clinical investigations in critical care

Why Do Physicians Not Follow Evidence-Based Guidelines for Preventing Ventilator-Associated Pneumonia?*
A Survey Based on the Opinions of an International Panel of Intensivists
Jordi Rello, MD, PhD; Carmen Lorente, MD; Maria Bod , MD; Emili Diaz, MD; Maite Ricart, RN; and Marin H. Kollef, MD, FCCP

Objective: Adherence to clinical practice guidelines is highly variable. Our objective was to review barriers to physicians adherence to evidence-based guidelines (EBGs) for preventing ventilator-associated pneumonia (VAP). Methods: A questionnaire was administered to 110 opinion leaders on VAP from 22 countries to indicate whether 33 pharmacologic and nonpharmacologic practices that had been listed in a recent publication had been implemented in their ICUs. If these prevention strategies were not used, the respondents were asked to indicate one of seven reasons for nonadherence, with the objective of identifying barriers to adherence to EBGs. Results: The overall nonadherence rate was 37.0%. The nonadherence rate was 25.2% for strategies recommended for clinical use, compared with 45.6% for strategies with less effectiveness (odds ratio [OR], 1.80). Pharmacologic strategies had a higher degree of nonadherence than nonpharmacologic strategies (OR, 2.92). Nonadherence to recommendations graded A, B, C, D, and U based on an objective assessment of the consistency of the supporting evidence was 41.3%, 35.7%, 16.0%, 45.7%, and 20.8%, respectively. The most common reasons for nonadherence were the following: disagreement with interpretation of clinical trials (35%); unavailability of resources (31.3%); and costs (16.9%). Conclusion: We conclude that nonadherence to EBGs for preventing VAP was common and largely uninfluenced by the degree of evidence. A rational approach toward improving VAP guideline adherence should take into account the heterogeneous factors that influence physician adherence to them. (CHEST 2002; 122:656 661)
Key words: evidence-based guidelines; ICU; nonadherence; prevention; ventilator-associated pneumonia Abbreviations: EBG evidence-based guideline; VAP ventilator-associated pneumonia

Centers for Disease Control and Prevention T he has been developing evidence-based guidelines (EBGs) for the prevention and control of nosocomial infections since the early 1980s. Clinical practice guidelines aim to improve the quality of care, to
*From the Critical Care Department (Drs. Rello, Lorente, Bod , and Diaz, and Ms. Ricart), University Hospital Joan XXIII, University Rovira and Virgili, Tarragona, Spain; and the Pulmonary and Critical Care Division (Dr. Kollef), Washington University School of Medicine, St. Louis, MO.
656

decrease costs, and to reduce inappropriate variation in decision making in this setting.13 The development of EBGs should be based on multidisciplinary
This study was supported by Comisio Interdepartamental de Recerca i Innovacio Technologica (grant No. 2000-SGR-128) and Distincio per la Promocio de la Recerca Universitaria. Manuscript received August 13, 2001; revision accepted February 13, 2002. Correspondence to: Jordi Rello, MD, PhD, Critical Care Department, University Hospital Joan XXIII, Dr Mallafre Guasch, 4, E43007 Tarragona, Spain; e-mail: jrc@hjxxiii.scs.es
Clinical Investigations in Critical Care

