Improving Medical ICU Outcomes: Promoting Respect and Ongoing Safety Through Patient Engagement Communication and Technology Study*
This study builds on previous studies that are related to parts of the intervention. There are some similarities of the work to prior studies involving use of data to guide decision making. For example, in the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment study (2), predictions of risk of mortality were given to physicians but not to patients or family. There are also similarities to work conducted with palliative care interventions to increase patient/care partner engagement. An early examples of that work is the study by Lilly et al (3). who developed structured frequent communication interventions for family members of critically ill patients both to inform and to gather information about goals and values. The trend to increased use of palliative care, which is patient-centered and family inclusive, in the ICU has furthered that work (4–7).
The authors cite examples of other work that informed the development of their intervention. This includes previous studies on improvement of safety when ICU patients, who are often unable to engage because of severity of illness or sedation, or their care partners are engaged (8, 9). Previous work with patient portals to increase this engagement is described (10–14), but this work has not been sited in ICUs previously.
The developers (1) creatively built on all this literature, as well as on use of checklists for safety, and developed an innovative intervention, which led to significantly improved outcomes when the postintervention group was compared with the preintervention baseline group. The intervention itself was complex and involved staff training both preintervention and real-time during implementation, development of a web-based toolkit with a safety checklist prepopulated in the electronic health record, and access through the patient portal to digital tools for shared provider-patient care planning and communication.
There are problems related to the complexity of the intervention. It would be difficult to replicate in another setting, and, as the authors note, it is impossible to tell what aspects of the intervention led to the improved outcomes. It is, and again the authors admit this, impossible to tell if simply the increased time spent with patients and care partners was the key aspect of the intervention. Lack of a control group is a key problem for interpreting the results, particularly important in any replication would be an attention control group.
Limitations also related to the selected outcomes. Patient adverse outcomes measured were blood stream infections, ventilator-associated events, catheter-associated urinary tract infections (CUTIs), falls, and pressure ulcers (PUs). Only two of these, CUTIs and PUs, changed significantly. The authors say that they may have lacked power to measure the other three outcomes, but discussion of trends is always suspect and best omitted. The sample size was adequate, per their power analysis.