Modern psychiatry has increasingly become an evidence-based scientific discipline. Expert consensus guidelines and treatment algorithms based on scientific data can now provide clinicians with stepwise recommendations for medication selection, dosing, treatment augmentation, and duration of treatment. Since algorithms, guidelines, and consensus statements are derived from research data and clinical expertise, they may also be used to teach the fundamentals of psychiatric drug treatment to students, residents, and trainees at all levels of experience. However, there is an art as well as science to prescribing psychiatric drugs. It is the thesis of this article that over reliance on these expert guidelines for teaching psychopharmacology may actually limit learning the art of prescribing psychotropic drugs, thereby providing trainees with only part of the information necessary to be a good psychopharmacologist. For purposes of this article, all types of treatment algorithms, guidelines, and expert consensus statements will be considered together as "expert guidelines" (EGs).

Expert guidelines are well accepted in contemporary psychiatric practice as encouraging high-quality care (+1). Expert guidelines help the clinician select appropriate medications for a patient’s treatment and promote useful treatments while discouraging ineffective ones. The use of EGs may also reduce costs of medical care by eliminating ineffective practices (+2) and improve consistency of care across geographical regions and among doctors and specialties (+3,+4). As noted by Kane (+5): "Within the limits of expert opinion and with the expectation that future research data will take precedence, these guidelines provide direction for addressing common clinical dilemmas that arise in the pharmacological treatment of (psychotic) disorders." Examples of such EGs that are well publicized and widely used in psychiatry include the Comprehensive Treatment Guidelines published by American Psychiatric Association (APA) and expert consensus guidelines for many disorders, including schizophrenia (+6), bipolar disorder (+7), obsessive-compulsive disorder (+8), agitation in older persons with dementia (+9), and posttraumatic stress disorder (PTSD)(+10). The Texas Medication Algorithm Project (+11) and the International Psychopharmacology Algorithm Project (+12) are two extensive and comprehensive series of medical algorithms that focus primarily on schizophrenia, bipolar disorder, and major depression. It is important to note that most expert guidelines recommend nonpharmacological treatment modalities in addition to the use of psychotropic drugs, and some recommend drug therapy as a first-line treatment (e.g., guidelines for the treatment of PTSD).

Despite their usefulness, several authors have called attention to potential drawbacks to EGs (+3,+13). In general, EGs become less useful for complex, treatment resistant, dual diagnosis, and comorbid patients (+2,+3). It has also been suggested that some guidelines might carry biases of the clinicians who developed them (+5). Even if unbiased, some have observed that EGs, whether from experts or published literature, carry a risk that they may "offer merely a ‘cookbook’ approach to psychiatric care" (+14). There is also the potential for EGs to become treatment standards with medico-legal implications. Expert guidelines may not apply equally to all age groups. Even within an age-based cohort (e.g., the elderly), great variability in psychotropic drug disposition and efficacy may severely limit the usefulness of treatment guidelines (+1). An expert guidelines-based treatment (not limited to psychopharmacology) may lack flexibility: applicable to broad categories of patients but insensitive to the unique characteristics of individual patients (+1). As applied to the use of psychotropic drugs, EGs may not leave sufficient room for a clinician to individually tailor care based on personal circumstances and medical history. Despite the best intentions of those who create guidelines and treatment algorithms, "the frequently touted benefit of clinical guidelines—more consistent practice patterns and reduced variation—may come at the expense of reducing individualized care for patients with special needs" (+3).

An EG is essentially a "menu-driven" guide to selecting psychotropic drugs. Initial treatment recommendations are largely based on diagnostic criteria: a particular diagnosis suggests a category of medication (e.g., an antidepressant is selected when the diagnosis is depression). Rigid adherence to the linkage between diagnosis and initial drug selection places undue emphasis on DSM diagnostic categories. Not all patients fit into clearly defined diagnostic categories, and when the diagnosis is not clear or when a patient does not respond well to a diagnosis-based treatment recommendation, the use of an EG may encourage a trainee to try to fit a patient into a DSM category in order to justify selection of a particular drug. Overemphasis on the algorithms for nonresponding patients may lead to teaching an inflexible approach that may not apply to all patients. Development of a trainee’s clinical curiosity and creative skills in using all forms of treatment for difficult and complex patients may not be enhanced by this "cookbook" style of treatment.

