Comanagement of Hospitalized Surgical Patients by Medicine Physicians in the United States FREE

Comanagement of surgical patients refers to patient care in which the medicine physician (generalist physician or internal medicine subspecialist) daily assesses acute issues, addresses medical comorbidities, communicates with surgeons, and facilitates patient care transition from the acute care hospital setting.¹ The benefits of comanagement include increased prescribing of evidence-based treatments²; reduced time to surgery³; fewer transfers to an intensive care unit for acute medical deterioration⁴; fewer postoperative complications^4- 6; increased likelihood of discharge to home⁴; reduced length of stay⁷; improved nurse and surgeon satisfaction⁵; and a lower 6-month readmission rate.²

Using a 5% national Medicare sample, we examined the rate of comanagement of surgical patients by generalist physicians or internal medicine subspecialists in US hospitals from 1996 through 2006. We also examined how comanagement by medicine physicians varied by type of surgery as well as by patient and hospital characteristics.

The study cohort consisted of 694 806 hospital admissions in the 5% Medicare sample who underwent inpatient surgery between 1996 and 2006 and were discharged with a surgical diagnosis related group (DRG) code associated with at least 1 of the following procedures: cholecystectomy (DRG 493, 195, 196, 197, and 198); resection for colorectal cancer (DRG 148 and 149); abdominal aortic aneurysm repair (DRG 110); lower-extremity revascularization (DRG 553, 554, and 478); major leg amputation (DRG 113, 213, and 285); coronary artery bypass grafting surgery (DRG 105, 547, 548, 549, and 550); aortic/mitral valve replacement (DRG 104 and 105); lung resection for cancer (DRG 75); radical prostatectomy (DRG 334); transurethral resection of the prostate for benign prostatic hypertrophy (DRG 476 and 306); radical nephrectomy for renal cancer (DRG 303); back surgery (DRG 496, 497, 498, and 499); knee replacement (DRG 544); hip replacement (DRG 544); and repair of hip fracture (DRG 210, 211, and 544). The surgical DRGs selected were those used by the Dartmouth Atlas of Healthcare⁷ for benchmarking US hospitals and associated with a mean length of stay of more than 3 days.

We identified 2 types of comanagement: that by a generalist physician (ie, general internist, geriatrician, family practitioner, or general practitioner) and that by an internal medicine subspecialist. Comanagement was defined by the relevant physician (generalist or internal medicine subspecialist) submitting a claim for evaluation and management services on 70% or more of the days that the patient was hospitalized, including partial days (ie, admission and discharge days). Inpatient physician claims were identified using American Medical Association–Common Procedure Terminology–Evaluation and Management codes 99221 to 99223 (initial hospital visit), 99251 to 99255 (inpatient consultation), and 99231 to 99233 (subsequent hospital visit). We also analyzed the effect of various cut points for the minimum percentage of total hospital days for which a medicine physician provided care. In some analyses, we examined the comanagement of surgical patients by hospitalist physicians, as previously defined.⁸

The proportion of admissions comanaged by any medicine physician was calculated and then stratified by patient and hospital characteristics. Linear trend in the percentage of patients comanaged from 1996 to 2006 was tested using a likelihood ratio test. Two trends were identified: during 1996 to 2000 and during 2001 to 2006. Hierarchical generalized linear models with a logistic link, after adjustment for clustering of admissions (level 1) within hospitals (level 2), were constructed to evaluate comanagement during 2001 to 2006 with any medicine physician or generalist physician. Analyses were performed with SAS version 9.1 (SAS Institute Inc, Cary, North Carolina); GLIMMIX (SAS Institute Inc) was used to conduct multilevel analyses.

Between 1996 and 2006, 35.2% of patients hospitalized for a common surgical procedure were comanaged by medicine physicians: 23.7% by a generalist physician and 14% by an internal medicine subspecialist (2.5% were comanaged by both). Comanagement by any medicine physician for patients hospitalized for a surgical procedure increased from 33.3% in 1996 to 40.8% in 2006 (P < .001). A likelihood ratio test showed 2 distinct time trends (P < .001). The percentage of surgical patients receiving comanagement changed little during the late 1990s and then increased in 2001 (Figure 1). The increase in comanagement was limited to comanagement by generalist physicians. Comanagement by generalist physicians increased from 20.5% in 1996 to 31.3% in 2006 (P < .001). This increase was entirely attributable to an increase in comanagement by generalist physicians who were hospitalists. Comanagement by hospitalists increased from 1.7% of patients in 1996 to 12.5% in 2006.

The percentage of patients comanaged by a medicine physician varied by type of surgery (Figure 2). For example, comanagement by a medicine physician increased from 28.6% in 1996 to 41.7% in 2006 (P < .001) for patients hospitalized for orthopedic surgery but actually decreased from 43.0% in 1996 to 39.9% in 2006 (P < .001) for patients hospitalized for cardiothoracic surgery.

Table 1 shows how comanagement varied by patient and hospital characteristics. Older adults, women, persons with low socioeconomic status, and those with more comorbidities were more likely to receive comanagement. Most comanaged patients were seen by a generalist physician, except for those undergoing cardiothoracic surgery, who were more likely to be comanaged by internal medicine subspecialists (almost entirely cardiologists or pulmonologists). Surgical patients cared for in nonteaching, midsize, and for-profit hospitals were more likely to receive medical comanagement.

