Association Between Treatment by Locum Tenens Internal Medicine Physicians and 30-Day Mortality Among Hospitalized Medicare Beneficiaries

Daniel M. Blumenthal; Andrew R. Olenski; Yusuke Tsugawa; Anupam B. Jena

doi:10.1001/jama.2017.17925

Key Points

Question听 Are there systematic differences in the quality of inpatient care provided by locum tenens vs non鈥搇ocum tenens internal medicine physicians?

Findings听 In this retrospective cohort analysis of 1鈥�818鈥�873 Medicare beneficiaries hospitalized between 2009 and 2014, there was no significant difference in 30-day mortality among patients treated by locum tenens physicians compared with those treated by non鈥搇ocum tenens physicians (8.83% vs 8.70%).

Meaning听 Among hospitalized Medicare beneficiaries, treatment by locum tenens physicians overall was not associated with excess mortality risk.

Abstract

Importance听 Use of locum tenens physicians has increased in the United States, but information about their quality and costs of care is lacking.

Objective听 To evaluate quality and costs of care among hospitalized Medicare beneficiaries treated by locum tenens vs non鈥搇ocum tenens physicians.

Design, Setting, and Participants听 A random sample of Medicare fee-for-service beneficiaries hospitalized during 2009-2014 was used to compare quality and costs of hospital care delivered by locum tenens and non鈥搇ocum tenens internal medicine physicians.

Exposures听 Treatment by locum tenens general internal medicine physicians.

Main Outcomes and Measures听 The primary outcome was 30-day mortality. Secondary outcomes included inpatient Medicare Part B spending, length of stay, and 30-day readmissions. Differences between locum tenens and non鈥搇ocum tenens physicians were estimated using multivariable logistic regression models adjusted for beneficiary clinical and demographic characteristics and hospital fixed effects, which enabled comparisons of clinical outcomes between physicians practicing within the same hospital. In prespecified subgroup analyses, outcomes were reevaluated among hospitals with different levels of intensity of locum tenens physician use.

Results听 Of 1鈥�818鈥�873 Medicare admissions treated by general internists, 38鈥�475 (2.1%) received care from a locum tenens physician; 9.3% (4123/44鈥�520) of general internists were temporarily covered by a locum tenens physician at some point. Differences in patient characteristics, demographics, comorbidities, and reason for admission between locum tenens and non鈥搇ocum tenens physicians were not clinically relevant. Treatment by locum tenens physicians, compared with treatment by non鈥搇ocum tenens physicians (n鈥�=鈥�44鈥�520 physicians), was not associated with a significant difference in 30-day mortality (8.83% vs 8.70%; adjusted difference, 0.14%; 95% CI, 鈭�0.18% to 0.45%). Patients treated by locum tenens physicians had significantly higher Part B spending ($1836 vs $1712; adjusted difference, $124; 95% CI, $93 to $154), significantly longer mean length of stay (5.64 days vs 5.21 days; adjusted difference, 0.43 days; 95% CI, 0.34 to 0.52), and significantly lower 30-day readmissions (22.80% vs 23.83%; adjusted difference, 鈭�1.00%; 95% CI 鈭�1.57% to 鈭�0.54%).

Conclusions and Relevance听 Among hospitalized Medicare beneficiaries treated by a general internist, there were no significant differences in overall 30-day mortality rates among patients treated by locum tenens compared with non鈥搇ocum tenens physicians. Additional research may help determine hospital-level factors associated with the quality and costs of care related to locum tenens physicians.

Introduction

The use of locum tenens, substitute physicians who cover established clinical positions when full-time physicians are temporarily unavailable, has increased significantly in the United States between 2000 and 2015.¹^-3 Several factors have spurred this demand for locum tenens physicians, including local and regional physician shortages; greater rates of employment by multispecialty group practices and hospitals, which have higher physician turnover than smaller practices; and increasing demand for clinical services among newly insured patients.¹^,2,4

At the same time, single, small group, and rural physician practices depend on locum tenens physicians for coverage during vacations and conferences.¹^,3^,5 Demand for locum tenens physicians has been highest in primary care, psychiatry, and hospitalist medicine.¹ Outside of the United States, hospitals and practices in several countries, including the United Kingdom and Australia, also rely heavily on locum tenens physicians.¹^,6^,7

Despite the importance of locum tenens physicians in US health care, comprehensive national data on the quality and costs of care of these physicians are lacking.³^,6 Because locum tenens physicians typically spend limited amounts of time at any single institution, their relative unfamiliarity with a hospital鈥檚 systems (including electronic health records), staff, hospital culture, and local postacute care options could, among other reasons, predispose to worse patient outcomes and increased costs of inpatient care.¹^,8^-10 Therefore, national data on Medicare beneficiaries hospitalized at acute care hospitals during 2009-2014 and treated by a general internist were used to evaluate the quality and costs of care provided by locum tenens physicians.

Methods

This study was approved by the Harvard Medical School institutional review board, which waived participant consent.

Study Sample and Data Sources

This study included all acute care hospitalizations identified in a random sample of Medicare fee-for-service beneficiaries aged 65 years and older during 2009-2014. For 2011-2014, these acute care hospitalizations were identified using the Medicare Provider Analysis and Review 20% files linked by beneficiary identification to the 20% Medicare Carrier files. For 2009-2010, 5% random Medicare claims samples were used because 20% Medicare files were not available. Hospital characteristics were obtained from the 2011 American Hospital Association Annual Survey.¹¹

Identification of Locum Tenens Hospitalizations

The study sample included hospitalizations in which the assigned attending physician was a general internal medicine physician. Physicians were assigned to each hospitalization based on the physician鈥檚 National Provider Identifier (NPI) in the Carrier File that accounted for the plurality of inpatient evaluation and management services during that hospitalization.¹² The study sample was restricted to hospitalizations in which the assigned physician had a designated specialty of 鈥渋nternal medicine,鈥� identified by NPI through a comprehensive physician database with specialty information obtained from multiple sources, including the National Plan and Provider Enumeration System Registry, the American Board of Medical Specialties, and state licensing boards.¹³^,14

Locum tenens hospitalizations were defined as those in which a locum tenens physician provided the plurality of evaluation and management visits. Locum tenens physicians鈥� evaluation and management claims were ascertained using the Healthcare Common Procedure Coding System modifier code Q6,¹⁵ which identifies services furnished by locum tenens physicians and is applied in the Carrier File to the NPI of the regular physician for whom the locum tenens physician is covering. Although the term locum tenens is often used broadly to refer to any temporary physician hire, Centers for Medicare & Medicaid Services (CMS) regulations specifically designate that the locum tenens modifier code only be used when (1) the locum tenens physician is covering an established clinical position that has been temporarily or permanently vacated, (2) the locum tenens physician provides 60 days or more of coverage, and (3) the locum tenens physician is not a member of the practice that is seeking coverage.¹⁵ Therefore, according to the definition by CMS, locum tenens physicians provide temporary coverage for non鈥搇ocum tenens physicians with established clinical positions. Physicians hired temporarily to fill new clinical positions (eg, to increase staffing during periods of excess seasonal demand or to temporarily fill a new clinical position) are not locum tenens.

Of note, the number of locum tenens physicians could not be identified because Medicare data do not include the NPI of the locum tenens physician, but rather the NPI of the physician who was temporarily covered by a locum tenens physician. Because a single locum tenens physician may temporarily cover the care of multiple non鈥搇ocum tenens physicians, the data can therefore only provide information on how many physicians are temporarily covered by a locum tenens physician, rather than the number of locum tenens physicians providing care.

Study Outcomes and Covariates

We compared quality and costs of care during hospitalizations involving locum tenens vs non鈥搇ocum tenens physicians. The primary outcome was 30-day mortality (death within 30 days of admission); secondary outcomes included inpatient costs of care, 30-day readmission rates, and length of stay. Costs were defined as total Medicare Part B spending per hospitalization because Part B spending encompasses professional and other services at a physician鈥檚 discretion and is therefore a proxy for inpatient resource use.¹²

Covariates included patient age, sex, race/ethnicity, Medicaid eligibility status (鈥渄ual eligibility鈥�), and patient-level Elixhauser comorbidity indicators for 31 conditions.¹⁶ Race and ethnicity were defined using standard fixed categories available in Medicare claims and originally provided to Medicare by the US Social Security Administration; these data were included because they are associated with care quality and spending, and therefore, they are important potential confounders.¹⁷^,18 The reported diagnosis-related group (DRG) was used to assign each admission into 1 of 25 mutually exclusive major diagnostic categories. Indicator variables for admission month were included to control for seasonal differences in locum tenens physician use that could correlate with outcomes. In addition, indicator variables for day of week of admission were added to control for factors associated with day of admission鈥攕uch as staffing patterns and patient severity of illness鈥攖hat might also correlate with outcomes. Furthermore, hospital fixed effects (ie, hospital-specific indicator variables) were included to account for hospital characteristics鈥攊ncluding unmeasured differences in patient populations and systematic differences in care quality鈥攖hat remain stable over time.

