The Department of State Hospitals (DSH) manages the California state hospital system, which provides mental health services to patients admitted into DSH facilities. The department strives to provide effective treatment in a safe environment and in a fiscally responsible manner.
In an effort to help the department achieve their goals, ISML is working with DSH to develop and provide an effective organizational management tool that allows for the accurate prediction/estimation of hospital bed capacities based on patient length of stay times.
The DSH population projection model uses a discrete-event simulation architecture to project referrals, admissions, and discharges over a three-year forecast horizon. Patients enter the model from counties and feed into distinct modules, corresponding to each of five DSH facilities. Each patient record includes standard demographic characteristics, subpopulation identifiers, and transition dates. Patients are held in their modules until release.
The model is designed to generate arbitrary patient subpopulation projections and queue statistics, such as the numbers of patients waiting admission at given any time. Like the LASD model, it can also be used to estimate the effects of policy interventions on subpopulations and queues.
Unlike the LASD model, however, admissions in the DSH model are driven by the availability of resources (i.e. beds) in addition to the forecasted referral of new patients.
The model produces DSH population projections three years into the future using referral forecasts generated from ARIMA models, a standing population database with time until discharge estimated using survival models, and sampled historical patients to approximate new admissions to be introduced weekly. It will also have the capabilities to provide subpopulation estimates over the same time frame in order to see change over time in important subpopulations.
This work is funded by the California Department State Hospitals (http://www.dsh.ca.gov/Hospitals/default.aspx) .