Downloaded From: https://journal.publications.chestnet.org/ on 01/21/2013

participation and should incorporate methods the efficacy and cost-effectiveness of which are supported by evidence-based clinical studies.4,5 Despite the rise in evidence-based medicine and the existence of clinical practice guidelines of proven utility, the use of these guidelines in daily clinical practice is still limited. Rates of adherence to guidelines are variable, ranging from 20% to nearly 100%, depending on the guideline.1 Moreover, the degree of evidence in favor of a strategy does not always guide practice.1,2 Despite the fact that physician adherence to guidelines is critical to their ultimate success, there have been few published studies of the factors that influence such adherence. To help prevent ventilator-associated pneumonia (VAP), several resources and strategies are available. Airway management is an important variable influencing the development of VAP.4 Kollef5 has identified 33 potentially useful strategies for the prevention of VAP. These strategies were taken from a systematic review of the literature, and the appropriate value of each recommendation was judged with respect to consistent methodological standards. A Franco-Canadian survey6 reported that such prevention guidelines were used in 64% and 30% of their ICUs, respectively (p 0.01). Decisional responsibility for each strategy differed among institutions. In addition, despite a large amount of evidence regarding the beneficial effects of hand washing and other infection control practices, studies in the literature have repeatedly documented that the importance of these procedures has not been sufficiently recognized by health-care workers,7 and compliance with hand-hygiene practices and other infection control measures is unacceptably low.7,8 The factors that influence physicians adherence to guidelines are likely to vary. Therefore, we performed an investigation with two main goals. The first goal of this study was to determine the level of compliance with EBGs for the prevention of VAP among an international panel of critical care clinicians. Our second objective was to identify the reasons for noncompliance with these EBGs. It was our intention that these data would assist in understanding how clinicians make decisions regarding the implementation of EBGs in the ICU setting. More importantly, such information also could be used in developing future programs that are aimed at more effectively implementing EBGs. Materials and Methods
We prepared a questionnaire that assessed 33 potential strategies for preventing VAP. The interventions examined on the questionnaire and the recommendation for clinical use (ie, effective, ineffective, or strategies of equivocal or undetermined
www.chestjournal.org

effectiveness) were taken directly from the review article by Kollef.5 The items listed in Tables 3 and 4 are taken directly from this publication. We asked all physicians who had previously been selected and invited to attend an international symposium on therapy for VAP, which was held in Barcelona on May 20 to 21, 2000, to answer the questionnaire. All participants were physicians working as intensivists and ICU directors in Europe, Asia, Africa, South America, and the United States. These clinicians were identified as being local opinion leaders in the management of ICU-acquired infections, including VAP. These clinicians had at least 5 years of experience working in ICUs. The session participants responded to the questionnaire without prior knowledge at the meeting of the recommendations for clinical use and the degree of evidence for each intervention as reported by Kollef.5 If these prevention strategies had not been used, the respondents were asked to indicate one of seven reasons for nonadherence. Only the most important reason was considered for analysis; the seven different possibilities for nonadherence to preventive strategies were as follows: disagreement with the results of previous studies; unavailability of resources; adverse effects; high costs; patient discomfort; convenience for nursing; and other unspecified causes. Participants answered anonymously but were asked to indicate their primary medical specialty and country of origin. All the strategies included in the questionnaire were backed up by studies of varying degrees of scientific-based evidence. Studies were graded as follows: those supported by two randomized, controlled investigations were classified as degree of evidence A; those supported by at least one randomized, controlled investigation were classified as degree if evidence B; those supported by nonrandomized, concurrent-cohort investigations, historical-cohort investigations, or case series as were classified degree of evidence C; those supported by randomized, controlled investigations of other nosocomial infections were classified as degree of evidence D; and those undetermined or not yet studied in clinical investigations were classified as degree of evidence U. The review manuscript5 recommended some of the procedures, while others were deemed less effective and, hence, were not recommended.

Results The survey was answered by 62 of the 110 physicians (56.4%) who were approached. Their countries of origin are shown in Table 1. Overall, 54.8% of responders were from member countries of the European Union. Other European countries represented 30.6% of responders, and 8% of responders were from South America. Russia was represented by four physicians, the Middle Orient was represented by two physicians, South Africa was represented by one physician, and North America by one physician. Their primary specialties were the following: critical care (33 physicians); anesthesiology (10 physicians); critical care and anesthesiology (5 physicians); critical care and respiratory (6 physicians); respiratory (3 physicians); internal medicine (29 physicians); infectious disease (2 physicians); and surgery (1 physician). All of these experts spent at least 25% of their time in the care of critically ill patients in an ICU setting.
CHEST / 122 / 2 / AUGUST, 2002