Statistical Analysis

Characteristics of patients treated by locum tenens and non鈥搇ocum tenens physicians were compared. Tests for observable differences in patient demographics and illness severity, including preexisting chronic conditions and the full distribution of admission diagnoses (DRGs) were performed. The geographic distribution and structural characteristics of hospitals where locum tenens physicians practice, including variations in locum tenens physician use across the 306 US hospital referral regions, which are regional markets for tertiary care,¹⁹ were also described.

A hospitalization-level multivariable logistic regression was estimated, modeling 30-day mortality as a function of whether care was provided by a locum tenens physician (indicator variable), patient covariates, indicator variables for day of the week and month of admission, and hospital fixed effects. Identical multivariable linear regressions were estimated for Medicare Part B spending and length of stay. Hospital fixed effects accounted for the possibility that locum tenens physicians may disproportionately practice in hospitals with systematically different quality of care, and indicator variables for month of admission controlled for seasonal differences, which could confound the analyses. This approach therefore compared differences in outcomes between locum tenens and non鈥搇ocum tenens hospitalizations within the same hospital and month.

All models included robust variance estimators to account for clustering of admissions within hospitals.²⁰ Adjusted patient outcomes for locum tenens and non鈥搇ocum tenens physicians were calculated using the marginal standardization form of predictive margins.²¹ We also assessed whether patient mortality was associated with locum tenens care in prespecified subgroups defined by hospital size, hospital profit status, and 鈥渓ocum intensity.鈥� Locum intensity was defined as the percentage of hospitalizations involving a locum tenens physician among hospitals with any locum use. This measure was designed to assess whether a hospital鈥檚 degree of experience with locum tenens physicians modified the association between treatment by a locum tenens physician and mortality. In each subgroup analysis, a formal test of interaction was conducted on the basis of standard errors of the regression coefficient on the interaction term between the locum tenens indicator variable and indicator variables for a given subgroup.

Additional Analyses

An analysis was performed to assess for unobservable, systematic differences in case mix between locum tenens and non鈥搇ocum tenens physicians. Some primary care physicians care for their primary care patients when they are hospitalized. If primary care physicians with sicker outpatient panels are more likely to use locum tenens physician coverage, then hospitalized patients cared for by covering locum tenens physicians could have a higher unobserved 30-day mortality risk. To account for this potential confounder, the mortality regression model was re-estimated using the NPI of the physician covered by the locum tenens physician to control for physician-level, rather than hospital-level, fixed effects. In this way, time-invariant characteristics of physicians covered by locum tenens physicians, including differences in outpatient case mix, were accounted for.

A trend analysis was also performed to evaluate for changes in the relationship between locum tenens physician care and patient 30-day mortality between 2009 and 2014. This analysis was performed for all patients in the study cohort and among patients treated at hospitals with high, medium, and low intensity of locum tenens physician use. For each group, the locum tenens indicator variable was interacted with year indicators, which allowed for the association between locum tenens care and a given outcome to vary nonmonotonically by year. All analyses were performed in R (version 3.2; R Foundation) and Stata (version 14; StataCorp). The 95% CIs around reported estimates reflect 0.025 in each tail or P鈥夆墹鈥�.05.

Results

The study sample included 1鈥�708鈥�398 non鈥搇ocum tenens hospitalizations and 38鈥�475 locum tenens hospitalizations (2.1% of hospitalizations) at 1590 hospitals. Approximately 9.3% of non鈥搇ocum tenens general internists in the study sample (4123/44鈥�520) were temporarily covered by a locum tenens physician at some point.

Intensity of locum tenens physician use varied substantially across hospital referral regions, ranging from close to zero percent of Medicare inpatient admissions in some hospital referral regions to more than 15% of admissions in others (Figure 1). Compared with non鈥搇ocum tenens physicians, locum tenens physicians disproportionately practiced in the South and West census regions, in smaller suburban and rural hospitals (vs urban), and in public hospitals (Table 1).

No clinically relevant differences were observed in the demographic and clinical characteristics of patients treated by locum tenens and non鈥搇ocum tenens physicians and statistically significant differences (including age and several chronic conditions) were small in magnitude (Table 1). After adjusting comparisons of patient characteristics for hospital fixed effects (ie, assessing the balance of patient characteristics within hospital), fewer significant differences between patients treated by locum tenens and non鈥搇ocum tenens physicians were observed (eTable 1 in Supplement). There were no significant differences in the distributions of DRG diagnoses and modal admission DRG diagnoses between patients cared for by locum tenens and non鈥搇ocum tenens physicians (Figure 2 and Figure 3).

Patient Outcomes

Unadjusted 30-day mortality was 9.01% (95% CI, 8.73%-9.30%; 3501/38鈥�475) for patients treated by locum tenens physicians vs 8.69% (95% CI, 8.64%-8.73%; 154鈥�717 /1鈥�780鈥�398) for patients treated by non鈥搇ocum tenens physicians). There was no significant difference in adjusted 30-day mortality between patients treated by locum tenens vs non鈥搇ocum tenens physicians (8.83% [95% CI, 8.52%-9.14%] vs 8.70% [95% CI, 8.69%-8.70%]; absolute adjusted difference, 0.14% [95% CI 鈭�0.18% to 0.45%]) (Table 2).

Patients treated by locum tenens physicians had significantly higher adjusted total Part B charges ($1836 [95% CI, $1806-$1866] vs $1712 [95% CI, $1712-$1713]; absolute adjusted difference, $124 [95% CI, $93-$154]), longer lengths of stay (adjusted mean length of stay, 5.64 days [95% CI, 5.55-5.73] vs 5.21 days [95% CI, 5.21-5.21]; adjusted difference, 0.43 days [95% CI, 0.34-0.52]), and significantly lower readmission rates (22.80% [95% CI, 22.26%-23.28%] vs 23.83% [95% CI, 23.82%-23.84%]; adjusted difference, 鈭�1.00% [95% CI, 鈭�1.57% to 鈭�0.54%]) than patients treated by non鈥搇ocum tenens physicians (Table 2).

Subgroup Analysis

In prespecified subgroup analyses over the full study period (2009-2014), adjusted 30-day mortality was significantly higher among patients treated by locum tenens vs non鈥搇ocum tenens physicians at hospitals in the lowest tercile of locum tenens intensity (11.63% [95% CI, 9.96%-13.30%] vs 8.53% [95% CI, 8.53%-8.53%]; absolute difference, 3.10% [95% CI, 1.43%-4.77%]; adjusted odds ratio, 1.47 [95% CI, 1.22-1.77]; both P values <.001 for interactions of locum tenens physicians with indicator variables for hospitals in the middle and upper terciles of locum tenens intensity, respectively, relative to hospitals in the lowest tercile of locum tenens intensity) (Table 3). However, in year-by-year analysis of patients admitted to hospitals in the lowest tercile of locum tenens intensity, mortality differences between patients treated by locum tenens vs non鈥搇ocum tenens physicians were largest during 2009-2012 (statistically significantly different from zero in 2011-2012 but not significant in 2009-2010) and nonsignificant and close to zero in magnitude during 2013-2014. Differences in annual 30-day mortality rates among all patients treated by locum tenens vs non鈥搇ocum tenens physicians, and among subgroups of patients treated at low, medium, and high locum tenens intensity hospitals, did not vary by year in a trend analysis (Table 4; eTable 2 in the Supplement). No effect modification was observed according to hospital size or for-profit status (Table 3).

Additional Analyses

In analyses adjusted for physician fixed effects, patients hospitalized during periods of locums tenens physician coverage were not systematically unhealthier than patients hospitalized during other periods, and care by locums tenens physicians was not associated with a difference in 30-day mortality (eTable 3 and eTable 4 in the Supplement).

