The Clinical Impact of Adverse Event Reporting

The FDA has the regulatory responsibility for ensuring the safety of all marketed medical products. Health professionals are critical to this process in that the first hint of a potential problem originates with the perceptive clinician who then reports the case to the appropriate source. It is important for all health professionals to be aware that some reporting is mandated by federal law and regulation while other reporting, although considered vital, is strictly voluntary.

By Health Professionals

Any postmarketing surveillance program depends on health professionals to report serious adverse events observed in the course of their everyday clinical work. Except for adverse events associated with specified vaccines, reporting by an individual health professional is voluntary.

Given the clinical importance of postmarketing surveillance, all healthcare providers (physicians, pharmacists, nurses, dentists and others) should look upon adverse event reporting as part of their professional responsibility. The American Medical Association⁽¹³⁾ and American Dental Association ⁽¹⁴⁾ advocate (respectively) physician and dentist participation in adverse event reporting systems as an obligation^(13,14). Further, The Journal of the American Medical Association instructs its authors that adverse drug or device reactions should be reported to the appropriate government agency, in addition to submitting such information for publication ⁽¹⁵⁾.

Health professionals can use the voluntary MedWatch form (PDF format) to report adverse events or product problems related to any medical product, with the exception of those occurring with vaccines. Reports can be sent to FDA either directly or, in most cases, via the manufacturer.

Reports concerning vaccines should be sent to the Vaccine Adverse Event Reporting System (VAERS), a joint program of the FDA and the Centers for Disease Control and Prevention⁽16). Certain events following immunization (e.g., paralytic poliomyelitis after oral poliovirus vaccine)⁽¹⁷⁾ are mandated by the National Childhood Vaccine Injury Act of 1986 to be reported but VAERS accepts all reports of suspected significant adverse events after any vaccine administration⁽¹⁶⁾. For more information on VAERS, call 1-800-822-7967.

Health professionals working in a hospital or other user facility (nursing home, ambulatory surgical facility, outpatient treatment facility and outpatient diagnostic facility) should be aware of the legal requirements for medical device-related reporting by user facilities mandated by the Safe Medical Devices Act of 1990 (SMDA) (see TABLE 2¹⁸). Under the SMDA, physicians' offices are excluded from the user facility definition and thus exempt from mandatory reporting requirements. The FDA likewise excludes other groups that perform similar functions to physicians' offices (e.g., dentists, optometrists, nurse practitioners) from mandatory reporting¹⁸. However, health professionals within a user facility should familiarize themselves with their institution's procedures for device-related reporting, and actively participate in the program.

Confidentiality: The FDA acknowledges that health professionals have concerns regarding their confidentiality as reporters, and that of the patients whose cases they report. In order to encourage reporting of adverse events, FDA regulations offer substantial protection against disclosure of the identities of both reporters and patients. This was further strengthened on July 3, 1995, when a regulation went into effect extending this protection against disclosure by preempting state discovery laws regarding voluntary reports held by pharmaceutical, biological and medical device manufacturers ⁽¹⁹⁾.

Return to the Table of Contents

By Hospitals

The FDA, recognizing the valuable role that hospitals play in the detection of adverse events and problems with medical products, views every active hospital monitoring program as a vital component of the national postmarketing surveillance system. Hospital reporting of adverse events, both within and outside an individual facility, is a mixture of mandatory and voluntary reporting.

Adverse event monitoring by hospitals is linked to Joint Commission on Accreditation of Healthcare Organizations (JCAHO) standards. In order to be accredited, JCAHO requires each hospital to monitor for adverse events involving pharmaceuticals and devices, with medication monitoring to be a continual collaborative function⁽²⁰⁾. JCAHO standards indicate that medical product adverse event reporting should be done per applicable law/regulation, including those of state/federal regulatory bodies⁽²⁰⁾.

The American Society of Health-System Pharmacists (ASHP) has also been instrumental in the evolution of active internal hospital adverse drug event (ADE) monitoring systems. ASHP guidelines include delineated criteria for classifying an adverse drug reaction (ADR) as significant⁽²¹⁾, unlike JCAHO standards, which do not mandate a specific definition for a serious ADE. ASHP guidelines specifically state serious or unexpected ADRs should be reported to FDA, manufacturer, or both⁽²¹⁾.

