Nurse staff allocation by nurse patient ratio vs. a computerized nurse dependency management system: a comparative cost analysis of Australian and New Zealand hospitals.
NPRs are known as occupancy-based nurse patient allocation. They generally ensure the number of nurses matches the number of patients requiring care and that nurses in similar settings care for the same number of patients. In the state of Victoria, Australia, NPRs were mandated in 2000 for the public sectors. They are based on a workload complexity model that incorporates the type of shift of care, such as morning or evening, and the level of the hospital, such as major or minor metropolitan and rural. In the United States, they were mandated in California in 2004.
Differences exist between rules of NPRs for Victoria and California (Gerdtz & Nelson, 2007). In comparison to Victoria, California has been the site for many studies concerning NPRs (Aiken et al., 2010; Chart, Killeen, Vilke, Marshall, & Castillo, 2010; Chapman et al., 2009; Donaldson et al., 2005; Sochalski, Konetzka, Zhu, & Volpp, 2008). A literature synthesis on the impact of Californian-mandated acute care hospital nurse staffing ratios found they reduced the number of patients per licensed nurses and increased the number of nursing hours per patient day in hospitals (Donaldson & Shapiro, 2010). A U.S. national survey of registered nurse perceptions of NPRs showed 62% wanted minimum nurse staff ratios mandated; however, RNs employed in California, where ratios are mandated, were significantly more likely to agree (Buerhaus, DesRoches, Donelan, & Hess, 2009).
Computerized nurse dependency workload systems consider patient dependency requirements for nurses by quantifying variables of direct nursing care. Since 1984, when the Australian Council for Health Care Standards introduced the requirement for hospitals to show evidence of a staffing pattern for nurses according to patient need, computerized nurse dependency systems were developed as a solution. They were neither widely adopted nor extensively integrated in hospitals. In the 1990s, some computerized nurse dependency workload systems used in several hospitals were accessed for data to support development of relative nursing costs. Similarly, for hospitals in many other countries, patient classification models based on dependency measures have been generally accepted to represent actual nursing resource requirements for patients.
But even with general consensus on their utility for nurse staff allocation, today's computerized nurse dependency systems for nurse staffing allocation purposes in acute hospitals vary across nations, states, and settings (Aschan, Junttila, Fagerstrom, Kanerva, & Rauhala, 2009; Sawatzky-Dickson & Bodnaryk 2009; Sundaramoorthi, Chen, Rosenberger, Kim, & Buckley-Behan, 2009; Welton, Zone-Smith, & Bandyopadhyay, 2009). In Sweden, the Zebra and Beakta systems are common in hospitals and are used with general satisfaction by nurse managers (Perroca, & Ek, 2007). A model of intensity measurement also underpins a nurse dependency system for one large 1,500 bed hospital in Singapore (Hoi, Ismail, Ong, & Kang, 2010). Systems of computerized nurse staffing allocation based on dependency models are also used outside acute care, such as community (Brady, Byrne, Horan, Macgregor, & Begley, 2008) and rural health care settings (Harper & McCully, 2007). An earlier U.S. study compared differences in nurse staff allocation systems based on patient classification and variances were shown in nursing hours of up to 4.33 hours per day for the same patient (O'BrienPallas, Cockerill, & Leatt, 1992).
Few studies report nursing costs related to dependency models of nurse staff allocation; though costing methodologies are subject to debate (Chiang, 2009; Welton, Fischer, DeGrace, & Zone-Smith, 2006). Costs related to nursing care intensity have been calculated using point scales such as the Finnish RAFAELA system (Fagerstr6m & Rauhala, 2007). Literature suggests empirical measurement of nurse dependency linked to nursing costs remains a formative science (Buerhaus, 2009, 2010; Welton, 2008; Welton, Zone-Smith, & Bandyopadhyay, 2009). It is timely that the need for nursing data has been listed as a key recommendation in the Institute of Medicine report (2010) as comprehensive data will help understand future nursing workforce needs. This study provides one of the first cost modeling and analyses that compares two nurse staff allocation practices. We include a range of scenarios to model selected parameters related to NPRs and a propriety computer nurse dependency management system.
