Family physicians rely on the PDR more than any other drug information resource. According to the Medical Economics Data Production Company, Montvale, NJ, which publishes the PDR: "Nine out of ten doctors consider the PDR to be their most used drug reference book. . . 97% of physicians say that the PDR is the book they go to when they're prescribing a drug they're not completely familiar with (Mr. Cy Caine, PDR Electronics, Medical Economics Data Production Co., Oral communication, September 12, 1994). Moreover, each year, approximately 500,000 PDR's are sent to physicians, and about 500,000 more are sold to other professionals and to the general public.
Even so, the PDR has received little objective scrutiny with regard to quality of its data. We will address the following questions: Is the information provided by the PDR comprehensive, current, and accurate? Does the cost-free mailing of annual editions of the PDR, underwritten by the drug industry, constitute conflict of interest? Does physicians' reliance on the PDR influence the quality and objectivity of their knowledge of drugs, and by extension, the quality given to their patients?
THE BASIS OF THE POPULARITY OF THE PDR
The popularity of the PDR among physicians is probably not a coincidence. First and perhaps foremost, for physicians, the PDR is delivered by mail free of charge. Unlike textbooks that can become quickly outdated, the PDR is updated by both periodic supplements during a given year and new annual editions.
The PDR enters the lives of physicians when they are young. The first free copy arrives during medical school. It is a powerful gift not only because it is free, but also, we believe, because receipt of the PDR has symbolic power, implying that a medical student is almost a physician. The result is that from the start, the PDR becomes identified as a natural part of being a physician. Medical students who begin clinical clerkships are often urged to consult the PDR to learn the brand names of drugs that preclinical pharmacology classes generally avoid teaching; yet, clinicians often favor brand names of drugs when they discuss and prescribe therapeutic agents. Moreover, the PDR provides information about dosage - a topic that is typically ignored in preclinical teaching.
Beyond cost and convenience, there are other reasons for the popularity of the PDR. The format of the PDR is superior to many other drug information resources. The organization of the PDR is simplified by 3 color-coded, easily identified indexes: (1) an alphabetical list of drugs that contain both the brand and generic names, (2) a listing of drugs by therapeutic category and symptoms, and (3) an index that provides the drug manufacturers' 1-800 telephone numbers for professional support. The "Product Identification Section" (ie, glossy pages of color pictures of drug formulations) is an excellent visual-identification section. Thus, its organization and presentation make the PDR very "user friendly". Package insert information in the PDR can thus be readily used to complement information that is provided by other sources of drug advertising.
As the PDR is well indexed, regularly updated, easy to use, and free, all these factors contribute to the popularity of the PDR. Moreover, because the PDR represents a compilation of package inserts that have been evaluated by the FDA, the PDR is perceived as government-approved information about drugs. Thus, it is commonly used to answer questions with regard to the "standard of care" of drug administration.
DEFICIENCIES IN THE PDR
The substance of the PDR lies in its "Product Information" section (ie, the manufacturer-written package inserts). It is in the largest and most important part of the book that we believe multiple problems exist: (1) In the PDR, dosage information and guidelines are inadequate and do not always account for interindividual variations. (2) Dosage methods that are recommended in the PDR are likely to be based on clinical studies that are skewed toward accelerated and higher dosing. (3) clinically relevant low-dose information, when present, tends to be lost amid the lengthy write-ups in the PDR. (4) Many Drug descriptions in the PDR provide no dosing adjustments for elderly patients. (5) Specific dosages for specific diagnoses are sometimes omitted. (6) Data with regard to side effects in the PDR tend to be unfocused, inaccurate, and inadequately updated. (7) The PDR sometimes contains outdated information.
Inadequate Dosage Information and Guidelines
It is a fundamental pharmacologic observation that normal subjects who are administered the same dose of a medication show variations in response. Depending on the drug, the range of interindividual variation in pharmacokinetics can be large, sometimes varying from 40-fold. If we add the factor of intraindividual variation, the implications become even greater. For this reason, textbooks of clinical pharmacology advise (and clinical experience dictates) individualization of drug dosages according to the requirements of each patient. Yet, in the guidelines in the PDR, many drugs are recommended in only 1 or 2 dosages and information with regard to dosing in relation to meals is not generally provided.
