Physician Opportunity Costs for Performing Practice-Based Research

In a companion article in this issue of JFP, Beasley and colleagues¹ address issues important to infrastructure development for the nascent enterprise of practice-based research networks (PBRNs). They note that academic family physician researchers estimated it took 25% of their time over 4 years to prepare for a first successful R01 grant to perform PBR. It is important to note that this estimate was for salaried academic primary care researchers who also received institutional support in the form of computers, secretarial assistance, and so forth. This essay documents comparable opportunity costs for a nonsalaried, nonacademic family practice clinician researcher whose income was dependent on patient volume and who did not receive any institutional support for performing PBR. The data suggest that personal costs will be greater for nonacademic community-based clinician researchers than for colleagues in academic institutions.

Case report

In 1986 a physician became aware of a novel respiratory pathogen, Chlamydia pneumoniae, that he believed could explain some clinical observations relating to unexpected antibiotic responses in “viral” illnesses.² While remaining in full-time practice he enrolled patients in several office-based research studies on this organism that spanned the decade from 1987 through 1996. This research resulted in the discovery of an association between infection and asthma³ that was subsequently confirmed.⁴ After 1996 he abruptly ceased to conduct office research because of increased pressures to perform patient care that coincided with a transition from patient recruitment in his own practice to designing a PBRN study. The Figure illustrates the effect of personally performing PBR on the number of outpatient encounters, a classic measure of physician productivity. Ten years of performing PBR resulted in lost productivity (equivalent to more than $220,000 of personal income) that increased to baseline after he ceased PBR. The cumulative decrease in production (measured by patient volume) represented an average productivity loss of 18% per year for the 10 years of performing PBR. Ninety percent of the total decrease in production was attributable to time spent enrolling patients. This “opportunity cost” of PBR does not include the time spent outside patient care hours performing literature review or designing, analyzing, and publishing the research results.

Discussion

If the out-of-office time and opportunity costs of lost production in this case example are summated, the yearly costs of PBR participation in the nonacademic setting were approximately equal to the estimates for academic family practice researchers provided by Beasley and coworkers. However, in the nonacademic example the costs are accumulated over 10 years, and the research has not yet resulted in a funded R01 grant. One possible reaction to this negative calculus is to conclude that practicing physicians who want to do research should leave practice and join academia. This conundrum (“I had to leave practice in order to study it”) is common but regrettable, in my opinion. In this example, a clinically important association between infection and asthma would not otherwise have been discovered as early (or ever), given the existing academic bias against such an association.⁵

It can be argued that study of chronic disease etiology is best performed during early disease stages presenting in the primary care setting, since initiating causes may be obscured in the patients with late-stage chronic disease that are more often referred to academic medical centers. Thus, I would add etiologic research to the list of acknowledged categories (theoretical and methodologic, health care, clinical and health systems research) suitable for primary care research.⁶ The need to provide patient-oriented evidence that matters to general patient populations underlies current efforts to develop credible PBRNs to perform widely generalizable clinical research (both efficacy and effectiveness studies) in community-based outpatient settings.⁷ These efforts will require the active participation of as many practicing clinician researchers as possible.

Although it can be argued that the development and financing of PBR is no different from that for other medical research, 2 characteristics of PBR development and funding are unique. First, the infrastructure for PBR is less well developed than existing academic medical research enterprises that have had a 50-year head start in infrastructure support. Second, academic research involves researchers in academic institutions, while PBR involves practicing nonacademic physicians who often do not receive institutional support for research activities. The first distinction is one of timing and results: Whether PBR deserves future support will depend on its perceived value and outcomes.^7,8 The second characteristic, that PBR must involve practicing physicians in the origination, design, and implementation of medical research, distinguishes these studies from traditional academic medical research enterprises.

Complex clinical research, including randomized controlled trials, is frequently conducted in nonacademic community practices. This research consists primarily of phase III studies funded by pharmaceutical companies in pursuit of new drug indications and is conducted successfully by practicing physicians aided by contractually well-funded staff and infrastructure. I refer to this type of community practice-based research as “for profit” to indicate clinicians’ primary motivation for participation, although secondary motives may include physician interest, enhanced institutional reputation, and patient access to free or novel treatments. Traditional PBRNs, by contrast, follow a “not-for-profit” model: developing their own research agenda, eschewing contractual pharmaceutical funding, and (similar to their academic counterparts) seeking to perform pilot projects in pursuit of extramural funding.