Discussion

Among Medicare beneficiaries hospitalized at acute care hospitals and treated by a general internist, no significant differences in 30-day mortality rates were observed among patients treated by locum tenens vs non鈥搇ocum tenens physicians. Overall, patients treated by locum tenens physicians had slightly higher Medicare Part B charges, longer lengths of stay, and lower 30-day readmission rates compared with patients within the same hospital who were treated by non鈥搇ocum tenens physicians. Most patients treated by locum tenens physicians received care at small- and medium-sized hospitals in rural and suburban regions. To our knowledge, this is the first study using national data to characterize locum tenens physicians鈥� patterns, quality, and costs of care.

A locum tenens physician is formally defined as a physician who provides temporary coverage for another physician from a different practice or institution. In the United States, the term locum tenens is often used to refer to all temporary physicians (eg, physicians hired temporarily when clinical volumes increase).³^,22 The definition of a locum tenens by CMS is consistent with the term鈥檚 formal definition. Accordingly, CMS restricts billing under the locum Q6 modifier to care episodes that meet this definition.¹⁵^,23 Thus, while Q6 is likely to be highly specific for identifying traditional locum tenens physicians, it will not capture other forms of services delivered by temporary physicians who do not meet the locum tenens definition of CMS.

While the CMS definition of locum tenens may not be perfectly sensitive for all temporary physician services, these findings may still have implications for care quality and spending given the number of patients treated by locum tenens physicians. In this study, 2.1% of Medicare inpatients treated by general internists received care from a locum tenens physician, as defined by CMS. Moreover, 9.3% of physicians were temporarily covered by a locum tenens physician during the study period. Furthermore, it is unclear why temporary physicians who do not meet the traditional locum tenens definition would have different mortality, readmissions, or spending patterns than traditional locum tenens physicians.

The lack of a significant overall difference in mortality rates between patients treated by locum tenens and non鈥搇ocum tenens physicians is reassuring, and it argues against the presence of systematic differences in the quality of care administered by these 2 groups of physicians. The analysis of year-by-year mortality among patients treated by locum tenens vs non鈥搇ocum tenens physicians did not reveal any statistically significant differences in mortality rates between these 2 patient populations. However, in 2009-2010, mortality rates among patients treated by locum tenens physicians were nonsignificantly higher than those treated by non鈥搇ocum tenens physicians. Given the study鈥檚 modest sample size, a small statistically significant difference in mortality during these earlier years cannot be ruled out. Even so, the lack of suggestion of a mortality difference during 2011-2014 indicates that any small differences in care quality between patients treated by locum tenens and non鈥搇ocum tenens physicians, and any systematic problems in the use of locum tenens physicians that could have predisposed to this difference in outcomes, in the early period of this study (2009-2010) has resolved. It is also possible that the small nonsignificant mortality differences observed in 2009-2010 reflect confounding due to unmeasured variables associated with the use of locum tenens physicians.

In subgroup analyses, treatment by a locum tenens physician was associated with significantly higher mortality among patients admitted to hospitals that used locum tenens physicians infrequently. It is plausible that hospitals that use locum tenens physicians infrequently may lack adequate support systems to help locum tenens physicians transition into temporary clinical positions and minimize any associated quality lapses, or hire less-skilled locum tenens physicians. However, this subgroup result should be considered to be hypothesis-generating given the retrospective nature of this study鈥檚 design, the lack of a difference in the primary mortality outcome in the overall sample, the small sample sizes in the low-intensity tercile subgroup, and the lack of evidence of a difference in mortality in the low-intensity tercile in the most recent years 2013-2014.

Several factors may predispose to higher spending and longer length of stay among patients treated by locum tenens physicians. Physicians may have little prior work experience at the institutions that hire them for locum tenens positions. A physician鈥檚 clinical performance may be affected not only by individual skill, overall experience, and organizational performance, but also by institution-specific experience.²⁴ Several facets of clinical care, including care protocols, resource availability, and clinical norms, vary substantially across health systems.⁸^,10^,25^-27 Therefore, temporary physicians, including locum tenens physicians, may initially struggle to efficiently and effectively deliver care or coordinate care transitions.²⁸ In a survey of 259 managers of health care facilities that use locum tenens physicians, among the most commonly cited drawbacks to hiring temporary physician staff were lack of familiarity with the department or practice (49%) and locum tenens physicians鈥� need to learn to use new equipment (32%).¹ The longer lengths of stay and higher Medicare Part B charges observed in this study among patients treated by locum tenens physicians are potentially consistent with this explanation. In addition, many hospitals deliver inpatient care in multidisciplinary teams.⁹^,29^-31 While high-functioning teams can improve clinical outcomes and reduce spending, they take time to develop,³² and team members鈥� familiarity with one another is an important determinant of team performance.⁹^,33

