Hospitalized patients often have co-morbid mental health disorders. Mental health disorders can contribute to the complexity of care and create disposition challenges. In addition, active behavioral issues may require enhanced supervision and disrupt a unit. As such, these patients pose unique staffing, care and financial challenges for hospitals.


The purpose of this pilot is to see if geographically cohorting these patients to a hospitalist run unit with enhanced, specially trained staff can decrease the need for continuous observation (CO), decrease disruptions on other units and improve the care we provide to these patients.


The hospital is a 500+ bed academic medical center in a major urban setting. A 15-bed medical unit was converted to a closed, med-psych unit. Strict admission criteria, such as the exclusion of patients with suicidal or homicidal ideation or advanced dementia, were created. The unit is managed by a dedicated hospitalist, nurse practitioner, psychologist, and nurse manager. All floor staff received aggression reduction training and an in-service on appropriate prn medication usage. A patient engagement specialist, trained at the neighboring inpatient psychiatric facility, is on the unit 24/7. There are daily morning interdisciplinary rounds and afternoon huddles that involve the entire care team, including case management, social work and pharmacy. The psychiatry consult service is on the unit or available 24/7. There is frequent rounding by security. Patient interventions include proactive ambulation and distraction modalities such as free TV, snacks, music and cards.

In the first 3 months of the pilot, there were 203 admissions, 100 of whom had been admitted previously to other units.  We used data from their prior admissions to assess outcomes. For the 100 readmitted patients, the need for CO dropped from 59% to 22.3%. The average length of CO dropped from 2.5 days to 0.79. The percentage of patients requiring restraints went from 18.2% to 1.7%. The 30-day readmission rate dropped from 27.6% to 18.2%. For the whole unit, patient satisfaction scores were comparable or better than other medicine units and have been on a steady incline. Length of stay did not show an appreciable change. 


Though the pilot has only been in existence for a short time, cohorting medically ill patients with co-morbid mental health disorders to a hybrid medicine and psychiatry unit, with a steady and specially trained staff, appears to reduce the need for CO. We are currently evaluating the potential cost savings associated with this reduction in staffing needs.  The consistency in staff and interdisciplinary approach may have contributed to the reduction in 30-day readmissions and positive trend in the patient satisfaction scores. Challenges include appropriate triage and identification of patients appropriate for the unit, night and weekend staffing, and identifying and obtaining additional outcomes data.