Effect of a Patient Care Partnership Project on cost and quality of care at an academic teaching hospital *. (For Debate).

ABSTRACT

Background. In 1998, the Patient Care Partnership Project was conducted by general internal medicine physicians and hospital administration in an academic health care center. The project was designed to optimize cost, quality, and service results to inpatients.

Methods. The project focused on improved communication among physicians, a nurse discharge planner, and hospital administration regarding appropriate resource utilization. The outcomes were average cost per inpatient, length of hospital stay, 30-day readmission rates, mortality rates, and resident and patient satisfaction. Comparisons were made with three control groups.

Results. The postintervention generalist-staffed services showed significant reductions in average costs per patient and length of stay. These parameters increased in the specialist group from 1997 to 1998. Readmission rates remained stable, and mortality rates actually decreased. Patient and resident satisfaction remained unchanged.

Conclusions. A collaborative effort between generalists and hospital administration led to a significant improvement in resource utilization compared with the three control groups, with no compromise in quality outcomes.

**********

IN 1995, nearly a trillion dollars was spent on health care in the United States.' This spending represented 14% of the gross domestic product, more than any other country in the world. (2,3) Despite these enormous expenditures, statistics of medical outcomes suggest that Americans are getting less than optimal return on their investment: there are between 39 million and 52 million uninsured persons in the United States, (4,5) life expectancy lags behind that of other industrialized countries, and infant mortality is only 19th best in the world. (2)

These public health trends have spurred debate for health care reform and encouraged the development of market driven health care organizations. Disease management programs are a relatively new concept designed to improve resource utilization. Disease management programs emphasize coordinated, comprehensive, evidence-based patient care along the disease continuum and across health care delivery systems. (6) Although such programs have shown positive effects in the areas of congestive heart failure, (7) asthma, (8,9) and hip surgery, (10) results from programs involving general medical care have been less encouraging. (8) Hospital charges, which represent the single largest component of health care cost, (1) may make a logical target for implementing a disease management program by generalists.

Ruby Memorial Hospital is an academic health care center affiliated with West Virginia University Hospitals. It was one of the highest cost hospitals in its market region because of its high volume of activity, training of residents and students, and availability of high technology testing. Through cost analyses, Ruby Memorial Hospital administrators were able to identify opportunities to reduce cost and improve quality via standardization and practice guidelines. In an effort to link hospital and provider outcomes and improve overall cost efficiency, the Patient Care Partnership Project (PCPP) was established in 1998 between the hospital administration and the faculty of the Section of General Internal Medicine. The PCPP was a multidisciplinary team designed to improve efficiency, utilization, and outcomes. The goal was to optimize costs and service without compromising quality of inpatient care. This report describes the effect of this case management program on the cost and quality of inpatient care relat ive to 1 concurrent control group and 2 retrospective control groups.

MATERIALS AND METHODS

Site and Subjects

Ruby Memorial Hospital at West Virginia University is a tertiary care referral center located in north central West Virginia. The inpatient general medicine service consisted of three teams, each admitting every third day. Each ward team was composed of an attending physician, a resident, 2 to 3 interns or sub-interns, 1 to 2 medical students, and a pharmacist. The ratio of generalists to specialists on the general medicine ward services was 1:1. Attending physicians staffed the wards 1 to 2 months per year. In 1998, 15 generalists attended on wards for 17 months, and 17 specialists attended for 19 months (for a total of 36 months on the 3 teams). Attending physicians of the same practice type followed each other. The 1997 generalist comparison group was comprised of the same physicians as the 1998 study group, with the addition of 3 new faculty members in 1998. Patients were admitted to these services on the basis of resident call schedule, not attending specialty. Although the specialist attending physicia ns did not participate in the project, there was ample opportunity for exchange of ideas regarding the project among services. All patients admitted until 8 PM were evaluated that day by the attending physician. Patients admitted after 8 PM were seen the following day by the attending physician on call. Separate cardiology and hematology-oncology services were not included in the analysis.

