| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Clinical Observations |
1 Department of Radiology, Beth Israel Deaconess Medical Center, 1 Deaconess
Rd., Clinical Center West, 3rd Fl., Boston, MA 02215.
2 Division of General Medicine and Primary Care, Beth Israel Deaconess Medical
Center, Boston, MA 02215.
Received April 18, 2005;
accepted after revision June 24, 2005.
Supported by the Stoneman Center for Quality Improvement in General
Medicine and Primary Care, Beth Israel Deaconess Medical Center, Boston,
MA.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CONCLUSION. We found no decreases in the utilization of radiology resources by hospitalists. Given the increasing volume and importance of radiology in clinical decision making, there will be increased pressure to justify the costs of these studies. We believe that more data should be collected about the use of radiology resources.
Keywords: diagnostic imaging • health care costs • hospitalist • imaging resources • radiology practice • radiology utilization • research
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Wachter and Goldman [4] coined the term "hospitalist" in 1996 to define a physician who spends more than 25% of his or her time in a hospital setting serving as the physician of record for a hospitalized patient after having accepted a "hand-off" from a primary care physician (PCP) and then returning care of that patient to the PCP at the time of patient discharge [4]. There have been numerous articles documenting improved efficiency in the management of inpatients by hospitalists. Hospitalists have been cited as being able to reduce length of stay (LOS), overall cost of care, and cost per admission [5-9].
When looking at reduction in inpatient costs by hospitalists, most of the reduction has been due to the decreased LOS. Although several studies have included various subcategories of resources (laboratory, pharmacy, radiology) in their cost analyses in addition to LOS calculations, to our knowledge, whether hospitalists specifically reduce inpatient radiology utilization independent of any reduction in LOS and whether any reduction in inpatient imaging is simply offset by an increase in hospital readmission have not previously been studied. Given the focus of hospitalists on improving inpatient efficiency, we hypothesized that hospitalists would use fewer radiology resources. We sought to determine whether hospitalists truly order fewer radiologic tests than their nonhospitalist counterparts and whether this purported reduction led to a reduction in overall radiology costs.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We reviewed all discharges for the 1-year period from October 2000 through September 2001. We determined the top-10 diagnosis-related groups (DRGs) on the basis of discharge volume for the hospitalist groups (Table 1). Our decision to use DRGs as a comparison tool was based on the fact that DRGs were easily identifiable for both hospitalists and nonhospitalists. Furthermore, these groups are well understood and categorized and are universally applicable to all providers and all patient populations. We chose the top-10 DRGs based on hospitalist discharge volume because hospitalists were our primary group of interest. Of note, eight of the top-10 hospitalist DRGs were also top discharge DRGs for the nonhospitalist group; only DRGs 320, kidney and urinary tract infection, and 183, gastroenteritis, esophagitis, and miscellaneous digestive disorders without complications, were not in the nonhospitalist top-10 discharge DRGs. Patients admitted to the hospital's transitional care unit, palliative care unit, clinical research center, and psychiatry ward were excluded.
|
We examined inpatient radiology utilization for eight hospitalists and 82 nonhospitalist physicians. All of the hospitalists were general internists. Nonhospitalists were physicians within the department of medicine who volunteered their services as teaching and clinical attending physicians for 1 month per year. Fifty-one of the nonhospitalists were internists, whereas the remaining 31 were subspecialists in the fields of cardiology, infectious disease, hematology and oncology, endocrinology, nephrology, gastroenterology, and reproductive medicine. The average age of the hospitalists was 34 years (range, 28-41 years), and the average age of the nonhospitalists was 45 years (range, 29-76 years). A medical house staff team treats all patients admitted to the medical service; there are no unassigned or attending physician-only patients. Typical hospitalist and nonhospitalist censuses range from eight to 15 patients and two to six patients, respectively. Of note, nonhospitalists admit patients only 3 days per week.
Radiology utilization for hospitalists and nonhospitalists was determined across the top-10 hospitalist DRGs. The techniques of diagnostic imaging examined for each DRG included general radiology, CT, MRI, nuclear medicine, sonography and vascular imaging, special procedures and angiography, and mammography. Each radiologic study was counted as an individual unit. Therefore, one patient may have undergone multiple tests within the same or different imaging categories. The number of tests and the cost per patient were determined within each category and across each DRG. Utilization data were determined by examining the radiology department's radiology information systems (RIS) databases and the hospital's admitting department databases. The RIS database supplied the numbers and types of examinations, and the admitting database supplied additional billing information and other admission-specific data.