As user facilities, hospitals are subject to mandatory federal medical device adverse event reporting. TABLE 2 outlines these requirements , which include reporting by the facility of suspected medical device-related deaths to both FDA and the manufacturer, and serious injuries/illnesses to the manufacturer or to the FDA, if the manufacturer is unknown.⁽¹⁸⁾ However, there are no federal laws or regulations that require hospitals to report pharmaceutical-related adverse events to the FDA, although they are strongly encouraged to do so regarding those events deemed serious.

Return to the Table of Contents

Reporting Required By Law or Regulation

Reporting by individual healthcare providers is essentially voluntary. However, manufacturers and distributors of FDA approved pharmaceuticals (drugs and biologics) and medical devices, plus pharmaceutical packers and device user facilities, all have mandatory reporting requirements.

TABLE 3 outlines mandatory reporting regarding pharmaceuticals^(22,23). By regulation, these companies are required to report all adverse events of which they are aware to the FDA and to provide as complete information as possible. As can be seen, mandated pharmaceutical reporting relies heavily on information provided by health professionals through both voluntary reporting and the scientific literature.

In the case of over-the-counter (OTC) drugs, reports are only required on OTC products marketed under an approved New Drug Application (NDA), including those prescription drugs that undergo a switch to OTC status. Reports are not required for other OTC drugs (i.e., older drug ingredients which are marketed without an NDA), although voluntary reporting is encouraged.

Both prescription and OTC drugs require FDA safety and efficacy review prior to marketing, unlike dietary supplements (which include vitamins, minerals, amino acids, botanicals and other substances used to increase total dietary intake). By law⁽²⁴⁾ the manufacturers of these products do not have to prove safety or efficacy, so the onus is on the FDA to prove that a particular product is unsafe. As a result, direct-to-FDA voluntary health professional reporting of serious adverse events possibly associated with dietary supplements is particularly important.

TABLE 2 lists the medical device-related reporting required of user facilities, manufacturers, and distributors¹⁸.

All unsolicited reports from health professionals received by FDA via either the voluntary or mandatory route are called spontaneous reports. A spontaneous report is a clinical observation that originates outside of a formal study²⁵. The combination of adverse event information generated by all reporting makes up the database upon which postmarketing surveillance depends.

Return to the Table of Contents

As with clinical trials, there are important limitations to consider when using spontaneously reported adverse event information. These limitations include difficulties with adverse event recognition, underreporting, biases, estimation of population exposure, and report quality.

Adverse Event Recognition

The recognition of ADEs [or any other medical product-associated adverse event] is quite subjective and imprecise⁽²⁶⁾. While an attribution between the medical product and the observed event is assumed with all spontaneously reported events, every effort is made to rule out other explanations for the event in question. It is well known that placebos⁽²⁷⁾ and even no treatment⁽²⁸⁾ can be associated with adverse events. In addition, there is almost always an underlying background rate for any clinical event in a population, regardless of whether there was exposure to a medical product.

Reaching a firm conclusion about the relationship between exposure to a medical product and the occurrence of an adverse event can be difficult. In one study, clinical pharmacologists and treating physicians showed complete agreement less than half the time when determining whether medication, alcohol or "recreational" drug use had caused hospitalization ⁽²⁹⁾.

Such considerations emphasize the crucial need for careful, thoughtful review of adverse event reports upon their receipt by FDA or the manufacturer. It is through this process that causality, or at least a high degree of suspicion for a product-adverse event association, is put to the test.

Underreporting

Another major concern with any spontaneous reporting system is underreporting of adverse events ^{(16, 30-32)}. It has been estimated that rarely more than 10% of serious ADRs, and 2-4% of non-serious reactions, are reported to the British spontaneous reporting program ⁽³⁰⁾. A similar estimate is that the FDA receives by direct report less than 1% of suspected serious ADRs⁽³²⁾. This means that cases spontaneously reported to any surveillance program, which comprise the numerator, generally represent only a small portion of the number that have actually occurred. The effect of underreporting can be somewhat lessened if submitted reports, irrespective of number, are of high quality.

Biases

Unlike clinical trial data, which are obtained under strictly controlled conditions, spontaneously reported information is uncontrolled, and therefore subject to the possible influence of a number of biases that can affect reporting. These biases include the length of time a product has been on the market, country, reporting environment, detailing time and quality of the data⁽³³⁾.

A striking illustration of the impact one such factor can have is the finding that the peak of spontaneous ADR reporting for a drug is at the end of the second year of marketing , with a subsequent precipitous decline in reporting ⁽³⁴⁾ despite a lack of apparent decline in usage or change in ADR incidence^{(34, 33)}. In addition to these biases, it is possible that reported cases might differ from nonreported cases in characteristics such as time to onset or severity ⁽³⁵⁾.