For example, terfenadine (Seldane) is recommended at 60 mg twice each day for all patients, regardless of factors such as age, size, gender, state of health, concomitant medications, or a history of medication intolerance. This one-size-fits-all approach might be appropriate if terfenadine were entirely benign, but substantial evidence implicates its role in life-threatening cardiac arrhythmias. This side effect appears to be, in part, dose-related. Moreover, data that preceded the marketing of terfenadine and that demonstrated effectiveness of the drug at dosages 50% lower than those recommended are not mentioned in its PDR description.
Product descriptions in the PDr often contain brief summaries of findings from selected clinical studies that support the dosages recommended by the manufacturer. In part, this is a consequence of the regulatory environment that demands agreement between manufacturer and the FDA with regard to the content of these descriptions. Studies that demonstrate effectiveness of lower than recommended dosages may be omitted from the PDR. examples include flurazapam hydrochloride (Dalmane), zolpidem tartrate (Ambien), ibuprofen (Motrin), doclofenac sodium (Voltraren), and omeprazole (Prilosec).
Similarly, most PDR drug descriptions provide pharmacokinetic data on certain factors (eg, peak plasma drug levels, mean elimination half-lives); however, in many cases, these data are expresses as statistical means without confidence limits. We believe that the absence of such data may inhibit the ability of physicians to understand unusual idiosyncratic reactions in some patients.
Guidelines for administration of medications at only 1 or 2 doses are sometimes medically irrational; yet, new drugs often are recommended in the PDR with this type of guideline. Perhaps this has more to do with economics than science. Drug manufacturers understand that penetrating an established market is difficult; thus, nonscientific issues may take precedence in carving out a niche for a new drug.
A drug for which the formulation and dosage are easy for physicians to remember (and for patients to use) will presumably have a better chance of being prescribed than will a drug with complex dosage guidelines that may better match the range of individual variation. Although flexible dosing schedules may be more medically sound, complicated dosing schedules are harder to remember, take more time to explain to patients, and may contribute to noncompliance. Thus, flexible dosing may be a liability in a competitive market for drug prescriptions.
Basis of PDR Recommended Dosage Methods
Because of economic and time considerations, Phase 1 and 2 studies may be relatively brief. For example, in the case of fluoxetine, studies were sometimes ended before this long-acting drug reached its peak blood levels. Studies that have been designed to define dosage may have been conducted with subjects who were younger, healthier, and thus potentially more tolerant of parmacologic effects than are patients. All of these factors favor higher dosages.
Recent efforts by the FDa to achieve a more balanced gender and racial representation in Phase 1, 2, and 3 studies will improve but may not completely remedy this situation. Even under the best of circumstances, studies that are conducted on a small sample are limited in their ability to determine side effects and optimal dosages in the larger population. Information about adverse drug reactions and the full range of dosing regimens are sometimes only learned after a drug is made available for prescribing, yet, the dosing recommendations in the PDR that are based almost entirely on data from Phase 1, 2, and 3 studies do not incorporate information obtained during Phase 4 (after drug release) unless a manufacturer applies to the FDA to change the package insert.
Clinically Relevant Low-Dose Information Amid Lengthy Write-ups in the PDR
The PDR sometimes offers efficacy and toxicity data associated with lower drug dosages, but these data can be difficult to find or recognize. In the description of cimetidine (Tagamet), one reads: "800 mg h.s. [at bedtime] is the dose of choice for most [acute duodenal ulcer] patients..."; yet, who these "most patients" are is not precisely defined. Elsewhere in the write-up, one finds results of a study in which cimetidene (400 mg at bedtime) healed 66% of patients with duodenal ulcers after 4 weeks of treatment; other data indicate that treatment with 800 and 1600 mg of cimetidine given at bedtime healed 75% and 81% of patients, respectively. Thus, most patients, healed in response to 400 mg of cimetidine that was taken at bedtime, but the recommendation ignores these results. A more prominent mention of the effectiveness of the lower dose in the dosage guidelines for cimetidine would offer physicians an alternative for use with patients who are prone to side effects.
