUs around the world participating in this survey. However, the Canadian CPGs are very similar to other CPGs developed by other countries or societies and only differ in recommendations regarding use of arginine-supplemented diets and timing of PN (7, 16, 17). 400

7 CONCLUSIONS We have shown that adhering to the recommendations of the Canadian Critical Care Nutrition CPGs is an achievable goal, and there are great opportunities for improvement of nutrition practices in ICUs. In 158 ICUs across the world, we observed that the use of EN is predominant, there is limited use of PN, the use of arginineenriched enteral formulas is appropriately low, and glycemic control is adequate. However, overall the adequacy of nutrition is suboptimal. Future quality improvement strategies should focus on ensuring that EN is started earlier, increasing the utilization of motility agents and small bowel feeding in patients with high gastric residual volumes, withholding soybean oil lipids in patients receiving PN, and improving the use of enteral formulas enriched with fish oils in patients with acute respiratory distress syndrome and glutamine in burn and trauma patients. The performance of the top-achieving sites should be the target for those engaged in these quality-improvement activities. Efforts to achieve these benchmarks may translate into reduced morbidity and mortality of critically ill patients. ACKNOWLEDGMENTS The authors are grateful to the critical care practitioners from all participating ICU sites for their dedication and commitment to collecting data for this study. REFERENCES 1. Institute of Medicine: Quality Assurance in Medicare. Washington, DC, National Academy Press, Rice TW, Swope T, Bozeman S, Wheeler AP: Variation in enteral nutrition delivery in mechanically ventilated patients. Nutrition 2005; 21: De Jonghe B, Appere-De-Vechi C, Fournier M: A prospective survey of nutritional support practices in intensive care unit patients: what is prescribed? What is delivered? Crit Care Med 2001; 29: O Meara D, Mireles-Cabodevila E, Frame F, et al: Evaluation of Delivery of Enteral Nutrition in Critically Ill Patients Receiving Mechanical Ventilation. Am J Crit Care 2008; 17: Jones NE, Heyland DK: Pharmaconutrition: a new emerging paradigm. Curr Opin Gastroenterol 2008; 24: Critical Care Nutrition: Available at: carenutrition.com. Accessed January 2, Heyland DK, Dhaliwal R, Drover JW, et al: Canadian Clinical Practice Guidelines for Nutrition Support in Mechanically Ventilated, Critically Ill Adult Patients. JPEN J Parenter Enteral Nutr 2003; 27: Heyland DK, Dhaliwal R, Day A, et al: Validation of the Canadian Clinical Practice Guidelines for Nutrition Support in Mechanically Ventilated, Critically Ill Adult Patients: Results of a prospective observational study. Crit Care Med 2004; 32: Heyland DK, Schtoter-Noppe D, Drover JW: Nutrition support in the critical care setting: Current practice in Canadian ICUs Opportunities for improvement. JPEN J Parenter Enteral Nutr 2003; 27: Jain M, Heyland DK, Dhaliwal R, et al: Dissemination of the Canadian Clinical Practice Guidelines for Nutrition Support: Results of a Cluster Randomized Trial. Crit Care Med 2006; 34: Knaus WA, Draper EA, Wagner DP, et al: APACHE II: a severity of disease classification system. Crit Care Med 1985; 13: Artigas A: Current definitions of acute lung injury and the acute respiratory distress syndrome. Intensive Care Med 2000; 26: Petros S, Engelmann L: Enteral nutrition delivery and energy expenditure in medical intensive care patients. Clin Nutr 2008; 25: Villet S, Chiolero RL, Bollman MD, et al: Negative impact of hypocaloric feeding and energy balance on clinical outcome in ICU patients. Clin Nutr 2005; 24: Rubinson L, Diette GB, Song X, et al: Low caloric intake is associated with nosocomial bloodstream infections in patients in the medical intensive care unit. Crit Care Med 2004; 32: Kreymann KG, Berger MM, Deutz NEP, et al: ESPEN Guidelines on Enteral Nutrition: Intensive Care. Clin Nutr 2006; 25: Doig GS, Simpson F: Evidence-Based Guidelines for Nutritional Support of the Critically Ill: Results of A Bi-National Guideline Development Conference. Sydney, EvidenceBased.net, Camp RC, Tweet AG: Benchmarking applied to health care. Jt Comm J Qual Improv 1994; 20: Jamtvedt G, Young JM, Kristoffersen DT, et al: Audit and feedback: Effects on professional practice and health care outcomes. Cochrane Database of Systematic Reviews 2006, Issue 1. Art. No.: CD DOI: / CD pub2 20. Kiefe CI, Allison JJ, Williams OD, et al: Improving Quality Improvement Using Achievable Benchmarks For Physician Feedback: A Randomized Controlled Trial. JAMA 2001; 285: Jones N, Suurdt J, Ouellette-Kuntz H, et al: Implementation of the Canadian Clinical Practice Guidelines for Nutrition Support: A multiple case study of barriers and enablers. Nutr Clin Pract 2007; 22: