PEDIATRICS Vol. 118 No. 4 October 2006, pp. 1327-1331 (doi:10.1542/peds.2005-3146)
ARTICLE |
Comparison of Outcome Measures for a Traditional Pediatric Faculty Service and Nonfaculty Hospitalist Services in a Community Teaching Hospital
Division of Pediatrics, St Joseph's Hospital and Medical Center, Phoenix, Arizona
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
OBJECTIVE. Several studies have found decreased cost and length of stay for patients who are cared for by pediatric hospitalists compared with traditional faculty models. The objective of this study was to compare cost and length of stay between a faculty group and 2 separate hospitalist groups in a community teaching hospital. This study differs from previous ones in that both the traditional faculty and hospitalist models were in place simultaneously, and the traditional faculty group was employed by the hospital, whereas the hospitalist groups were in private practice.
METHODS. A total of 1009 pediatric patients with any of the 11 most frequent diagnosis-related groups were analyzed according to the admitting physician group. Total direct costs and length of stay were computed for 3 separate groups (faculty group, hospitalist group 1, and hospitalist group 2). Linear regression models were used to compare total direct costs and length of stay among the groups. Each model accounted for age, severity index, and payer source.
RESULTS. Age, severity index, and physician group were predictive in determining total direct costs and length of stay. There was no significant difference in patient age among the groups, but the faculty group had significantly increased severity indices compared with hospitalist groups 1 and 2 (1.6 ± 0.7 vs 1.3 ± 0.6 vs 1.4 ± 0.6, mean ± SD). The faculty group had statistically significantly lower total direct costs compared with hospitalist groups 1 and 2 ($1781 ± $1449 vs $1954 ± $1212 vs $1964 ± $1495, mean ± SD). The faculty group had shorter average length of stay compared with hospitalist groups 1 and 2 (2.6 ± 2.0 vs 3.1 ± 2.6 vs 2.9 ± 2.3, mean ± SD). The readmission rates among the groups were similar.
CONCLUSIONS. Traditional faculty models can be as efficient in terms of total direct costs and length of stay as evolving hospitalist models. This study's results may be unique because the traditional faculty model was composed of general pediatricians instead of a blend of generalists and subspecialists. In addition, the traditional faculty physicians concentrated almost entirely on the care of inpatients while engaged in hospital care.
Key Words: hospitalists system • length of stay • cost analysis
Abbreviations: LOS—length of stay • TDC—total direct costs • SJHMC—St Joseph's Hospital and Medical Center • DRG—diagnosis-related group • CRS—Chronic Rehabilitative Services
The term "hospitalist" rapidly became a part of the medical vernacular after Wachter and Goldman first described hospitalists in 1996.1 They postulated that hospitalists could meet managed care organizations' increased demand for efficiency by creating a group of physicians with increased availability and expertise to respond to the needs of hospitalized patients. Five years after their original report, Wachter and Goldman performed a systematic review of the literature to determine the effect of the adult hospitalist model on resource utilization, length of stay (LOS), quality of care, satisfaction, and teaching. Almost all of the reviewed studies found that the hospitalist programs significantly reduced hospital costs and LOS without impairing quality of care or patient satisfaction.2
The word "hospitalist" has many different meanings, depending on the model in place. Most chairs of pediatric departments defined hospitalists as physicians who spend a minimum of 25% to 50% of their time caring for inpatients.3 The Society of Hospital Medicine identified hospitalists as "physicians whose primary professional focus is the general medical care of hospitalized patients. Their activities include patient care, teaching, research, and leadership related to hospital care."4 Conversely, for the purposes of this article, a traditional model is a system in which physicians rotate coverage of hospitalized patients while continuing with other types of patient care responsibilities.
Fewer studies have compared the efficiency of pediatric hospitalists with that of traditional models, especially at a community teaching hospital. The majority of studies demonstrated improved cost and/or LOS for pediatric hospitalists.5–9 Ogershok et al5 compared a restructured pediatric ward service, which was organized as a hospitalist model, with a traditional ward service. They found that the restructured pediatric ward service had significantly lower costs and resource utilization. There was no difference in LOS between the groups. A limitation of this study was that the groups were compared during different years. Therefore, background changes in hospital practices could have confounded the results.
Landrigan et al6 resolved the problem of comparing different time periods by controlling for baseline rates of change with quarterly time-series analysis. The authors found that the hospitalists had significantly decreased LOS and hospital charges compared with nonhospitalist physicians. There was no difference in mortality, readmission, or postdischarge follow-up rates.
Bellet and Whitaker8 compared cost, LOS, and readmission rates for a hospitalist and a traditional model at a large academic children's hospital. The hospitalist model had significantly decreased cost and LOS but significantly increased readmission rates. The added cost for the readmitted patients was not accounted for in the cost comparison between the groups. The groups were compared during separate time periods because the hospitalist model had replaced the traditional model.