Concerns have been raised abroad that locum tenens physicians may have different clinical competencies than the physicians whom they replace.³ In the United Kingdom in the early 1990s, concerns about locum tenens physicians鈥� clinical competency resulted in the creation of a Working Group on Locum Physicians and the development of competency standards and evaluation guidelines for locum tenens physicians.²² In the United States, many locum tenens positions are filled through national physician staffing agencies, which vet candidates before placing them.³⁴ Although this study found no evidence of a statistically significant overall difference in mortality rates of patients treated by locum tenens physicians, the United States lacks overarching competency and quality standards for these physicians. Furthermore, because locum tenens physicians bill for clinical services under the NPI of the physician whom they replace, no mechanism currently exists for using administrative data to serially evaluate individual locum tenens physicians鈥� quality and costs of care.

This study also found that receiving treatment from a locum tenens physician was associated with lower 30-day readmission rates. The discordance between this finding and the lack of association between treatment by a locum tenens physician and 30-day mortality is consistent with previous work demonstrating that, across hospitals, differences in 30-day readmission rates do not correlate well with differences in 30-day mortality.³⁵ Moreover, the only previous study to evaluate physician-level 30-day mortality rates and 30-day readmissions rates found no relationship between these 2 outcomes at the physician level.³⁶ Furthermore, this study鈥檚 findings also suggest that, if locum tenens physicians deliver more expensive inpatient care than non鈥搇ocum tenens physicians, this higher spending may be offset by lower spending on 30-day readmissions.

Limitations

This study has several limitations. First, the results should be interpreted cautiously given the limitations inherent in retrospective study designs. Prospective comparisons of outcomes among patients treated by locum tenens and non鈥搇ocum tenens physicians are necessary to corroborate this study鈥檚 findings. Moreover, additional work is needed to understand the characteristics of locum tenens physicians; hospital-specific and experiential determinants of the quality of care that they deliver; and the quality and costs of outpatient, pediatric, and surgical services delivered by locum tenens physicians.⁴^,37 CMS and private insurers could facilitate evaluations of physician-level variations in quality and spending by requiring that locum tenens physicians submit identifying information with the billing claim. Second, it is possible that unmeasured patient confounders that correlate with locum tenens care were not accounted for. Efforts to account for unmeasured confounding included analyzing outcomes of locum tenens and non鈥搇ocum tenens physicians within the same hospital, demonstrating similarities of patient characteristics between the 2 groups of physicians, and conducting sensitivity analyses. Third, no information on locum tenens physicians鈥� characteristics, including age, training, and board certification, was available. Fourth, information was lacking on how hospitals attempt to orient locum tenens physicians to, and support, their clinical positions, which may influence care quality. Fifth, this study evaluated inpatient care delivered by general internists. The study鈥檚 findings may not generalize to other specialties or clinical settings. Sixth, this study did not evaluate care delivered by non鈥搇ocums tenens temporary physician staff. Nonetheless, factors that could account for differences in outcomes between patients treated by locum tenens and non鈥搇ocum tenens physicians鈥攊ncluding challenges delivering care effectively in new health systems鈥攎ay also deleteriously affect care quality among other temporary physicians. If non鈥搇ocums tenens temporary physicians have similar outcomes to locum tenens physicians evaluated in this study, reclassifying them as locum tenens physicians could result in a statistically significant overall difference in mortality between locum tenens and non鈥搇ocum tenens physicians. However, the small nonsignificant difference in mortality between patients treated by locum tenens and non鈥搇ocum tenens physicians observed in this study, and the narrow confidence intervals around this estimate, suggest that if a true difference in mortality exists, it is likely to be very small in absolute terms. Seventh, because Medicare data do not include specific identifiers for locum tenens physicians, our study could not determine the number of physicians who provide locum tenens care to Medicare beneficiaries.

Conclusions

Among hospitalized Medicare beneficiaries treated by a general internist, there were no significant differences in overall 30-day mortality rates among patients treated by locum tenens compared with non鈥搇ocum tenens physicians. Additional research may help determine hospital-level factors associated with the quality and costs of care related to locum tenens physicians.