Intervention

The PCPP was developed in 1997 and implemented January 1, 1998. The goal of the PCPP was to align the interests of the hospital and physician providers via a multidisciplinary team to optimize cost, quality, and service to hospitalized patients. The PCPP was jointly developed by the hospital administration (vice president, administrative coordinators, decision support personnel, case managers), and the general internal medicine (GIM) faculty. Physician participation in the project was voluntary. The key interventions of the project were as follows:

First, the GIM faculty were encouraged to develop clinical and operational process modification to reduce cost and maintain quality and satisfaction. Before implementing the project, a series of meetings between hospital administration and the physicians were conducted to establish means of accomplishing the interventions. The major goal was to ensure appropriate resource utilization, but specific attention was given to the 10 most frequent diagnosis related groups (DRGs) (11):

DRG No.  Category

  79     Respiratory infection
  88     Chronic obstructive pulmonary disease
  89     Simple pneumonia
  96     Bronchitis and asthma
 127     Heart failure
 130     Peripheral vascular disorder
 174     Gastrointestinal hemorrhage
 277     Cellulitis
 320     Kidney and urinary tract infection
 449     Poisoning and toxic ingestion

Each physician agreed to contribute a minimum of 8 hours per month to the project to develop ways to be more cost effective (ie, developing practice guidelines for common DRGs, assessing costs and utility of tests related to specific DRGs, and emphasizing cost containment and discharge planning to the residents).

Second, monthly feedback sessions between GIM faculty and hospital administrators were conducted to give the physicians direct feedback regarding project effects (eg, monthly costs and length of stay comparisons with the control groups). Hospital administrators presented data on issues such as barriers to hospital discharge, practice guidelines, and new areas of focus.

Third, residents were educated regarding cost-effective medicine. Topics such as evidence-based medicine, practice guidelines, and resource utilization were emphasized to the house staff via a variety of venues such as morning reports, monthly reminder sessions, and daily teaching rounds. House staff were encouraged to explain the need for each test before ordering it. Tests that were considered unnecessary by the attending physician were not ordered.

Fourth, a nurse discharge planner was added to the GIM staffed services. She was a registered nurse trained in case management who made rounds daily with the GIM staffed ward teams during the project. In addition to aiding in discharge planning, she also helped educate the house staff regarding the cost of tests and negotiated with insurance companies for optimal case reimbursement.

Measurements

Outcome measurements included average cost per inpatient stay, average charge of hospitalization, length of hospital stay, and specific diagnostic test ordering (laboratory tests and radiology tests). Inpatient costs reflected any expense that the hospital had to pay while the patient was in the hospital, including resident and support staff salaries, but not staff physician fees. Charges reflected the bill generated by the hospital and submitted to the insurance carrier or patient. The cost and charge data were obtained via Clinical Cost Accounting Systems for Trendstar. Costs, charges, and length of stay and diagnostic test ordering for 1998 GIM faculty were compared with those of three comparison groups: GIM faculty in 1997 and specialist faculty in 1997 and 1998. Comparisons were made between the four groups to determine project effects overall and for the top 10 DRGs. Overall costs, charges, and length of stay were controlled for patients' age, sex, and payer status and case mix index. Case mix index, wh ich is the weighted DRG average for each service, was determined as a measurement of disease severity. (11) Average costs of each specific DRG were also analyzed.

All patients admitted to West Virginia University Hospital internal medicine services during this period were included in the analysis with two exceptions: Medicare length of stay outliers (11) and patients who had "major" surgical procedures while in the hospital (ie, patients who went to the operating room for any procedure other than a central line placement, tracheostomy, or wound debridement). Determinations regarding exclusion criteria were made a priori by hospital administration. These exceptions were applied to reduce variation caused by high-cost patients and to better reflect issues controllable by medicine attending physicians. Physicians were blinded to the specifics of these exclusionary criteria during the study.

