A little-known fact about health care is that hospitals frequently hire substitute doctors – or, as they are called in the profession, locum tenens physicians (from the Latin for “to hold the place, to substitute for”) – to temporarily cover for doctors who are vacationing, sick, attending conferences, or on leave for other reasons.

In the United States, which first started using Locum Tenens in the 1970s, demand for these temporary physicians has quietly exploded over the past 15 years – and the number of U.S. physicians working as locum tenens has risen steadily from an estimated 26,000 physicians in 2002 to 48,000 physicians in 2016, or approximately 5% of the physician workforce.

At the same time, there has been concern over whether locum tenens physicians provide lower quality care compared to permanent staff. Particularly in the UK, which relies heavily upon locum physicians, concerns about their quality of care have grown as demand for temporary physician coverage has outstripped the supply of qualified locum candidates, leading, at times, to locum tenens being hired to cover medical specialties they’re not trained in.

While surveys of administrators at U.S. hospitals indicate that they are generally satisfied with the quality of locum tenens’ work, little empirical data exists on the quality and costs of care delivered by locum tenens physicians.

We attempted to address this important knowledge gap by evaluating the quality and costs of inpatient care delivered by substitute physicians to Medicare beneficiaries hospitalized with a general medical condition (e.g., pneumonia, heart failure, shortness of breath) between 2009 and 2014. We compared both clinical outcomes and spending among inpatients treated by locum tenens and non-locum tenens general internal medicine physicians. Our study was recently published in JAMA.

The Rise of Locum Tenens

A number of factors have stimulated demand for locum tenens, including increasing demand for clinical services from newly insured patients, physician shortages, and increasing consolidation of physician practices and health systems (this results in more physicians being employed by multi-specialty group practices and hospitals, which have higher physician turnover than smaller practices).

Locum tenens are most widely used in primary care, psychiatry, and hospitalist medicine, but are regularly utilized across several other specialties.

A 2017 survey of roughly 900 locum tenens performed by Staffcare, a U.S.-based locum tenens staffing agency, provides some basic insights into who locum tenens are: Three-quarters of surveyed physicians were 51 years of age or older, 65% had been in practice for at least 21 years.  Nine of ten physicians surveyed had worked in permanent practice at some point during their careers, while 8% indicated that they had only worked as locum tenens. Approximately half of those surveyed reported that they began working as locum tenens during the middle of their careers, while 36% started taking locum tenens positions after retiring from permanent practice.  The most common factors influencing respondents’ decisions about which locum tenens positions to take were location (89%), pay rate (67%), length of assignment (60%), and patient load (38%).

What Our Data Says

Our overall findings should reassure patients, hospital and clinic administrators, clinicians, and policy makers that temporary, contracted general internists deliver inpatient care of similar quality to their non-locum tenens colleagues. However, our analysis did find that patients treated by locum tenens physicians have slightly higher costs of care and longer lengths of stay, which raises the possibility that they deliver modestly less efficient care.

Our Methodology

We found that patients treated by locum tenens and non-locum tenens physicians had similar 30-day mortality rates (8.83% for locum tenens’ patients versus 8.70% for non-locum tenens’ patients). Thirty day readmission rates were 22.80% for patients treated by locum tenens physicians, and 23.83% for patients treated by non-locum tenens physicians—a difference that is small in absolute terms but did achieve statistical significance.

On average, patients treated by locum tenens physicians cost $124 more than patients treated by non-locum tenens physicians ($1836 vs. $1712) and spent an additional 0.4 days in the hospital (mean length of stay was 5.6 days for locum tenens’ patients and 5.2 days for non-locum tenens’ patients). Both of these findings were also statistically significant, but were not large in practical terms.

