For years, those on the ground level at addiction treatment facilities saw a problem boiling up among their patients: prescription opioid addiction. As the issue became increasingly commonplace, law enforcement and medical examiners also began to see increasing evidence of the situation in crime statistics and on the morgue slab.
But larger governments, despite vast connections into communities, were slower to realize the full gravity of opioid addiction in America without hard data and analytics systems in place. The problem now maintains a full boil, fluctuating between the successes of state and local programs and spikes in prescription drug use.
While many hear the phrase “opioid addiction” and immediately conjure mental images of crime-addled streets and pill-popping junkies, what’s not often understood is how unintentional overprescribing at the hands of busy and under-informed doctors is contributing to its rise. And while illicit drug use is a very real problem, spotting it has proven just as vexing.
Tom Clark, of the Brandeis University Prescription Drug Monitoring Center of Excellence, finds that prescription drug monitoring programs and the data they produce are an essential tool in the fight against the misuse of prescription opioids. Photo by Brian Smith.
To date, getting the health-care professionals behind the prescriptions the information they need about their patients has been an uphill battle. To determine whether a patient has a legitimate need or is just doctor-shopping to feed an addiction has meant digging into often incomplete records spread across a host of platforms and maintained by various state agencies.
Tom Clark, an associate researcher with the Brandeis University Prescription Drug Monitoring Center of Excellence, studies how states and agencies are approaching their respective prescription drug monitoring programs (PDMPs) and common prescribing habits. From his perspective, PDMPs and the data they produce are an essential tool in the fight against the misuse and abuse of prescription opioids. State and national monitoring systems allow public health agencies to keep a pulse on the trends of controlled substances. For example, a rise in so-called “treatment drugs,” like buprenorphine, coupled with a decrease in powerful prescription opioids, like oxycodone or hydrocodone, could signal a shift in the right direction for a jurisdiction’s policies and practices.
“It’s important to realize that it’s a fairly new problem over the last 15 years that opioids have become so prevalent,” Clark said. “And that’s because the pharmaceutical industry marketed them so heavily and changed norms among prescribers such that pain became what they called the ‘fifth vital sign.’”
When governors took to their respective podiums earlier this year during State of the State addresses, the call for better drug monitoring was hard to ignore. In the northeastern U.S. and down the Eastern Seaboard, tackling opioid addiction has been a serious challenge over the last several years. The closely clustered states are prime targets for border-jumping doctor-shoppers, who travel from doctor to doctor in search of prescriptions.
With state systems often operating independently of one another, verifying the validity of patients’ dosage needs or the legitimacy of their condition is a systematic gap many are working to close. At a more widespread level, the National Association of Boards of Pharmacy (NABP) is working to connect state programs and jump-start data sharing.
Dr. Peter Kreiner, a senior scientist with the Brandeis University Institute of Behavioral Health, and his colleagues have been working extensively with state and federal partners to improve public health surveillance tools in the prescription drug space since 2011. Kreiner said the reliance on old data streams put the Centers for Disease Control and Prevention (CDC) and Food and Drug Administration (FDA) substantially behind when it came to identifying trends playing out nationwide.
“They noted that death certificate data and hospitalization data takes a minimum of one to two years to become available. So by the time you see what might be happening … you’re already a year or two behind what is actually happening, whereas PDMP data are basically available right away,” he said. “From the CDC and FDA perspective, such a tool could serve as an evaluation tool to look at the effects of state policy or law changes or sub-state initiatives.”
Dr. Peter Kreiner, a senior scientist with the Brandeis University Institute of Behavioral Health. Photo by Brian Smith.
The Prescription Behavior Surveillance System, a longitudinal database that uses de-identified patient data, is updated quarterly and produces a report with 43 descriptive measures and risk indicators, which are then distributed to the program’s 12 partner states. Kreiner said although data is not directly attributable to a patient by name, the unique identifying number they are assigned within the system allows researchers to build and examine patient profiles and point to the risk factors in their prescription drug behavior.
For Kentucky, a state bordered by seven others, the close proximity and ease of access to doctors across state lines posed a substantial hurdle to accurate tracking of the signs of prescription drug abuse.
Dave Hopkins, with the state’s Office of the Inspector General, said the enhanced Kentucky All Schedule Prescription Electronic Reporting (eKASPER) system and policies have helped fight the epidemic. Kentucky has seen a nearly 5 percent decline in controlled substances, he said, with some opioid painkiller prescription rates dropping as much as 24.6 percent in a study that examined 2011-12 data against 2014-15 data. The original KASPER system was launched in 1999 and then reworked in 2005 to handle an increase in report requests.