Quality measurements included 30-day readmission rates, inpatient mortality, patient satisfaction, and resident satisfaction. Patient surveys were mailed to the patients after discharge. Resident surveys were sent to the residents after the rotation. Patient surveys focused on 9 questions related to physician availability, communication, and care coordination. Resident surveys focused on 8 questions related to cost-effective care, education, and overall experience. Surveys were scored on 5-point Likert anchors of scale. Patient and resident satisfaction surveys were compared between the 1998 GIM service and 1997 GIM service. West Virginia University Institutional Review Board approval for this project was obtained.

Data Analysis

Data from the PCPP were analyzed by the decision support and statistical support teams with JMP SAS software (version 4.04, SAS Institute, Cary, NC). We used ANOVA to compare continuous outcome variables such as cost, charges, length of stay, and laboratory tests. General linear regression models with continuous and discrete predictors were developed for multiple regression analysis. Multiple regression analyses allowed us to control for the effect of case mix index, age, sex, and payer status on outcomes. We made direct comparisons between the groups with the Fisher Least Significant Difference test. We also evaluated for interactions between groups and project years with the tests for interaction. Dichotomous outcome variables (ie, payer status, readmission rates, and mortality) were evaluated with chi-square. Cell chi-square was used to estimate differences between payer groups. Satisfaction surveys (resident, student, patient) were scored on a quasi-continuous 5-point Likert scale and analyzed with Studen t's t test.

RESULTS

The baseline characteristics of each of the 4 groups are shown in Table 1. There were no significant differences between patients admitted to the services in terms of age (P = .4) or sex (P = .6). However, with regard to payer status 1997 and 1998, specialist staffed services tended to have fewer managed care patients than the generalist groups (P < .0001). Also, the 1997 specialist group had a higher case mix index than the 1998 generalist group (P = .03) . The proportion of patients excluded was significantly higher in the generalist groups during both years (P < .001), but the difference in exclusion rates was significantly greater in 1997 than in 1998 (P = .001).

The average charges, costs, and length of stay for the 4 services are shown in Table 2. For overall comparisons, general internists had significantly lower costs, charges, and lengths of stay between 1998 and 1997 as compared with the specialty attending physicians. Compared directly with 1997 generalists, 1998 generalist attending physicians were able to reduce average cost per patient by $419 (P = .11), average charge by $1,785 (P = .04), and length of stay by one half day (P = .03). These parameters increased in the specialist attended services from 1997 to 1998. Similar trends were seen regarding average costs and length of stay for the top 10 most common DRGs.

Figures 1 and 2 show the consistency of the GIM physicians compared with specialists in terms of overall average monthly cost and length of stay per patient in 1998. In 11 of the 12 months studied, GIM attending physicians had the lowest average cost per patient. For all 12 months, the length of stay was shortest on the GIM services.

These trends toward reduction of cost and length of stay were also seen in the top 10 most common DRGs, which represented 31% of the patients. The values for the individual top 10 DRGs are shown in Table 3. Between 1997 and 1998, generalists were able to show cost reduction trends relative to specialists in 9 of the 10 DRGs. The exception was DRG 79 (respiratory infection). Between 1997 and 1998, generalists were also able to show a trend to-ward-reduced diagnostic test ordering relative to specialists (Table 4). These differences were most apparent in blood chemistries (P = .048) and computed tomography (P= .04).

The primary quality control measures of this project were in-house mortality, readmission rates, and overall satisfaction. The mortality and readmission rates for the services are shown in Table 5. Despite overall cost and length of stay reductions, the 1998 GIM service had a significantly lower in-house mortality rate (2.7%, relative to the 1997 [4.7%] and 1998 [4.9%] specialist services, P = .03). Readmission rates were lower in 1998 than in 1997 (P = .01) but similar among generalists and specialists each year (P = .7).

