Medicaid's Most Costly Outpatient Drugs
Appendix B: Methodology
For our analysis of the 50 most costly drugs in the Medicaid program, we used 2014 and quarters one and two of 2015 State Drug Utilization Data available from CMS merged with data from Wolters Kluwer Clinical Drug Information, Inc (“WKCDI”).1 The State Drug Utilization Data is publicly available data used as part of the Medicaid Drug Rebate Program. This data provides information on the number of prescriptions, Medicaid spending, and cost-sharing for rebate-eligible Medicaid outpatient drugs at the National Drug Code (NDC) level. The WKCDI data provides product information for drug products. We accessed State Drug Utilization Data in February 2016 and used the most recent data available for all states. The WKCDI data is also from February 2016. The use of WKCDI data does not represent and should not be characterized as a WKCDI endorsement of any data, findings, or other content presented in this report.
We merged the State Drug Utilization Data and the WKCDI data at the NDC-level to consistently identify the drug name, as well as to incorporate brand versus generic status and the WKCDI Therapeutic Classifications System’s drug group. We classified single-source and multi-source, originator drugs as brand drugs. If a drug was available as both a brand and a generic, we categorized it as a brand when summarizing how many of the most costly drugs were brands and how many were generics. Using the Center for Drug Evaluation and Research List of Licensed Biological Products and the Center for Biologic Evaluation and Research List of Licensed Biological Products2 as of February 2016, we identified all biologics in the State Drug Utilization Data based on drug name. We looked up the orphan drug status of the 50 most costly drugs in Medicaid in the Orphan Drug Product designation database.3 To identify specialty drugs in the State Drug Utilization Data, we compiled a list using formularies from the top Medicaid pharmacy benefit managers.4 We identified specialty drugs using the drug name.
To determine the 50 most costly drugs to Medicaid, we summed total Medicaid spending before rebates by drug name over the 2014 quarter one through 2015 quarter two period. Due to data reliability, we were unable to include 2014 quarter two data from Kansas or Virginia. We also calculated average Medicaid spending per prescription and summed total prescriptions by drug name over this period. We ranked the drugs by spending per prescription and total prescriptions, calculated their percentiles, and identified a drug as “frequently prescribed” or “expensive at the prescription level” if it was in the top 10th percentile of either. We included the 50 most expensive drugs by spending per prescription and the 50 most prescribed drugs over the period in the Appendix tables. When reporting the former, we only included drugs with ten or more prescriptions to avoid any outliers in the data.
This analysis does not include rebates, because this data is unavailable to the public at the NDC level. Rebates have a considerable effect on Medicaid drugs spending overall, but lower spending at the drug level at different rates.
Additionally, although Medicaid beneficiaries largely self-administer drugs that are prescribed in an outpatient setting, medical practitioners must administer some drugs. Although states are to collect drug rebates on all reimbursed outpatient drugs, regardless of whether they are physician- or self-administered, research has shown that not all states are collecting rebates on physician-administered drugs.5 Because biologics and other specialty drugs are often physician-administered, it is possible that the data reflects lower Medicaid spending and utilization of certain drugs of this kind.Appendix A: Tables