Ondansetron hydrochloride (Zofran) provides another example. If one reviews a detailed table that is published in the PDR and that summarizes the efficacy of the recommended dosage (8 mg 3 times daily) of ondansetron hydrochloride, one is likely to miss the effectiveness of another dosage (4 mg 3 times daily), unless one reads a footnote to the table. In addition, the effectiveness of a dosage of 1 mg 3 times daily in some subjects is not mentioned at all.
With ranitidine hydrochloride, the PDR acknowledges the effectiveness of treating a duodenal ulcer with a dosage of 100 mg twice daily, which is 33% lower than the recommended dosage of 150 mg twice daily; yet, a preparation is not marketed for treatment with the lower dosage.
Absence of Dosing Adjustments for Elderly Patients in Drug Descriptions
Many drugs demonstrate altered pharmacokinetics in elderly patients, owing sometimes to hepatic, but more commonly, to decreased renal excretion as a consequence of a decreased glomerular filtration rate and renal blood flow. As a result, elderly patients should receive lower doses of certain drugs than do other adult subjects to achieve equivalent efficacy and to prevent unnecessary toxic effects. Lower doses may be safer, too, because the responses of older patients to drugs can be more unpredictable than for other patient populations. this unpredictability, coupled with a higher drug use than in other age groups, has resulted in a higher incidence of side effects in the elderly compared with that in the younger adult population.
Until recently, the FDA did not require package inserts to include dosage adjustments for elderly patients. In 1989, for 425 drugs that are often used by elderly patients, only about half had information in the PDR about use in this population. Thus, information in the PDR may be misleading with regard to the safety of "standard adult" dosages in elderly patients. Given the growing number of elderly patients in the United States, and the possibility that interindividual variation in these patients may be even broader than among the general population, the PDR should provide dosage guidelines for elderly patients for all drugs that are used in this population.
Omission of Specific Dosages for Specific Diagnoses
Drug manufacturers are not required to study every possible use of a new drug. Approval by the FDA requires proof of effectiveness in the treatment of 1 definable disorder. Yet, once a drug is approved, physicians can (and often do) use it for any condition that they choose. In some cases, these "new" uses prove to be helpful to patients; thus, medical practice is not necessarily reflected by what is contained in the PDR. For a manufacturer to add a new indication to the PDR would require the filing of a New Drug Application with the FDA. As a result, guidelines for use and dosage in the PDR may not be accurate for certain conditions. Examples include drugs such as diclofenac for pain or tendinitis, propranolol hydrochloride (Inderal) for mitral valve prolapse syndrome, trazodone hydrochloride (Desyrel) for insomnia, and topical tretinoin (Retin-A).
This may be a particular problem for certain classes of therapeutic agents. One example relates to antidepressant drugs, for which most dosage guidelines in the PDR make no reference to specific diagnoses. Although fluoxetine, paroxetine hydrochloride (Paxil), sertraline hydrochloride (Zoloft), venlafaxine hydrochloride (Effexor), amitryptyline hydrochloriede (Elavil), imipramine hydrochloride (Tofranil), and other antidepressants were initially approved for treating major depression, the guidelines do not indicate that the recommended dosages are specific for this condition. This is important because these drugs are commonly utilized in treating milder conditions (ie, dysthymic panic, and obsessive-compulsive disorders, pain an premenstrual syndromes, insomnia, and other types of depression) that may respond to lower dosages than those recommended for major depression. The result is that physicians tend to prescribe the PDR-recommended dosages that befit major depression to patients with milder conditions; this practice may play a role in the high incidence of side effects that are seen with antidepressant drugs.
Unfocused, Inaccurate and Inadequately Updated Data for Side Effects in the PDR
The range and rate of side effects listed in the PDR are derived from findings from both preclinical and clinical studies, which are, by necessity, limited in the number of subjects and duration of treatment. Thus, previously unrecognized side effects and varying incedences of already recognized side effects may be discovered during Phase 4, as drugs are administered to a larger number of patients with a wide variety of diseases and conditions and for longer terms than when those drugs were used in Phase 2 and 3 trials. Typically, though, a package insert (PDR description) for a new drug is written before the drug is able to be prescribed. The result is that the PDR tends to present potentially inaccurate, imprecise, and incomplete data with regard to side effects.