Overall patient satisfaction scores were high on the 1998 generalist service (average composite score was 4.31 [+ or -] 0.12 on a 5-point Likert scale), with no significant change from 1997 (average composite score = 4.40 [+ or -] 0.12, P = .59). In 1998, 78% of the patients rated their physician as very good to excellent, compared with 86% in 1997 (P = .16). Scores were also consistently high on the resident surveys (average composite score in 1998 was 3.96 [+ or -] 0.10, compared with 3.75 [+ or -] 0.12 in 1997, P = .21). In 1998, 74% of the residents rated their attending physicians as very good to excellent (compared with 64% in 1997, P = .15). Residents believed that practice changes toward more cost-effective care were more influenced by hospital administration, relative to attending physicians, residents, and third party payers, in 1998 than 1997 (average score 4.5 in 1998 versus 3.5 in 1997, P = .003).

DISCUSSION

The PCPP was developed in 1997 and implemented January 1, 1998. The goal was to align the interests of the hospital and physician providers via a multidisciplinary team. The project focused on optimizing cost, quality, and service to hospitalized patients. The project led to significant reductions in average hospital cost per patient and average length of stay on the generalist staffed services without compromising readmission rates, mortality; or overall satisfaction. Evidence-based medicine, cost-containment strategies, and efficient discharge planning were well received by house staff and attending physicians. Patient and resident satisfaction remained high on generalist staffed services between 1997 and 1998.

Our study has several strengths. First, the initial design was developed with input from hospital administration, participating physicians, and nurses. Physician participation in the project was entirely voluntary. Patients were alternately assigned to the services according to call schedule, which should have avoided sample bias. Comparisons were made both retrospectively and prospectively to determine trends among the comparison groups. Finally, quality control measures were in place throughout the study.

Cost reductions were seen both globally, and in 9 of the 10 most common DRGs. The primary reason for cost reduction appears to be reduced length of stay. The average cost for a nonacute hospital bed without telemetry in our hospital is $326. Substantially higher costs are involved for telemetry beds and stepdown unit beds. The approximate one-half day reduction in average length of stay would account for a substantial portion of the $419 average cost reduction per patient. Some of the cost reductions were also due to decreased diagnostic tests ordered.

These results are similar to those described at the University of California-San Francisco (UCSF). (12) In the UCSF project, the academic medical service was reorganized into managed care service teams focused on cost-effective medicine, and traditional ward service teams. The average hospital cost for patients admitted to the managed care service was nearly $800 less than for patients admitted to the traditional ward service ($7,007 and $7,777 respectively, P = .05). The higher cost per patient in the UCSF study relative to the PCPP is likely explained by the higher case mix index (1.29 [+ or -] 1.2). Average length of hospital stay for the managed care service was 4.3 days, compared with 4.9 days on the traditional service, comparable to our study. Quality of care measurements (ie mortality rates, readmission rates, and overall satisfaction rates) were not compromised.

The PCPP has several limitations that need to be addressed. First, this single site project involved 4 major components: hospital administration--physician meetings, feedback sessions, resident education, and establishment of a nurse discharge planner. Although all interventions appeared to be important, the project was not designed to determine the magnitude of each component. The case manager no doubt played a major role in discharge planning on these services and may have provided an advantage as compared with specialist staffed services. It is difficult to say whether these results are generalizable to other sections or departments. We hope this type of project can be undertaken at other institutions to further define and validate these results.

Second, regarding comparisons between generalists and specialists, the specialist attending services tended to have a case mix that was slightly higher but statistically significant, particularly in 1997. Specialist staffed services also had fewer managed care patients and overall higher mortality rates. These findings suggest that severity of illness on specialist staffed services may have been higher. However, on multiple regression analysis (which controlled for case mix index, payer status, age, and sex), the differences in cost and length of stay persisted. We believe that the differences between the generalists and specialists represent spurious findings rather than a flaw in patient assignment or quality of care.