Back to top

Article Information

Corresponding Author: Anupam B. Jena, MD, PhD, Department of Health Care Policy, Harvard Medical School, 180 Longwood Ave, Boston, MA 02115 (jena@hcp.med.harvard.edu).

Accepted for Publication: November 1, 2017.

Author Contributions: Drs Blumenthal and Jena had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Blumenthal, Olenski, Jena.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: All authors.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: All authors.

Obtained funding: Jena.

Administrative, technical, or material support: All authors.

Supervision: Blumenthal, Jena.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Blumenthal reported receiving consulting fees unrelated to this work from Precision Health Economics and Novartis Pharmaceuticals. Dr Jena reported receiving consulting fees unrelated to this work from Pfizer, Hill Rom Services Inc, Bristol Myers Squibb, Novartis Pharmaceuticals, Vertex Pharmaceuticals, and Precision Health Economics. No other disclosures were reported.

Funding/Support: This study was supported by grant 1DP5OD017897 from the Office of the Director, National Institutes of Health (Dr Jena, NIH Early Independence Award). Dr Blumenthal is supported by a grant from the John S. LaDue Memorial Fellowship at Harvard Medical School.

Role of the Funder/Sponsor: The study sponsor had no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

References

1.

Staff Care. 听2015 Survey of Temporary Physician Staffing Trends Based on 2014 Data. . Published 2015. Accessed November 3, 2017.

2.

Zimlich 听R. 听The rise of locum tenens among primary care physicians.听听Med Econ. 2014;91(7):58.

3.

Simon 听AB, Alonzo 听AA. 听The demography, career pattern, and motivation of locum tenens physicians in the United States.听听J Healthc Manag. 2004;49(6):363-375.

4.

Staff Care. 听Nurse Practitioners and Physician Assistants: Supply, Distribution, and Scope of Practice Considerations. . Published 2015. Accessed November 3, 2017.

5.

Poole 听SR, Efird 听D, Wera 听T, Fox-Gliessman 听D, Hill 听K. 听Pediatric locum tenens provided by an academic center.听听笔别诲颈补迟谤颈肠蝉. 1996;98(3, pt 1):403-409.

6.

Doty 听B, Andres 听M, Zuckerman 听R, Borgstrom 听D. 听Use of locum tenens surgeons to provide surgical care in small rural hospitals.听听World J Surg. 2009;33(2):228-232.

7.

Sim 听AJ. 听Locum tenens consultant doctors in a rural general hospital: an essential part of the medical workforce or an expensive stopgap?听听Rural Remote Health. 2011;11(4):1594.

8.

Speroff 听T, Nwosu 听S, Greevy 听R, 听et al. 听Organisational culture: variation across hospitals and connection to patient safety climate.听听Qual Saf Health Care. 2010;19(6):592-596.

9.

Xu 听R, Carty 听MJ, Orgill 听DP, Lipsitz 听SR, Duclos 听A. 听The teaming curve: a longitudinal study of the influence of surgical team familiarity on operative time.听听Ann Surg. 2013;258(6):953-957.

10.

Campbell 听EG, Singer 听S, Kitch 听BT, Iezzoni 听LI, Meyer 听GS. 听Patient safety climate in hospitals: act locally on variation across units.听听Jt Comm J Qual Patient Saf. 2010;36(7):319-326.

11.

Health Forum. AHA Annual Survey Database. . Accessed November 3, 2017.

12.

Tsugawa 听Y, Jena 听AB, Figueroa 听JF, Orav 听EJ, Blumenthal 听DM, Jha 听AK. 听Comparison of hospital mortality and readmission rates for Medicare patients treated by male vs female physicians.听听闯础惭础 Intern Med. 2017;177(2):206-213.

13.

Jena 听AB, Khullar 听D, Ho 听O, Olenski 听AR, Blumenthal 听DM. 听Sex differences in academic rank in us medical schools in 2014.听听闯础惭础. 2015;314(11):1149-1158.

14.

Jena 听AB, Olenski 听AR, Blumenthal 听DM. 听Sex differences in physician salary in us public medical schools.听听闯础惭础 Intern Med. 2016;176(9):1294-1304.

15.

Center for Medicare & Medicaid Services. 2016 Alpha-Numeric HCPCS File. . Accessed November 3, 2017.

16.