657

Downloaded From: https://journal.publications.chestnet.org/ on 01/21/2013

Table 1List of Physicians by Country

Countries Argentina Austria Belgium Czech Republic Colombia Croatia Estonia Germany France Greece Italy Portugal Poland Russia Saudi Arabia Slovenia South Africa Spain Sweden Switzerland Uruguay United States Physicians, No. 3 4 5 1 1 1 2 2 6 2 2 1 1 4 2 6 1 9 3 4 1 1

use compared with those with less effectiveness (16.4% vs 23.4%, respectively). The correlation between the degree of evidence and the rate of adherence to the EBGs is shown in Table 2. Nonadherence with recommendations graded A, B, C, D, and U based on an objective assessment of the consistency of the evidence was 41.3%, 35.7%, 16.0%, 45.7%, and 20.8%, respectively. Physicians reasons for nonadherence and the rates of nonadherence are shown in Tables 3 and 4. Overall, the most important barriers to adherence were the following: disagreement with the interpretation of reported trials (35%); the unavailability of resources (31.3%); high costs (16.9%); nursing convenience (3.7%); fear of potential adverse events (2.2%); potential patient discomfort (0.8%); and miscellaneous indications for nonadherence (9.7%). Discussion Little is known about barriers to physicians adherence to clinical practice guidelines. There is very scarce information, to our knowledge, to assess why physicians may not follow EBGs for preventing VAP. We assessed six potential barriers and found the most important to be disagreement with the interpretation of reported trials and lack of resources. In contrast, potential patient discomfort, adverse effects, and convenience for nursing staff were rarely mentioned. In 2000, a Franco-Canadian survey6 evaluated the use of ventilator circuits and secretion management by ICU directors. A unique feature of our study is the inclusion of physicians with an international focus, representing 22 different countries. In addition, we incorporated the role of broader infection control issues in our questionnaire. Consistent with the findings of Cook et al,6 factors other than the degree of evidence in support of specific strategies appeared to be the important barriers to their ad-

The overall rate of nonadherence was 37.0%. For strategies recommended for clinical use (as used by Kollef5 to identify more clinically useful strategies), the nonadherence rate was 25.2%, compared with 45.6% (odds ratio, 1.80) for strategies classified as being less effective. This difference was due to a twofold increase in the percentage of nonadherence for pharmacologic measures classified as being ineffective or of equivocal effectiveness compared with those classified as being effective (78.9% vs 38.9%, respectively). Pharmacologic strategies had a significantly higher degree of nonadherence (odds ratio, 2.92) compared with nonpharmacologic strategies (57.4% vs 19.6%, respectively). In contrast, similar percentages were documented for nonpharmacologic strategies that were recommended for clinical

Table 2Correlation Between the Degree of Evidence and the Rate of Adherence*
Strategy Recommended for Clinical Use, % Variables Implementation Disagree Not available Adverse effects Costs Patient discomfort Nursing convenience Other A 11.7 16.7 50.0 0 10.0 0 3.3 3.2 B 80.0 7.6 3.9 1.0 0 0.4 3.4 3.8 C 84.0 4.3 4.6 0.4 0.8 0 2.1 3.8 D 65.8 9.6 12.6 1.3 5.7 1.0 0 3.9 U 0 0 0 0 0 0 0 0 Overall 48.3 7.4 14.2 0.5 3.3 0.3 1.8 2.9 Strategy Ineffective or of Undetermined Effectiveness, % A 66.5 13.8 9.0 2.3 4.1 0.3 0.9 3.2 B 48.5 25.8 13.5 3.1 2.7 0.3 2.0 4.1 C 0 0 0 0 0 0 0 0 D 8.6 39.7 12.1 0 31.0 0 0 8.6 U 79.2 6.0 4.3 0 5.2 0.9 0.9 4.2 Overall 40.6 17.1 7.8 1.1 8.6 0.3 0.8 4.0

*Disagree disagree with the results. 658

Clinical Investigations in Critical Care

Downloaded From: https://journal.publications.chestnet.org/ on 01/21/2013

Table 3Adherence to EBGs and Reason for Nonadherence to Nonpharmacologic Strategies*