This study used statistical analysis and cost modeling of retrospective data obtained from a cohort of electronic nursing workload management reports from hospital information systems. Data were collected over a 4-month period in 2003 and 2004. Hospitals chosen to supply the workload management reports were users of a propriety-owned Computer Nursing Dependency Management System (CNDMS). At the time of the study more than 100 Australian and 14 New Zealand hospitals were system users. This system was chosen as it had capacity to simultaneously measure nursing workload by a dependency method of nursing hours per patient day (HPPD) by various patient types and categories, as well as by NPR. The authors have no pecuniary interest in the CNDMS mentioned in this study.
Patient dependency is measured in the CNDMS as nursing time required to care for specific patient types and for specific dependency categories (Lowe & Dunigan, 2001). The range of patient types amount to over 40 and include, for example, medical, surgical, maternity, labor ward, special care nursery, coronary care, pediatrics, pediatric intensive care, rehabilitation, day surgery, short stay, psychiatric, emergency department, palliative care, and aged care patients. There are five dependency categories that include every aspect of nursing care for every shift. These aspects of care are referred to as indicators. Examples of indicators include mobility, nutrition, hygiene, and thought processes. In addition, the system accounts for other aspects of care that include, for example, nurses' documentation, doctors' rounds, patient enquiries, simple medications, routine teaching, and counseling. The CNDMS is underpinned by more than a decade of development. The timings for patient types and specific dependency categories have been validated every 2-3 years based on user feedback. Functionally, the system allocates a higher category for complex care, depending on the patient type and category. For example, the indicator "mobility" has less time allocated for care of day surgery patient types than for rehabilitation patient types.
The hospitals selected were users of the propriety CNDMS and included 13 Australian hospitals and 8 from New Zealand. Just over 70% of shifts of care in the sample were provided by Australian hospitals and the remainder were from New Zealand. In the sample were 11 private hospitals and 10 public hospitals with 16 located in metropolitan and 6 in rural areas. The sample size was 103,269 shifts of care. Nurses work three shifts of care on a daily basis. Table 1 shows hospital variables based on country, region status and level, and the frequency and percentage of shifts by hospital variables. In Australia, level denotes the size and complexity of hospital services. In the state of Victoria where nurse patient ratios are mandated and fixed according to the hospital level, each hospital has a designated level. For example, a level 1 hospital is a major metropolitan provider of health services.
The units of measure were predicted nursing hours based on HPPD. HPPD refers to the total number of productive hours worked by nursing staff with direct care responsibilities, per patient per day. HPPD is the sum of the total nursing hours worked (or predicted) for 24 hours, divided by the number of patients occupying beds at midnight, plus other separations during the previous 24 hours (discharges, deaths, transfers). The unit of measure "predicted hours" is a function of the propriety system where the nurse reviews and updates indicators during a shift so that hours predicted for nursing care are automatically adjusted to reflect actual hours currently being worked. A feature of the system is that both "predicted" and "actual" nursing care requirements for acute inpatient care are recorded. Predicted nursing hours arise from the input of nurses as they review and update the CNDMS indicators so that hours predicted for nursing care are automatically adjusted to reflect actual nursing hours currently being worked for the patient category.
Several reports from the propriety system were requested from hospitals to capture the required data. The reports included predicted and actualized hours by category and summaries of the number of patients in each category. Statistical tests included descriptive statistics conducted on the raw data where an analysis of the characteristics of the sample was conducted, and costing analysis where the analysis of the costs for the sample was conducted on a range of variables, by both nurse staff allocation practices. Cost modeling techniques appropriate for different scenarios were used to inform the analysis.
Subsets of data were analyzed on scenarios that closely matched Victorian public hospital profiles to ensure equivalence for patient, ward, and hospital types in the sample. Scenarios were developed by examining frequencies of the shifts by patient type. Medical/ surgical wards were most frequent in level 2 Australian metropolitan public hospitals and formed the most common scenario. In other words, our analysis was undertaken using common scenarios to determine if the cost of care provided by the mandated NPR model was a close match with the cost of nursing care required, how this compared with the cost of care measured by the CDNMS and, importantly, if the costs approximated or were close to the "actual" for the hospital types. Analyses of the scenarios for all patient categories and for the different hospital groups would indicate which method performs best at predicting actual clinical requirements. The mean and standard deviation were calculated in each of these cases. The means are in hours per ward/per shift.
