December 04, 2008

Don't Just Do Something, Sit There

All over the country, hospitals are setting up “rapid response teams.” The idea is to intervene quickly when a patient looks as though he might be going downhill—for example if his blood pressure is falling or he’s breathing very rapidly. The expectation is that this preventive approach will avoid cardiac arrests, emergencies in which a patient’s heart stops and respiration ceases. Massachusetts recently passed a law that requires hospitals to put rapid response teams in place (The Act to Promote Cost Containment, Transparency, and Efficiency in the Delivery of Quality Health Care). The only problem is that there is now pretty good evidence that the teams don’t save lives.

A study in this week’s Journal of the American Medical Association is the first large, carefully conducted study of the effect of a rapid response team, taking into consideration normal seasonal variability in the rate of cardiac arrest and death. Carried out at a medium sized tertiary care hospital in Kansas City, Missouri, the investigators found there was no decrease in the frequency of cardiac arrest after the intervention was introduced and no change in the mortality rate at the hospital.

What was particularly striking about the study is that at first glance, it looked as though the rapid response teams were effective: the crude code rate (number of cardiac arrests per 1000 hospital admissions) fell from 11.2/1000 to 7.5/1000. But when the physicians carrying out the study adjusted for normal month-to-month changes in the frequency of cardiac arrests, the differences vanished. Failure to look at the consequences of rapid response teams over a long enough period of time and to calculate the effect of seasonal changes may have accounted for the positive results in earlier, preliminary studies.

The other startling finding in this study is that of the 70 patients who died despite the rapid response team interventions, fully 46 decided on a do-not-resuscitate status after their evaluation. In other words, the rapid response team stimulated a discussion among physicians, patients, and families about the fact that the patient was very ill and doing poorly. The result of this discussion, in many cases, was an explicit decision to limit invasive and painful interventions that had only a small chance of success.

What can we learn from this provocative study? The first is that it’s a bad idea to institute sweeping changes in medicine without good evidence that the changes will have the desired effect. Several physicians expressed skepticism about rapid response teams when they were becoming the vogue (for example, "Rapid Response Teams—Walk Don’t Run," JAMA 2006). A review of the subject done by the Cochrane Collaborative, a group that undertakes systematic reviews of all kinds of questions in medicine to determine whether the preponderance of evidence support their use, found no evidence of benefit from rapid response teams. But despite these calls for caution, the Institute for Healthcare Improvement recommended that all hospitals adopt rapid response teams. The intervention may turn out to be more effective than the current study suggests—the JAMA study was at a single hospital and used a particular combination of two ICU nurses and a respiratory therapist with ICU physician backup; other institutions using other types of clinicians may achieve different results. But what is clear is that sweeping policy changes in medicine, changes that involve many resources and great expense, should be preceded by correspondingly extensive evaluation.

The second conclusion to be drawn is that the best time to engage patients and families in advance care planning—talking about what approach to medical care they want if they are extremely ill—may not be when they are doing well and may not be when they are in the middle of a crisis. Rather, the best time to consider what they hope treatment will accomplish for them, and what treatments are consistent with their goals, may be when they have just come through a serious illness. We know, for example, that patients are particularly susceptible to advice to stop smoking right after they had had a heart attack. That, of course, is a bit late, but a recent study found that after any hospitalization, whatever the medical problem, patients are good candidates for giving up smoking (Rigotti et al, “Smoking Cessation Intervention for Hospitalized Smokers," Archives of Internal Medicine, October 2008). While it would be premature to conclude that advance care planning is best done right after a close call—just as it was premature to conclude that rapid response teams are the way to go based on very preliminary data—it’s an area to explore further.