Estimation of Population Exposure

Compounding these limitations is the lack of denominator data, such as user population and drug exposure patterns⁽³⁵⁾, that would provide the exact number of patients exposed to the medical product, and thus at risk for the adverse event of interest. Numerator and denominator limitations make incidence rates computed from spontaneously reported data problematic ⁽³⁵⁾, if not completely baseless. However, even if the exposed patient population is not precisely known, estimation of the exposure can be attempted through the use of drug utilization data. ⁽³⁶⁾.

This approach, whose basic methodologies are applicable to medical products in general, can be of great utility. Major sources of data on the use of drugs by a defined population include market surveys based on sales or prescription data, third-party payers or health maintenance organizations, institutional/ambulatory settings or specific pharmacoepidemiological studies⁽³⁶⁾. Cooperative agreements and contracts with outside researchers enable FDA to utilize such databases in its investigations. Device utilization studies employ the same sources of data, as well as Medicare-derived information.

Care must be taken in interpreting results from studies utilizing these databases. That drug prescribing does not necessarily equal drug usage⁽³⁶⁾, and the applicability of results derived from a specific population (such as Medicaid recipients) to the population at large, need to be weighed carefully.

Report Quality

The ability to assess, analyze and act on safety issues based on spontaneous reporting is dependent on the quality of information submitted by health professionals in their reports. A complete adverse event report should include the following:

• product name (and information such as model and serial numbers in the case of medical devices)

• demographic data

• succinct clinical description of adverse event, including confirmatory/relevant test/laboratory results

• confounding factors (such as concomitant medical products and medical history)

  temporal information, including the date of event onset and start/stop dates for use of medical product

• dose/frequency of use (as applicable)

• biopsy/autopsy results (as applicable)

• dechallenge/rechallenge information (if available)

• outcome

Given the limitations of spontaneously reported data, what are its strengths?

Return to the Table of Contents

---

STRENGTHS

Large-Scale and Cost-Effective

Two vital advantages of surveillance systems based on spontaneous reports are that they potentially maintain ongoing surveillance of all patients, and are relatively inexpensive⁽³⁷⁾. In fact, they are probably the most cost-effective way to detect rare, serious adverse events not discovered during clinical trials.

Generation of Hypotheses and Signals

Making the best possible use of the data obtained through monitoring underlies postmarketing surveillance ⁽³⁸⁾. Toward that goal, the great utility of spontaneous reports lies in hypothesis generation⁽³¹⁾, with need to explore possible explanations for the adverse event in question. By fostering suspicions⁽³⁹⁾, spontaneous report-based surveillance programs perform an important function, which is to generate signals of potential problems that warrant further investigation.

Assessment of the medical product adverse event relationship for a particular report or series of reports can be quite difficult. TABLE 4 lists factors that are helpful in evaluating the strength of association between a drug and a reported adverse event⁽⁴⁰⁾.

The stronger the drug-event relationship in each case and the lower the incidence of the adverse event occurring spontaneously, the fewer case reports are needed to perceive causality⁽⁴¹⁾. It has been found that for rare events, coincidental drug-event associations are so unlikely that they merit little concern, with greater than three reports constituting a signal requiring further study⁽³⁵⁾. In fact, it has been suggested that a temporal relationship between medical product and adverse event, coupled with positive dechallenge and rechallenge, can make isolated reports conclusive as to a product-event association⁽⁴²⁾. Biological plausibility and reasonable strength of association aid in deeming any association as causal ⁽³⁰⁾.

However, achieving certain proof of causality through postmarketing surveillance is unusual ⁽⁴¹⁾. Attaining a prominent degree of suspicion is much more likely, and may be considered a sufficient basis for regulatory decisions⁽⁴¹⁾.

Clinician Contribution

The reliance of postmarketing surveillance systems on health professional reporting enables an individual to help improve public health. This is demonstrated by one study that found direct practitioner participation in the FDA spontaneous reporting system was the most effective source of new ADR reports that led to changes in labeling ⁽⁴³⁾. Ensuring that the information provided in the adverse event report is as complete and in depth as possible further enhances postmarketing surveillance.

Thus, while possessing inherent limitations, postmarketing surveillance based on spontaneous reports data is a powerful tool for detecting adverse event signals of direct clinical impact. It is dependent not only on health professional participation, but also on the quality of the reports that are submitted.

Return to the Table of Contents