Cost prediction comparison. This part of the analysis gives the Australian dollar values to actual inpatient hours (total hospital direct care nursing costs), propriety system predicted hours, and NPR predicted hours. The analyses are all hospitals (see Table 2); all hospitals, by patient type category (see Table 3); Australian, public, metropolitan, level 2 (see Table 4); and Australian, public, metropolitan, level 1 or level 2, by patient type category (see Table 5).
Patient type categories refer to the following six categories: medical/surgical, pediatric, ante/postnatal, adult critical care, non-adult critical care, and other. Forty-one CDNMS patient types were mapped to these six patient type categories. The term hospital level refers to the manner in which hospitals were categorized at the time of the study and were based on the Victorian Department of Health classification of health services. The three levels represent general types of hospital. For example, level 1 hospitals are large and metropolitan, level 2 include smaller metropolitan and large rural hospitals, and level 3 comprises smaller hospitals in both metropolitan and rural settings. We undertook detailed mapping of non-Victorian hospitals to determine equivalence for these levels.
The dollar value is based on an industry-accepted working measure of $40 per nursing hour in Australia for the year 2003. This amount is derived from all nursing wage levels but excludes executive managers. The amount includes compulsory on-costs of superannuation, shift penalties, annual leave, long service leave, payroll tax, orientation, and safety training. It excludes sick leave and other training. Table 2 shows that, for our sample, the CNDMS system under-predicted the total actual inpatient hours by about as much as NPR seemed to over-predict.
There were also several cases where NPR over-predicted the actual total more than the propriety system, and vice-versa. For example, in Table 3, where there is further breakdown in patient type such as medical/surgical and pediatric, variation is shown within the overall sample and the CNDMS is not always lower for different patient types.
Tables 4 and 5 show the analysis undertaken by selected patient type and hospital variables. For example, Table 4 illustrates selected patient types and hospital variables of designated hospital level.
Additional analyses using scenarios. Data were analyzed further to consider four possible scenarios. The mean and standard deviation were calculated for each scenario. The means are presented in hours per ward, per shift. The scenarios are shifts of care where (a) both CNDMS and NPRs over-predict, (b) both CNDMS and NPRs under-predict, (c) CNDMS over-predicts but NPR is satisfactory, and (d) NPR over-predicts and CNDMS is satisfactory. The scenarios were analyzed on a subset of data which we selected through a rigorous process to closely match shifts of care in Victorian public hospitals. Hence the additional cost analysis was conducted on data from Australian public hospitals, for three hospital levels and two regions, with a focus on medical/surgical, pediatric, and maternity patient types.
The parameters selected are where (a) CNDMS predicted less than actual and NPR predicted less than actual, (b) CNDMS predicted less than actual and NPR predicted at least as much as actual, (c) CNDMS predicted at least as much as actual and NPR predicted less than actual, and (d) CNDMS predicted at least as much as actual and NPR predicted at least as much as actual. In other words, the parameters selected are where (a) both CNDMS and NPR "underbudgeted" nurses for care actually required, (b) CNDMS "under-budgeted" for nurses, but NPR did not, (c) NPR "under-budgeted" for nurses, but CNDMS did not, and (d) neither CNDMS nor NPR "underbudgeted" for nurses. In each of these cases, the focus was on (a) the CNDMS variance (actual--CNDMS predicted); (b) the NPR variance (actual--NPR predicted); and (c) which prediction method was closest to actual. To investigate (a) and (b), we considered four scenarios for all patient categories and for each of the patient type categories and for the different hospital groups. Tables 6 and 7 provide analyses of two scenarios.
For all hospitals and all cases, the scenario when both CNDMS and NPR under-predicted actual hours, the average number of hours per ward and per shift that CNDMS under-predicted was 4.0273 hours (see Table 7). This corresponds to $40 per hour or $161 per ward, per shift, on average for this sample. In comparison, NPR under-predicts by 5.4519 hours, or $218 per ward, per shift, on average.
The hours for tables in the additional analysis are reported in terms of "per cent of actual." This is because for some wards, under-predicting by 4 hours might not be a lot whereas for other wards it might be very significant. It is clear that, on average, for all hospitals and wards, when both CNDMS and NPR under-predict actual need, CNDMS under-predicts by 27.8% of actual, whereas NPR under-predicts by 35% of actual.
