Thursday, May 7, 2015

If it is Worth Having, Is it Worth Checking?

Data, the world seems to revolve around data these days. I spent a day at the WCRI conference in Boston this spring. If you are ever looking for data, WCRI is a great place to start. NCCI has a wealth of information accumulated also. I recently wrote a paper on comparisons of workers' compensation insurance premiums, and NCCI was a major data source for that research.

In another blog post on the WCRI Conference, I addressed a problem with data. Citing some WorkCompCentral stories, I noted that different conclusions can be drawn from data. The data itself may be different because of the way a sample was defined or the way data was selected. The conclusions may be different because of the way in which the data was analyzed. These seem axiomatic. 

Data can be powerful if it is used appropriately. The ether is replete with jokes demeaning both data and the people who gather it. 

Mark Twain said "there are lies, damned lies and statistics." Some credit that quote to Benjamin Disraeli. Ernest Rutherford said "if your experiment needs a statistician, you need a better experiment." And Josefina Mota said "there are two ways of lying. One, not telling the truth and the other making up statistics." The interpretation or description effect is sometimes cast humorously. The former Mayor of Washington D.C., Marion Barry, was quoted as saying "outside of the killings, DC has one of the lowest crime rates in the country;" by redefining what crime means, the data can be refocused. 

So data is perhaps not the be all and end all. I was thinking of the usefulness of data recently because of a story out of Oklahoma. Governor Fallin there signed a bill that will take effect next November. This law requires that "doctors in Oklahoma" must "check a new prescription drug database before prescribing certain addictive drugs" according to the Insurance Journal

Oklahoma has designed a database (these are generally referred to as "prescription drug monitoring program" or "PDMP" databases) that "includes real-time information on whether that patient has obtained prescriptions from another doctor."  The goal is to prevent "doctor-shopping" meaning "the practice of drug seekers going to multiple physicians to acquire prescription narcotics." 

According to the Centers for Disease Control, prescription painkillers prescribed in the United States has quadrupled since 1999. 

The CDC touts Florida as a success story. It credits the 2010 statutory changes to regulate pain clinics and eliminate physician dispensing of painkillers with a 50% decrease in oxycodone overdose deaths. Florida also has a PDMP. Reporting prescriptions into the database is mandatory. But there is no requirement for doctors to consult it before writing a script. Before handing someone access to opium, shouldn't you want to see if they are already getting opium from some other source?

Other states have deployed PDMPs and required prescribers to check them before prescribing. The CDC credits the New York PDMP requirement with "a 75% reduction in patients' seeking multiple prescribers." It credits the Tennessee requirement with a similar 36% decline.  According to a Brandeis University study updated in 2013, 49 states "had passed legislation authorizing a PDMP, and 43 states had an operating PDMP." This report cites various statistics on PDMP success from states like California, Kentucky, Maine, and more. Kentucky gets praise for their system in many reports and conferences.

Two Physicians published a road map to the "Ideal" PDMP in the New England Journal of Medicine in 2012. They "advocate for more informed prescribing." They identify hurdles to the acceptance of PDMPs by prescribers, but note that technology is helping with prescriber concerns. They note that logging the prescriptions is time-consuming, and that "data are uploaded to the prescription database at variable intervals." Another challenge is metropolitan areas that are close to or even straddle state lines, leading to a call for intrastate information sharing. The article cites one network that currently facilitates sharing of data among 20 states. 

These physicians also caution regarding the interpretation of the data from the PDMPs. They note that there is a "required balance between optimizing pain management and the excessive accessibility of opiods." They caution that "lingering concerns about punitive action for perceived misprescribing may lead to reluctance among physicians and medical boards to adopt PDMPs." In other words, what interpretation is placed on the gathered statistics may affect the willingness of prescribers to use the data, or in some instances perhaps even to voluntarily contribute to the data. 

So it appears that the states are conclusively onboard with collecting the prescription data. The advocates are convinced that having the data can save lives. But, what is the point of collecting this data if physicians will not check it before writing a prescription? With opium the need is fairly clear. But is not as clearly beneficial to know with certainty what else the patient is taking, from other providers, before you prescribe a chemical for them to ingest? Cannot various drugs interact, cannot that present dangers?

WorkCompCentral reports May 7, 2015 that Nevada and New Jersey are "poised to become the 9th and 10th states to require doctors to check" the PDMP before writing scripts for certain medications, particularly opiod pain medication. California is also considering this move. 

Collecting data may be interesting. But what is it worth if you do not use it? 


No comments:

Post a Comment