This current study differs from previous studies primarily in regard to the characteristics of the traditional and hospitalist physicians as described in Methods. Unlike previous studies, both models were in place simultaneously, eliminating the potential bias of comparing groups during different time periods. Despite the momentum of hospitalist models, some have concerns regarding the ability of hospitalists to balance the possibility of increased efficiency with the disruption in continuity of care. Because there is great diversity in hospitalist and traditional models, the authors of this study believed that it was important to compare total direct costs (TDC) and LOS for the groups within their own institution.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
St Joseph's Hospital and Medical Center (SJHMC) is a not-for-profit, 535-bed tertiary teaching hospital in Phoenix, Arizona. The pediatric ward has 64 licensed beds. Each year,
3500 patients are admitted to the pediatric ward. Three study groups were compared: 2 hospitalist groups and 1 traditional faculty group. The hospitalist physicians were in private practice and had arrangements with community pediatricians to care for their patients who were admitted to the hospital. Hospitalists devoted themselves entirely to the care of patients in the hospital, including both a pediatric inpatient and a newborn nursery service. Hospitalist group 1 had 4 pediatricians, and hospitalist group 2 had 5 pediatricians. Both hospitalist groups admitted to several hospitals in the Phoenix metropolitan area. At any 1 time, 1 to 2 pediatricians from each of the hospitalist groups devoted themselves to care for patients at SJHMC. There were 8 pediatricians in the traditional faculty group, and they were employed by SJHMC. They cared for both inpatients and outpatients. They rotated coverage of their inpatient service at SJHMC bimonthly. Patients who were admitted to the faculty group were from SJHMC's pediatric residency teaching practice, those without an identified primary care physician, and those with a primary care physician who was not covered by the hospitalists' arrangements. While caring for inpatients, the faculty staffed 1 half-day continuity clinic per week and continued ongoing teaching and administrative duties. They did not have additional outpatient or nursery responsibilities.
Residents were involved in the care of patients from all study groups. A team that was composed of 1 or more attending physicians, residents, interns, and often medical students cared for patients in all groups. This study was approved by the institutional review board at SJHMC.
Data Collection
Information regarding patients was obtained from SJMHC’s McKesson TRENDSTAR Clinical Cost Accounting System. Because this is a retrospective study, physicians in the study groups were blinded when the data were generated. Patients were included in the study when they were admitted to physicians in the study groups from January 1, 2003, through December 31, 2003, with the 11 most frequent diagnosis-related groups (DRGs) as defined by the 3M All Patient Refined DRG system.10 A total of 1132 patients met these criteria. Patients who were admitted to the ICU, surgical cases, and patients who were admitted to pediatric subspecialty services were excluded because decisions that affect TDC and LOS were not made primarily by the study groups. After these exclusions, 1009 cases remained.
The study groups were compared in terms of their patients' TDC and LOS. Direct cost is the standardized cost that is allocated by the hospital for a specific service. TDC were used because they were costs over which physicians have control, because an order must be written for the services to be rendered. LOS was measured as calendar days with each day accruing when the patient was hospitalized at 11:59 PM. Patient age, payer source, and severity index were compared to determine whether the populations of the study groups were similar. Age was calculated in months. There were 3 payer group categories: Medicaid, managed care/private insurance, and other (Chronic Rehabilitative Services [CRS], Champus, self-pay). CRS is a program that is administered by SJHMC and funded by the State of Arizona's Department of Health Services to provide specialized medical care to children who have chronic and disabling or potentially disabling conditions. The faculty group exclusively cared for patients who were covered by CRS. The severity index was assigned by a standardized coding system on the basis of comorbidities. Severity was ranked 1 to 4, where 4 represented the highest level of complexity.10 For comparison of the experience level for the groups, the mean number of years both after residency and at SJHMC was calculated for each of the groups' members.
Analysis
To evaluate TDC, a logarithmic transformation of the data was used to normalize them. A linear regression model was used to determine which input variables had a significant effect on predicting TDC. The distribution for LOS was skewed because the majority of patients were discharged within the first several days. Therefore, a square-root transformation was used to normalize the data for analysis. A linear regression model was used to determine which of the input variables were significant in predicting LOS. The independent input variables in the regression models were patient age, payer group, severity index, and physician group. Analysis of variance was used to compare patient age and the patient severity between the groups. Statistical analysis was performed using SAS statistical software, version 9.1 (SAS Institute, Cary, NC).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In this study, 1009 inpatient pediatric cases were analyzed. The composition of DRGs that created each study group's panel of patients is shown in Table 1. A comparison of patients who were cared for by the study groups is displayed in Table 2. The data demonstrate that there was no difference with respect to patient age among all groups. Patients who were cared for by the faculty group had significantly higher severity indices as compared with hospitalist groups 1 and 2 (P < .0001 and P = .0004, respectively). There was no significant difference in severity index between the hospitalist groups. Table 3 illustrates the percentage of various payer sources for patients who were cared for by each physician group. Payer status did not have a statistically significant effect on TDC or LOS. There was a trend toward greater experience after residency for the faculty group as shown in Table 4. The results were not statistically significant. The mean years working at SJHMC were similar for all groups.
