Thank you for the opportunity to respond to the Safe Use of Opioids - Concurrent Prescribing Proposed eCQM specifications. Memorial Hermann is the largest not-for-profit healthcare system in Texas with a total of 3,557 beds and 5,500 affiliated physicians across our organization. We are an organization that has invested over $100 million in our EHR technologies and infrastructure since 2000 to successfully achieve our clinical quality and patient safety goals. In 2009, we were recognized by the National Quality Forum with the National Healthcare Quality Award for our efforts. We would like to provide the following comments;

General Comments

• We recommend sub-dividing the controlled substances into categories of opioids, benzodiazepines and ADHD meds to identify patterns of potential abuse for each category. In addition, we recommend that the population be based on all patients receiving scripts for a given provider stratified by these categories. This would be useful for identifying the percentage of patients that are receiving even a single script of controlled substance. We believe this will better facilitate 1) Identifying providers that may need additional education on appropriate controlled substance prescribing practices and 2) Assist physicians to identify patients that may be at risk for adverse events and/or substance abuse issues. We do not believe the same information will be provided if the patient population is based strictly on encounters.
• We agree with and fully support the need for a quality measure to support the safe use of opioids and prevent concurrent subscribing practices. We also realize that in the electronic health record landscape, there remain many siloed sources of health care data not only across the nation, but even within our local system. It is because of this that the barriers to incorporating data across multiple types of encounters in a yet to be defined time period, will make capturing the data for this measure very difficult. While we are not discouraging CMS to pursue this quality measure, we are requesting that CMS and the eMeasure developer work diligently to adjust the logic of this measure to be able to capture the data available in light of current EHRs limited ability to share data.
  Initial Population
• Healthcare Encounter- While we understand and agree with the need to view this population across multiple encounters, accomplishing this presents numerous challenges and will be extremely difficult to achieve accuracy. In addition, there is no guidance on the timespan of the healthcare encounter. We request more clarity around the expected start date:time as it relates to the end of the measurement period which we assume is the inpatient encounter discharge date:time. How or who will determine this time period?
  Denominator Exclusions
• We are requesting more clarity on the denominator exclusions in relation to the Healthcare Encounter. It appears that the diagnosis or palliative care can occur during anytime of the defined time period. So if for example, the time period is defined at 6 months and multiple opioid prescriptions were present throughout that time period but the last date of the encounter a cancer diagnosis occurs; is that patient excluded from the population for the entire time period?
  Numerator
• It appears from the logic that the measure is looking for any opioid/benzodiazepine prescription during the healthcare encounter when a similar prescription is ordered at discharge for an inpatient encounter. It does not appear to take into account the date:time or quantity of the prescriptions to determine if the patient is actively taking the medication. For example, if the healthcare encounter is defined as a 2 month look back for all encounters prior to inpatient discharge and the patient was prescribed an opioid with a 2 week supply at the beginning of that time period, it appears that the patient would be pulled into the numerator because there is no logic or data element identifying that the prescription time period ended and the patient is not actively taking it. If the logic is to assume that the medication is not active because it did not appear on the admitting medication list, then what is the purpose of the look back period?