The difference between hospitalist and nonhospitalist utilization for each DRG and each technique was then calculated. In addition, outcome-of-care measurements—LOS and 30-day readmission rates—were determined. At our institution, the expectation is that patients will be evaluated by the admitting attending physician within 24 hours of admission. However, examinations are often performed in the emergency department before the admitting attending physician's evaluation of the patient. We attempted to account for these studies by eliminating all studies performed within 24 hours of admission and recalculated our data for this patient subset. From the total groups of 1,173 hospitalist and 743 nonhospitalist patients, this resulted in subgroups of 504 hospitalist and 313 nonhospitalist patients.
Cost data reflected total hospital cost—that is, fixed plus variable costs. These were determined using the hospital's financial application database (Decision Support Manager, Eclipsys). Payor information was determined using the hospital's billing system (Plus 2000 Patient Accounting System, McKesson). LOS and readmission measurements were made using the hospital's comprehensive clinical computing information systems. Readmissions data reflect only readmissions to our medical center and do not account for admissions to other facilities because that information was not available. We used Wilcoxon's rank sum test to compare continuous outcomes measures such as radiology costs and LOS for patients under the care of hospitalists and nonhospitalists. We compared readmission rates using Fisher's exact test. All analyses were conducted with statistical software (SAS [version 8.1], SAS Institute for Windows [Microsoft]).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
The total cost of all radiologic examinations in all DRGs combined was $552,265 for hospitalists and $312,459 for nonhospitalists. These figures represented 11% and 10% of the total hospital costs for hospitalists and nonhospitalists, respectively. The total number of imaging examinations across all DRGs was 1,605 and 1,005 for hospitalists and nonhospitalists, respectively. The per-patient values and differences between hospitalists and nonhospitalists for radiologic cost and number of imaging examinations are shown in Table 3. The cost per patient, stratified per DRG and per imaging technique, for hospitalists and nonhospitalists is shown in Table 4.
|
|
There were no statistical differences between hospitalists and nonhospitalists in the number of tests ordered per DRG or the proportion of tests ordered by technique. In addition, there were no significant differences in cost per patient across any of the 10 DRGs studied (Table 4). The total overall cost of radiologic studies per patient was higher for hospitalists than nonhospitalists ($470.54 vs $420.54, respectively), but this difference was not statistically significant. The DRG with the greatest difference in cost was 182, gastroenteritis, esophagitis, and miscellaneous digestive disorders with complications, for which the hospitalists' cost was $186 more per patient. For treatment of patients in this DRG, hospitalists used a slightly higher percentage of CT examinations (36% vs 29%).
When evaluating the hospitalist and nonhospitalist subgroups in which the examinations performed during the first 24 hours of admission were excluded, there were no statistically significant demographic differences between the groups. The proportion of costs for each DRG as a percentage of the total cost did not differ from those in the larger groups of all hospitalist and nonhospitalist patients. Although the per-patient cost of radiologic studies across all DRGs was again higher for hospitalists ($695) than for nonhospitalists ($574), this difference was not statistically significant (p = 0.12). The bulk of the difference was noted in DRGs 174 (gastrointestinal hemorrhage) and 182 (gastroenteritis, esophagitis, and miscellaneous digestive disorders with complications), for which the hospitalists' costs per patient were $505 and $386 more than those of the nonhospitalists, respectively. Of note, hospitalists' cost per patient was $195 less for DRG 296 (nutritional and miscellaneous metabolic disorders), but this difference was not statistically significant.
The mean LOS for both hospitalist and nonhospitalist patients was 3.2 days. In a subanalysis of LOS for patients stratified by each individual DRG, hospitalist patients had a shorter LOS than nonhospitalist patients for only DRG 088 (chronic obstructive pulmonary disease): 3.0 versus 4.0 days, respectively (p = 0.04). The overall range of LOS was 1-86 days and 1-20 days for hospitalist and nonhospitalist patients, respectively.