|
|
|
|
Age, severity index, and physician group were statistically significant in determining TDC (P < .0001 for age, P < .0001 for severity scores, P = .004 for physician group). TDC for the faculty group was significantly less than that of hospitalist group 1 and hospitalist group 2 (P = .0005 and P = .0054, respectively) as seen in Table 2. No other statistically significant differences in TDC between the groups were demonstrated in the model.
Age, severity index, and physician group also were significant predictors of LOS. There was a significantly shorter LOS for the faculty group compared with hospitalist groups 1 and 2 (P < .0001 in both cases) as shown in Table 2. No statistically significant differences for LOS were found when the hospitalist groups were compared with each other.
Readmission rates for the 10 days after hospitalization were analyzed. The faculty group had 1 (0.37%) readmission from its 269 cases. Hospitalist group 1 had no readmissions in 220 cases. Hospitalist group 2 had 6 (1.15%) readmissions from 520 cases.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Accounting for patients with higher severity indices, the traditional faculty group had decreased costs and a shorter average LOS compared with the hospitalist groups. The majority of the pediatric literature supports the hypothesis that hospitalists are more efficient than the traditional model in terms of cost and LOS.5–9 One study demonstrated no significant difference between traditional faculty and hospitalists.11 Unlike others previously, the current study shows significantly decreased cost and LOS for the traditional faculty model.
Several hypotheses may be suggested to explain why the present study differs from other reported comparisons. The traditional faculty in fact may have operated like hospitalists. During the traditional faculty members' inpatient rotations, their primary focus was the care of hospitalized patients. They spent the majority of their time caring for inpatients. However, the faculty group spent only 15% to 17% of their time caring for inpatients during the year. The efficiency of the faculty group may be attributed to the fact that they maintained a variety of skill sets but were given sufficient time to focus on inpatient responsibilities when caring for hospitalized patients. In addition, the faculty group had continuity of care with patients who were admitted from the pediatric residency teaching practice, whereas the hospitalists had no direct connection to patients once they were discharged.
In previous studies, the traditional-model physicians often were a blend of general pediatricians and subspecialists. In this study, the faculty group was composed entirely of general pediatricians. General pediatricians may be better suited to manage efficiently a broad range of pediatric problems. In several of the previous studies, the number of physicians in the traditional model often was large, thereby limiting the amount of inpatient experience for each member. With only 8 pediatricians in the faculty group, 6 of whom provided the majority of inpatient care, the faculty group may have had the exposure to inpatient care that was necessary to develop additional expertise.
Another factor that may have influenced the results is that the faculty group may have had a closer working relationship with the residents. The faculty group was available throughout the day to facilitate moment-to-moment patient treatment and discharges. The residents had more scheduled time for discussing patient care with the faculty group. This faculty–resident relationship may have fostered better communication regarding decisions that affect TDC and LOS. The hospitalists covered small newborn services that were not included in the model. During the study period, hospitalist group 1 provided care for 28 newborns, and hospitalist group 2 cared for 234 newborns. On rare occasions, the hospitalists cared for inpatients at >1 hospital. With these additional responsibilities, they may not have been as available to assist residents in management decisions. Therefore, differences in attending supervision of residents' decisions may have affected the TDC and LOS.
Two areas with uncertain impact on the results are the workload of the attending physicians and physician experience. Having a large volume of patients helps physicians to develop efficiency, yet there likely is a point of diminishing returns. Given restrictions of the data, an average daily census for the groups was not computed. Therefore, it is difficult to determine how differences in the groups' workload contributed to their efficiency. The faculty group's increased efficiency may have correlated with their increased experience. The effect of physician experience on TDC and LOS is difficult to determine on the basis of previous literature. Intuitively, physicians become more efficient with experience, and studies suggest that hospitalist programs gain efficiency over time. Conversely, Dwight et al7 demonstrated that the younger physicians in their hospitalist model were more efficient than those in the traditional model. The impact of experience in this study is even more difficult to ascertain because the groups were small and the differences between them did not reach statistical significance.
The limitations of this study were that LOS was measured in days instead of hours in the data set, an average daily census was not measured, and the number of physicians studied was small. Other metrics, such as variation in patient care and patient satisfaction, were not available to the researchers.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
General pediatricians who focus on inpatient care seem to be more efficient in terms of TDC and LOS compared with those who do not. Currently, many hospitals are restructuring their inpatient services with the goal of providing high-quality patient care at the lowest cost. Not surprising, the hospitalist model is becoming increasingly popular because it provides a way for pediatricians to focus on caring for inpatients. This study suggests that there are other options for providing efficient inpatient care.
