The trusted source for
healthcare information and
By Gordon D. Rubenfeld, MD, MSc
A nurse manager in an 18-bed medical-surgical ICU wants to implement a sedation-holiday protocol in his ICU. An intensivist in a 14-bed academic medical ICU wants every patient in her ICU with acute lung injury to be identified and receive lung protective ventilation with low tidal volumes. The infection control committee at a large community hospital wants every central venous catheter placed with appropriate barrier technique.
Proponents of these initiatives will encounter a common barrier: to accomplish their goals they will have to figure out how to get clinicians to change their behavior.1-6 It may seem that the biggest impediment to quality care, particularly in critical care, is knowing what the best practices should be. Once clinicians know what should be done, they will simply follow the Nike dictum, Just do it. Unfortunately, as anyone knows who has tried to quit smoking, lose weight, or remember to take a morning vitamin—or as anyone knows who has worked with patients to try to improve adherence to a medical therapy—knowing and doing are only distantly related.
In a strange way, critical care has the advantage of only having to worry about clinician compliance to practice recommendations. Patient adherence, an important barrier to improving outcomes in ambulatory medicine, is generally not an ICU problem. However, critical care does pose unique barriers to changing behavior. The heterogeneous group of clinicians who work in ICUs and come from different clinical backgrounds with varying levels of experience in critical care are a difficult audience to reach with any intervention designed to change behavior. A lack of outcome data on treatment efficacy for many of the syndromes routinely encountered in the ICU leaves a great deal of room for debate on which practices to promote. However, recent studies on ventilatory strategies in ARDS, sedation use by protocol, noninvasive ventilation, the use of red blood cell transfusions, and the pharmacologic therapy of sepsis place a new burden on our shoulders: can we practice what we read?
There are two essential factors that must be in place before implementing a program to change clinical behavior. First, there must be some agreement on the behavior that needs changing. It doesn’t make much sense to put together an extensive program to increase the use of nitric oxide in adults with ARDS, when there is no agreement that this therapy benefits patients with the disease. The process of identifying the clinical priorities for a given ICU and agreeing on what should be done is itself a behavioral change intervention.
This process can be facilitated by the second essential factor in implementing change: data. Knowing what you’re doing and how well it is working is essential to programs designed to change practice. What percentage of patients in your ICU receive DVT prophylaxis? What percentage of patients with ARDS are on appropriately low tidal volumes? How often are central lines placed without full barrier precautions? How often do attending physicians meet with family members to update them on patient condition? How much indiscriminate use of vancomycin is there in the ICU? Obviously, it is much easier to gather data on antibiotic use than on quality of communication, but, without data it will be difficult to convince your colleagues that there is a need for change and it may be even harder to show them the benefits of change.
Given the number of years that clinicians spend in school and going to class, continuing medical education would seem an ideal opportunity to inform clinicians and change behavior. Unfortunately, the evidence shows that the simplest forms of education—lectures and journal articles, including, unfortunately, what you are reading right now—have a negligible effect on practice.1 There are a number of problems with these forms of education: they are too passive, they do not specifically address the learner’s needs, and they do not incorporate a plan for action. This is rather distressing news for those of us who give lectures and write articles, but it does not mean that all education is ineffective. Small group and "hands-on" sessions that engage clinicians in active learning and provide an opportunity to practice and build confidence can be more effective. You should not expect much from having your critical care committee adopt a guideline and place it in your official book of clinical practice guidelines which sits unread in the physicians’ workroom.
Pharmaceutical companies have known for years that direct, one-on-one "detailing," by sales people armed with articles, graphics, and perhaps a few trinkets, can have a profound effect on clinician practice. In an effort to duplicate this success, several investigators have shown that "academic detailing" by pharmacists or other clinicians, similarly armed with articles and graphics, can have a strong effect on clinical practice. Another approach is to enlist the teaching and support of a local opinion leader. Every ICU has them. This is the physician, nurse, or respiratory therapist that everyone turns to for questions on difficult cases. When information is delivered by this trusted individual, clinicians tend to listen and change their practice.
It may seem mercenary, but an obvious way to change behavior is to provide appropriate incentives. If you were paid $500 for every patient who successfully quit smoking, imagine how you might change your practice. Payment incentives can have a profound effect on physician practice, sometimes leading to rather perverse incentives for physicians to generate demand for their own procedures. Many insurance companies are beginning to link reimbursement to achieving certain population health goals. A number of managed care organizations explored a variety of "payment withhold" strategies where physicians were paid a bonus out of the money that was not expended on patient care. Financial incentives for clinical care need to be carefully designed, precisely because they are so potent. The key is not to provide incentives to do "more" or to do "less" but to do what’s right.
Incentives need not be strictly financial. Clinicians will respond to peer pressure as a form of reward and punishment. For example, a growing body of evidence points to the benefits of keeping mechanically ventilated patients at a 45° angle to reduce ventilator associated pneumonia. The ICU nurse plays an important role in patient positioning. What if you had spot position checks on mechanically ventilated patients and gave coffee coupons to the nurses whose mechanically ventilated patients were at 45°?
