High Utilizer Committee:

It is widely known that a small percentage of patients account for an outsized share of healthcare cost. A similar phenomenon exists for emergency room visits and hospital readmissions - a small group of patients account for an outsized number of visits each year. While complex medical and psychosocial factors undoubtedly play a role in this phenomenon, physicians must also examine how health systems could serve these patients more effectively.

Inspired by work from John Nelson at Overlake Hospital Medical Center in Washington State, our Division formed a High-Utilizer Committee to design care plans for patients with high use of ER and inpatient services. The goals of the project are as follow:

  • Improve patient care by avoiding unnecessary or inappropriate diagnostic or therapeutic interventions by physicians unfamiliar with the patient
  • Improve patient satisfaction by bringing consistency to a patient's care (i.e. avoid "reinventing the wheel" each time a patient is seen)
  • Decrease ER visits and hospital readmissions by more effectively linking inpatient care to outpatient care plans and outpatient resources
  • Improve provider satisfaction by providing clear guidance and rules for a patient's care

  • The Committee is multi-disciplinary and includes representatives from nursing, psychiatry, social work, emergency medicine, and hospital medicine. All care plans will also include input from the patient's primary care physician
  • Care plans will cover key elements of an encounter, including ER care, admission criteria, inpatient care plan, discharge criteria, and transition to outpatient care
  • Harness the EMR to put care plans in a location clearly visible to all providers who see the patient
  • Re-evaluate care plans at regular intervals to keep them up-to-date

Data Dashboard:


Audit and feedback of data is a critical component of any quality improvement (QI) endeavor. However, at many institutions, the process for performing audit and feedback to inpatient clinicians does not provide data that is sufficiently accurate, timely, and actionable to make it effective for driving change.

In the Division of Hospital Medicine, a pioneering group of physicians and data scientists have developed a QI Data Dashboard that fulfills these important criteria.

  • Accurate: the addition of metrics to the dashboard is moderated by a governance committee that scrutinizes the data to make sure it is accurate and reflects a valid measurement for the process or outcome under consideration.
  • Timely: data is automatically pulled from the electronic medical record (EMR) every day. It is then filtered and visualized in real-time by a system that was developed in-house.
  • Actionable: since the unit of care delivery on the wards is the ward team (attending, resident, interns, and students), the Data Dashboard performs QI metric attribution by team. The visualization software allows ward teams to pull up their team-level data and dive deeper into certain metrics to understand their practice patterns in more detail. Teams can even use links to the patient's chart in the software so they can write orders to correct their deficiencies in real-time.

Rather than simply roll out this software tool to the entire Division, the leadership of the program decided to study it in a randomized controlled trial. The Data Trial splits the ward teams at UCSF into two groups. One group meets briefly each week with a physician or data scientist coach to go over their dashboard results and discuss ways to improve. The other group is usual practice and does not receive access to the data or weekly coaching. Results from the trial will be published later this year.

Team Members: Alvin Rajkomar, Victoria Valencia, Sumant Ranji, Michelle Mourad, and Sajan Patel

Promoting Sleep Project:

less sleep

If you or your loved one has spent a night in a hospital, you know firsthand how challenging it can be to get a good night's sleep. Whether it is fluorescent lighting, the noise or the constant interruptions due to medication administration, vitals, or blood draws, we often hear about these common disruptors of sleep from our patients. While few studies specifically link sleep and patient outcomes, most physicians agree that the connection is obvious:

"Patients need sleep. If they get more of it, they'll likely recover faster."

In addition, the "quietness of the hospital environment" is a patient experience question in the nationally administered HCAHPS survey, so health systems have a keen interest in tackling this important issue from the perspective of value-based purchasing.

The good news is there are several low-hanging fruit solutions that we are implementing at our institution to promote sleep and an optimal healing environment for our patients. Interventions include:

  • Modifying the admission order set to allow reduced nighttime vital sign checks.
  • Creating a view in the electronic medical record of all interventions (including lab checks, respiratory treatments, and vital signs) that might awaken a patient at night in order to identify discontinuation opportunities.
  • Developing a bedtime checklist for nurses and patient care assistants to assess patients for sleep readiness in the hospital, including turning off the TV and lights and providing ear plugs if necessary.

Team Members: Michelle Mourad, Sasha Morduchowicz, Timothy Judson, and Dan Hausrath