A cost estimation model for measuring professional practice.
In concert with the model development component of the project, a comprehensive program evaluation was designed to test the effect of the professional practice model on nurse retention, patient outcome, and cost. The model was expected to decrease registered nurse (RN) turnover, improve patient outcome, and increase cost effectiveness of care delivery. To test the model's effects, within-unit change was measured over time. These changes were then compared to data collected from comparison units matched for size, patient mix, care delivery system, and nurse attrition rate.
Description of the Study
The professional practice model designed for the study was based on previous research concerning nurse satisfaction and retention. Principal among these were studies by Alexander, Weisman, and Chase (1982), Alexander (1988), Deets and Froebe (1987), Donovan (1980), Froebe, Deets, and Knox (1983), Hinshaw, Smeltzer, and Atwood (1987), McCloskey (1974), McClure, Poulin, Sovie, and Wandelt (1983), Price and Mueller (1981; 1986), Stamps and Piedmonte (1986), and Weisman, Alexander, and Chase (1980; 1981). When findings from these studies were reviewed for common themes, five major concepts emerged. These conceptual elements were hypothesized to individually and collectively contribute to nurse satisfaction and retention and included: staff control over practice, collaborative work relationships, continuity of care, rewards and compensation reflective of education and experience, and continuing education based on unit need.
During the 3-year implementation period of the study, EPPM experimental units used the five conceptual elements to create models that met the needs of the nurses working there. Units were expected to incorporate all five conceptual elements into their model implementation plan, with the understanding that some unit-specific modifications might be needed. These modifications allowed for progress to be made on all units, where previous staff experience in planned change efforts varied significantly. A project director skilled in organizational change and group process assisted with the model implementation on each unit.
The Enhanced Professional Practice Model was introduced into five general medical-surgical units -- two in an urban medical center, two in a medium-sized community hospital, and one in a small rural hospital. Data pertaining to characteristics of the nurse and the environment were collected at baseline and annually for the duration of the 5-year project. Multiple triangulation was used to maximize successful measurement of model effect. Multiple theories, multiple researchers, multiple research methods, and multiple data collection procedures were used. In addition, two advisory committees (one comprising national experts in health care, evaluation, and nursing and one comprising chief executives and chief nurse officers of study hospitals) offered recommendations about data collection procedures.
For variables pertaining to nurse retention and patient outcome, within unit and across unit changes were measured. Original plans to compare experimental and comparison units on cost outcomes were abandoned because of inability to adequately match cost analysis variables. In addition, although all units were categorized as general medical-surgical, patient populations differed significantly, making comparisons of patient mix difficult.
Findings pertaining to the EPPM's effect on nurse retention and patient outcome are reported elsewhere (Ingersoll, Schultz, Hoffart, & Ryan, 1996). This discussion focuses on the cost analysis model and the cost outcomes for the study.
Developing the Cost Analysis Model
An assessment of the cost of the EPPM was considered an essential component of the overall professional practice model evaluation. Without a clear indication of the cost of the model, administrators and policymakers would be unlikely to implement the model regardless of its perceived benefit to patients and staff.
Previous studies of nursing costs have used varied methods for measuring the effect of the intervention. More traditional methods for examining nursing costs have used cost per patient day, or the per diem method. Unfortunately, this approach fails to recognize the unique needs of individual patients who require differing amounts of care during their hospital stay. Moreover, many argue that the cost per patient day allocation factor ignores case mix differences and assumes all patients consume the same amount of nursing resources for each day in the hospital (Caterinicchio & Davies, 1983; Mowry & Korpman, 1985; Wilson, Prescott, & Aleksandrowicz, 1988). It also fails to measure actual resource use. To correct for this problem, cost researchers have recommended including some measure of patient illness such as patient severity or nursing patient classification (Wilson et al., 1988).
Development of the cost model for the EPPM program was complicated by the inclusion of three different types of hospitals -- an academic medical center hospital, a community hospital, and a rural hospital. In addition, each of these hospitals had unique organizational characteristics (see Table 1). Moreover, they employed a variety of financial data reporting and cost accounting systems.
