Monitoring risk: Post marketing surveillance and signal detection
Introduction to risk management
Risk management is the process of assessing a product's benefits and risks, followed by developing and implementing tools to minimize its risk. The goal is to maintain the benefits of a drug while reducing the risks as much as possible. Ultimately, the balance of benefits and risks determines whether the drug will be withdrawn from the commercial market, or in the case of new drug applications, approved or disapproved.
The importance of risk management has been recognized by the U.S. Food and Drug Administration (FDA) for many years. The FDA defines three phases of risk management: (1) risk assessment, (2) surveillance, and (3) intervention. These phases form a cycle that is repeated to provide continual assessment and intervention and optimize the benefit-risk ratio of a drug.
Often, the importance of surveillance in the process of risk management is overlooked. At the time of initial marketing, the data available to evaluate the risks of a medication are limited. Information regarding chemical structure, abuse liability in animals and humans, and clinical trial data allow useful characterization of many risks. Pre-marketing risk assessment before drug approval could be improved by development of standards for assessing tamper-resistance, improved animal models that can address formulation-related variables (e.g., onset, duration), and the redesign of human laboratory studies providing appropriate models for comparing formulations. Although risks can be identified before marketing and addressed in a risk management strategy, some risks will not become apparent until after the drug is in widespread use in multiple populations. These events are detected by postmarketing surveillance, which provides data for intervention as well as informing potential revisions to the risk management assessment.
Postmarketing surveillance is a core component of risk management because without data to guide interventions, risk assessment is based largely on expert assessment, which is based in turn on pre-marketing information. While a useful and important adjunct in decision making, expert assessment can be misleading. Numerous medications have been approved for marketing, for example, only to receive warnings or be withdrawn because expert assessment did not identify specific problems during the pre-marketing phase. For examples, see the FDA list of new warnings and withdrawals of marketed drugs.
Postmarketing surveillance
Postmarketing surveillance is the practice of monitoring a pharmaceutical product or device after it has been released on the general market. Pharmaceutical products have been subjected to various forms of postmarketing surveillance for as long as drugs have existed. Today, some type of postmarketing surveillance is required for all pharmaceutical products in the U.S.
Pharmaceutical manufacturers have developed extensive adverse event reporting systems that accept spontaneous reports of adverse events from patients or health care professionals. In some cases, adverse event surveillance can identify a new, but real, risk when an unexpected event is reported and, perhaps followed by reports of a cluster of similar or related events. For example, when a report to a pharmaceutical manufacturer revealed that a patient developed fulminant hepatic failure during treatment with troglitazone (Rezulin), a hypoglycemic agent for the treatment of diabetes mellitus, this previously unrecognized risk was documented and triggered a regulatory response. Detailed analysis of individual cases and perhaps the discovery of other cases allow the identification of potential associations between drug exposure and the event. After a relationship has been identified, more focused data collection and analysis allow the manufacturer and the FDA to investigate the events and take appropriate action.
The term adverse event has typically indicated undesired effects that occur during intended use of a product. The discovery of congenital malformations after the use of thalidomide during pregnancy occurred during the normal therapeutic use of the drug in pregnant women. Conversely, adverse drug effects that occur with non-therapeutic intent are often not considered adverse events. For example, toxic epidermal necrolysis (TEN) during the use of phenytoin is a well-known adverse event, but respiratory arrest and coma following abuse of an opioid is generally considered separately because it is the expected pharmacological effect of an intentional overdose. In truth, all unintended effects of medication use, even abuse and intentional overdose, are adverse events that should be reported.
Surveillance systems for misuse, abuse and diversion of prescription drugs
An effective surveillance system informs risk management decisions by providing current, as well as sensitive and specific information. The timeliness of surveillance data is crucial. However, classic epidemiology studies are carried out over longer periods of time and the results are often reported a year or more after the research was performed. These time frames are very long in terms of drug abuse surveillance. For instance, by the time the results are reported, the abuse conditions in the area may have changed. While traditional epidemiologic studies provide crucial information in terms of general characteristics of abuse, they do not provide information that a risk management program can act upon in a timely manner. Thus, the FDA often requests that surveillance data be provided on a quarterly basis although more frequent monitoring may be needed for specific reasons such as the introduction of a new product. Data for surveillance should be available within weeks or months depending on the precise application of those data.
A surveillance program should also cover all regions within the United States because drug abuse of specific drugs and products can be localized. Again, data generated for a valid national estimate are important, but do not provide all the information needed to allow targeted investigation and intervention. Specific information for every 3-digit ZIP code (or similar specific geopolitical unit) should be available.
Product specificity is also important because not all products are equal in terms of abuse. Generic versions of products may be abused more, less, or the same as their branded counterparts. In some cases, abusers may prefer certain formulations or brand names. Using brand names, for instance, allows an abuser to be more certain that they have not purchased a counterfeit drug. On the other hand, a generic drug may be easier or less expensive to obtain. Further, certain formulations may be easier to abuse. Surveillance systems must be able to identify the specific product accurately and consistently.
