Opioids--are we throwing out the benefits with the risks?

I have been curious about opioid legislation, media obsession with aggregated overdose rates (prescribed, diverted, heroin), and the myriad of solutions hurled at the "problem". Let's throw a bunch of jello at the wall and see if anything sticks. My fear is similar to the age old adage about babies and bathwater--what if we are looking at the wrong data? Are we throwing out the benefits with the statistics?

"Statistics are like swimwear - what they reveal is suggestive but what they conceal is vital." 
-Ashish Mahajan, Lancet 2007

The article published by govtech.com features an interactive map to encourage a granular review of aggregated data informing the public and government about the opioid crisis. Colleagues often underestimate the amount of data freely available for story generation. Off hand, I typically know of at least 3 resources for each type of data question--and the access keeps improving. 

Choosing Wisely, The American Society of Addiction Medicine, and others have guidance for both patients and providers--a place to start for important evidence.

There are powerful graphics in the Solving the Opiate Crisis Through Mapping and Data Analytics--prescription claims, drug poisoning deaths, prescriptions per provider.

West Virginia reported the highest rate of overdose in the country. Lets take a quick look at the data. 

Fayette County, West Virginia is highlighted in blue indicating the percentage of claims higher than the national average. Nine percent (n=30,109) of prescription claims (n=333,803) are opioid claims compared to the national average of 6%. These prescriptions are written by 124 providers. A total of 30,109 prescriptions are written by 124 providers. You can take the math from here...makes me curious--how many actual patients?

When you zoom in on the interactive map you discover Fayette County has a drug overdose death rate of > 20 individuals per 100,00 indicating a class 11 category--estimated to be fatal at a dose of 5 to 30 grams.. Places in red indicate "very high" drug poisoning rates.

The graphic below shows large circles that represent areas where there are relatively few providers but a large amount of opioid claims. When you select for the largest bubble in West Virginia you may be surprised what you find. It isn't Fayette County but  Wyoming County. Although Fayette leads West Virigina in precription claims and overdoses, Wyoming County with 8% (n=14,215) of prescription claims (n=167,818) in this county are opioid claims--but are written by 31 providers--highest county of prescriptions per provider.

Wyoming County also has a drug poisoning death rate of >20 individuals per 100,000, this county falls under the class 11 category of drug poisoning deaths. 

Returning to the Red Lake County, Minnesota in the prescription claims data there is another interesting pocket of activity. Although total prescription claims are only 109, 37% (n=40) are opioid claims compared to the state average of 5%--all prescribed by 2 providers.   

Making use of publically available The Treatment Episode Data Set TEDS 2014 Q1 data (TEDS) contains the demographic characteristics and substance abuse problems of admissions to treatment facilities in the United States. You can spotlight Minnesota at the state level (West Virginia data is not available at this time.) 

When exploring data it can be eye-opening to locate "small" data as well. TEDS has a category for other opiates that includes admissions for non-prescription use of methadone, codeine, morphine, oxycodone, hydromorphone, meperidine, opium, and other drugs with morphine-like effects.

This data seems to capture the diverted use of this category although I did not query the specific definition of "non-prescription" use. It is important to notice that admission data is not the same as patient-level--individuals may have more than one admission.

We only know one main fact. It is the 21-30 year old white males (73%) that are being admitted for treatment.

The moral of the data? The aggregate statistics may distract from identifying where physician education programs, drug category regulation, treatment programs, and obvious abuse/criminal charges may be solutions for several distinct problems.

Are you interested in Medical Expenditure Panel Survey data? MEPS is the most complete source of data on the cost and use of health care and health insurance coverage.

I will be onsite at the Agency for Healthcare Research and Quality for a 2-day interactive workshop. Join us for lectures on survey design and healthcare utilization, expenditures, and medical conditions.

You can also follow  LinkedIn, Data & Donuts, Twitter, or our Newsletter...

Thoughtful discussions about content development and outcomes analytics that apply the principles and frameworks of health policy and economics to persistent and perplexing health and health care problems...