To determine part (c), "ABS (NPR variance)--ABS (CNDMS variance)" was used, which is simply the difference in the absolute values of the two different variance measures. In Table 7, where both CNDMS and NPR under-predicted actual, NPR under-predicted by 7.2% of actual more than CNDMS did. This means that if budgets are developed on predicted hours, then hospitals using CNDMS would have paid out more (7.2% more) for these cases than hospitals using NPR. If the NPR variance is larger in absolute value than the CNDMS variance, then this measure will be positive. In all but two cases, this number was positive, meaning CNDMS almost always appears to outperform the NPR method for predicting actual clinical need.
Strengths and limitations. We pursued national representation by conducting the study with support from many CNDMS administrators spread across Australia and New Zealand with variation in hospital categories and location. The study is based on a large sample of shifts of care.
The analysis focused on cost comparison using selected data from a propriety information system. We do not claim the CNDMS-predicted HPPD are a reliable measure of nursing care requirements. Still, the CNDMS remains accepted in practice as it is used in over 100 Australian hospitals for nurse allocation and related patient management. We have presented a study that compares and models cost variations in two primary nurse staff allocation practices. To our knowledge, there is no published literature that has compared variations between two nurse staffing allocation systems endorsed in practice using cost analysis and modeling.
Interpretation of the results. The findings are that the budgeted cost of nursing care is less if hospitals used the CNDMS for allocation of nurse staffing rather than NPR. We have shown that the CNDMS under-predicts the costs of the actual hours of care by as much as NPR over-predicts those costs. This demonstrates that the budgeted cost of nursing care would be less for hospitals using the CNDMS.
The most important interpretation of the cost analysis is the provision of nursing care at lower cost may be possible through use of a computerized nurse dependency system that allows for staff allocation to meet patients' varying requirements for nursing hours and skill mix. This is an important outcome for knowing costs of nursing care and for distribution of limited nursing resources.
Recommendations for practice and further research. Too few studies are available on the topic of acuity-based nurse costing. We recommend further analyses of nurse costing and development of costing methodologies. Our study suggests that in comparison to NPRs, the CNDMS provides managers with continuous patient dependency data for allocating and predicting staff allocation based on nursing care requirements. It helps managers prepare dependency-based budgets, and provides a tool for staff allocation. The CNDMS measures with detail the complexity of patient needs and their dependency on nurses. It records in electronic form the actual care and predicts the care required for individual patients following each shift. The system has capacity to measure trends in patient dependency and nursing resource consumption efficiency because the system quantities, classifies, and codes nursing care and makes, to some extent, the elements of nursing care visible in an information system. We do not seek to promote a particular computerized nurse dependency system; nevertheless, the system used in this study provided a database where we showed differences in cost for two methods of nurse allocation.
The analysis has increased the technological capacity of a health care information system to show differences in nursing costs between two nurse allocation practices. Computerized nurse dependency management systems may trim costs by assisting unit and other managers to allocate nursing staff by hours and skill mix and may improve nurse staff allocation relative to patient dependency requirements. Nursing and quality managers form an integral part of hospital governance and require local and timely data to inform decision making. This study has identified patient-level nursing workforce and staffing allocation data that can provide quality coordinators with a valuable resource to predict nursing staff requirements. Nevertheless, national implications are that costing measures of nurse workload require development and refinement in Australian and New Zealand public and private hospitals. This research makes a significant contribution to the nursing administrative literature. It has applications for government, businesses, health care organizations, professional associations, and the insurance industry. It furthers understanding of ways in which predictive models have national significance for workforce development.
Over recent decades, computerized nurse dependency systems based on patient classification and acuity systems have developed to support nurse management and, in particular, staff allocation. The study sought to determine if it would cost hospitals more to provide nursing care requirements by NPR practices or by a computerized nurse dependency system. The costing analysis showed that NPR allocation practices were less predictable than CNDMS and were a higher cost compared to similar costs for CNDMS.
Australian public and private hospitals require capacity to attribute direct nursing care costs to patient-level data. The states of Victoria, Australia, and California, United States, have set benchmarks for NPRs and have established, to some extent, their favorable impact on nurse practice environments in acute care; however, information system data structures need development to support nurse staff allocation at the operational level. The computerized systems to transform nurse staff allocation could be NPRs or dependency/acuity based but they need sufficient detail at the patient level to enable accurate nurse costing analyses.