Nonpharmacologic Strategies Removal of nasogastric or endotracheal tube as soon as clinically feasible Avoidance of unnecessary reintubation Avoidance of gastric overdistension Provision of adequate nutritional support Dedicated use of disposable suction catheters Semirecumbent positioning of the patient Oral (non-nasal) intubation Maintenance of adequate pressure in endotracheal tube cuff Postural changes Adequate handwashing between patient contacts Chest physiotherapy Use of a formal infection-control program Humidification with head and moisture exchanger Use of protective gowns and gloves Scheduled drainage of condensate from ventilator circuits Routine changes of ventilator circuit Routine changes of in-line suction catheter Humidification with head and moisture exchanger with bacteriologic filter Daily changes of heat and moisture exchangers Continuous subglottic suctioning Overall Grade Implementation, % D, % NA, % AE, % C, % PD, % Nurs, % Other, % NR, No. C C B C A B D C B B A C A B C A B U A A 100 100 100 95.1 93.1 91.8 91.8 90.0 88.1 85.0 83.3 78.3 77.2 76.7 75.4 71.2 70.9 70.0 58.9 11.7 80.4 0 0 0 1.6 0 3.3 1.6 0 1.7 1.7 3.3 1.7 1.8 5.0 1.6 15.3 5.5 6.9 7.1 16.7 3.7 0 0 0 0 1.7 1.6 0 5.0 3.4 0 8.3 13.3 10.5 3.3 4.9 5.1 18.2 8.6 14.3 50.0 7.4 0 0 0 0 1.7 0 1.6 0 0 0 0 0 1.8 0 1.6 0 0 0 1.8 0 0.4 0 0 0 1.6 1.7 0 0 1.7 0 0 1.7 1.7 3.5 5.0 0 3.4 3.6 10.3 12.5 10.0 2.8 0 0 0 0 0 0 4.9 0 1.7 0 0 0 1.8 0 0 0 0 1.7 0 0 0.5 0 0 0 0 1.7 1.6 0 3.3 5.1 10.0 1.7 0 0 5.0 13.1 1.7 0 1.7 0 3.3 2.4 0 0 0 1.6 0 1.6 0 0 0 3.3 1.7 5.0 3.5 5.0 3.3 3.4 1.8 1.7 5.4 8.3 2.3 1 1 0 1 4 1 1 2 3 2 2 2 5 2 1 3 7 4 6 2

*D disagreement with results; NA not available; AE adverse effects; C costs; PD patient discomfort; Nurs nursing convenience; NR no response.

herence. Pharmacologic strategies had a higher degree of nonadherence than nonpharmacologic ones. In fact, eight of the proven nonpharmacologic strategies had a rate of nonadherence of 10%, and only two proven pharmacologic strategies (ie, avoidance of unnecessary antibiotics and use of antibiotics for neutropenic fever) had a rate of nonadherence of 10%. Moreover, the degree of nonadherence seems to be independent of the strength of the evidence in support of specific interventions that have been reported in previous trials. There was great variability between strategies, ranging from 0 to 88.3% for strategies with grade A evidence. The degree of nonadherence was also independent of the grade of effectiveness. For instance, some strategies considered to be ineffective by trials classified as degree of evidence A had rates of adherence of 50%. These strategies included the following: dedicated use of disposable suction catheters; chest physiotherapy; routine changes of ventilator circuit; and daily
www.chestjournal.org

changes of heat and moisture exchangers. At the other extreme, the continuous aspiration of subglottic secretions, which had been found to be effective in trials with an A degree of evidence, had a rate of nonadherence of 88%. In this case, the most important reason for nonadherence was lack of availability. Selective decontamination, which was the only pharmacologic strategy with strong evidence (class A) against clinical use according to the report by Kollef,5 was implemented by only 16% of opinion leaders, whereas 55% disagreed with this recommendation and based their decision on whether to implement this technique on the circumstances of the individual patient. Based on these findings, interventions to improve adherence to EBGs regarding VAP that ignore the variety of barriers to their implementation are unlikely to be successful. Our findings have broad policy implications. Standard infection control procedures remain the cornerstone for the prevention of nosocomial and devicerelated infections. Interestingly, 78.3% of opinion
CHEST / 122 / 2 / AUGUST, 2002

