The best way to project inpatient demand for hospitalist services is to identify and quantify what may change in the next year: what trends could increase or decrease the number of cases that will need to be treated? These change factors include the following:
- Population trends: Is the community growing? It there an influx of new residents? Is the community aging? Is it likely that there will be more seniors requiring inpatient services? Health plans and medical groups often can more easily assess population trends because they treat an enrolled population.
- Local health care factors: Will a hospital in the region be closing, resulting in additional inpatient demand? Is there a shortage of nursing home beds in the community that may affect the need for inpatient care? Is Medicaid reducing the number of covered recipients, potentially increasing the demand from uninsured patients?
- Changing referral patterns from community physicians: Do you expect additional community physicians to stop/start referring patients to the hospital medicine program? Are referring medical groups increasing or decreasing in size?
- Institution-specific factors: Does the hospital medicine program expect to assume new responsibilities in the next year – e.g., in the emergency department (ED), in the intensive care unit (ICU), providing night coverage, doing surgical co‑management, etc.?
Work
The best practices for measuring hospitalist output (work) are summarized in Box 2.
Determining how to quantify the labor of hospitalists can be the most controversial component of developing a staffing model. To ensure buy-in of these modeling decisions, participation by hospitalists and other key players (e.g., other physicians, physician leadership, and hospital/medical group administration) is crucial. Hospitalists and other key individuals must understand and agree on the quantification of time and labor.
It is critical that the analysis include ALL elements of work. Brainstorming with hospitalists can be helpful in this process. To build physician acceptance of and trust in the model, it is important to acknowledge the full set of hospitalist responsibilities in the initial stages of model development.
The services provided by a hospitalist team can vary from program to program and hospital to hospital. For example, at Kaiser Permanente-Hawaii, the dedicated hospitalist triage physician may direct patients coming from the clinic or ED to the ambulatory treatment center. A hospitalist then sees the patient in the center and an admission is often avoided. This physician labor must be captured in the model even though an admission did not occur. If your program includes a day team and a night team, you may want to handle these two teams as separate models.
Based on an analysis performed at Kaiser Permanente-Hawaii, some examples of hospitalist labor components are noted in Box 3 (page 50).