The measure criteria and analysis methods used in our recent among Washington's Medicare beneficiaries is described below. (See results for the Medicare population and for healthcare professionals.)

Project Measures

Qualis Health analyzed the following metrics for Washington's Medicare population:

Any beneficiaries, regardless of opioid use level

• Non-fatal opioid overdose (with an additional focus on opioids prescribed shortly before an overdose)
• Beneficiaries prescribed any opioid
   

Beneficiaries with a new opioid prescription (no prescription during the previous quarter)

• Opioid prescription for beneficiaries with new low back pain
• High-dosage opioid prescription (>210 MMEs) for opioid-naïve beneficiaries
   

Beneficiaries who use opioids at a chronic level

• Beneficiaries with chronic opioid prescriptions
• High-dosage opioid prescription (>50 MMEs per day) for beneficiaries who use opioids at a chronic level
• Chronic benzodiazepine prescription for beneficiaries using opioids at a chronic level
   

The overdose measure complements the work being done in the state by the Bree Collaborative, and the remaining measures are directly related to the CDC's prescribing guidelines.
 

Data Sources & Patient Population

Analyses were conducted using Medicare Part A, Part B, and Part D claims. The most recent 12-month period of data available at the start of this work was July 2016 - June 2017.

Medicare beneficiaries were included in the initial data set if they met all of the following criteria:

• Must have been enrolled in Medicare Part A or B during part of the measurement year (July 2016 - June 2017)
• Must have been enrolled in Medicare Part D during part of the measurement year, and for at least one quarter before the event that triggered our study criteria
• Must have a Washington address

Data Selection

The steps taken by Qualis Health to select the appropriate data sets are outlined in the following table.

Analytic Approach

A two-phase approach was used to analyze the data—first in several aggregate groupings focused on the beneficiary, then by individual provider.

By beneficiary location and demographics

The results were aggregated by several levels of location (state, Accountable Community of Health, county). It is important to note that these locations were determined by the beneficiary's residence, not where the care was provided.
 

By provider

Using the Medicare claims data, Qualis Health identified the healthcare professionals (primary care, emergency care, surgeons, dentists, and others) who wrote the prescriptions included in our data set. In some situations, there were multiple providers who prescribed opioids and/or benzodiazepines to a single beneficiary.
 

Outreach Methodology

In order to increase the number of prescribers who consistently use the Prescription Monitoring Program and are following the latest clinical guidelines regarding opioids, as well as to alert providers that some of their patients experienced an opioid-related overdose, we are contacting a subset of the prescribers identified from our data set. 

We mailed the following subset of providers their individual results:

• Providers who prescribed an opioid to one or more Washington Medicare beneficiaries who later suffered a non-fatal opioid overdose. (The relevant time period was defined as the number of days' supply for the selected prescription, plus 10%. For example, the risk period for a 30-day supply would be 33 days after the prescription was written.)
   
• Providers whose rate of prescribing opioids for low back pain was higher, by a statistically significant margin, than the statewide average.
   
• All providers who prescribed opioids to at least one beneficiary whose total prescriptions reached the high-dosage threshold, for either the opioid-naive or chronic opioid metric.
   
• Providers who prescribed chronic benzodiazepines to 10 or more of their patients who were also using opioids at a chronic level.
   

The individualized results were mailed in February 2018. In the spring of 2018, a sampling of these providers will also be contacted by phone for further discussion.