ACKNOWLEDGMENTS: The authors thank Emeritus Professor Donna Diers, Dr. Lee Seldon, Ms. Cherrie Lowe, and Mr. Dean Athan for providing advice and guidance for this study; Dr Kathryn Forbes for assisting with data analysis; and the hospital teams who assisted with data transfer. Part of this research was supported by Monash University.
Aiken, L.H., Sloane, D.M., Cimiotti, J.P., Clarke, S.P., Flynn, L., Seago, J.A., ... Smith, H.L. (2010). Implications of the California nurse staffing mandate for other states: Nursing and home care. Health Services Research, 45, 904-921.
Aschan, H., Junttila, K., Fagerstrom, L., Kanerva, A., & Rauhala, A. (2009). RAFAELA[TM] patient classification system as a tool for management. Studies in Health Technology and Informatics, 146, 478-82.
Brady, A.M., Byrne, G., Horan, P., Macgregor, C., & Begley, C. (2008). Reliability and validity of the CCNCS: A dependency workload measurement system. Journal Clinical Nursing, 17, 1351-60.
Buerhaus, P.I. (2009). Avoiding mandatory hospital nurse staffing ratios: An economic commentary. Nursing Outlook, 57, 107-12.
Buerhaus, P.I., DesRoches, C., Donelan, K., & Hess, R. (2009). Registered nurses' perceptions of nurse staffing ratios and new hospital payment regulations. Nursing EconomicS, 27(6), 372-376.
Buerhaus, P.I. (2010). What is the harm in imposing mandatory hospital nurse staffing regulations? Nursing Economic$, 28(2), 87-93.
Chan, T.C., Killeen, J.P., Vilke, G.M., Marshall, J.B., & Castillo, E.M. (2010). Effect of mandated nurse-patient ratios on patient wait time and care time in the emergency department. Academic Emergency Medicine, 17, 545-552.
Chapman S.A., Spetz, J., Seago, J.A., Kaiser, J., Dower, C., & Herrera, C. (2009). How have mandated nurse staffing ratios affected hospitals? Perspectives from California hospital leaders. Journal of Healthcare Management, 54, 321-333.
Chiang, B. (2009). Estimating nursing costs --A methodological review. International Journal Nursing Studies, 46, 716-722.
Donaldson, N., Bolton, L.B., Aydin, C., Brown, D., Elashoff, J.D., & Sandhu, M. (2005). Impact of California's licensed nurse-patient ratios on unit-level nurse staffing and patient outcomes. Policy, Politics & Nursing Practice, 6, 198-210.
Donaldson, N., & Shapiro, S. (2010). Impact of California mandated acute care hospital nurse staffing ratios: A literature synthesis. Policy Politics and Nursing Practice, 11,184-201.
Duffield, C., Roche, M., Diers, D., Catling-Paull, C., & Blay, N. (2010}. Staffing, skill mix and the model of care. Journal of Clinical Nursing, 19(15-16), 2242-2251.
Fagerstrom, L., & Rauhala, A. (2007). Benchmarking in nursing care by the RAFAELA patient classification system --A possibility for nurse managers. Journal Nursing Management, 15, 683-692.
Gerdtz, M., & Nelson, S. (2007). A model of minimum nurse-to-patient ratios in Victoria. Australia. Journal Nursing Management, 15, 64-71.
Harper, K., & McCully, C. (2007). Acuity systems dialogue and patient classification system essentials. Nursing Administration Quarterly, 31, 284299.
Hoi, S.Y., Ismail, N., Ong, L.C., & Kang, J. (2010). Determining nurse staffing needs: The workload intensity measurement system. Journal Nursing Management, 18, 44-53.
Institute of Medicine. (2010). The future of nursing: Leading change, advancing health. Retrieved from http://iom. edu/Reports/2010/The-Future-ofNursing-Leading-Change-AdvancingHealth.aspx
Lowe, C., & Dunigan, R. (2001). Introduction to the TrendCare Management System. Cardindale, Queensland, Australia: TrendCare Pty Ltd.
O'Brien-Pallas, L., Cockerill, R., & Leatt, P. (1992). Different systems--Different costs. Journal of Nursing Administration, 22, 17-22.
Perroca, M.G., & Ek, A. (2007). Utilization of patient classification systems in Swedish hospitals and the degree of satisfaction among nursing staff. Journal Nursing Management, 15, 472-480.
Sawatzky-Dickson, D., & Bodnaryk, K. (2009). Validation of a tool to measure neonatal nursing workload. Journal Nursing Management, 17, 84-91.