Table 1. Hospital Characteristics Medical Community Community Base Year 1988 Center Experimental Comparison # Beds 722 113 102 # Admissions 27,675 4,307 5,395 Hospital Occupancy Rate 82.5% 78.2% 70% Rural Rural Base Year 1988 Comparison Experimental # Beds 53 49 # Admissions 1,416 1,868 Hospital Occupancy Rate 36% 63%
Although each hospital used a measure of patient illness, the measures were not comparable across hospitals. Consequently, the APACHE II (Knaus, Draper, Wagner, & Zimmerman, 1985) was used as an illness severity indicator for monitoring quality of care data abstracted from a random sample of medical records. This scale, developed originally for use in critical care areas, has been used successfully in other studies of noncritical care patients (Cassard, Weisman, Gordon, & Wong, 1994; Larvin & McMahon, 1989). Data analysis pertaining to illness severity found no differences between experimental and comparison units at any time during the study. However, since these data were not available for all patients in the sample, this indicator was not used for the cost analysis component of the evaluation. Nevertheless, the absence of illness severity findings between experimental and comparison units suggests that if any differences were seen in costs, they would not be the result of disparities between types of patients seen.
Because inter-hospital comparisons of nursing costs differ according to patient illness indicator (Wilson et al., 1988), the lack of consistent patient illness measure in combination with differences in hospital characteristics prevented the use of a nursing cost model that compared costs across hospitals. As a result, a decision was made to monitor change in cost per unit over time.
The model developed for the study examined intra-hospital costs using a time series approach, with costs annualized for each year of the study. The model, depicted in Figure 1, examined total nursing cost as derived from two components -- direct and indirect nursing costs.
Figure 1. Nursing Cost Model
Year 1 Year 4
Total Annual = Annual Direct + Annual Indirect Nursing Costs Nursing Costs Nursing Costs
Direct nursing costs were computed for all direct care providers -- RN, licensed practical nurse (LPN), aide/technician and float/agency/per diem nurse. Salary and benefit costs were summed for all care providers. Indirect nursing costs included all other nursing costs associated with care delivery on the unit (see Table 2). Standardized data collection tools were provided to assist the units in collecting nursing cost data.
Table 2. Components of Nursing Costs Direct Nursing Costs Indirect Nursing Costs Salary & Benefit Costs for: Salary & Benefit Costs for: RN Nurse Manager LPN Unit Secretary Aide/Technician Other Administrative and Educational Costs Float/Agency/Per Diem
The EPPM cost model examined trends in total annual nursing costs by unit across time. Two additional monitoring statistics (see Table 3), average total nursing costs per patient day and average RN costs per patient day, were calculated and trended across 1989 (base year), 1990, 1991, and 1992. Tracking of RN costs was viewed as an important component of the cost analysis since the objectives of the EPPM project included recruiting and retaining RNs. All cost data were examined in current and constant 1989 dollars to control for inflation.
Table 3. Cost Monitoring Statistics Calculated for All Years Total Annual / # Annual Patient = Average Nursing Nursing Costs Days Costs per Patient Day Total RN Costs / # Annual Patient = Average RN Costs Days per Patient Day
A second component of the cost analysis was the collection and monitoring of costs associated with model implementation. These data were collected because of the concern that the benefits of the model would be dismissed if the costs of implementing the model were prohibitive. Although a data collection tool was designed and implemented on each of the units, significant differences in how the tool was interpreted and inconsistencies in data collection practices made these data unacceptable for cost analysis purposes. As a result, the assumption was made that model implementation costs were consistent across experimental units. This assumption was supported by qualitative data describing unit-related activities and the number of individuals involved in the model implementation process.
Findings for hospitals are presented according to hospital type (medical center, community hospital, or rural hospital) and study unit (experimental and comparison). Where data were missing or incomplete, years were excluded from the analysis.
Rural hospitals: Experimental and comparison units. Total nursing costs for the experimental unit increased by 11.2% between 1989 and 1991 and by 36% between 1989 and 1992. Comparison unit total nursing costs increased by an estimated 17.2% between 1989 and 1991. (Data are incomplete for the comparison unit for 1992.) Average total nursing costs per patient day increased from $88.30 to $105.92 for the experimental unit and from $90.59 to $98.88 for the comparison unit. In constant dollars, the experimental unit increased from $88.30 to $93.21, whereas the comparison unit decreased from $90.59 to $89.98.
The experimental unit RN costs per patient day increased from $46.64 in 1989 to $63.05 ($55.48 constant dollars) in 1992. The RN costs per patient day for the comparison unit decreased from $56.45 in 1989 to $55.76 ($50.74 constant dollars) in 1991. The significant increase in nursing costs demonstrated by the rural experimental unit is attributable to several factors. During the course of the study, the experimental unit moved to a new area and added five beds. In response to this move, staffing increased by 3.2 RN (from 16 to 19.2) and 3.2 aide/tech (from 1.6 to 4.8) full-time equivalents (FTEs), while LPN FTEs were reduced by 3.2 (from 17.6 to 14.4). These changes resulted in a 20% increase in RN complement in 1992. In addition, the hospital paid a retroactive pay increase between 1989 and 1990 that increased total nursing costs by 6%. When total nursing costs are adjusted for these factors, the cost increases are comparable between rural hospital experimental and comparison units.