While we couldn’t evaluate why patients treated by locum tenens had higher costs of care and longer lengths of stay, one potential explanation is that because locum tenens aren’t as familiar with the work environment, they’re not as efficient. It takes time to learn how to operate efficiently in a new clinical delivery system, and higher costs of care and longer lengths of stay for locum tenens’ patients could reflect this learning curve. This explanation would be consistent with research by HBS professors Rob Huckman and Gary Pisano on cardiothoracic surgeons who operate at multiple hospitals, which suggests that a physician’s clinical experience at a given hospital has a greater impact on their patients’ outcomes than their cumulative clinical experience across multiple hospitals.  Alternatively, locum tenens could simply be less efficient clinicians.

Unfortunately, we couldn’t determine whether locum tenens were unfamiliar with where they worked. Because locum tenens bill under the provider identifier of the physicians whom they are covering, they remain completely anonymous. Thus, we were unable to see how many times physicians had worked as locum tenens or whether hospitals repeatedly hired the same locum tenens.

But we do know that some hospitals also use locum tenens physicians more frequently than others. We wondered whether this influences their patients’ outcomes. For instance, if hospitals rely on substitute doctors infrequently—which we defined as hospitals where locum tenens care for less than .45% (1 in 220) of hospitalized patients on general medical wards—they may have less experience vetting these physicians or preparing them to provide temporary coverage.

To further investigate this hypothesis, we stratified the hospitals in our data into three groups (terciles): Hospitals in the upper third of our sample relied on locum tenens to deliver care for more than 2.5% of all general inpatient admissions, while hospitals in the lowest third used locum tenens to care for less than 0.45% of all general inpatient hospitalizations. When we looked at outcomes in the upper and middle terciles, we found no statistically significant differences in clinical outcomes between patients treated by locum tenens and non-locum tenens physicians.

However, for hospitals in the lowest tercile in terms of locums use, we found that adjusted 30-day mortality rates were substantially higher (11.6% vs. 8.5%) among patients treated by locum tenens versus non-locum tenens.  In other words, at hospitals that use locum tenens doctors infrequently, the patients treated by those doctors appear to have higher mortality than non-locums doctors.

We can’t explain why this is the case. But a few potential explanations deserve mention. Perhaps hospitals that use locum tenens infrequently lack access to high quality locum tenens, and employ lower quality clinicians to serve in these roles; or perhaps they do not invest in the types of onboarding and integration programs that help integrate locum tenens into their organizations. According to one UK-based surgeon who has worked as a locum tenens at 20 different hospitals, rates that locum tenens staffing agencies charge hospitals vary widely, and some hospitals simply don’t have the budgets to pay to pay “top dollar” for the best contract physicians. This surgeon also reported that only 9 of the 20 facilities provided him with any onboarding or introduction to the organization, while only 6 facilities offered instruction in how to use their computer system.

It is important to acknowledge potential limitations of our work. While we accounted for numerous variables that could have influenced our outcomes of interest, we cannot be certain that we accounted for all potential confounders. Our study also only included locum tenens physicians who met CMS’ formal definition for locum tenens, so we cannot be certain that our results generalize to other temporary physician staff that fall outside of this definition—such as physicians hired to assist with seasonal spikes in clinical demand or to fill new clinical positions until a permanent hire is made. And our findings may not apply to locum tenens in other medical specialties—including emergency medicine, surgery, and anesthesiology—or to non-Medicare patients who receive treatment from general medicine locum tenens physicians. Additional research will be extremely helpful for understanding whether the practice patterns of other temporary physicians are consistent with our observations.

With demand for temporary physicians expected to keep growing in the United States and abroad, expected to keep growing, it will become more important to learn about physicians who choose to work as locum tenens, what their training and qualifications are, and how hospitals and clinics onboard, evaluate, and manage them.

Right now, there are no national guidelines on “best practices” for hiring, training, and integrating locum tenens. Developing these could help hospitals and clinics better identify, onboard, and manage highly qualified locum tenens. Ultimately, our goal should be to build systems that enable substitute physicians to consistently deliver high quality, high value care, that does not differ from their colleagues’.