IN THIS ISSUE OF JAMA, BERWANGER AND COLLEAGUES 1 REport a teaching-case–perfect cluster-randomized trial demonstrating the positive effects of a “multifaceted intervention” to improve the reliability of evidencebased management of acute coronary systems in general hospitals in Brazil. With a combination of reminders, a checklist, case management, and staff education, the intervention hospitals used all eligible acute therapies in the first 24 hours after admission 37% more often than control hospitals (67.9% vs 49.5%) and used all evidence-based therapies (ie, not just acute management) 67% more reliably (50.9% vs 31.9%). The authors’ attention to detail in the design and execution of their study is admirable, the presentation is lucid, and the statistical analysis is solid. Subject to their own caveat about generalizability, the investigators have demonstrated convincingly that this portfolio of changes worked well to give patients a better opportunity to receive the care that science says they should. However, the authors take a step onto thin ice when they refer to their work as test of a “quality improvement [QI] intervention” and conclude that their results suggest that “QI interventions may be feasible and effective” in settings like the resource-constrained public hospitals of lowand middle-income countries such as Brazil. A closer look is in order. First, is it fair to call their intervention an example of “QI”? Or, better, is it helpful to do so? That would depend on what “QI” is. Because I have spent the better part of the last 30 years of my career trying to improve health care, I may sound coy if I write, “I don’t know.” But I do not know. It is the capital letters that I question, which imply that “QI” is some sort of contained, classifiable mechanism, like a drug or a device. If the capital letters were dropped and the intervention were described simply as “activities that help performance get better in complex systems;” ie, “quality improvement”—I would feel better. Entropy is at work—often viciously so—when frail humans try to make something good happen over and over again in circumstances characterized by interdependency, many components, unknown risks, and effects remote in time and place, eg, circumstances like health care delivery. For reasons that are often puzzling, the result is not good, despite good intentions—and, confusingly, trying harder makes things even worse. Who would not want patients with acute coronary syndrome always, without fail, to receive exactly the therapies that help them the most? But there it is: 50.5% of the time (in this study), they did not. The aim—improvement—may be clear. But the aim is not enough for progress to occur; it must be linked to a method for improvement. The methods that actually produce new, better results when the old results are not satisfactory are those through which the processes of production—the processes of work—change. Their name is legion. Methods for inventing better work have been codified, expanded, and denominated a thousand ways, with terms such as “science” and “teaching,” “practice” and “industrial engineering.” In the investigation from Brazil, Berwanger et al embraced some good ideas for making a system of work better— like education, standardization, reminders, and job redesign—and applied them in real hospitals. As a result, patient care got better. But in what sense does this happy story represent quality improvement? And did the authors actually test “QI in lowand middle-income countries”? I think it is a disservice to the sciences of improvement to reify the term “quality improvement” as if it were a device or even a stable methodology. Making patient care better is always a good idea, and there is no harm at all in using the term “improvement” to describe that quest. However, treating the pursuit of improvement (no initial caps) by searching for a boxable, boundable formula, let alone canonizing it with a proper-noun label—“Quality Improvement” (initial caps)—is misleading. The ways in which people and organizations try to overcome the destructive forces of entropy in complex systems and to continually improve the work that they do on behalf of patients are numerous and, thank goodness, will forever evolve. At a high level of abstraction, useful principles for improvement in complex systems do seem to exist, although even those principles should be negotiable as more knowledge and learning emerge. Some principles that seem robust, for example, are (1) knowledge of systems helps people
Search from 250 M+ research papersSearch from 250 M+ research papers
Mar 28, 2019
Arnab Ghosh + 6
Open Access