Finally, any program that is designed to reduce the cost of patient care is controversial in that it may create a conflict of interest for physicians (ie, between patient care quality and third party goals). However, quality control measures were in place throughout this project regarding patient care, resident education, and patient satisfaction. None of these parameters appeared to be compromised.

Academic health care centers have traditionally been recognized for their three administrative lines: teaching, research, and clinical service. Much of the funding for academic centers has been from the public sector, industry, and philanthropy. However, many centers increasingly rely on clinical income to support operating costs. (13) Although the image associated with an academic medical center is considered an asset in the eyes of many managed care markets, there is concern regarding the perceived excessive use of medical resources in these centers. This cost inefficiency could potentially lower overall competitiveness in managed care markets. (14) To remain competitive, academic internal medicine must remain true to the educational and research missions while becoming cost competitive. (15) Multidisciplinary disease management programs have shown promise in areas such as congestive heart failure management (7) and beta-blocker use after myocardial infarction. (16) Shared goals for improvement, administrat ive support, strong physician leadership, and feedback data have been identified as characteristics associated with quality improvement. (16) Physicians can be involved in new areas of interest and research such as quality assurance, utilization review, and practice guideline development with disease management programs.

This project describes a fairly simple collaboration between hospital administration and physicians focused on improving patient care and resource utilization with the use of evidence-based treatment guidelines. The results were encouraging and maintained high quality medical care. We hope to further develop and evaluate this project at West Virginia University hospitals. Similar projects might be considered at other academic institutions.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

TABLE 1

Characteristics of Patients Admitted to the Internal Medicine Services
at West Virginia University Hospital in 1997 and 1998

Characteristic      GIM * 1998          GIM * 1997

Number of patients  1,248               966
Number excluded +   188                 147
Age (years)         57.2 [+ or -] 18.8  58.4 [+ or -] 19.4
Sex (% male)        49.4                47.5
Case mix index      1.01 [+ or -] 0.67  1.03 [+ or -] 0.50
Insurance (%)
  Commercial        17.9                18.0
  Managed care      6.2                 6.1
  Medicare          51.4                54.6
  Private pay       6.7                 7.4
  Medicaid          17.8                14.0

Characteristic      SS * 1998           SS * 1997           P Value

Number of patients  1,014               1,145
Number excluded +   86                  50                  < .01
Age (years)         57.4 [+ or -] 19.4  58.2 [+ or -] 20.0    .4
Sex (% male)        49.1                47.4                  .6
Case mix index      1.07 [+ or -] 0.63  1.09 [+ or -] 0.85    .03 **
Insurance (%)                                                <.0001 ++
  Commercial        15.8                16.9
  Managed care      2.3                 2.2
  Medicare          54.8                54.2
  Private pay       8.7                 9.8
  Medicaid          18.4                17.0

* GIM refers to generalist staffed services; SS refers to the specialist
staffed services.

+ The number of patients excluded was influenced by group and year.

** GIM 1998 had a lower case mix index than SS 1997.

++ SS 1997 and SS 1998 had fewer patients with managed care than GIM
1997 and GIM 1998, based on chi-square.

TABLE 2

Hospital Cost Comparisons Between Generalist (GIM) and Specialist (SS)
Attended Services for 1997 and 1998

Outcome *                 GIM 1998             GIM 1997

Global Cost ($)      4,266 [+ or -] 201   4,685 [+ or -] 222
Global length of       4.8 [+ or -] 0.20    5.3 [+ or -] 0.20
stay (days)
Global charges ($)   5,874 [+ or -] 295   6,659 [+ or -] 326
Top 10 DRG cost ($)  3,654 [+ or -] 380   4,583 [+ or -] 388
Top 10 DRG length      4.3 [+ or -] 0.4     5.4 [+ or -] 0.4
of stay (days)

Outcome *                 SS 1998               SS 1997        P Value +

Global Cost ($)      6,159 [+ or -] 224   5,249 [+ or -] 212     .003
Global length of       6.3 [+ or -] 0.21    5.7 [+ or -] 0.20    .002
stay (days)
Global charges ($)   8,248 [+ or -] 328   7,566 [+ or -] 3.11    .007
Top 10 DRG cost ($)  5,142 [+ or -] 505   4,345 [+ or -] 547     .09
Top 10 DRG length      5.8 [+ or -] 0.5     4.8 [+ or -] 0.5     .09
of stay (days)

* Outcomes controlled for case mix index, age, sex and payer status (+
or -] standard error).