Community hospitals: Experimental and comparison units. Total nursing costs for the community hospital experimental units increased by 11.6% between 1989 and 1992. This increase equates to a 1.8% decrease in constant dollars. The comparison units' total nursing costs remained relatively unchanged and experienced a 12% decrease in constant dollar nursing costs between 1989 and 1992.
Average total nursing costs per patient day for the community hospital experimental units ranged from $106.36 in 1989 to $103.21 ($90.82 constant dollars) in 1992. The comparison units' average total nursing costs per patient day was $89.07 in 1989 and $98.86 in 1992 ($86.64 constant dollars).
RN costs per patient day for the community hospital experimental units ranged from $68.95 in 1989 to $65.49 ($57.63 constant dollars) in 1992. The comparison units' RN costs per patient day were $39.38 in 1989 and $50.09 ($44.08 constant dollars) in 1992.
The community hospital experimental units reported an 11.7% increase in total nursing costs between 1989 and 1992 and a 1% increase in constant dollars. During the same period, the comparison units experienced no growth in total nursing costs in real terms and a 13% decline in constant dollars. These reductions were the result of a major reorganization and downsizing program initiated at the hospital during the course of the study. These activities were external to the EPPM project and cannot be considered indicative of response to model implementation. Overall, the experimental units' total nursing costs indicate no major changes in expenditures for nursing resources during the study.
Urban medical center: Experimental and comparison units. The medical center surgery experimental unit total nursing costs decreased by 6.6% from $832,000 in 1989 to $777,200 in 1992 ($683,928 constant dollars). The surgery comparison unit total nursing costs increased by 37% from $622,000 in 1989 to $852,000 in 1992 ($749,763 constant dollars). The surgery experimental unit average nursing costs per patient day ranged from $114.92 in 1989 to $91.32 in 1992 ($80.36 in constant dollars), while the surgery comparison unit average nursing costs per patient day ranged from $84.09 in 1989 to $101.01 in 1992 ($88.89 constant dollars). The surgery experimental unit's average RN costs per patient day was $99.70 in 1989 and $77.19 in 1992 ($67.93 constant dollars). The surgery comparison unit's average RN costs per patient day increased from $67.84 in 1989 to $85.10 in 1992 ($74.89 constant dollars).
The medicine experimental unit total nursing costs increased by 34.1% from $753,600 in 1989 to $1,010,350 in 1992 ($889,100 constant dollars). The medicine comparison unit total nursing costs increased by 27% from $695,100 in 1989 to $883,760 in 1992 ($777,700 constant dollars). The medicine experimental unit's average nursing costs per patient day increased from $89.76 in 1989 to $116.20 in 1992 ($102.26 constant dollars), while the medicine comparison unit's average nursing costs per patient day increased from $96.63 in 1989 to $100.97 in 1992 ($88.85 constant dollars).
The average RN costs per patient day for the medicine experimental unit increased from $75.41 in 1989 to $88.42 in 1992 ($77.81 constant dollars). The average RN costs per patient day for the medicine comparison unit went from $78.43 in 1989 to $76.06 in 1992 ($66.94 constant dollars). Both the experimental and comparison units at the medical center experienced severe RN vacancies in 1990, which resulted in significant use of agency staff. The RN vacancy problem in medicine was so severe in 1990-1991 that beds were closed on the experimental and comparison units. Also, during this period, the medical center changed its staff mix by decreasing RN positions and increasing aide/technician FTEs. In 1992, the medical units' budgeted positions were increased by an average of 5.5 RN FTEs and 1.5 aide/technician FTEs.
The surgery experimental unit was the only medical center unit that actually decreased total nursing costs in terms of both current and constant dollars. The likely reason for this effect was consistency in nursing leadership. Unlike the medical experimental unit in which several changes in nursing leadership occurred, the same nurse manager oversaw the surgery unit for the duration of the study. The instability on the medical experimental unit resulted in greater use of nursing resources and costs. See Table 4 for a summary of experimental and comparison unit costs over time.
