An improved goldmine of information not otherwise available to the public and a flood of opioids.
On August 18, CMS released the second iteration of its cost-and-utilization database of prescription drugs written for patients covered by Medicare Part-D Medicare which includes beneficiaries in Part-C Managed Care and stand-alone Medicare Part-D Prescription Drug Plans. Compared to the updates of Medicare’s other 2013 public-use healthcare utilization files, this release seemed delayed and I feared the program’s continuing implementation had been quashed by the pharmaceutical industry lobby and its friends as was a 2013 initiative to provide the public with average retail drug prices. The wait was worth it.
Although restricted to the two specific Medicare populations mentioned above comprising some 70% of all Medicare beneficiaries, the database is unique in that the names and other identifying information about individual prescribers are disclosed to the public. Although the drug companies themselves know who is prescribing their products, to my knowledge, this degree of transparency for the public is unique – and therefore doubly valuable. The top 25 opioid prescribers to this population are listed below.
I have written several articles in this series over the last year based on national data from the 2013 utilization data release and in Kentucky. It is easy to find examples of how broken our healthcare system is with regard to the selection of drugs used and how much we pay for them. I also attempted to learn what this Medicare data might tell us about our current national opioid addiction crisis. To summarize, I determined to my satisfaction that for opioids, the Medicare data serves as a valuable proxy for prescribing patterns to the public at large. It seems to me that a more valuable public health or even law enforcement tool would be hard to find!
The database proper is huge, comprising millions of lines of data each with several dozen fields, one line for each unique drug prescribed more than 10 times by each and every individual prescriber in the country. I can deal with this, but it will take some time to unpack the data, to ask the right questions, and to make any answers understandable. Medicare helps a lot by providing summary data for individual drugs and for individual providers. This latter summary table of a mere million or so lines collates all drugs prescribed by each individual provider to a single line. It includes all varieties of healthcare providers and specialties. It is still too large to analyze with Excel or similar programs. However, this year’s provider summary release has been augmented by breaking out utilization and cost data for each individual prescriber with respect to the following factors:
• Disabled vs. Elderly (older than 65) Medicare Beneficiaries.
• Part-C Managed Care vs. Stand-alone Drug Plan.
• Medicare Participating Provider or not.
• Brand name vs. Generic drug.
• Low income subsidy eligibility.
• Antibiotics.
• Anti-psychotics.
• High-risk drugs for the elderly.
and,
• Opioids!
These data will be very useful for health policy makers and scientists.
Looking at Opioids.
Because I have been focusing lately on opioid use and abuse, my first plunge into the 2014 data was to extract the number of opioid prescriptions and refills written in the USA and in Kentucky. [Use the Search Site box for the key word “opioid.”] I should note, that CHS/Medicare used a slightly larger selection of opioid drugs than I did. The major difference is that Medicare includes tramadol (Ultram), a modestly weaker opioid that until recently was not a categorized as a controlled substance but is widely used, and whose abuse potential has become apparent. The Medicare data includes buprenorphine (Suboxone) and methadone, drugs increasingly used to treat opioid addiction.
Colossal numbers of prescriptions.
As we have come to expect, the numbers of prescriptions written by individual professionals varies by orders of magnitudes. For example, for all physicians prescribing more than 10 prescriptions for any given opioid, the average number written was 159 with a standard deviation of 368. The median number written was only 52 per prescriber. There are however, hundreds of prescribers writing thousands of opioid prescriptions per year. The highest number of opioid prescriptions written by an individual was 26,449. Some 1335 prescribers wrote over 3000 opioid prescriptions each.
Previous analysis of the 2013 data predicts that the highest prescribers will by no means be limited to pain management specialists, nor will they be exclusively physicians. The names of the top 25 in terms of absolute numbers of opioid prescriptions and refills written are listed below.
Number of Opioid Prescriptions in Medicare Parts-C&D in 2014
Name NPI City State #Rx
MURPHY, MARK -1326019712 DECATUR, AL 26,449
GARLAPATI, BUTCHAIAH -1326035163 N LITTLE ROCK, AR 23,474
COUCH, JOHN -1053372201 MOBILE, AL 19,970
JAMES, OLIVER -1194727701 LEXINGTON, KY 16,330
RAMOS, FABIAN -1831150630 BRADENTON, FL 15,548
DOLE, MICHAEL -1750364121 ALEXANDRIA, LA 15,268
ELLIS, MARK -1730150194 ATHENS, GA 15,139
BALLARD, KATHERINE -1871521252 LEXINGTON, KY 14,578
MCGEE, DARIUS -1619004744 LAUREL, MS 14,033
RUAN, XIULU -1023079274 MOBILE, AL 13,689
SIDHOM, GEORGE -1003814435 SPRING HILL, FL 12,661
MOREE, LAMAR -1306853510 ALBANY, GA 12,658
MURCIANO, ENRIQUE -1750347605 SOUTH MIAMI, FL 12,465
LOVE, RILEY -1851342372 PADUCAH, KY 12,454
REVELIS, ANDREAS -1902809080 TULSA, OK 12,275
PETERS, LAWRENCE -1417927229 LOUISVILLE, KY 11,995
GANTT, RODDIE -1912996059 HUNTSVILLE, AL 11,890
REHEEM, MEDHAT -1073555397 BROOKSVILLE, FL 11,772
NELSON, CHRISTOPHER -1386639474 MIRAMAR BEACH, FL 11,610
RATINO, THOMAS -1104030782 FORT WORTH, TX 11,462
BARNGROVER, KENNETH -1588787162 COLUMBUS, GA 11,199
MANCHIKANTI, LAXMAIAH -1730178799 PADUCAH, KY 11,071
BAKER, JAY -1972509974 SPRINGFIELD, MO 11,067
DIAMOND, HOWARD -1396772737 RICHARDSON, TX 11,035
MARTUCCI, MARTIN -1962405035 TULSA, OK 10,991
Data from CMS, Edited by KHPI 18 Aug 2016
In my opinion, surely it is fair to label these professionals as statistical super-outliers in this database. Of particular concern for Kentucky, of the top 25 prescribers of opioids to this population of Medicare beneficiaries nationally, 5 of them practice in the Commonwealth. Kentucky has by its policy choices effectively set the bar for an acceptable upper limit of opioid prescription very high. There may be medically appropriate and defendable reasons to be on this list, but in my opinion and in the interest of public health and safety, the reasons deserve to be looked at.
More to come.
The above was just an initial peek. I will have more to say about the latest CMS release in future weeks and will share links to Medicare that allow you to do specific look-ups yourself. If there are questions you would like me to attempt to ask of the data, please feel free to share your expertise with me. Use the link in the sidebar or at the page-bottom on your device to contact me confidentially at KHPI.
Peter Hasselbacher, MD
President, KHPI
Emeritus Professor of Medicine, UofL
August 19, 2016
In working up the data, I prepared the map below of the rate of opioid prescriptions per Medicare Part-D Beneficiary who actually utilized their drug benefit in any way, by their state of residence. The usual three-fold variation among states is present. Rate of opioid prescription correlates well with the proportion of disabled Medicare beneficiaries.
As with an earlier map of rate of prescription in 2013 which was adjusted for total state population, Kentucky lies in the heartland of opioid availability. Another difference between the analyses is that the 2014 data set provided by Medicare includes the relatively weaker opioid tramadol (Ultram). I will redo a new analysis from scratch including separate consideration of the “hard” opioids and those used in medication-assisted treatment of opioid addiction.