You, the medical director of an intensive care unit (ICU), discover that mortality has increased for patients with sepsis. You are considering a quality improvement (QI) initiative to improve the care and outcomes of your patients. However, you are concerned that many QI studies have weak designs, poor data quality, and often overestimate potential benefits. Before beginning, you decide to identify and evaluate existing QI interventions.
You perform a literature search using PubMed and identify a before-after study that evaluated an educational QI program for sepsis in 59 medical-surgical ICUs in Spain.1 This program trained clinicians to recognize and treat severe sepsis based on evidence-based guidelines from the Surviving Sepsis Campaign.2 The program implemented 2 guideline-based treatment bundles: a resuscitation bundle (ie, a group of interventions that constitute optimal care that need to be implemented together—in this case, 6 tasks started at sepsis recognition and completed within 6 hours) and a management bundle (4 tasks completed within 24 hours). The program looks very promising in the setting of the study, and you wonder about the value of taking this approach in your institution. Your next step is to critically appraise the report.
Quality Improvement—An Overview
Opportunities for QI are common.3,4 Patients frequently do not receive evidence-based treatments,5 and more than 9% of hospitalized patients are harmed by adverse events.6 Quality improvement interventions attempt to change clinician behavior and, through those changes, lead to more consistent, appropriate, and efficient application of established clinical interventions, resulting in improved care and patient outcomes.7 The target of the intervention in QI research is not ascertaining the efficacy of theintervention but rather determining the effect of the intervention on behavior change, typically manifest as adherence to an optimal process of care.
Traditional evaluative clinical research (eg, estimating the efficacy of therapies) typically evaluates interventions provided in well-controlled environments that ensure, in the context of, for instance, the clinical trial setting, that patients receive the intervention.3,4 In contrast, QI interventions are often designed to enhance the implementation of proven therapies and use data routinely collected in clinical practice. As a result, QI efforts are not always considered research8; in that case, an institutional review board may agree to waive informed consent and detailed review because the study exposes patients to minimal risks beyond those involved in standard clinical practice.9
Quality improvement interventions are frequently context dependent, complex, and iterative, seeking to address barriers to and facilitators of QI.10 When high-quality evidence has securely established substantial net benefit of existing therapies, measuring the processes of care (ie, the incremental implementation of therapies) may be sufficient to establish the benefit of QI interventions. When the net benefit of therapies is less securely established, measurement of improved patient-important ...