I had the pleasure of interviewing Dana Edelson, a Hospitalist at the University of Chicago, about her team’s recent publication in the Journal of Hospital Medicine (JHM), Comparison of mental status scales for predicting mortality on the general wards. We know from previous literature that altered mental status is a strong predictor for inpatient mortality. But it is not clear which mental status assessment tool is best in detecting the risk of mortality. This study retrospectively compared the accuracy of 3 mental status scales [Glasgow coma scale (GCS), the Richmond Agitation and Sedation Scale (RASS) and Alert, responsive to Verbal stimuli, responsive to Painful stimuli, and Unresponsive (AVPU)] in predicting inpatient mortality. The authors found RASS to be the most predictive, followed by GCS, followed by AVPU. The combination of RASS + GCS was the most predictive, compared to any of the scales alone.
What is your background and how did you become interested in evaluating this topic?
I am a hospitalist at the University of Chicago and Medical Director of Rescue Care and Resiliency. Within this role, I focus on how to identify deteriorating patients and how to intervene quickly. My clinical and research interest is in predicting and preventing deterioration within the hospital. At the University of Chicago, we have a model to predict deterioration in real time, but it is only as good as the metric “inputs”. We know that here and elsewhere, mental status is notoriously poorly tracked in clinical documentation. The question we had for this study was what is the best way to document and track mental status. Historically here, mental status was documented via a multi-select drop down menu, which was not quantifiable. So we developed ways to take that data and convert it into a more reliable scale (AVPU) and we wanted to see how it compares to the RASS and/or the GCS.
How should hospitalists use this information to change their practice?
We should make sure our nurses document mental status using RASS and/or GCS on the wards; these scales are well known, previously validated, and easy to learn and use. Then we should use mental status as a vital sign, because we know it can help determine how “sick” a patient is. Mental status should be routinely incorporated into risk prediction algorithms (e.g. early warning scores) that can alert hospitalists to how sick their patients are (e.g. who needs additional support, a higher level of care, etc.).
Where there any unexpected findings in your study?
I was surprised that the combination of the scales was better than either alone; the good news is these scales are easy to learn and use, so most hospitals should be able to have them both documented and used routinely.
Where does this take you with respect to future research efforts?
My next immediate step is to incorporate these scales into our predictive modeling, then we will move onto other documentation input improvements, such as respiratory rate. With these efforts, we continually tweak our algorithms to make them better. Future efforts will focus on how to get clinicians to use the information to improve care on a large scale.
In summary, hospitalists should ensure that either the RASS or CGS scale (or both) is universally incorporated into their “vital signs” assessments; this vital sign can help hospitalists determine which patients need to be seen early and often.
Dana Edelson, MD is assistant professor of medicine at The University of Chicago Medicine. Dr. Edelson completed her residency at The University of Chicago Medicine. She is an expert in cardiac resuscitation, with a focus on prevention and treatment of in-hospital cardiac arrest. Her efforts in measuring CPR quality and using team training and post-event debriefing to improve resuscitation quality has been widely cited. Dr. Edelson has also utilized predictive analytics to harnesses electronic health record data in order to identify subtle changes in clinical stability and alert clinicians, in real-time, to patients at risk for cardiac arrest.