Outliers were considered to be patients with an LOS of greater than 10 days, which represented an LOS of 2 SDs from the mean. If outliers in the hospitalist and nonhospitalist groups are excluded, then the mean LOS data for hospitalist and nonhospitalist patients are 2.7 and 3.0 days, respectively. Sixteen percent of hospitalist patients were readmitted within 30 days of discharge, 35% of whom were readmitted for the same diagnosis. Seventeen percent of nonhospitalist patients were readmitted within 30 days of discharge, 30% of whom were readmitted for the same diagnosis. These differences are not statistically significant. However, in the DRG subgroup analysis, hospitalists did have a statistically significant (p = 0.01) higher rate of readmission for patients in DRG 277, cellulitis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Second, there were no significant differences in the overall LOS for any of the DRGs between hospitalists and their nonhospitalist counterparts who were serving as the teaching attending physician. However, the range in LOS was significantly wider for hospitalist patients.
Third, there were no differences between hospitalist and nonhospitalist patients in overall 30-day readmission rates.
Our study has several limitations. The first is that this study was a retrospective observational study conducted at a major academic teaching hospital. Second, all patients were taken care of by the same pool of house staff. Often, examinations are ordered by house staff without the attending physician's knowledge or approval. This fact holds true for both hospitalists and nonhospitalists and therefore may have obscured any true differences between them. Furthermore, hospitalist practice patterns that may have led to an improvement in LOS efficiency may also have led to changes in the practice pattern of the house staff, which would then also impact patients cared for by nonhospitalists. A third limitation is that we were not able to adjust our data to account for case severity differences between the hospitalist and nonhospitalist patients. It is possible that hospitalist patients were, in general, sicker than nonhospitalist patients and that, therefore, the range of LOS was wider, thereby increasing the overall apparent LOS for the hospitalist patients. Other possible reasons for the lack of difference in LOS between the two groups include the organizational scheme providing coverage for the patient. Patients were admitted to the hospitalist service 4 out of 5 weekdays. On weekends, patients on the hospitalist service were managed by nonhospitalists.
Another limitation is that we were unable to separate costs of radiologic tests performed before the patients were evaluated by an attending physician at the time of admission. We attempted to account for the cost of these examinations by removing patients with radiology charges within the first 24 hours of admission. When examining this subgroup of patients, hospitalists again had a higher cost of imaging per patient that was, on average, $121 more than that of nonhospitalists, but this difference was not statistically significant. A potential reason for costs being greater for patients treated by hospitalists than for those treated by nonhospitalists is that the hospitalists are general internists, tend to be younger, and are closer to their medical training than their nonhospitalist counterparts. These factors may have caused the hospitalists to rely more on imaging studies for diagnostic purposes than nonhospitalists.
Over the past two decades, the rate of increase in the yearly expenditure for medical care has been greater than the yearly increase in the GNP [10]. This rise in costs led to the growth of managed care, capitation, and HMOs to achieve the goals of cost control and containment. More recently, medical centers began using hospitalists to improve their efficiency and quality of care and to achieve cost savings for their institutions. The hospitalist model allows physicians to focus on and specialize in the delivery of care in the inpatient setting without the competing pressures of managing an outpatient practice. It is estimated that there are currently approximately 5,000 hospitalists nationwide and that the total projected workforce may range from 10,000 to 30,000 [11]. The potential impact of decreased resource utilization by hospitalists is therefore quite large.
There is a general notion that an increased volume of diagnostic imaging in the inpatient setting has added to the cost of medical care. Although some studies have not found the total number of inpatient studies to have increased [12], most studies have shown a change in the distribution of examinations in that the volume of CT and MRI examinations has increased at a greater rate than other imaging techniques [1, 12-14]. General diagnostic radiologic studies, however, continue to remain the largest-volume category of examination.
Although our data apply specifically to our institution and to our specific model of hospitalist care at the time of the study, we found no decreases in radiology utilization by hospitalists; in fact, we observed a tendency toward greater cost per patient for hospitalists. We believe that as imaging volume increases, there will also be increased pressure on clinicians and radiologists to justify the costs of the imaging procedures. Given the increasing importance of radiology in clinical decision making, we believe that more data should be collected about the utilization of radiology resources and other specific hospital resources.
Acknowledgments
We thank Gail Piatkowski, without whom the data collection and preparation
would not have been possible.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