+ Least squares means contrast.

TABLE 3

Hospital Cost Comparisons Between Generalist (GIM) and Specialist (SS)
Attended Services for 1997 and 1998 for the Top 10 DRGs

Specific DRG No. *        GIM 1998               GIM 1997

       79           10,705 [+ or -] 1,819  7,979 [+ or -] 1,604
       88            3,204 [+ or -] 448    3,816 [+ or -] 589
       89            5,334 [+ or -] 549    5,728 [+ or -] 590
       96            2,695 [+ or -] 629    3,548 [+ or -] 693
      127            3,177 [+ or -] 665    4,891 [+ or -] 821
      130            3,613 [+ or -] 1,980  5,062 [+ or -] 1,400
      174            4,119 [+ or -] 610    5,170 [+ or -] 665
      277            3,391 [+ or -] 951    4,814 [+ or -] 971
      320            3,134 [+ or -] 485    3,418 [+ or -] 548
      449            2,163 [+ or -] 906    3,202 [+ or -] 785

Specific DRG No. *         SS 1998               SS 1997

       79           10,836 [+ or -] 1,702  7,793 [+ or -] 1,485
       88            5,517 [+ or -] 518    4,090 [+ or -] 461
       89            6,223 [+ or -] 691    4,927 [+ or -] 605
       96            3,226 [+ or -] 781    3,683 [+ or -] 580
      127            4,220 [+ or -] 812    5,045 [+ or -] 670
      130            6,290 [+ or -] 1,755  5,790 [+ or -] 1,548
      174            5,183 [+ or -] 694    4,723 [+ or -] 739
      277            4,720 [+ or -] 104    3,948 [+ or -] 1,319
      320            5,169 [+ or -] 632    4,386 [+ or -] 522
      449            6,442 [+ or -] 1,110  4,040 [+ or -] 799

Specific DRG No. *  P Value +

       79             .9
       88             .046
       89             .17
       96             .77
      127             .55
      130             .57
      174             .27
      277             .31
      320             .33
      449             .06

* Parameter refers to average cost per diagnosis related group ([+ or -]
standard error).

+ Least squares means contrast.

TABLE 4

Laboratory and Radiologic Test Ordering Comparison Between Generalist
(GIM) and Specialist (SS) Attended Services

     Parameter *          GIM 1998            GIM 1997

Chemistries          20.7 [+ or -] 23.6  22.6 [+ or -] 24.5
Hematology            7.0 [+ or -] 8.7    8.2 [+ or -] 9.5
Radiology tests       1.9 [+ or -] 2.7    2.2 [+ or -] 3.1
Computed tomography   .28 [+ or -] .63    .32 [+ or -] .73

     Parameter *          SS 1998             SS 1997        P Value +

Chemistries          28.6 [+ or -] 31.3  27.2 [+ or -] 31.1    .048
Hematology            8.9 [+ or -] 9.3    9.4 [+ or -] 11.2    .22
Radiology tests       2.5 [+ or -] 3.2    2.5 [+ or -] 3.8     .30
Computed tomography   .38 [+ or -] .80    .34 [+ or -] .71     .04

* Parameter refers to the number of individual tests ordered during the
hospitalization ([+ or -] standard deviation).

+ Least squares means contrast.