Prompts are an effective strategy for changing behavior. Having another clinician or a computer keeping an eye on practice and providing a reminder can work wonders. Imagine a morning phone call from a respiratory therapist to an ICU physician with a reminder that a patient meets criteria for ARDS but still has static airway pressures of 40 cm H2O. Or a computer generated page that informs the physician that a patient is on 3 drugs that all prolong the QT interval. One of the clinical pharmacists at an institution at which I worked used to leave a note in the chart "reminding" the ordering physician about the appropriate use of therapeutic drug level tests. A bright orange sticker on a central venous catheter kit might remind clinicians about appropriate barrier technique.
The problem with this form of concurrent feedback is that it can be very expensive. Expanding the role of other clinicians to track care and provide reminders can cost a lot of money. Computerized prompts have been shown to be effective and can be much less expensive. Computerized prompts particularly lend themselves to evaluation of drug interactions and fairly simple algorithmic evaluations.
Perhaps one of the most effective techniques for changing behavior and one that should be part of care in every ICU involves strategies that basically make it easy to do the right thing. The most common example is pre-written admission orders. Physicians can certainly deviate from the check boxes and prewritten doses and should be encouraged to do so when appropriate. However, it is much easier to remember to write for prophylaxis against deep venous thrombosis and gastric ulcers, for raising the head of the bed 45°, and for drawing appropriate daily laboratory tests when the orders for these things are prewritten.
Treatments that require titration, such as sedation, lung-protective ventilation for ARDS, anticoagulation, and insulin dosing are all better performed when a protocol is used to guide care that removes the physician from every step of the process. Empowering other members of the ICU team to guide care ensures that changes occur rapidly and smoothly, freeing physicians to address other aspects of care. Perhaps the area that has been most influenced by this approach in critical care has been the benefits attributed to weaning protocols. The focus of the weaning literature has been away from novel strategies of exercising respiratory muscles toward effective and universal deployment of protocols to evaluate whether patients are ready for extubation.
Academic clinicians may be concerned that all of these preprinted orders and nonphysician clinician based protocols will remove an essential aspect of training from house officers. Physicians-in-training may not learn important skills by losing the valuable experience of choosing a heparin dose or interpreting a set of weaning parameters. Given the wealth of data showing the benefits of protocolized care in specific areas and the value of learning to work on a multidisciplinary critical care team, not to mention a growing body of literature that shows that protocols actually contribute to house staff learning, I suspect that these concerns are unwarranted compared to the demonstrated benefits. House officers need to learn to work with other clinicians in using protocols and to learn when to over-ride protocols since they will use, or should be using, them when they leave the academic medical center. One excellent approach is to involve house officers in the development and implementation of protocols. This teaches them 2 important lessons: the value of protocolized care and how to implement it in the critical care setting.
There are other interventions that fall under the category of "Make it easy for them." A number of organizational changes in the ICU have been shown to change behavior and outcomes. The most drastic of these is closing the ICU to all admissions except from a selected attending staff who are committed to critical care. Other options are developing teams with specific expertise. For example, a large hospital with many critical care beds may have a line placement team consisting of an intensivist and nurse whose only responsibility is placing and caring for central venous catheters. Other examples include special care units for "chronically critically ill" patients or for hopelessly ill patients. The rationale for these special care units and teams is that clinicians who see a high volume of patients with a specific illness tend to have better outcomes caring for them.
Which of the above interventions is the most effective at changing behavior? The evidence base is not mature enough to make this judgment, but most authors believe that incorporating more than one approach into a multifaceted intervention is most effective. For example, in an attempt to increase the use of lung-protective ventilation for patients with ARDS an ICU might develop a low-tidal volume ventilation protocol adapted from those presented in clinical trials; educate nurses, physicians, and respiratory therapists on the use and benefits of lung-protective ventilation; and train respiratory therapists to screen patients’ physiology and chest radiographs for evidence of ARDS and to place reminder calls to physicians about the ventilator protocol. This incorporates education, reminders, and a protocol.
Ultimately, the best research in the world does patients little good unless it can be translated into practice. The ICU presents unique challenges and opportunities for implementing best practices. Fortunately, committed practitioners can adopt proven techniques to change practice and improve outcomes in their own ICUs.
1. Davis DA, Taylor-Vaisey A. CMAJ. 1997;157(4): 408-416.
2. Davis DA, et al. JAMA. 1995;274(9):700-705.
3. Smith WR. Chest. 2000;118(2 Suppl):8S-17S.
4. Curry SJ. Chest. 2000;118(2 Suppl):40S-46S.
5. Weingarten S. Chest. 2000;118(90020):4S-7.
6. Weiss KB, et al. Chest. 2000;118(90020):53S-58.