Sochalski, J., Konetzka, R.T., Zhu, J., & Volpp, K. (2008). Will mandated minimum nurse staffing ratios lead to better patient outcomes? Medical Care, 46, 606-613.
Sundaramoorthi, D., Chen, V.C.P., Rosenberger, J.M., Kim, S.B., & Buckley-Behan, D.F. (2009). A data-integrated simulation model to evaluate nurse-patient assignments. Health Care Management Science, 12, 252-268.
Twigg, D., Duffield, C., Bremner, A., Rapley, P., & Finn, J. (2011). The impact of the nursing hours per patient day (NHPPD) staffing method on patient outcomes: A retrospective analysis of patient and staffing data. International Journal of Nursing Studies, 5, 540-548.
Welton, J.M. (2008). Mandatory hospital nurse to patient staffing ratios: Time to take a different approach. Online Journal of Issues in Nursing, 12(3).
Welton, J.M., Fischer, M.H., DeGrace, S., & Zone-Smith, L. (2006). Hospital nursing costs, billing, and reimbursement. Nursing Economic$, 24(5), 239-245, 262,227.
Welton, J.M., Zone-Smith, L., & Bandyopadhyay, D. (2009). Estimating nursing intensity and direct cost using the nurse-patient assignment. Journal Nursing Administration, 39, 276-284.
LIZA HESLOR PhD, RN, is Professor, Faculty of Health, Science, and Engineering, Victoria University, Melbourne, Victoria, Australia.
VIRGINIA PLUMMER, PhD, RN, is a Course Coordinator, Monash University, Melbourne, Victoria, Australia.
Table 1. Hospital Variables and Frequency of Shifts by Hospital Variables Variable Frequency Percentage Australia 74,299 71.9 Country New Zealand 28,970 28.1 Total 103,269 100.0 Metropolitan 75,047 72.7 Region Rural 28,222 27.3 Total 103,269 100.0 Public 45,363 43.9 Status Private 57,906 56.1 Total 103,269 100.0 1 22,555 21.8 Level 2 62,060 60.1 3 18,654 18.1 Total 103,269 100.0 Table 2 Cost Analysis: All Hospitals, All Shifts N Sum Mean Actual inpatient hours 103,269 2,051,049.23 19.8612 CNDMS-predicted hours 103,269 2,019,200.30 19.5528 NPR -predicted hours 103,269 2,088,085.03 20.2199 Standard Deviation Total Cost Actual inpatient hours 17.73146 $82,041,969 CNDMS-predicted hours 19.12870 $80,768,012 NPR -predicted hours 19.86600 $83,523,401 Table 3. Cost Analysis: All Hospitals, All Shifts by Patient Type Category Patient Type Category N Actual inpatient hours 63,363 Medical/Surgical CNDMS-predicted hours 63,363 NPR-predicted hours 63,363 Actual inpatient hours 7,209 Pediatric CNDMS-predicted hours 7,209 NPR-predicted hours 7,209 Actual inpatient hours 15,197 Ante/Post Natal CNDMS-predicted hours 15,197 NPR-predicted hours 15,197 Actual inpatient hours 9,319 Adult Critical Care CNDMS-predicted hours 9,319 NPR-predicted hours 9,319 Actual inpatient hours 1,431 Non-Adult Critical Care CNDMS-predicted hours 1,431 NPR-predicted hours 1,431 Actual inpatient hours 6,750 Other CNDMS-predicted hours 6,750 NPR-predicted hours 6,750 Patient Type Category Sum Actual inpatient hours 1,279,878.05 Medical/Surgical CNDMS-predicted hours 1,283,228.98 NPR-predicted hours 1,360,295.40 Actual inpatient hours 118,770.15 Pediatric CNDMS-predicted hours 112,598.27 NPR-predicted hours 95,865.87 Actual inpatient hours 234,189.23 Ante/Post Natal CNDMS-predicted hours 207,556.28 NPR-predicted hours 176,592.10 Actual inpatient hours 223,229.87 Adult Critical Care CNDMS-predicted hours 225,356.83 NPR-predicted hours 263,585.88 Actual inpatient hours 47,746.48 Non-Adult Critical Care CNDMS-predicted hours 47,417.85 NPR-predicted hours 30,112.08 Actual inpatient hours 147,235.45 Other