Elixhauser 听A, Steiner 听C, Harris 听DR, Coffey 听RM. 听Comorbidity measures for use with administrative data.听听Med Care. 1998;36(1):8-27.

17.

Popescu 听I, Vaughan-Sarrazin 听MS, Rosenthal 听GE. 听Differences in mortality and use of revascularization in black and white patients with acute MI admitted to hospitals with and without revascularization services.听听闯础惭础. 2007;297(22):2489-2495.

18.

Filice 听CE, Joynt 听KE. 听Examining race and ethnicity information in Medicare administrative data.听听Med Care. 2016.

19.

Wennberg 听JE, Cooper 听MM, Birkmeyer 听JD, 听et al. 听The Dartmouth Atlas of Healthcare, 1999. Hanover, NH: Health Forum Inc; 1999.

20.

White 听H. 听A heteroskedasticity-consistent covariance matrix and a direct test for heteroskedasticity.听听贰肠辞苍辞尘别迟谤颈肠补. 1980;48(4):817-838. doi:

21.

Williams 听R. 听Using the margins command to estimate and interpret adjusted predictions and marginal effects.听听Stata J. 2012;12(2):308-331.

22.

NHS Employers. Guidance on the appointment and employment of NHS locum doctors. . Accessed November 3, 2017

23.

Center for Medicare & Medicaid Services. Medicare claims processing manual. . Accessed November 3, 2017.

24.

Huckman 听RS, Pisano 听GP. 听The firm specificity of individual performance: evidence from cardiac surgery.听听Manage Sci. 2006;52(4):473-488. doi:

25.

Bradley 听EH, Herrin 听J, Curry 听L, 听et al. 听Variation in hospital mortality rates for patients with acute myocardial infarction.听听Am J Cardiol. 2010;106(8):1108-1112.

26.

Shah 听B, Hernandez 听AF, Liang 听L, 听et al; Get With The Guidelines Steering Committee. 听Hospital variation and characteristics of implantable cardioverter-defibrillator use in patients with heart failure: data from the GWTG-HF (Get With The Guidelines-Heart Failure) registry.听听J Am Coll Cardiol. 2009;53(5):416-422.

27.

Patel 听MR, Chen 听AY, Roe 听MT, 听et al. 听A comparison of acute coronary syndrome care at academic and nonacademic hospitals.听听Am J Med. 2007;120(1):40-46.

28.

Jennison 听T. 听Locum doctors: patient safety is more important than the cost.听听Int J Surg. 2013;11(10):1141-1142.

29.

Dietz 听AS, Pronovost 听PJ, Mendez-Tellez 听PA, 听et al. 听A systematic review of teamwork in the intensive care unit: what do we know about teamwork, team tasks, and improvement strategies?听听J Crit Care. 2014;29(6):908-914.

30.

Mills 听P, Neily 听J, Dunn 听E. 听Teamwork and communication in surgical teams: implications for patient safety.听听J Am Coll Surg. 2008;206(1):107-112.

31.

O鈥橪eary 听KJ, Sehgal 听NL, Terrell 听G, Williams 听MV; High Performance Teams and the Hospital of the Future Project Team. 听Interdisciplinary teamwork in hospitals: a review and practical recommendations for improvement.听听J Hosp Med. 2012;7(1):48-54.

32.

Hackman 听JR. 听Leading Teams: Setting the Stage for Great Performances. Boston, MA: Harvard Business Press; 2002.

33.

Huckman 听RS, Staats 听BR, Upton 听DM. 听Team familiarity, role experience, and performance: evidence from Indian software services.听听Manage Sci. 2008;55(1):85-100.

34.

Lowes 听R. 听Locum tenens: when you need one, how to get one.听听Med Econ. 2007;84(9):38, 40, 42 passim.

35.

Krumholz 听HM, Lin 听Z, Keenan 听PS, 听et al. 听Relationship between hospital readmission and mortality rates for patients hospitalized with acute myocardial infarction, heart failure, or pneumonia.听听闯础惭础. 2013;309(6):587-593.

36.

Tsugawa 听Y, Jha 听AK, Newhouse 听JP, Zaslavsky 听AM, Jena 听AB. 听Variation in physician spending and association with patient outcomes.听听闯础惭础 Intern Med. 2017;177(5):675-682.

37.

Staff Care. Survey of nurse practitioners: practice trends and perspectives. . Accessed November 3, 2017.

糖心vlog