Different prescription drugs may also be abused by different portions of the population. For example, individuals initiating abuse of drugs may prefer certain drug products specifically because of their brand name, as noted above. Drug dealers may prefer formulations that contain a large amount of the drug so that this can be divided and sold profitably. New initiates often prefer to use a product orally but with time and experience may switch to more potent and riskier injection routes of administration leading them to choose a different formulation.
There are also unintended victims of prescription drug abuse, and surveillance systems may need to include these individuals. For example, pediatric deaths have occurred in association with the use or abuse of prescription opioid products by their caregiver (e.g. pills left within reach of young child). Thus, surveillance systems may need to include pediatric cases. Other potential victims include family members and friends, who may suffer theft, loss of financial support, or potentially infectious complications like hepatitis or human immunodeficiency virus.
The role of quality in data collection and analysis is often neglected. The processes needed to ensure data integrity and accuracy are not used in some systems. Quality control systems are necessary to assure that data are collected appropriately, managed cleanly and reported accurately without sacrificing time. Attentive quality control always reveals sources of errors that can affect the impact of surveillance data. A quality improvement program is needed to address quality issues and assure that the system consistently improves itself.
Groups of surveillance systems have emerged due to the need for multiple perspectives in the surveillance of CNS drugs, For example, the Researched Abuse, Diversion and Addiction-Related Surveillance System (RADARS? System) includes multiple simultaneous perspectives on prescription opioid and stimulant abuse. The RADARS? System's Drug Diversion System provides a criminal justice perspective on prescription drug abuse by surveying more than 300 diversion investigators from jurisdictions in all 50 states. The Key Informant System provides the perspective of substance abuse treatment professionals by surveying approximately 160 designated key informants from a variety of treatment facilities across the United States. The Poison Control System provides information on acute incidents involving prescription drug both in the home and in the emergency care system. Data include acute and chronic exposures for all ages from 47 poison centers, including rural, urban, and suburban areas of the United States. The Opioid Treatment Program System provides the perspective of opioid dependent patients by surveying patients in 75 methadone treatment programs about their individual drug use. The Survey of Key Informant Patients (SKIP) complements the Opioid Treatment Program by anonymously surveying approximately 300 patients nationwide. The College Survey provides information about young new abusers of prescription drugs by surveying 2000 college and university students across the nation. Finally, the Health Care Professional System aggregates information from the other systems that are specific to health care professionals. All of these RADARS System components provide current data every quarter or more frequently.
The Navippro? system consists of two major proprietary databases and draws from other public databases across the United States. The proprietary databases are ASI-MV Connect and Web informed Services (WIS): internet monitoring and surveys on prescription drug misuse. The public databases used include American Association of Poison Control Centers' New Core System Database (NCSBeta), FDA-Adverse Event Reporting System (AERS), Drug Abuse Warning Network (DAWN Live!), DEA news, and Medline articles. Other systems are also available. Often, manufacturers construct their own systems by melding data from one or more surveillance systems with their own monitoring of adverse events and media surveillance. In some cases, unique approaches have yielded insights into abuser's perceptions of prescription drugs.
Conclusion and recommendations
Assessment of misuse, abuse, and diversion of prescription drugs must address its occult nature and involves fundamentally different surveillance compared to other pharmaceutical products. Current best practice is to use multiple detection systems to assess misuse, abuse and diversion of CNS active drugs by various populations in a timely, sensitive, and specific manner. Data should be available within weeks, to able to detect reasonable changes in abuse rates, and must accurately identify the product(s) involved as well as the specific geographic location. Furthermore, the best surveillance systems also can provide preliminary information about further research and intervention needed in those areas.
Several surveillance challenges persist for manufacturers and regulators of CNS active drugs. The acceptable level of misuse, abuse and diversion has not been defined. The rate of abuse, or the change in the rate of abuse, that warrants intervention has not been defined. Since some level of misuse and abuse occur with all CNS drugs, the FDA and other authorities should begin the process of explicitly designating tolerable levels of misuse and abuse. The tolerable level of diversion must also be defined, but raises different issues because it involves the criminal justice system.
An important emerging issue involves identifying specific drugs that should always require surveillance. The list of drugs reported to be abused is long and many contrasts and contradictions are apparent. While extended-release opioid medications are scrutinized closely, other opioid products seem less closely examined. It seems logical that branded products and products from generic drug manufacturers should be required to have equally rigorous risk management programs. Benzodiazepine drugs seem to receive little attention despite widespread evidence of abuse. Similarly, anticholinergic drugs like diphenhydramine are sometimes abused by adolescents. The effects can be dramatic (hallucinations), but outcomes are generally mild. Perhaps risk management should be required for these drugs as well. Another issue concerns the precise elements that should be required in a surveillance system for abuse of prescription drugs? In the past, a wide variety of systems have been accepted by the FDA. While general risk management guidance has been developed FDA, more specific analyses and guidance are needed to improve surveillance methodology for drugs which are misused, abused, diverted and these activities.