Table 4. Experimental and Comparison Units' Care Delivery Costs Over Time Average Total Costs Constant RN Costs in in Constant Dollars Dollar Change Constant Dollars 1989 1992 1989 1992 RHE $88.30 $93.21 +$4.91 $46.64 $55.48 RHC $90.59 $89.98 -$.61 $56.45 $50.74 CHE $106.36 $90.82 -$15.54 $68.95 $57.63 CHC $89.07 $86.64 -$2.43 $39.38 $44.08 MCEs $114.92 $80.36 $34.56 $99.70 $67.93 MCCs $84.09 $88.89 +$4.80 $67.84 $74.89 MCEm $89.76 $102.26 +$12.50 $75.41 $77.81 MCCm $96.63 $88.85 +$7.78 $78.43 $66.94 Constant Dollar Change RHE +$8.84 RHC -$5.71 CHE -$11.32 CHC +$4.70 MCEs -$31.77 MCCs +$7.05 MCEm +$2.40 MCCm -$11.49
RHE = Rural Hospital Experimental; RHC = Rural Hospital Comparison; CHE = Community Hospital Experimental CHC = Community Hospital Comparison; MCEs = Medical Center Experimental Surgery; MCCs = Medical Center Comparison Surgery; MCEM = Medical Center Experimental Medicine; MCCM = Medical Center Comparison Medical
Nursing, as the largest component of the hospital labor force, accounts for a significant proportion of the personnel budget. Typically, nursing payroll costs represent 50% of the total labor budget (Reitz, 1985) and ranges from 20%-30% of total institutional costs (Wilson et al., 1988). Consequently, nursing costs are frequent targets for cost-efficiency strategies. As a result, a comprehensive cost analysis will be required for any intervention designed to alter the way care is delivered to patients.
In this study, care delivery costs associated with an enhanced professional practice model were comparable to pre-existing models of care delivery. These findings suggest that models designed to increase unit-level decision making and accountability for practice are no more costly than other, more centralized models. Nonetheless, some aspects of this study warrant mention. First, the study was conducted during a period when hospitals were experiencing a significant nursing shortage. By the time the study ended, hospitals were beginning to see improvements in recruitment rates and reductions in vacancies. Consequently, some of the cost-related changes in this study may have been the result of factors outside the project's control. Even so, the fact that the experimental units were no more costly than comparison units suggests that the intervention added, the EPPM, produced no additional variance in changes.
Second, the professional practice model in this study was targeted for RNs and was based on previous research concerning the needs of professional staff. Although data were collected for all unit employees, the focus of the attention was the RN staff. This approach may not be as acceptable in today's hospitals, where increased numbers of nonprofessional staff have been added to unit staffing mix. Despite this study's focus on professional staff, however, the model at all times incorporated a combination of professional and nonprofessional staff. Moreover, the number of nonprofessional staff on all units climbed by the end of the project. Even with the increased numbers of nonprofessional staff and the continued focus on professional staff, differences in cost outcome were not seen. This suggests that efforts to maximize RN staff control over the work setting can be successful even when higher numbers of nonprofessional staff work within the unit. Additional research is needed to determine whether the findings in this study continue when greater numbers of nonprofessional staff are added. Moreover, the targeted focus on professional staff, alone, may be problematic for units where the focus of care delivery is on a team approach rather than on any one specific care provider group.
The failure to use a common indicator of patient mix or intensity also warrants attention when comparing the findings of this study to others. Original plans to use diagnosis-related groups as an indicator of patient type proved impossible because of data reporting and access problems. Consequently, specific information for this component of the cost analysis was absent. Future studies should assure in advance that some method is included for monitoring patient acuity.
In addition, the unavailability of implementation cost data restricts the generalizability of the findings of this investigation. Implementation cost data would have been useful for conducting a comprehensive cost-effectiveness analysis of the model. They also could have been used to further clarify the differences seen for experimental units at the medical center. Future researchers should conduct comprehensive orientation sessions to assure that those responsible for providing data have a common understanding of terms and activities. Data submission should be frequent enough to prevent lengthy periods during which no data are being recorded. The availability of computer entry software would be desirable, with screens and prompts developed to assure comparability of data. These data-recording methods should be supported by close interaction between data collectors and those responsible for overseeing the model implementation activities. Although some efforts were initiated in the EPPM study, these were insufficient to prevent problems with accuracy and consistency of data.
Finally, attention must be paid to the potential for multiple interaction effects in complex organizational studies such as this. This is particularly true when interventions are targeted at the patient care unit level where the individual and combined effects of unit-specific and organizational factors are likely to contribute to both cost and cost savings.
The cost-analysis model for the EPPM project demonstrates the model was useful for measuring effect of a professional practice model on cost of care in three different hospitals. The rural and community hospitals' experimental units performed similarly to their respective comparison units, while the experimental surgery unit proved to be the most cost efficient at the urban medical center.$
Alexander, C.S., Weisman, C.S., & Chase, G.A. (1982). Determinants of staff nurses' perceptions of autonomy within different clinical contexts. Nursing Research, 31, 48-52.
