The approval of new drugs by the US Food and Drug Administration (FDA) depends on demonstration of their safety and efficacy through clinical studies. To best inform clinical practice, these pivotal trials would ideally be conducted in care settings where the drugs will ultimately be used. For example, drugs used to treat diabetes mellitus are most frequently prescribed to patients by primary care physicians in office-based settings, yet trials used to support their approval may be conducted among patients receiving care from endocrinologists, in acute care settings and/or in non-US-based healthcare delivery systems 1. Our objective was to characterize the patient populations and settings of pivotal trials supporting FDA approvals from 2005 to 2012 for new drugs used for indications most commonly treated in the primary care setting.


Using data previously collected for all new drugs approved by the FDA between 2005 and 2012 2, we identified drugs originally approved for use for indications most commonly treated in primary care, along with the pivotal trials that supported their approvals. Using publicly available FDA documents, including medical officer reviews, and publications 3, four independent reviewers collected the following characteristics for each trial: recruitment setting, principal investigator (PI) specialty, and study country. Discrepancies were resolved by discussion. We used descriptive statistics to quantify these trial characteristics.


Forty new drugs approved between 2005 and 2012 were determined as being most commonly used in primary care settings (Table 1). These drugs were approved on the basis of 151 pivotal trials, for which 129 corresponding publications were identified. Of these publications, we were unable to ascertain the trial setting for 96 (74%), 22 (17%) were conducted in non-primary care settings, 11 (8%) in both primary care and non-primary care settings, and none exclusively in primary care settings (Table 2). PI specialty could be determined for 121 (89.4%), of which 116 (95.9%) were specialty physicians and five were primary care physicians (internal medicine, 4; family medicine, 1). Among the 129 publications, 38 (30%) were conducted exclusively in the USA, 24 (19%) exclusively outside the USA (19%), 39 (31%) both inside and outside the USA, and 27 (21%) had no discernible study locale.

Table 1 Characteristics of new drugs for indications most commonly treated in primary care approved by the FDA from 2005 to 2012 (n = 40).
Table 2 Patient recruitment setting and country for pivotal trials of new drugs for indications most commonly treated in primary care supporting the approval by the FDA from 2005 to 2012 (n = 129).


Among the new drugs approved between 2005 and 2012 that are most commonly used in the primary care setting, three-quarters of pivotal trials supporting FDA approval did not provide sufficient information in publicly available FDA documents and corresponding publications to determine if they were conducted in primary care settings. Based on the information available, we determined that 8% of trials were partially conducted in primary care settings—none exclusively so. Our results suggest that the approval of these new drugs were predominantly supported by evidence generated in subspecialty settings, despite that most end users of these therapies will likely be patients and prescribers in primary care settings.

Evaluating drugs in settings that do not resemble where they will most often be used may result in outsized estimates of benefit—because of the risk-treatment paradox 4—as well as safety risk, since patients with more comorbidities who are taking more medications—and thus more likely to require specialty care—are more likely to experience adverse drug events 5. The Randomized Aldactone Evaluation Study was one instance; following publication, the rate of hospitalization for hyperkalemia increased from 2.4 to 11.0 per 1000 patients (P < 0.001), and the associated mortality increased from 0.3 to 2.0 per 1000 patients (P < 0.001) from 1994 to 2001, presumably due to the different patient populations seen in trial and practice settings 6. There are other important differences in care processes between primary care and specialty care settings that may affect medication management, including visit length, ancillary staff availability, and educational support.

Our estimates are deliberately conservative, as the descriptive text provided in FDA documents and publications was often vague. It is possible that settings described ambiguously as “clinics” were in fact primary care clinics, but without explicit detail; we categorized these as indeterminate. This information could be clarified through the FDA’s Drug Trials Snapshots initiative (, which provides demographic characteristics of trial participants of newly approved drugs, offering additional regulatory transparency and information to patients and clinicians. Future efforts should encourage trial recruitment in clinical settings in which the drugs will most commonly be used, ensuring new drug evaluations that can best inform patient care.