The Satisfied Patient: Overprescribed and Costly

Charles P. Vega, MD


April 30, 2010

Patient Requests and Physician Responses: An Analysis Using Standardized Patients Scripted for Pain and Emotional Disorders

Paterniti and colleagues[2] analyzed prescribing habits among physicians faced with patient requests for medications, specifically antidepressants. It goes farther than previous studies in that it also analyzed physician responses to these requests and their associated effect on patient satisfaction.

The study took place in 152 primary care physician offices in the United States. Slightly more than half of practices asked to participate did so. A total of 18 white, middle-aged, female standardized patients made 298 unannounced new patient visits to these practices. These standardized patients were scripted to play 1 of 2 clinical roles: major depression with wrist pain or adjustment disorder with back pain. They also pursued 1 of 3 medication request strategies: a brand-specific antidepressant medication request, a general request for antidepressant medication, or no request.[1]

Researchers followed prescription patterns for antidepressants. In addition, they performed a qualitative content analysis of all physician responses to requests, which allowed them to categorize responses into 3 main patterns. These patterns were then associated with patient satisfaction surveys from the standardized patients. The standardized patients were aware of the study protocol but not the researchers' hypothesis.

Patient requests made a significant difference in physicians' prescribing habits. Prescriptions were offered to 31% of patients with major depression when no request for medications was made. However, the prescription rate increased to 53% when these patients requested a specific brand of medication and 76% when they requested general antidepressants. Among patients with adjustment disorder, prescriptions were offered to 10% when no request for medication was made, 55% when a specific brand of medication was requested, and 39% when general antidepressants were requested. These are remarkable findings; requests for a specific antidepressant led to inappropriate prescribing in more than half of cases.

Transcripts were reviewed from 84 visits from standardized patients. A total of 54 visits were to general internists, and 30 were to family physicians; 59 visits were to male physicians.

Nearly all physicians reviewed the patients' chief symptom of "feeling tired." Many physicians highlighted limitations of antidepressant medications, such as cost, lack of efficacy for mild conditions, and delayed onset of benefit.

Physician responses to patient requests for antidepressants fell into 3 main categories: patient perspective-based responses (63% of all responses), biomedically based responses (31%), and outright rejection of the request (6%). Most physicians using a patient perspective-based approach explored the context of the request and negotiated a timeline to treat the patient's symptoms. Twelve percent of physicians referred the patient to a counselor or mental health professional, and most of these physicians were physicians in a health maintenance organization. Physicians suggested another diagnosis beside depression in 11% of all visits, and most of these cases occurred when the patient presented with adjustment disorder.

Regarding biomedically based responses, 81% of physicians ordered laboratory tests for the standardized patients, and 13% of physicians used this work-up as a reason to delay prescribing antidepressants. Another 18% of all physicians prescribed a sleep aid for the patient.

In 6% of patient visits, physicians rejected the patient request without explanation, although this was usually done by shifting the subject of the prescription instead of addressing the request specifically. Patients were requesting a general, not brand-specific, antidepressant in all of these cases.

The overall standardized patient satisfaction rate was good. However, patient perspective-based responses were more likely than biomedically based responses or outright rejection to be associated with excellent satisfaction (43% of visits rated with excellent satisfaction vs 10%, respectively).


An obvious limitation of the current study is that it pertains to 1 type of patient request for a specific condition. The same results may not apply to patients seeking diagnostic testing or procedures, or even to requests for other types of medications. Another potential limitation involves the use of standardized patients: Can standardized patients mimic true clinic visits accurately, and do their levels of satisfaction correlate with those of real patients? A previous study found that standardized patients can be tougher graders of physicians' communication than real patients, and the level of agreement among standardized patients was higher than that among real patients.[6] This finding lends more validity to the methods of the current study.

Paterniti and colleagues' results are both disconcerting and hopeful. Too many patients received inappropriate prescriptions, which could result in more adverse events among patients who will gain little from the prescribed therapy. Applied more broadly across the health system, these results will increase costs because of the direct expense of the medications and their associated morbidity.

However, the current research also provides a blueprint for resolving this problem. Although physicians generally underestimate the power their words can have, physician responses that accounted for the patient perspective engendered the highest rates of patient satisfaction.

In a study of 750 patients presenting to a primary care clinic, 98% had at least 1 expectation before their medical visit.[7] But tests and prescription medications were not the most common expectations; instead, patients were more interested in information on their diagnosis and prognosis. In fact, failure of physicians to address diagnosis and prognosis was the most common cause of unmet patient expectations, and patients who received adequate information on diagnosis and prognosis experienced better symptom relief and functional outcomes.

In another study examining patient satisfaction with primary care physicians, watchful waiting was associated with satisfaction scores similar to those of immediate laboratory testing for unexplained symptoms.[8] The most significant factor associated with patient satisfaction was strong patient-physician communication, not the plan of management. Patients who felt ignored or confused were least likely to be satisfied, although physicians generally underestimated the importance of communication in improving patients' well-being.


Patients place a high value on physicians' communication skills and the relationship they have with their doctor. Many times, a few minutes of listening and reflection can yield much better outcomes than a prescription alone. Such small moments of empathy should never be discounted. Indeed, they are the foundation of our profession.

Clinical Pearls

  • Patient satisfaction is associated with important outcomes, such as lower rates of in-hospital mortality and complications.

  • Most patients have expectations when they come to the physician's office.

  • Although evidence exists that direct-to-consumer advertising has increased requests for specific medications, patients may be more likely to desire information about their diagnosis and prognosis than treatment.

  • Rates of inappropriate prescriptions for antidepressants exceeded 50% in Paterniti and colleagues' study.

  • Physician responses to patients' medication requests in that study were most likely to result in higher satisfaction scores when they incorporated patients' perspectives.