INTRODUCTION
Both general internists and specialists serve as attending physicians on general internal medicine (GIM) wards in hospitals.1, 2 As attending physicians on GIM wards, specialists often care for patients with conditions outside their domain of expertise, which may lead to differences in efficiency and quality of care compared to general internists.3 This study’s objective was to examine the relationship between receipt of care from general internists versus specialists as attending physicians and outcomes including hospital length of stay (LOS), readmissions, and inpatient mortality.
METHODS
We conducted a retrospective cohort study at Kingston General Hospital, a large Canadian tertiary care hospital. Consecutive adult patients admitted to GIM wards from January 1, 2015, to January 1, 2016, were included. Patient data were obtained from patient electronic medical records and the discharge abstract database.
Attending physician was the main responsible physician for the GIM inpatient at time of discharge or death. A general internist was defined as a faculty member of the GIM division with a practicing license listed as GIM or internal medicine without subspecialty whose main responsibilities included GIM ward service and GIM outpatient clinics. A specialist was defined as a faculty member of an internal medicine subspecialty division with a practicing license listed as subspecialty whose main responsibilities included subspecialty inpatient consult service and subspecialty outpatient clinics.
Patient outcomes included hospital LOS, readmission, and all-cause in-hospital mortality. LOS was compared using the Wilcoxon signed-rank test. Readmission and mortality rates were compared using Fisher’s exact test.
As a secondary analysis, propensity score matching based on patient baseline characteristics was used to improve comparability between patients under general internists and those under specialists. The two propensity-matched groups were compared in terms of LOS, readmission, and in-hospital mortality as described above.
RESULTS
Attending Physicians
From January 1, 2015, to January 1, 2016, 34 staff physicians attended on the 4 GIM teams, including 9 (26%) general internists and 25 (74%) specialists (Table 1).
Table 1.
Characteristics and Outcomes of General Internist and Specialist Attending Physicians
| General internists (N = 9) | Specialists (N = 25) | P-value | |
|---|---|---|---|
| Male, N (%) | 4 (44%) | 17 (68%) | 0.25 |
| Years post graduation, N (%) | 0.06 | ||
| 0–10 years | 6 (67%) | 5 (20%) | |
| 11–20 years | 1 (11%) | 9 (36%) | |
| > 20 years | 2 (22%) | 11 (44%) | |
| Specialty, N (%) | GIM, 9 (100%) | Pulmonary, 6 (24%) | |
| Nephrology, 4 (16%) | |||
| Intensive care, 3 (12%) | |||
| Cardiology, 3 (12%) | |||
| Endocrinology, 3 (12%) | |||
| Rheumatology, 3 (12%) | |||
| Gastroenterology, 2 (8%) | |||
| Allergy, 1 (4%) | |||
| Weeks on GIM ward in study year, median (IQR) | 11 (7–13) | 4 (2–6) | < 0.001 |
| Deaths in hospital | |||
| Rate of deaths/100 patient days | 0.96 | 0.98 | |
| Unadjusted IRR (95% CI) | 0.98 (0.76–1.25) | Reference | 0.87 |
| Adjusted IRR (95% CI)* | 0.87 (0.60–1.24) | Reference | 0.44 |
| Discharges from hospital | |||
| Rate of discharges/100 patient days | 14.22 | 12.42 | |
| Unadjusted IRR (95% CI) | 1.14 (1.07–1.22) | Reference | < 0.001 |
| Adjusted IRR (95% CI)* | 1.11 (1.01–1.22) | Reference | 0.03 |
GIM, general internal medicine; IRR, incidence rate ratio
*Adjusted incidence rate ratio is estimated using a multivariable Poisson regression model that includes physician specialty (general internist versus specialist) along with other physician characteristics including sex, years post graduation, and number of weeks on GIM ward
Patient Outcomes
Of the 3782 consecutive GIM inpatients, 1747 (46%) and 2035 (54%) received care from general internists and specialists respectively (Table 2).
Table 2.
