CLASSIC LITERATURE
Fiscal Benefits of Hospitalists
Wachter RM, Katz P, Showstack J, et al. Reorganizing an academic medical service: Impact on cost, quality, patient satisfaction, and education. JAMA. 1998;279:1560-1565.
Background: In the 1990s the expansion of managed care insurance programs was placing large financial pressures on academic medical centers. Attempts at managing these pressures had previously focused on using house staff feedback, resource utilization professionals, or creating non-teaching faculty services, each of which has potential negative implications for training programs.
Purpose: To determine if an academic medical service led by faculty members who attended more frequently, became involved in the care of patients earlier, and had an explicit mandate to “increase quality and decrease costs” would lower costs without affecting clinical or educational quality.
Methods: On July 1, 1995, the general medicine service at Moffitt-Long Hospital (San Francisco) was reorganized into two services of two teams each. These services were the managed care service (MCS) and the traditional service (TS). Major differences between the groups included:
- MCS faculty attended more frequently (57% of MCS faculty attended two or more months);
- MCS attendings examined or discussed patients at time of admission;
- MCS physicians became involved in quality improvement activities surrounding inpatient activity; and
- MCS attendings were given an explicit mandate to increase quality and decrease costs.
MCS attendings were chosen by their interest and availability to participate. The intervention was an alternate-day controlled trial. Patients were assigned to the MCS or TS by day of admission. House officers, nurses, and ward of admission were identical for the MCS and TS teams. Clinical outcomes included in-hospital mortality, post-discharge mortality, hospital readmissions within 10 days, and functional status. Resource utilization and cost outcomes included LOS, total hospital costs, and subspecialty consultations.
Results: Fourteen attendings covered 24 months on the MCS and 26 attendings covered 24 months on the TS. There were 1,623 total admissions to the general medicine service from July 1, 1995, to June 30, 1996. Of these, 817 went to the TS and 806 went to the MCS.
For those patients with follow-up data available, there were no differences in clinical outcomes; however, there were significant differences in cost and resource utilization outcomes. The adjusted average LOS on the MCS was 4.3 days and was significantly lower than the 4.9 days on the TS (p=0.01). Additionally, the total hospital costs were lower for the MCS ($7,007/admission) compared to the TS ($7,777/admission, p=0.05). Most of the reduction in hospital costs was accounted for by LOS. There were no significant differences in patient satisfaction or faculty satisfaction with either model.
Discussion: This seminal article describes the effects on costs and resource utilization for a reorganization of an academic general medicine service that would evolve into the hospitalist program. In this study, there were significant cost and resource savings without any effect (positive or negative) on clinical quality. Wachter and colleagues postulated that these cost-saving effects were most likely related to the earlier involvement of attendings in the care of the patients and less likely due to factors such as experience, involvement with quality improvement, or a focus on cost-effectiveness.
There are some important limitations to this study, particularly if interpreted as to the effects of hospitalists. First, although the MCS attendings were on service more than the TS, only 57% did more than one month of service and 21% did three or four months of service. Clearly, most hospitalists currently attend more frequently than even those on the MCS in this study. This factor may have limited the ability to find an effect of experience on outcomes.
Second, although more than 1,600 patients were enrolled, the study was limited in its ability to detect differences in clinical outcomes as evidenced by the wide confidence intervals.
Third, although not chosen for their ability, the MCS attendings were chosen by their interest and availability. Although hospitalists are a self-selected group as well, the effects of this self-selection are not well known.
It is important to revisit this article only seven years after being published. In those years, many studies have supported that reorganizations of medical services similar to that described may in fact save money. There is also some evidence to suggest that there may be some positive effect on clinical outcomes as well. However, there are still many unanswered questions, particularly the mechanism(s) of effects.
Meltzer’s research suggests that experience may be an important factor. Included in this article is a review of the Halasyamani and colleagues study that suggests the structure of a hospitalist service may affect outcomes as well. While the field of hospital medicine continues to grow, ongoing research into the mechanism of the effects, both positive and negative, of hospitalist programs is essential for the field’s long-term success.
—David Meltzer, MD,
associate professor of medicine,
General Internal Medicine,
University of Chicago