Moreover, a review of the profile of side effects of a given drug in the PDR can often be so extensive that clinically important side effects may be difficult for the physicians to discern from those that are less frequent and less severe. This is particularly the case with older medications, for which the frequency of side effects is generally not provided in the PDR. Because this information is lacking in the PDR, physicians may not appropriately warn patients of side effects that are most serious and most likely to occur, and when a side effect occurs, physicians may overlook the drug as the culprit.
For example, since 1990, the PDR has listed the rate of fluxoxetine-induced sexual impairments (reduced libido, impaired ability for ejaculation, or orgasm) at about 2% each whereas studies have reported the combined incidences as high at 34%. Physicians who have prescribed fluoxetine may not have been aware of these facts, in part because the PDR was not updated to reflect this information. Interestingly, because the manufacturer sought FDA approval for this drug in treating obsessive-compulsive disorders, a new side-effect table occurs rates of 11% for decreased libido plus 7% for abnormal ejaculation in males; the lower incidences listed in the table regarding use in depression remain the same.
Outdated Information in the PDR
Even though a new edition of the PDR is published annually, the information contained therein is not necessarily up-to-date. If one compares the drug descriptions in the 1995 PDR with versions published a decade (or, in some cases, 2 decades) earlier, many of the descriptions are virtually identical in the 2 versions. Examples include the descriptions of Benadryl (diphenhydramine hydrochloride), Esidrix (hydrochlorthiazide), Dalmane (flurazepam hydrochloride), and Elavil (amitriptyline).
With most medications, post marketing studies and clinical experience reveal unanticipated changes in uses and dosing patterns. These discoveries are reported in journals and some textbooks, but physicians may not consult these sources as often as they consult the PDR. Consequently, the PDR does not keep physicians informed about new uses and new concerns. This issue is related to the general lack of incorporation of information obtained from postmarketing (Phase 4) studies into the PDR.
AN INHERENT CONFLICT OF INTEREST?
The poor prescribing habits of physicians and the relationship between these habits and iatrogenic illnesses have been documented. We believe that such habits in part relate to the information on which physicians base therapeutic decisions. Where do physicians turn for this information? Ideally, this information would be objective and would derive from medical school course work, textbooks, journals, postgraduate training, and continuing education courses. The reality, though, is that a substantial portion derives from the PDR. In essence, the industry that develops medications is a principal source of information for physicians who prescribe and choose among these products, thus presenting a problematic conflict of interest.
Although it is expected that physicians should play an independent role in the evaluation and utilization of the medications that they prescribe, reliance on the PDR (and other pharmaceutical company-derived sources of information) will almost certainly comprise independent judgements. In addition, as a company-derived collection of package inserts, the PDR provides neither adequate comparative analyses among drugs in similar or different classes nor information about cost, which is an issue of immense importance in the current era of health care provision.
THE ROLE OF THE FDA IN THE CONTENT OF THE PDR
Although the information contained in the PDR is generated by the pharmaceutical manufacturers, the individual drug descriptions must be approved by the FDA. Generally, the approved process is negotiated by representatives of the manufacturers and the FDA officials, often with input from external advisers to the FDA. This helps to provide a consensus for what is included in the package insert, but it may limit the information that manufacturers can include in their drug descriptions. For example, FDA regulations that require manufacturers to establish the safety and efficacy of a given drug and dosage may preclude a manufacturer's ability to include information about lower doses that have been shown to be effective in premarketing studies but have not been shown conclusively enough or in enough subjects to meet FDA standards. Similarly, requirements by the FDA may make it difficult for manufacturers to add information about previously unproved uses for their drugs despite findings during Phase 4 studies. Thus, procedural features and statutory rules probably contribute to shortcomings in the PDR.