Baseline Characteristics of Patients Under General Internists and Specialists
| All patients | Patients matched by propensity score | |||||
|---|---|---|---|---|---|---|
| General internists (N = 1747) | Specialists (N = 2035) | Std. diff | General internists (N = 1339) | Specialists (N = 1339) | Std. diff | |
| Male | 878 (50%) | 1050 (52%) | 0.03 | 687 (51%) | 682 (51%) | 0.007 |
| Age, median (IQR) | 71 (58–82) | 71 (57–81) | 0.04 | 71 (59–82) | 71 (57–81) | 0.05 |
| Admission on weekend | 438 (25%) | 544 (27%) | 0.04 | 329 (25%) | 344 (26%) | 0.03 |
| Charlson Comorbidity Index | ||||||
| 0 | 480 (27%) | 585 (29%) | 0.03 | 374 (28%) | 404 (30%) | 0.05 |
| 1 | 406 (23%) | 440 (22%) | 0.04 | 312 (23%) | 271 (20%) | 0.07 |
| ≥ 2 | 861 (49%) | 1010 (50%) | 0.007 | 653 (49%) | 664 (50%) | 0.02 |
| Top 7 admitting diagnoses | ||||||
| COPD | 176 (10%) | 190 (9%) | 0.02 | 134 (10%) | 126 (9%) | 0.02 |
| Viral/unspecified pneumonia | 97 (6%) | 133 (7%) | 0.04 | 71 (5%) | 77 (6%) | 0.02 |
| Lower urinary tract infection | 79 (5%) | 91 (4%) | 0.002 | 64 (5%) | 57 (4%) | 0.03 |
| Heart failure without cardiac catheter | 68 (4%) | 75 (4%) | 0.01 | 51 (4%) | 55 (4%) | 0.02 |
| Renal failure | 58 (3%) | 60 (3%) | 0.02 | 43 (3%) | 49 (4%) | 0.02 |
| Gastrointestinal hemorrhage | 51 (3%) | 59 (3%) | 0.001 | 38 (3%) | 39 (3%) | 0.004 |
| Diabetes | 59 (3%) | 46 (2%) | 0.07 | 39 (3%) | 34 (3%) | 0.02 |
COPD, chronic obstructive pulmonary disease; Std. diff, standardized difference
The propensity score matched groups were also matched on admission months, general internal medicine teams, individual Charlson comorbidity score items, and the top 52 most common admitting diagnoses. These variables are not shown in the table
The median LOS was 4.89 (IQR 2.62–9.52) days for patients under general internists and 5.25 (IQR 2.75–10.28) days for those under specialists (p = 0.049). Of the 3516 discharges, 667 (19%) patients were readmitted within 30 days: 326/1634 (20%) discharged by general internists and 341/1882 (18%) discharged by specialists (p = 0.17). During their hospital stay, 266 (7%) patients died: 113/1747 (6%) patients under general internists and 153/2035 (8%) under specialists (p = 0.23).
Secondary Analysis
Using propensity scores, 1339 patients under general internists were matched to 1339 patients under specialists (Table 2). There was excellent balance on a large number of baseline characteristics (standardized differences for these characteristics were all < 10%).
The median LOS was 4.98 (IQR 2.63–9.71) days for patients under general internists and 5.41 (IQR 2.85–10.77) days for those under specialists (p = 0.04). Of the discharges, 259/1250 (21%) patients discharged by general internists and 236/1241 (19%) patients discharged by specialists were readmitted to the hospital within 30 days (p = 0.29). During their hospital stay, 89/1339 (7%) patients under general internists and 98/1339 (7%) under specialists died (p = 0.54).
DISCUSSION
In this study, GIM inpatients under the care of general internists had a modestly shorter LOS relative to those under the care of specialists without significant differences in readmissions or mortality. These results complement and extend findings of an earlier study that showed shorter LOS for GIM inpatients receiving care from general internists as compared to endocrinologists and rheumatologists.2
Study limitations include possible residual confounding from not adjusting for disease severity. As well, our study used data from a single academic acute care hospital, which may limit the generalizability of the results.
General internists may provide more efficient care because of their greater familiarity with inpatient GIM wards and related hospital processes. Note that general internists made up only 26% of the physicians in this study but provided care to 46% of GIM ward patients. Their experience may help expedite work-up, procedures, consultations with other specialty services, and collaboration with the interdisciplinary team for discharge planning, resulting in shorter LOS.
In conclusion, our study findings endorse the value of general internists for hospital GIM, which may not be replaceable by other internal medicine specialists.
Acknowledgments
Parts of this manuscript are based on data and information compiled and provided by the Canadian Institute for Health Information (CIHI). However, the analyses, conclusions, opinions, and statements expressed herein are those of the authors and not necessarily those of CIHI. No endorsement by CIHI is intended or should be inferred. We are indebted to Rod Albrough who extracted the data from the hospital database.
Compliance with Ethical Standards
Conflict of Interest
AD Bai: None to declare.
S Srivastava: None to declare.
CA Smith: CA Smith was one of the general internists included in this study.
SS Gill: None to declare.
References
- 1.Hauer KE, Wachter RM, McCulloch CE, Woo GA, Auerbach AD. Effects of hospitalist attending physicians on trainee satisfaction with teaching and with internal medicine rotations. Arch Intern Med. 2004;164(17):1866–71. doi: 10.1001/archinte.164.17.1866. [DOI] [PubMed] [Google Scholar]
- 2.Parekh V, Saint S, Furney S, Kaufman S, McMahon L. What effect does inpatient physician specialty and experience have on clinical outcomes and resource utilization on a general medical service? J Gen Intern Med. 2004;19(5p1):395–401. doi: 10.1111/j.1525-1497.2004.30298.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Weingarten SR, Lloyd L, Chiou CF, Braunstein GD. Do subspecialists working outside of their specialty provide less efficient and lower-quality care to hospitalized patients than do primary care physicians? Arch Intern Med. 2002;162(5):527–32. doi: 10.1001/archinte.162.5.527. [DOI] [PubMed] [Google Scholar]