Alexander, J.A. (1988). The effects of patient care unit organization on nursing turnover. Health Care Management Review, 13(2), 61-72.
Cassard, S.D., Weisman, C.S., Gordon, D.L., & Wong, R. (1994). The impact of unit-based self-management by nurses on patient outcomes. Health Services Research, 29, 415-433.
Caterinicchio, R.P., & Davies, R.H. (1983). Developing a client-focused allocation statistic of inpatient nursing resource use: An alternative to the patient day. Social Science Medicine, 17, 259-272.
Deets, C., & Froebe, D.J. (1987). Incentives for nurse employment. Nursing Research, 33, 242-246.
Donovan, L. (1980). What nurses want. (And what they're getting.) RN, 43(4), 22-30.
Froebe, D., Deets, C., & Knox, S. (1983). What motivates nurses to join and remain with an organization? Nursing Leadership, 6, 22-33.
Hinshaw, A.S., Smeltzer, C.H., & Atwood, J.R. (1987). Innovative retention strategies for nursing staff. Journal of Nursing Administration, 17(6), 6-16.
Ingersoll, G.L., Schultz, A.W., Hoffart, N., & Ryan, S.A. (1996). The effect of a professional practice model on staff perception of work groups and nurse leaders. Journal of Nursing Administration, 26(5), 52-60.
Knaus, W.A., Draper, E.A., Wagner, D.R, & Zimmerman, J.E. (1985). APACHE II: A severity of disease classification system. Critical Care Medicine, 13, 818-829.
Larvin, M.P., & McMahon, M.J. (1989). APACHE-II score for assessment and monitoring of acute pancreatitis. The Lancet, 22, 201-205.
McCloskey, J.C. (1974). Influence of rewards and incentives on staff nurse turnover rate. Nursing Research, 23, 239-247.
McClure, M.L., Poulin, M.A., Sovie, M.D., & Wandelt, M.A. (1983). Magnet hospitals. Attraction and retention of professional nurses. Kansas City, MO: American Nurses Association.
Mowry, M., & Korpman, R.A. (1985). Do DRG reimbursement rates reflect nursing costs? The Journal of Nursing Administration, 15(7,8), 2935.
Price, J.L., & Mueller, C.W. (1986). Absenteeism and turnover of hospital employees. JAI Press, Inc.
Price, J.L., & Mueller, C.W. (1981). Professional turnover: The case of nurses. New York: SP Medical & Scientific Books.
Reitz, J.A. (1985). The development of a cost-allocation statistic for nursing. In F.A. Shaffer (Ed.), Costing out nursing: Pricing our product (pp. 123-133). New York: National League for Nursing.
Stamps, RL., & Piedmonte, E.B. (1986). Nurses and work satisfaction. An index for measurement. Ann Arbor, MI: Health Administration Press Perspectives.
Weisman, C.S., Alexander, C.S., & Chase, G.A. (1981). Determinants of staff nurse turnover. Medical Care, 19, 431-443.
Weisman, C.S., Alexander, C.S., & Chase, G.A. (1980). Job satisfaction among hospital nurses: A longitudinal study. Health Services Research, 15, 341-364
Wilson, L., Prescott, P.A., & Aleksandrowicz, L. (1988). Nursing: A major hospital cost component. Health Services Research. 22, 773-795.
PATRICIA A. WITZEL, MS, MBA, RN, is Assistant Director for Applications, Information Systems Division, Strong Memorial Hospital, Rochester, NY.
GAIL L. INGERSOLL, EdD, RN, FAAN, is Associate Dean for Research, Vanderbilt University School of Nursing, Nashville, TN.
ALISON W. SCHULTZ, EdD, RN, is Clinical Associate Professor, University of Rochester School of Nursing, Rochester, NY.
SHEILA A. RYAN, PhD, RN, FAAN, is Dean and Professor, University of Rochester School of Nursing, Rochester, NY.
NOTE: This study was supported by Cooperative Agreement Award #U01 NR02156 from the National Institute of Nursing Research, National Institutes of Health, and the Division of Nursing, Health Resources and Services Administration.
|Printer friendly Cite/link Email Feedback|
|Author:||Witzel, Patricia A.; Ingersoll, Gail L.; Schultz, Alison W.; Ryan, Sheila A.|
|Date:||Sep 1, 1996|
|Previous Article:||Partners in patient care: measuring the effects on patient satisfaction and other quality indicators.|
|Next Article:||Economic awareness of nurses: relationship to budgetary control.|