SOR Project Activities

Measurable objectives to reduce Opioid Use Disorders (OUDs) and opioid-related deaths in Arizona:

  • Number of naloxone kits disseminated; Medication Assisted Treatment (MAT) utilization and retention
  • Use of The Arizona State Board of Pharmacy Controlled Substances Prescription Monitoring Program (CSPMP)
  • Rates of opioid prescribing
  • Rates of individuals in prescribed doses in excess of 50 morphine equivalent daily doses (MEDDs)
  • Rates of new opioid prescriptions in excess of five-day supplies
  • Community knowledge and prevention behavior
  • Emergency Department (ED) utilization
  • Rates of fatal and non-fatal overdose

Study Design

  • Exploration of recent spikes in OUD to identify critical gaps in treatment
  • Predict opioid related events geographically
  • Employ what-if scenarios to quantify the impact of interventions in the full system

The planned models will consider data on individuals (such as age, ethnicity, opioid use, services, geography, etc.), service provider data (such as facilities, specialty services, location, medication stocking, first responders, MAT, etc.), and additional data related to economic, housing, behavioral risk factor surveillance etc.

Note: Models and graphics below will be updated and refined as more information is available and will provide resources and prevalence assessments.

Data Sources:

  • Arizona Medicaid Claims Data - AHCCCS is the state Medicaid provider.  The data include health care transactions (claims) on all members, patients receiving inpatient, emergency department, or other outpatient care in the state.
  • Death Records - The Arizona Department of Health Services (ADHS) provides vital records (registered birth or death certificates) through the Bureau of Vital Records, which includes all vital records from county offices statewide. 
  • Other Data – This includes Medication stocking, first responders, MAT, and additional data related to economic, housing, behavioral risk factor surveillance etc. (publicly available data)

In order to identify future heat-maps from past opioid incidence data, models are under development that incorporate the features of past incidences of OUD. Potential features to formulate the model include the following: age, gender, race/ethnicity, income, education and location, EMS incidence information

Pathways to Addiction Diagram

 

The pathways to addiction model goes from Non-Opioid Addiction State(green), organizes the type of event that the people go to(blue), and then uses the substance used, amount used, and length of use (purple), to create a patient story or population story of the Pathway to Addiction.  A person may or may not reach the Opioid Addiction State (Red). It is reasonable to assume that the Pathway to Addiction story has an event that precedes the use of opioids. Risk is assessed by the length of use recommendations from the CDC.

Agent Based Modeling

The agent-based models to be developed will model the following:

  • Addiction through prescription (changing rates) 
  • Addiction due to illicit sources (rates) 
  • Treatments (rates) 
  • Relapse (Rates) 
  • Rate of illicit drug dependence 
  • Overdose risk 
  • Optimal control 

Agent Based Modeling Simulation

Baseline Simulation: Agents, Environment, Rules

Starting Cohort: N individuals, K neighborhoods

  • Subset of individuals on pain medication
  • Subset of addicts
  • Subset overdosed -- non-fatal, fatal
  • Subset under intervention
  • Subset recovered

Based on rates of incidences:

  • Monitor system evolution
  • Detect emergent patterns

 

Initial State

 

Scenario I