659

Downloaded From: https://journal.publications.chestnet.org/ on 01/21/2013

Table 4 Adherence and Reasons for Nonadherence for Pharmacologic Strategies*

Pharmacologic Strategies Avoidance of unnecessary antibiotics Antibiotics for neutropenic fever Combination antibiotic therapy Limitation of stress-ulcer prophylaxis to high-risk patients Chlorhexidine oral rinse Granulocyte colony-stimulating factor for neutropenic fever Antibiotic-class rotation Vaccines against Streptococcus pneumoniae, Haemophilus influenzae type b strains and influenza virus Selective digestive decontamination Prophylactic immune globulin Acidification of enteral feeding solutions Routine parenteral prophylactic antibiotics for patients with coma Aerosolized antibiotic prophylaxis Overall Grade Implementation, % D, % NA, % AE, % C, % PD, % Nurs, % Other, % NR, No. C D U B B D C D 96.6 91.5 88.3 67.8 55.4 44.1 36.2 35.7 1.7 1.7 5.0 15.3 17.9 15.3 27.6 19.6 0 3.4 0 1.7 16.1 22.1 13.8 25.0 0 1.7 0 5.1 0 0 1.7 1.8 0 0 0 0 0 10.2 1.7 12.5 0 0 0 0 1.8 0 0 0 0 0 0 1.7 3.6 0 0 0 1.7 1.7 6.7 8.5 5.4 8.5 19.0 5.4 3 3 2 3 6 3 4 6

A D U B B

15.5 8.6 6.9 5.1 1.7 42.6

55.2 39.7 32.8 66.1 50.8 26.8

13.8 12.1 29.3 6.8 35.6 13.8

8.6 0 5.2 8.5 6.8 3

1.7 31.0 0 5.1 0 18.1

0 0 0 0 0 0.1

0 0 5.2 0 0 0.8

5.2 8.6 20.7 8.5 5.1 8.1

4 4 4 3 3

*See legend of Table 3 for abbreviations not used in the text.

leaders reported that the lack of a formal infection control program was, in most cases, due to a lack of availability. Most interviewees agreed on the importance of using protective gowns and gloves (unproven measures) and on the need for adequate hand washing between patient contacts (77% and 85%, respectively) to reduce the rate of acquired nosocomial infections in ICUs. This contrasts with repeated reports that compliance with hand washing in the ICU setting remains low.710 These observations, and the fact that cost was not often cited, stress the importance of educational measures to improve the prevention of VAP and other nosocomial infections. For this reason, these prevention strategies should be disseminated by an active educational process in order to be successful in changing health-care providers behaviors. As for our study design, the sampling frame was based on a selected international group of intensivists who were invited to a conference on VAP. The strengths of our study include the representation of different practices in different countries, which provides a good baseline of the perceptions of these professionals regarding various infection control practices and strategies to prevent hospital-acquired pneumonia. However, we should emphasize that a bias toward responders from European countries is present and that we cannot draw any conclusion from this survey on practices within specific countries. Several other limitations of our study are worth noting. First, the relatively small size of our sample limited our ability to detect weak associations. Second, classifications of these measures were based on
660

a study published in 1999,5 and later reports could modify the classification of the different strategies or increase the list of potentially useful measures. Thus, potential disagreement between expert opinion and EBGs should be interpreted with caution. It is possible that opinion leaders were wrong as well as that new data may have emerged changing the evidence for selected practices. Third, we did not assess the influence of experience on general attitudes toward preventing VAP. It is possible that experience may influence clinicians behaviors regarding EBGs for the prevention of VAP. Fourth, although all participants were considered to be opinion leaders in their respective countries, some selection bias may exist because no objective definition of opinion leaders was employed. Fifth, it remains unclear how many experts were unaware of data regarding a specific intervention, because this response category was not explored. Finally, the implementation of guidelines for preventing VAP is a multidisciplinary task. Thus, it would be interesting to know the opinion of other caregivers (eg, nurses, respiratory therapists, and physiotherapists), which we did not assess. Other interesting follow-up studies should be undertaken based on our findings. For example, this article reflects the opinions of responders but does not indicate what is actually done at the bedside. For example, for almost 150 years most health-care workers (85% in our survey) have been taught that the most effective way to prevent cross-infection is with hand washing before and after every patient contact, although they do not do it.10,11 Thus, reClinical Investigations in Critical Care

