Table 4. Annual Adjusted Difference in 30-Day Mortality Among Patients Treated by Locum Tenens and Non–Locum Tenens Physicians, Overall and Stratified by Tercile of Intensity of Locum Tenens Physician Usea.
| Locum Tenens Intensity, by Tercileb | Mortality Difference, % (95% CI) | |||||
|---|---|---|---|---|---|---|
| 2009 | 2010 | 2011 | 2012 | 2013 | 2014 | |
| Overall | 0.6 (−0.5 to 1.7) | 0.8 (−0.2 to 1.8) | −0.4 (−1.0 to 0.3) | 0.4 (−0.2 to 1.1) | −0.2 (−0.9 to 0.4) | 0.1 (−0.6 to 0.8) |
| Lower | 3.6 (−0.1 to 7.4) | 2.7 (−2.0 to 7.3) | 5.1 (1.2 to 8.9) | 7.0 (2.6 to 11.3) | −1.4 (−4.5 to 1.7) | 1.8 (−1.9 to 5.4) |
| Middle | −0.1 (−2.0 to 2.3) | −0.4 (−2.6 to 1.8) | 0.2 (−1.2 to 1.5) | 0.0 (−1.7 to 1.6) | −0.3 (−1.8 to 1.1) | −0.2 (−1.6 to 1.2) |
| Upper | 0.1 (−1.1 to 1.5) | 0.9 (−0.2 to 2.0) | −0.7 (−1.4 to 0.04) | 0.3 (−0.5 to 1.1) | 0.2 (−1.0 to 0.7) | 0.7 (−0.2 to 1.5) |
Table presents the annual adjusted difference in mortality between patients treated by locum tenens vs non–locum tenens physicians in a given year. Positive adjusted differences indicate higher adjusted 30-day mortality among patients treated by locum tenens vs non–locum tenens physicians. All estimates were adjusted for patient age, sex, race/ethnicity, month of year of admission, day of week of admission, Medicaid eligibility, indicators for 31 Elixhauser conditions, and the admitting Major Diagnostic Category, and hospital fixed effects, with robust standard errors clustered at the hospital level. Trend tests were performed to assess for trends in adjusted differences in mortality among all patients and among each locum intensity tercile subgroup. These tests were nonsignificant (using P < .10 as cutoff for statistical significance).
Among hospitals with any locum tenens physician use, we computed the percentage of a hospital’s patients treated by a locum tenens physician, and divided hospitals into terciles along that metric. The lower tercile hospitals involved locum tenens physicians in 0.01% to less than 0.45% of their admissions, the middle tercile in 0.45% to less than 2.5% of admissions, and the upper tercile in at least 2.5% of all admissions.