Daly D, Pugh CB, Rangappa S, Cotter J, Smith WR; Academy for Health Services Research and Health Policy. Meeting.
Abstr Acad Health Serv Res Health Policy Meet. 2001; 18: 14.
Division of Quality Health Care, Department of Internal Medicine, Virginia Commonwealth University, Medical Campus, PO Box 980306, Richmond, VA 23298-0306, Phone: (804) 828-6938, Fax: (804) 828-4862, E-mail: ddaly@hsc.vcu.edu
RESEARCH OBJECTIVE: During recent years, a growing number of health systems nationwide have implemented hospitalist programs. There are, however, few reports that describe program evaluation methods or demonstrate quality and efficiency of care. An inner-city academic medical center recently implemented a hospitalist program; measurement of program impact is described below.STUDY DESIGN: An evaluation team (composed of a hospitalist, Department of Internal Medicine administrative staff, and several health services researchers in the Division of Quality Health Care) designed an evaluation plan during implementation of the hospitalist program. A pre-/post-research design comparing matched quarters was used. Indicators from quarters 2-4 of FY99 (before program implementation) were compared to those collected for quarters 2-4 of FY00 (after implementation). A 3-month wash-in period (1st quarter of FY99) was included, to accommodate for the transition from six traditional teaching services (TTS) and one non-teaching service (TNTS) to five traditional and two hospitalist teams. Each hospitalist team used a different model; one was a teaching service (HTS) and the other was not (HNTS). The algorithm to admit patients to adult general medicine services changed during the evaluation period. Since evaluation activities were developed during program implementation, it was not possible to conduct primary data collection before and after it began. Descriptive analyses were conducted on the available demographic characteristics (i.e., race, age, zip code, gender), utilization, and costs. Retrospective analyses were conducted to evaluate program impact using data on hospital admissions and post-discharge emergency department (ED) encounters from the hospital's administrative databases. These data were supplemented with information from other hospital databases containing more details about each encounter. Chi-square and Wilcoxon rank-sum tests were performed using SAS to calculate between- and within-comparisons, as well as to compare indicators for hospitalists vs. non-hospitalists overall.POPULATION STUDIED: All discharges from seven adult general medicine services in an inner-city university teaching hospital.PRINCIPAL FINDINGS: The total cost savings attributed to the hospitalist program was $271,432. HNTS patients were more likely to be African-American, city residents, more likely indigent, and less likely privately insured than FY99 TNTS patients. They also had a longer length of stay. There were no differences between TTS and HTS patient demographics or utilization. Overall, post discharge ED rates were higher in FY00. Hospitalists billed for higher complexity of care than their non-hospitalist colleagues did. Hospitalists were also better at writing discharge orders by 11 a.m., a productivity measure, than non-hospitalists.CONCLUSIONS: The evaluation team was able to accomplish the bulk of its original objectives; results were adequate for hospital administrators to make informed decisions regarding program continuation. It may be too early to tell the true impact of hospitalists, since no specific guideline implementation programs were in effect and efficiencies may still be improving among hospitalists. As we are able to plan for and conduct prospective data collection, a more comprehensive evaluation will be possible.IMPLICATIONS FOR POLICY, DELIVERY, OR PRACTICE: A post hoc evaluation design can be used to generate reliable and functional data to evaluate and make decisions about programs operating in the real world.PRIMARY FUNDING SOURCE: Virginia Commonwealth University Health System
Publication Types:
Keywords:
- Adult
- Case-Control Studies
- Family Practice
- Hospitalists
- Hospitalization
- Hospitals, Teaching
- Hospitals, University
- Humans
- Internal Medicine
- Longitudinal Studies
- Patient Discharge
- Program Evaluation
- Teaching
- Virginia
- methods
- hsrmtgs
Other ID:
UI: 102273069
From Meeting Abstracts