In addition to overemphasizing the menu-driven approach to psychiatric drug selection, the use of EGs to teach principles of psychiatric drug treatment raises questions about the reliability and validity of the data on which the EGs are based. It is increasingly evident that clinical trials data may be faulty or incomplete. Data may be derived from small samples, inadequate treatment periods, dosages that do not apply to clinical practice, and based on inappropriate outcome measures or rating-scale scores. Furthermore, recent publications (+15) have focused on problems with drug company-sponsored clinical trial outcome data. For example, attention has been directed toward the lack of published negative studies (+16), as well as the promulgation of statistically significant but clinically meaningless differences between drugs. Furthermore, there are few (if any) randomized, double-blind, controlled, prospective trials of patients with multiple psychiatric disorders, medical disorders, substance abuse, and lack of response to standard treatments (+17), populations likely to be treated by trainees. Some clinical trials data, therefore, may lack generalizability from research subjects to clinical populations (+18). It is possible, therefore, that using EGs as a primary source of psychopharmacology teaching may impart incomplete or even erroneous information to a trainee.

The "art" of psychopharmacology treatment encompasses all nonpharmacological aspects of drug treatment as well as the wide range of attitudes and responses that individuals with similar diagnoses have toward psychiatric medications. Learning when and how to prescribe psychotropic drugs may be as important as learning which drug and dose to select. Examples of the art of prescribing psychotropic medications include the importance of forming a therapeutic alliance; learning to communicate with psychotic, anxious, or affectively ill patients about medications and their side effects; the use of medications in noncompliant patients; and careful attention to the impact of side effects on treatment outcome. Expert guidelines rarely address these topics. For example, EGs do not teach a trainee how to approach a frightened or paranoid patient or one who is hesitant about medication because of previous bad experience. EGs do not provide guidelines about the psychological meaning of medications to a frightened, depressed or psychotic individual. Although practicing psychopharmacology includes "listening with a third ear, staying attuned to the nuances of human communications, and using empathy and compassion" (+17), the use of an EG to teach psychopharmacology does not contribute to the development of these important skills.

The art of psychopharmacology is enhanced by familiarity with the use of a broad range of medications. By emphasizing the role of newer medications in preference to older and more traditional drugs, EGs may limit the breadth of a developing psychopharmacologist. It is not unusual to encounter psychiatrists who have recently completed their training with little or no experience using older medications such as lithium, tricyclic antidepressants, monoamine oxidase inhibitors, and conventional antipsychotics. Lack of broad-based training in the use of all available psychotropic drugs appears to be a growing problem in training programs (Annual Meeting, American College of Neuropsychopharmacology, San Juan, Puerto Rico, December 2003).

Thus, EGs, may contribute to a trainee erroneously learning that psychiatric treatment is based only on DSM diagnosis, that psychopharmacology is the dominant treatment modality for all patients, and that treatment consists primarily of finding the "right" medication or combination of medications. It is not unusual to encounter a resident who has an excellent grasp of a patient’s medical and pharmacological history and DSM diagnosis, but who, by strictly following an EG, keeps adding and/or switching medications, having completely failed to understand the psychological, social, and contextual elements of the patient’s symptoms. It is possible that failure to learn the importance of these nonbiological and nonpharmacological factors may actually encourage a trainee to employ a more aggressive use of psychotropic drugs (i.e., increased dosages, more medication switches, or inappropriate polypharmacy for insufficient treatment response). It is not unusual to encounter a trainee who presents a case of inadequate treatment response to a supervisor, with a long list of failed medication trials, asking only for advice on the next medication.

It is clear that EGs may be quite helpful for teaching basic psychotropic drug use. They provide guidelines and a rationale approach to selection of psychotropic drugs. Effective psychopharmacology training, however, rests not only learning to correctly prescribe psychotropic drugs, but on many other factors, including skills in communication with a distressed patient, formation of a therapeutic alliance, and the integration of medications together with other treatment modalities. Residents must also learn to critically judge the evidence upon which EGs are based as applied to their own patients. Training in critically evaluating clinical trials research should be included as part of psychopharmacology training. Expert guidelines may serve as a starting point for the relationship between diagnosis and drug selection, but residents must also be taught the subtleties and variability of patient response to medications.

As neurobiological research continues to progress, development of drugs with more specific receptor targets in addition to increased identification of genetically responsive patients will further define the selection of psychotropic drugs. "One size" treatment will suit fewer and fewer patients, and improved diagnostic criteria, along with biological markers, will guide drug selection with increased accuracy. Even then, however, there will continue to be an "art" to selecting and prescribing psychotropic drugs. Well-trained psychiatrists will still need to learn how to talk with disordered and upset patients and use this information, along with research-based data, to guide their prescribing. Clinical experience and wisdom will continue to develop with practice and patience complemented by use of prescribed menu-driven guidelines. At present, EGs can provide a scientific framework for the art of prescribing, but they must not supercede the development of other clinical skills during psychiatric training.