Many questions are generated as a result of this study. There is uncertainty regarding how the faculty–resident relationship, the experience of physicians, and the volume of patients affect TDC and LOS. Future studies are needed to determine how the structure of inpatient services and the amount of attending supervision affect the quality of resident education. Studies should be designed to determine whether there are differences in patient and physician satisfaction between hospitalist and faculty models.
Although there are many unanswered questions, this study suggests that one does not necessarily need to be a pediatric hospitalist to provide effective and efficient care. A traditional model, in which general pediatricians have dedicated time to care for inpatients, is another option for providing proficient and efficient care.
|
ACKNOWLEDGMENTS |
|---|
This study was developed during the Health Research Design Certificate Program, which was developed for SJHMC by the School of Health Management and Policy at the W. P. Carey School of Business, Arizona State University (Tempe, AZ).
We are grateful for all of the guidance that we received from the instructors involved in this project. We recognize in particular Jeffrey Wilson, PhD, who directed the program, and Mark Reiser, PhD, who assisted with the statistical analysis. We appreciate greatly the support of Celia Barbieri, MS, CPHQ, who extracted the data from the hospital's McKesson TRENDSTAR Clinical Cost Accounting System. We also appreciate review of the manuscript and helpful comments provided by John Bodensteiner, MD.
|
FOOTNOTES |
|---|
Accepted May 24, 2006.
Address correspondence to Kristen Samaddar, MD, St Joseph's Hospital and Medical Center, 124 W Thomas Rd, Phoenix, AZ 85013. E-mail: kristen.samaddar{at}chw.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
- Wachter RM, Goldman L. The emerging role of "hospitalists" in the American health care system.
N Engl J Med. 1996;335
:514
–517
[Free Full Text] - Wachter RM, Goldman L. The hospitalist movement 5 years later.
JAMA. 2002;287
:487
–494
[Abstract/Free Full Text] - Srivastava R, Landrigan C, Gidwani P, Harary OH, Muret-Wagstaff S, Homer CJ. Pediatric hospitalists in Canada and the United States: a survey of pediatric academic department chairs. Ambul Pediatr. 2001;1 :338 –339[CrossRef][Web of Science][Medline]
- Society of Hospital Medicine. Definition of a Hospitalist; 2005. Available at: www.hospitalmedicine.org/Content/NavigationMenu/AboutSHM/DefinitionofaHospitalist/Definition_of_a_Hosp.htm. Accessed March 22, 2005
- Ogershok PR, Xiaoming L, Hugh CP, Renee SM, Martin EW, Norman DR. Restructuring an academic pediatric inpatient service using concepts developed by hospitalists.
Clin Pediatr (Phila). 2001;40
:653
–660
[Abstract/Free Full Text] - Landrigan CP, Srivastava R, Muret-Wagstaff S, et al. Impact of a health maintenance organization system in academic pediatrics.
Pediatrics. 2002;110
:720
–728
[Abstract/Free Full Text] - Dwight P, MacArthur C, Friedman JN, Parkin PC. Evaluation of a staff-only hospitalist system in a tertiary care, academic children's hospital.
Pediatrics. 2004;114
:1545
–1549
[Abstract/Free Full Text] - Bellet PS, Whitaker RC. Evaluation of a pediatric hospitalist service: impact on length of stay and hospital charges.
Pediatrics. 2000;105
:478
–484
[Abstract/Free Full Text] - Rogers JC. Pediatric hospitalist programs offer chance to improve quality and cost. Health Care Strateg Manage. 2003;21 :12 –15[Medline]
- Averill RF, Goldfield N, Steinbeck B, et al. Development of the all patient refined DRGs. 3M HIS Res Rep. 1997;1 –22
- Seid M, Quinn K, Kurtin PS. Hospitalist and community pediatricians: clinical and financial outcomes for asthma and bronchiolitis. J Clin Outcomes Manage. 1997;4 :21 –24
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Bekmezian, P. J. Chung, and S. Yazdani Staff-Only Pediatric Hospitalist Care of Patients With Medically Complex Subspecialty Conditions in a Major Teaching Hospital Arch Pediatr Adolesc Med, October 1, 2008; 162(10): 975 - 980. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Srivastava, C. P. Landrigan, D. Ross-Degnan, S. B. Soumerai, C. J. Homer, D. A. Goldmann, and S. Muret-Wagstaff Impact of a Hospitalist System on Length of Stay and Cost for Children With Common Conditions Pediatrics, August 1, 2007; 120(2): 267 - 274. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. C. Smith Residency Faculty as Hospitalists Pediatrics, February 1, 2007; 119(2): 412 - 412. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
