BACKGROUND/RATIONALE:
Veterans using specialty mental health services at VA has been increasing 7% annually. Increased patient load and limited resource capacity has reduced timely access and intensity of care availability. Currently, there are no analytical planning tools that support determining the resource needs to improve timely access, efficiency and effectiveness of mental health services.

OBJECTIVE(S):
The proposed project aims to develop a planning method using simulation-based decision support tools to allow consideration of redesign and implementation changes to improve access to appropriate levels of mental health care. We hypothesized a simulation-based planning method would facilitate operational changes by assisting decision-makers to better understand the system, build consensus about the nature of problems and solutions, make changes based on the findings, and continuously monitor system performance.

Our aims: 1: Formulate and implement a set of procedures for a collaborative planning approach using simulation models as a decision support tool; 2: Construct and validate simulation models; and 3: Evaluate the utility of the collaborative planning approach.

METHODS:
3.1 Study overview
We sought to model mental health care delivery at a single VAMC, and identify strategies for resource allocation to improve key performance measures. We use key informant interviews, statistical analyses of utilization data, and computer simulation and refinement by experts. This study was approved by the IRB and Research Committee.

3.2 Sample population
To determine mental health services demand, we developed a database of patients with at least two outpatient visits to a mental health clinic and/or one inpatient hospitalization with a mental health diagnosis (ICD-9 codes 290-319) in FY10 and FY11. Patients with prior mental health contact in FY09 were excluded from the sample to identify new patients and patients not seen for at least a year. The sample included 7,389 patients. Data including patient demographics, diagnosis, health status, appointment history, emergency department visits and hospitalizations, were retrieved from an electronic medical record system.

3.3 Settings
In this study, we included all four mental health clinics at the Indianapolis VAMC: i) Psychiatric Ambulatory Care Clinic (PACC), ii) Psychosocial Rehabilitation and Recovery Center (PRRC), iii) Substance Abuse Treatment Services (SATS), and iv) Mental Health Services for Returning Service Members (OEF/OIF/OND). We included information from the ED and inpatient unit, as well. Veterans may be seen in several clinics according to their care needs (i.e. PRRC and SATS for a dual diagnosis) or be transferred from one clinic to another (i.e. transferred from SATS to PACC once treatment completed). Utilization data assessed included: services provided, provider FTEs, individual and group therapy rooms, and capacity of therapy rooms.

3.4 Data and Simulation Modeling
We performed semi-structured interviews with unit managers and clinicians to determine patient flow. In the agent-based simulation model, each patient was represented as an agent showing the patient's stage in the flow process (triage, intake, assessment, enrolled, discharged), location of the health services provided (emergency department, inpatient unit, outpatient), resources in each service, and patient health status (acute, stabilization, stable). The processes involved include triage, intake, assessment, and treatment. The veteran walks-in for triage and intake, and schedules an appointment for assessment and treatment. The services provided, treatment length, and frequency change according to factors including diagnosis, patient health status, severity, and treatment guidelines. Though guidelines determine treatment duration and frequency, the patient may not adhere to the treatment plan. We used data to determine the utilization of services. Patients were clustered according to their diagnoses, the mental health clinics seen, and frequency of clinic visits. Visit frequencies to mental health clinics determined the demand for clinic services and ED visits and hospitalizations were determined acute care utilization.

FINDINGS/RESULTS:
Our main product is the simulation model, which represents patient flow within and between clinics. Demand for mental health services is generated according to the care needs of the veteran population. Population characteristics, diagnoses, and health status of the patients determined visit frequencies in each clinic. Current resource capacities are determined according to provider availabilities and clinic services.

The simulation model allows the user (decision maker) to alter the availability of resources (eg, change provider FTEs), or consider different policies (eg, the time allocated for new and existing patients, and individual and group therapies). The model presents key performance measures identified, including waiting time for an appointment, number of missed opportunities, workload (# of individual and group therapies), and # of ED visits and hospitalizations. The decision maker can analyze the impact of limited resource capacity on key performance measures, determine the providers needed to meet the increasing demand for services, and allocate available capacity for different patient groups. Mental health provider leaders reviewed and validated the simulation model, and provided usability feedback.

IMPACT:
Our impact is in demonstrating the value of assessing policy and resource changes, in order to consider policy and practice alternatives. Our model can help provide improved access to mental health services through better capacity planning and resource allocation. The model should help clinic managers quantify the change in key performance measures for the increasing number of veterans seeking care and assist in capacity planning decisions to meet veteran care needs in a timely manner.

---

External Links for this Project

---

PUBLICATIONS:

None at this time.

DRA: Mental, Cognitive and Behavioral Disorders, Health Systems
DRE: Treatment - Observational
Keywords: Access, Management, Severe mental illness
MeSH Terms: none