After other variables were adjusted for, comanagement by a generalist physician increased at 11.4% per year and overall comanagement by any medicine physician increased 7.8% per year during 2001 to 2006 (Table 2). Advanced age, emergency admissions, and increasing comorbidities were all strong predictors of comanagement. Patients cared for in major teaching hospitals were substantially less likely to receive comanagement. Comanagement varied widely by region, with patients in New England much less likely than others to be comanaged.

In these analyses, we defined comanagement as participation of a medicine physician on 70% or more of total hospital days. Using different cut points (eg, ≥50% or ≥80% of hospital days) changed the estimates of the percentage of patients receiving comanagement. However, the pattern of increase in comanagement over time and the association of comanagement with patient and hospital characteristics did not change appreciably by cut point.

We found a rapid rise in the percentage of hospitalized surgical patients who were comanaged by a medicine physician. The increase, begun in 2001, was caused by more comanagement by generalist physicians who are hospitalists. The percentage of patients who were comanaged by internal medicine subspecialists or nonhospitalist generalist physicians was essentially unchanged from 1996 through 2006. Orthopedic surgery patients experienced the fastest growth in medical comanagement as well as the greatest overall use of comanagement by generalist physicians. Almost all studies of comanagement are in orthopedic patients.^2- 6,9 Indeed, the rapid growth in medical comanagement coincided with the first randomized controlled trials published in 2001, showing benefits of comanagement in orthopedic patients.⁹ Clearly, prospective trials of medical comanagement are needed in other surgical disciplines.

The growth in care of surgical patients by medicine physicians raises the issue of appropriate training. A cross-sectional survey of generalist physicians who devoted 25% or more time to inpatient care revealed that perioperative management was underemphasized during their training.¹⁰ The American Council of Graduate Medical Education currently does not list competencies in perioperative management as a core requirement for internal medicine training.¹¹ The growth in comanagement by medicine physicians in our study was attributed to increased care by hospitalist physicians. Hospitalists are well suited to respond quickly to changes in postoperative patients. A recent survey found that 91% of hospitalists have cared for surgical patients.¹² The Society of Hospital Medicine recognizes perioperative management as a key skill for hospitalists and lists competencies in perioperative medicine as core requirements.¹³

Older adults and persons with comorbidities are more likely to receive comanagement. These patients are at higher risk for complications of surgery and will more likely benefit from comanagement. In a recent study of Medicare beneficiaries, of patients rehospitalized within 30 days of a surgical discharge, 70.5% were rehospitalized for a medical condition.¹⁴ Closer attention to medical comorbidities during the initial hospitalization might be expected to reduce this rate. The increase in comanagement of surgical patients by hospitalists has implications for the number of hospitalists that are needed. If we assume that 100% of Medicare patients who are hospitalized for surgical procedures are to be followed up by hospitalists, and taking into account the current workload of hospitalists, an additional 2500 to 3000 full-time equivalent hospitalists would be required.¹⁵

Our study has several limitations. First, we examined comanagement only in a fee-for-service Medicare population, and our findings may not be generalizable to non-Medicare patients. We studied 15 common inpatient surgical procedures that were performed in this population, and the results may not apply to other types of surgical procedures. These 15 procedures represent 39.1% of all operations performed in this population. Our definition of comanagement—evaluation and management claims submitted on at least 70% of all hospital days by a medicine physician—is arbitrary. Using different cut points changed the proportion of patients comanaged but not the increasing trend. A further limitation is that we did not assess processes or outcomes of care and therefore cannot comment on any benefits of comanagement.

In summary, comanagement of surgical patients by medicine physicians is increasing. To meet this need, training in internal medicine should include medical management of surgical patients. Further prospective trials of comanagement in surgical patients in specialties other than orthopedic surgery are clearly needed.

Correspondence: Gulshan Sharma, MD, MPH, Department of Internal Medicine, 301 University Blvd, JSA-5.112, University of Texas Medical Branch, Galveston, TX 77555-0561 (gulshan.sharma@utmb.edu).

Accepted for Publication: November 15, 2009.

Author Contributions:Study concept and design: Sharma, Kuo, Freeman, and Goodwin. Acquisition of data: Kuo, Freeman, Zhang, and Goodwin. Analysis and interpretation of data: Sharma, Kuo, Zhang, and Goodwin. Drafting of the manuscript: Sharma, Kuo, Freeman, and Goodwin. Critical revision of the manuscript for important intellectual content: Sharma, Kuo, Freeman, Zhang, and Goodwin. Statistical analysis: Sharma, Kuo, and Zhang. Obtained funding: Goodwin. Administrative, technical, and material support: Sharma, Kuo, Zhang, and Goodwin. Study supervision: Sharma, Kuo, Freeman, and Goodwin.

Financial Disclosure: None reported.

Funding/Support: This work was supported by grants R01 AG 033134, K05 CA 134923, K08 AG 031583, and P30 AG 024832 from the National Institutes of Health.

Additional Contributions: Sarah Toombs Smith, PhD, helped in the preparation of the manuscript.