Can IT Stop the Off-label Prescription Hemmorhage?

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 - Electronic RX
Documentation of treatment indication in the Medical Office of the XXI Century (MOXXI) electronic prescribing system.
Source: Arch Intern Med 2012;172(10):781-788.

Off-label prescribing, the use of drugs for indications that have not received regulatory approval, occurs with up to 21 percent of prescribed drugs among office-based physicians (Arch Intern Med 2006;166(9):1021-1026). Yet, the problem isn't unique to the ambulatory setting. While individual providers are investigating the problem internally, comprehensive inpatient data aren't available. Off-label usage of prescription drugs is not illegal, but it has the potential to lead to wasteful or even harmful care in some instances, yet can be advantageous in others. Regardless, most experts agree in the value of tracking and properly assessing this practice.  

Many administrators and physicians are starting to recognize the value of various health IT systems, such as EHRs, e-prescribing, computerized physician order entry (CPOE) and clinical decision support (CDS) as a means to track, collate data on and potentially, prevent widespread use of the more harmful practices.

First is data collection. Evaluating the MOXXI primary care EHR network in Canada between January 2005 and December 2009, Tewodros Eguale, MD, MSc, of McGill University in Montreal, et al found that 113 primary care physicians wrote 253,347 e-prescriptions for 50,823 patients. They classified each drug indication as on-label or off-label based on the Health Canada drug database, and identified off-label uses lacking "strong" scientific evidence (Arch Intern Med 2012;172(10):781-788).

"Our overall objective was to see how computerization helps the day-to-day activities of primary care physicians by giving them drug information and patient information," Eguale says. "Instead of relying on somewhat unreliable survey data, we implemented a treatment indication chapter within the EHR system, enabling us to capture very fine details of treatment indications."

In the study, the prevalence of off-label use was 11 percent, with 79 percent of the off-label prescriptions lacking strong scientific evidence. Off-label usage was highest for central nervous system drugs (26.3 percent), including anticonvulsants (66.6 percent), antipsychotics (43.8 percent) and antidepressants (33.4 percent). Sicker patients were less likely to receive off-label drugs, which "may be the result of their poor health creating less room to 'experiment' with a drug." This trend also has been observed in children.

The bigger question may lie in understanding why and how physicians make their clinical decisions. "Traditionally, when physicians have written prescriptions, they do not have to write down an indication," explains Surrey Walton, PhD, of the University of Illinois at Chicago. "Indications are typically coded for billing purposes with ICD-9 codes tending to be broad, and therefore, they aren't very helpful for assessing clinical diagnostics or when trying to ascertain why a physician used one drug over another."

For improved prescribing, Eguale says that physicians need to be made aware of three things at the time of decision-making: whether the drug is approved; whether there is strong evidence to use a particular drug for a particular indication; whether there is any report of adverse drug reactions and the severity of that reaction.

The solution may be as simple as color coding on-label and off-label status in the CDS, or another technological or visual method of informing the physician of a potential error, suggests Eguale.

"Linking a prescribed drug with an indication could be a meaningful use objective, and vendors could easily incorporate this feature into EHR systems," the study authors wrote. "EHRs can be used to document treatment indication at the time of prescribing." 

Inpatient introspection

Recognizing that off-label use might not be unique to the outpatient environment, a team at the University of Illinois at Chicago launched a pilot project to develop an electronic-based intervention for gathering data regarding off-label drug use in the inpatient setting.  The project is seeking to identify and focus on specific drugs in an inpatient setting where there are particular clinical and economic concerns regarding off-label use. Also, the researchers are collecting data on the accuracy of physicians' answers when prompted about why they prescribed a particular medication.  

Their single-center, inpatient study assessed a CDS system that was designed to obtain indications and document during CPOE ordering of drugs, which are frequently used off-label, namely the proton pump inhibitor (PPI), lansoprazole, the IVIG Flebogamma and the Factor VIIa Novoseven (Appl Clin Inf 2011;2:94-103).

More specifically, the study examined "what would happen if an IT system required physicians to explain more explicitly why they used certain drugs, and if they took a little longer to question their own reflex of prescribing a particular drug," says Randall S. Stafford, MD, PhD, of the Stanford University School of Medicine in Stanford, Calif. He adds that the factor VIIa use is particularly "concerning." The drug is indicated for bleeding episodes in hemophilia A or B patients, but is diffused into widespread use for a number of other leading situations, says Stafford. It also is very expensive at $40,000 to $50,000 for every use.  

The PPI intervention generated 873 alerts during 60 days of operation; IVIG 55 alerts during 93 days; and Factor VIIa 25 alerts during 175 days. Agreement between indications entered and chart review was 63 percent for PPI, 49 percent for IVIG and 29 percent for Factor VIIa. The alerts for PPI, IVIG and Factor VIIa produced accurate diagnoses for the problem list 9 percent of the time, 16 percent and 24 percent, respectively.

"This study illustrates substantial off-label use for certain medications and many challenges for obtaining indication information. In particular, the indication data generated in this pilot study were not highly accurate," Walton et al wrote. Yet, because the Joint Commission requires connection between indications for inpatient medication use and indications approved by a Pharmacy and Therapeutics Committee, "this type of CDS may help institutions comply" with these regulations. The authors added that prompts during CPOE to remind clinicians of a medication's indications may improve evidence-based use of medications and can improve problem list documentation.  

Data collection is critical

The consensus on the need for improved data in this area might be best summarized by a recent editorial by Stafford. "Objective, comprehensive and comparative syntheses of data relating to off-label use of drugs would be valuable to prescribers, patients and healthcare payors," he wrote (Nature 2012;91(5)920-925). "Data from EHRs relating to drug prescriptions could provide an early alert if off-label use of a particular drug is increasing. In addition, data mining of claims data from multiple healthcare payors could help identify potential safety issues associated with off-label use."

Stafford says that the primary goal in attaining these metrics is to ensure patient safety by reducing the risk that somebody has an adverse effect of a medication, to improve cost-efficiencies by using cheaper drugs "if those drugs do pretty much the same thing as a more expensive drug."   

As with many quality improvement initiatives, progress may come on a case-by-case basis. For instance, Eguale says that their MOXXI EHR system is tracking drug indications during treatment, as well as what happens with treatment down the line. Therefore, if a physician stops administering or change dose a particular medication, the EHR system has a mandatory requirement to input a reason (e.g. adverse drug reaction, ineffectiveness)—a novel method of generating data for CDS and pharmacosurveillance.

"With the treatment indication and treatment discontinuation features of the EHR, we have created a longitudinal record of 'drug-treatment indication-reason for drug discontinuation.'" Eguale says. Also, these data then could be filtered to the physicians or drug regulatory bodies as CDS or comparative safety/effectiveness profiles of drugs, respectively.

"Ultimately, the benefits reaped from IT systems will always come down to the physicians' willingness to input the appropriate information about their decision making," Walton says. This input may require a helpful nudge from a CMIO.