“State PDMPs are a very valuable prevention tool,” said Hopkins. “If we can get the information to our prescribers and pharmacists, and they can identify the patient that is at risk earlier on, there is a much better chance of getting them into treatment.”
The data goes beyond profiling a potential pill shopper — it also speaks volumes as to who is being put in danger because of mixed medications from multiple prescribers and other vital parameters needed to assess risk to an individual. A prime example is the combination of benzodiazepines and opioid painkillers, which has proved time and again to be deadly.
By mandating that prescribers maintain, report and query eKASPER under certain conditions, Hopkins said the interaction with other state data systems through the NABP multi-state InterConnect system is providing more accurate patient and prescriber information. Kentucky plans to test and implement the national hub in the near future.
“It’s really helping us to give our health-care providers who use these systems a complete and accurate picture of what a patient might be getting,” he said.
Six of Kentucky’s surrounding states share monitoring program data access through InterConnect. Missouri, which borders Kentucky to the west, is the only state in the country without a PDMP.
Maryland’s PDMP manager, Kate Jackson, said the state’s analytics are not showing an abundance of “pill mills” or careless prescription practices, but rather what are referred to as “multiple provider episodes” — essentially doctor or pharmacy shopping.
“The takeaway from some of our initial work is that … we don’t in Maryland have a ton of pill mills operating, or prescribers who appear to be willfully skirting best practices and putting their patients in danger,” she said. “A lot of what we see is an individual patient-provider interaction that is probably completely legitimate and within clinical standards absent having looked at the PDMP.”
Maryland’s program joined the InterConnect hub in August 2015, but Jackson said, like in so many other states, there are limits to how the sensitive prescription drug data is shared from a legal/policy and tech standpoint. In addition to the challenges facing PDMPs from an implementation and data sharing perspective, the question of which data belongs to whom poses opportunity for inaccuracies and the potential for missed risk factors at the clinical level.
Maryland uses what Jackson called a “conservative” probabilistic algorithm to match patient data with appropriate profiles in a Master Patient Index. Kreiner said California leverages a similar algorithm. Even with misspelled names or addresses, these tools can match information based on probabilities or create a new user profile for the patient. This process greatly cuts down on mismatched records and missing patient information.
Once entered into state monitoring systems, data goes on to play several roles throughout the prescription process. On the front end, patient data can be compiled into solicited and unsolicited reports for physicians making an initial decision about who gets a particular medication. On the back end, the data can be de-identified and shared with researchers or used to track trends in the larger population.
In Virginia, PDMP data has been used effectively as a tool for the Office of the Chief Medical Examiner in determining cause of death. It provides valuable information such as doctor interactions and prescription medications, and helps to interpret toxicology results, according to Dr. Amy Tharp, assistant chief medical examiner.
“We use it most importantly in interpreting toxicology results — the level of a drug that may be lethal in someone who is not tolerant to the medication can be very similar or equal to the level that is expected (not lethal or therapeutic) in someone who has been prescribed the medication for a long period of time and is therefore tolerant to its effects,” she said via email.
Throughout the larger PDMP space, the call for greater connectivity and real-time reporting access is prominent, and the need for better information sharing among states and prescribers seems clear. Officials, like Jackson, said putting the information in the hands of prescribers through the electronic health records platforms they use on a daily basis would be a progressive step in the right direction.
Indiana’s Drug Dashboard visualization tool is a nod to PDMP tracking systems around the U.S., but pulls data sets from state crime labs, not clinical sources, to paint a picture of where drug activity is taking place and being treated.
Indiana CIO Dewand Neely. Photo by David Kidd.
“You can begin to see where the most drugs are coming in off the streets at the county level, where there is the highest percentage of deaths per capita related to drugs, where the treatment centers are available, and you can very quickly begin to look at where there are gaps in Indiana in terms of treatment and the amount of drugs coming off the streets,” said David Matusoff, executive director of the Indiana Management and Performance Hub.
State CIO Dewand Neely said the conversation around prescription drug abuse is one that has spilled over into a top state committee on counterterrorism and is a priority for state officials.
Matusoff and Neely said integrating state PDMP, known as INSPECT, data into the Drug Dashboard would greatly widen the effect of what has already proven to be a valuable visualization tool in Indiana. Before that can happen, however, Matusoff said the legal mechanisms need to be in place to allow an agency access to information typically reserved for law enforcement and public health professionals.
Making the data and analytics tools available to more prescribers, pharmacies and public health professionals has been the goal of PDMPs all along, but it has taken valuable time.
As Kentucky, Maryland and other states move forward in their efforts, both Hopkins and Jackson said they are working to connect the state systems directly to the electronic health records platforms doctors use when working face to face with patients. Officials believe putting the information right in front of clinical users will help to improve the use and accuracy of these critical health systems.