Downloaded From: https://journal.publications.chestnet.org/ on 01/21/2013

search is also required to evaluate what is actually done at the bedside for the prevention of VAP. In addition, it would be interesting to know the rates of compliance with the individual strategies adhered to by the responders since these practices may not be fully carried out in their own ICUs. Finally, physicians practice patterns and how to change them has been the subject of considerable research.12 Our findings confirm that published guidelines and clinical trials alone cannot be expected to change practices at the bedside. More active measures are needed to modify practice patterns.1315 Twenty years ago, Eisenberg and Williams16 suggested six ways to change physicians behavior: education; feedback; financial rewards; financial penalties; administrative changes; and physician participation. It is now recognized that interventions combining more than one strategy (eg, education plus feedback) will yield better success rates than single methods.17,18 Moreover, the best way to influence hospital physician behavior may be to identify a problem and then involve the most respected staff as champions of the intervention to gain overall support for its implementation.19,20 In summary, this opinion poll of experts from 22 countries found nonadherence to clinical practice guidelines for the prevention of VAP to be associated with a variety of factors, including the nature of the strategy and its local availability. The most obvious barrier to implementation of the guidelines was disagreement with the interpretation of reported trials. Therefore, successful implementation strategies for the prevention of VAP should take into account the heterogeneous reasons that influence medical decision making.
ACKNOWLEDGMENT: We are indebted to Montse Olona, MD, for statistical and methodological advice.

References
1 Halm EA, Atlas SJ, Borowsky LH, et al. Understanding physicians adherence with a pneumonia practice guideline. Arch Intern Med 2000; 160:98 104

2 Cabana MD, Rand CS, Powe NR, et al. Why dont physicians follow clinical practice guidelines? JAMA 1999; 282:1458 1465 3 Gundersen L. The effect of clinical practice guidelines on variations in care. Ann Intern Med 2000; 133:317318 4 Cook D, De Jonghe B, Brochard L, et al. Influence of airway management on ventilator-associated pneumonia. JAMA 1998; 279:781787 5 Kollef MH. The prevention of ventilator-associated pneumonia. N Engl J Med 1999; 340:627 634 6 Cook D, Ricard JD, Reeve B, et al. Ventilator circuit and secretion management strategies: a Franco-Canadian survey. Crit Care Med 2000; 28:35473554 7 Jarvis WR. Handwashing: the Semmelweis lesson forgotten? Lancet 1994; 344:13111312 8 Pittet D, Mourouga P, Perneger TV. Compliance with handwashing in a teaching hospital: Infection Control Program. Ann Intern Med 1999; 130:126 130 9 Farr BM. Reasons for noncompliance with infection control guidelines. Infect Control Hosp Epidemiol 2000; 21:411 416 10 Heseltine P. Why dont doctors and nurses wash their hands? Infect Control Hosp Epidemiol 2001; 22:199 200 11 Manangan LP, Pugliese G, Jackson M, et al. Infection control dogma: top ten suspects. Infect Control Hosp Epidemiol 2001; 22:242247 12 Cohen MM, Rose DK, Yee DA. Changing anesthesiologists practice patterns: can it be done? Anesthesiology 1996; 85:260 269 13 Nathwani D, Rubinstein E, Barlow G, et al. Do guidelines for community-acquired pneumonia improve the cost-effectiveness of hospital care? Clin Infect Dis 2001; 32:728 741 14 Centers for Disease Control and Prevention. Guidelines: improving the quality. Washington, DC: US Department of Health and Human Services, 1996 15 Lomas J, Haynes RB. A taxonomy and clinical review of tested strategies for the application of clinical practice recomendations: from official to individual clinical policy. Am J Prev Med 1998; 4(suppl):7794 16 Eisenberg JM, Williams SV. Cost containment and changing physicians practice behavior. JAMA 1981; 246:21952201 17 Greco PI, Eisenberg JM. Changing physician practices. N Engl J Med 1993; 329:12711274 18 Solomon DH, Hashimoto H, Daltroy L, et al. Techniques to improve physicians use of diagnostic test. JAMA 1998; 280:2020 2027 19 Lundberg GD. Perseverance of laboratory test ordering: a syndrome affecting clinicians [editorial]. JAMA 1983; 249:639 20 Lunberg GD. Changing physicians behavior in ordering diagnostic tests [editorial]. JAMA 1998; 280:2036

www.chestjournal.org

CHEST / 122 / 2 / AUGUST, 2002

661

Downloaded From: https://journal.publications.chestnet.org/ on 01/21/2013