ASSURANCE OF THE RELIABLE SOURCE OF DRUG INFORMATION FOR PHYSICIANS
The ultimate responsibility for therapeutic decisions lies with the medical community, but physicians rely on the data that are made available to them. Without the PDR or a similar equivalent, where would physicians derive drug information, and would it be an improvement?
One possibility for ensuring a more objective source of drug information might be the dissemination to all physicians on an annual basis of 1 or more monograph, that would provide such information. The principal impediment to the dissemination of this type of monograph to all practicing physicians is financial, since the cost of preparation of an annual mailing would be large. Although the advent of computer-based technology should allow development of online, readily updated, objective, and inexpensive guides for the prescribing of drugs that include comparative information about the pharmacology and cost of different drugs, it is not clear that an interest group is committed to the development of such a guide. Nevertheless, this alternative deserves further consideration. We believe that a more feasible alternative is to improve the PDR, thereby maintaining the benefit of its already established system of annual dissemination and its long history of acceptance by physicians. Improvement of the PDR could be accomplished by having it meet well-defined medical and pharmacologic standards including (1) regular updating of drug descriptions to incorporate new data, in particular, data obtained in Phase 4 studies with regard to clinical uses, side effects, and doses; (2) recommendations about adjustments in dosing required for both younger adult and elderly patients and, to the extent possible, for pediatric patients as well; (3) hightlighted sections to illustrate changes in annual editions; (4) more precise information with regard to the frequency of side effects and results of studies with lower than "standard" doses; and (5) the current wholesale cost of a drug.
Are the changes realistic? Could pharmaceutical manufacturers, the medical profession, and the FDA work cooperatively to fashion such a PDR? In fact, this potential has already been met in isolated cases. For example, the pharmacokinetic data on desipramine hydrochloride (Norpramin) mentions that a 36-fold difference in plasma levels has been seen in individuals who take identical oral doses of this drug, thus alerting physicians to the wide range of clinical responses that may occur.
The oral anticoagulant warafrin sodium (Coumadin) demonstrates that dosage guidelines can be updated. For 2 decades, warfarin has sometimes been prescribed in unnecessarily high dosages, as reflected by a high incidence of side effects. In response to this problem, in 1990, the manufacturer updated the dosage recommendations for this drug. The new dosages are lower, the range is wider, and the recommendations include the need to individualize dosages, in particular with lower dosages for elderly and debilitated patients.
The PDR dosage recommendations can be changed when it has been deemed worthwhile or necessary. For example, triazolam (Halcion) and temazepam (Restoril) have had a lowering of dosages recommended in the PDR.
Indeed, if the potential of a "new" PDR is to be maximized, a reevaluation of the role of the FDA must also be part of the process. A requirement of more information about clinical responses a different dosing levels for different diseases, an assessment of drugs so as to define appropriate dosages for geriatric patients, and a regular updating of dosage guidelines and of prevalence rates of side effects would seem to be appropriate stipulations. In addition, revision of regulations by the FDA so that manufacturers would add newly discovered uses and results of efficacy and toxicity in postmarketing studies. By requiring that manufacturers add dates to all drug descriptions, the FDA would alert readers to the timeliness of the data in the PDR.
CONCLUSIONS
The PDR is a mainstay of drug information for American physicians and probably the profession's leading source of drug data and prescribing guidelines. Despite this key role as a source of drug information, we believe that a number of aspects of the PDR require reappraisal in light of the evolving demography of the American population, pharmacokinetic and pharmacodynamic discoveries, clinical experiences with marketed drugs, and the reality of cost containment in medicine today.
Because physicians, as well as informed patients, need a source of drug information in 1 form or another, there is a role for a PDR - an objective, comprehensive, carefully conceived collection of descriptions of available drugs that can provide guidance about prescribing and utilizing medications. Although ultimately we believe that computer-based, online sources will become the standard for drug information to physicians (and, indeed, one of these sources may well be the PDR), at the present time we propose that appropriate revisions of what is in part an imprecise and outdated PDR could prove to be beneficial to the prescribing of drugs and, thus, to patient care.
Return to Dr. Hookman's Home Page