The trusted source for
healthcare information and
Abstract & Commentary
The purpose of this study was twofold: first, to increase the use of ketoconazole (KZ) prophylaxis and second, to identify the effect of KZ prophylaxis on the development of acute respiratory distress syndrome (ARDS). Historically, the use of KZ prophylaxis was not common (7.5%) in the test hospital. A multidisciplinary team developed a protocol to identify high-risk patients, detail treatment, and list contraindications to KZ prophylaxis. In addition to developing and publicizing the guideline, Sinuff and colleagues discussed the guideline frequently with house officers and nurses, circulated a summary of the studies used to support the use of KZ prophylaxis, displayed the actual KZ prophylaxis performance data, confronted physicians failing to follow the guideline, and presented the ARDS incidence data. Another ICU in the same hospital system served as a contemporaneous control.
Patients at risk were given 200 mg KZ by tube or mouth two hours following feeding once a day for 21 days or until they were discharged from the ICU. Pregnancy, age less than 16, congestive heart failure, esophageal surgery, cirrhosis, liver enzymes elevated beyond two times the normal values, established ARDS at ICU admission, or strict nothing-by-mouth status resulted in disqualification from the study. Forty consecutive high-risk patients were entered into the study, 20 in each hospital. There were no differences at entry, and most patients had severe sepsis or pneumonia. The mortality was 55% in the guideline hospital and 50% in the control. No patients with exclusion criteria were placed on KZ; two patients receiving cisapride were started on KZ, but after intervention it was stopped. Twelve of 20 patients in the test hospital (60%) were placed on KZ, while only one of 20 (5%) was started on KZ in the control hospital (P < 0.0001). ARDS, as defined by a shunt greater than 15% or an oxygenation ratio of less than 150 in a patient requiring ventilatory support for at least 48 hours, absence of heart failure, and bilateral infiltrates on chest radiograph, occurred in only one of the 20 patients at the test hospital and in seven (35%) of the patients at the control hospital. (Sinuff T, et al. J Crit Care 1999;14:1-11.)
COMMENT BY CHARLES G. DURBIN, Jr., MD, FCCM
This study shows how difficult it is to change physician behavior. Using a consensus model and including all involved disciplines, a practice guideline was created with the intent that all eligible patients would receive KZ prophylaxis. The guideline was circulated, advertised, discussed, and displayed. Sinuff et al devoted substantial time and effort to educating and re-educating caregivers on the benefits of this practice change. The staff was provided continuous feedback about performance, and interventions were carried out when discrepancies were seen. Despite these efforts, only 60% of eligible patients received the appropriate care regimen.
What we don’t know from this study is the time course of change in practice. Others have noted that changes in caregiver behavior increase with time if the wanted behavior is continually reinforced. It is also easy to demonstrate that without continued reinforcement, behavior reverts quickly to the status quo. The important message is that if one wants to produce a lasting change in behavior with an educational intervention, such as with a practice guideline, simply issuing the guideline is only the beginning. Continued education through feedback and direct intervention is necessary to achieve the desired change.
The effectiveness of KZ in preventing ARDS is demonstrated in this study. The impressive reduction from 35% to 5% of patients developing ARDS is remarkable. The lack of effect on mortality is surprising and concerning, however. As with any study, control group selection and bias is a concern. That the control group was treated at a different hospital by different clinicians may indicate that different treatment strategies were used. Since Sinuff et al made the determination of the diagnosis of ARDS in both groups, the difference in ARDS incidence is probably real. A large NIH study addressing the role of KZ prophylaxis and treatment in ARDS is in progress and should shed light on this problem.
a. results in immediate change in practice.
b. increases the legal liability to caregivers.
c. should be reserved for professional societies.
d. may have little effect without continuous education.
e. is best accomplished by making a change in the hospital bylaws.