TABLE 5

Mortality and Readmission Rate Comparison Between Generalist (GIM) and
Specialist (SS) Attended Services

Outcome *                GIM 1998  GIM 1997  SS 1998  SS 1997  P Value

Mortality rate             2.7       3.8       4.9      4.7     .02 +
30-Day readmission rate    5.5       7.6       5.8      7.9     .7 **

* For outcomes, mortality rate and readmission rate refer to the number
per hundred.

+ SS 1997 and SS 1998 a had higher mortality rate than GIM 1997 and GIM
1998, based on chi-square

** For readmission rates, year (P=.01), not group, was significant.

Acknowledgments. We thank William T. Shockcor, MD, Department of Medicine, and Teresa S. Dunsworth, Pharm D, Department of Clinical Pharmacy, West Virginia University, for editorial support.

References

(1.) Michota F, Lewis T, Gash J: The hospitalist: will inpatient specialists improve care? Cleve Clin J Med 1998;65:297-300

(2.) Wachter RM: Rationing health care: preparing for a new era. South Med J 1995;88:25-31

(3.) Iglehart JK: Germany's health care system. N Engl J Med 1991;324(pt 1):503-508

(4.) Jackson J: The German health care system. Arch Intern Med 1997;157:155-160

(5.) Greenberg W: Competition, Regulation and Rationing in Health Care. Ann Arbor, Mich, Health Administration Press, 1991

(6.) Ellrodt G, Cook DJ, Lee J, et al: Evidence-based disease management. JAMA 1997;278:1687-1692

(7.) Rich MW, Beckham V, Wittenberg C, et al: A multidisciplinary interaction to prevent the readmission of elderly patients with congestive heart failure. N Engl J Med 1995;333:1190-1195

(8.) Ferguson JA, Weinberger M: Case management programs in primary care. J Gen Intern Med 1998;13:123-126

(9.) Mayo PH, Richman J, Harris HW: Results of a program to reduce admissions for adult asthma. Ann Intern Med 1990;112:854-871

(10.) Weingarten S, Riedinger MS, Sandhu M, et al: Can practice guidelines safely reduce hospital length of stay? results from a multicenter interventional study. Am J Med 1998;105:33-40

(11.) DRG Guidebook. Reston, Va, Saint Anthony's Publishing, 15th Ed, 1997

(12.) Wachter RM, Katz PK, Showstack J, et al: Reorganizing an academic medical service: impact on cost, quality, patient satisfaction, and education. JAMA 1998;279:1560-1565

(13.) Allcorn S, Winship DH: Restructuring medical schools to better manage their three missions in the face of financial scarcity. Acad Med l996;71:846-857

(14.) Culbertson RA: How successfully can academic faculty practices compete in developing managed care markets? Acad Med 1996;71:858-870

(15.) Fogelman A, Campbell K, Holmes E, et al: How academic internal medicine can succeed in managed care and capitated environments. Am J Med 1995;98:321-327

(16.) Bradley EH, Holmboe ES, Mattera JA, et al: A qualitative study of increasing beta-blocker use after myocardial infarction. why do some hospitals succees? JAMA 2001;285:2604-2611

RELATED ARTICLE: KEY POINTS

* The Patient Care Partnership Project led to a significant improvement in resource utilization, without compromising patient care.

* The participation of a nurse discharge planner had a favorable effect on the coordination of patient care.

* Evidence-based medicine practice, cost containment strategies and efficient discharge planning were well received by house staff and attending physicians.

* Voluntary projects between physicians and hospital administration, with the focus of improved resource utilization, should be considered at other academic institutions.

* Presented as a poster at the National Society of General Internal Medicine Meeting, San Francisco, Calif, April 1999.

From the Departments of Medicine and Hospital Administration, west Virginia University School of Medicine, Morgantown.

Reprint requests to H. Carlton Palmer, Jr., MD, west Virginia University School of Medicine, Section of General Internal Medicine, Department of Medicine, HSC PO Box 9160, Morgantown, W 26506.