CNDMS-predicted hours 143,042.08 NPR-predicted hours 161,633.70 Patient Type Category Mean Actual inpatient hours 20.1991 Medical/Surgical CNDMS-predicted hours 20.2520 NPR-predicted hours 21.4683 Actual inpatient hours 16.4753 Pediatric CNDMS-predicted hours 15.6191 NPR-predicted hours 13.2981 Actual inpatient hours 15.4102 Ante/Post Natal CNDMS-predicted hours 13.6577 NPR-predicted hours 11.6202 Actual inpatient hours 23.9543 Adult Critical Care CNDMS-predicted hours 24.1825 NPR-predicted hours 28.2848 Actual inpatient hours 33.3658 Non-Adult Critical Care CNDMS-predicted hours 33.1362 NPR-predicted hours 21.0427 Actual inpatient hours 21.8127 Other CNDMS-predicted hours 21.1914 NPR-predicted hours 23.9457 Standard Patient Type Category Deviation Actual inpatient hours 18.45949 Medical/Surgical CNDMS-predicted hours 19.90268 NPR-predicted hours 20.37116 Actual inpatient hours 13.69405 Pediatric CNDMS-predicted hours 14.80118 NPR-predicted hours 12.39210 Actual inpatient hours 13.37345 Ante/Post Natal CNDMS-predicted hours 14.27377 NPR-predicted hours 12.34282 Actual inpatient hours 17.88241 Adult Critical Care CNDMS-predicted hours 19.64746 NPR-predicted hours 24.19768 Actual inpatient hours 28.50000 Non-Adult Critical Care CNDMS-predicted hours 28.21746 NPR-predicted hours 26.08304 Actual inpatient hours 16.47870 Other CNDMS-predicted hours 17.74230 NPR-predicted hours 18.62731 Patient Type Category Total Cost Actual inpatient hours $51,1951,22 Medical/Surgical CNDMS-predicted hours $51,329,159 NPR-predicted hours $54,411,816 Actual inpatient hours $4,750,806 Pediatric CNDMS-predicted hours $4,503,931 NPR-predicted hours $3,834,635 Actual inpatient hours $9,367,569 Ante/Post Natal CNDMS-predicted hours $8,302,251 NPR-predicted hours $7,063,684 Actual inpatient hours $8,929,195 Adult Critical Care CNDMS-predicted hours $9,014,273 NPR-predicted hours $10,543,435 Actual inpatient hours $1,909,859 Non-Adult Critical Care CNDMS-predicted hours $1,896,714 NPR-predicted hours $1,204,483 Actual inpatient hours $5,889,418 Other CNDMS-predicted hours $5,721,683 NPR-predicted hours $6,465,348 Table 4. Cost Analysis: Australian Metropolitan Public Level 2 Hospitals by Patient Category Patient Type Category N Sum Actual inpatient hours 4,023 82,599.97 Medical/Surgical CNDMS-predicted hours 4,023 87,474.97 NPR-predicted hours 4,023 88,337.00 Actual inpatient hours 716 16,235.35 Adult Critical Care CNDMS-predicted hours 716 14,489.62 NPR predicted hours 716 14,497.02 Standard Patient Type Category Mean Deviation Actual inpatient hours 20.5319 16.10417 Medical/Surgical CNDMS-predicted hours 21.7437 17.85130 NPR-predicted hours 21.9580 17.85618 Actual inpatient hours 22.6751 14.58144 Adult Critical Care CNDMS-predicted hours 20.2369 14.50061 NPR predicted hours 20.2472 13.36028 Patient Type Category Total Cost Actual inpatient hours $3,303,999 Medical/Surgical CNDMS-predicted hours $3,498,999 NPR-predicted hours $3,533,480 Actual inpatient hours $649,414 Adult Critical Care CNDMS-predicted hours $579,585 NPR predicted hours $579,881 Table 5. Cost Analysis: Australian Metropolitan Public Level 1 and 2 Hospitals by Patient Type Category Patient Type Category N Sum Actual inpatient hours 4023 82,599.97 Medical/Surgical CNDMS-predicted hours 4023 87,474.97 NPR-predicted hours 4023 88,337.00 Actual inpatient hours 2419 64,215.23 Pediatric CNDMS-predicted hours 2419 65,291.67 NPR-predicted hours 2419 53,571.00 Actual inpatient hours 2518 55,512.13 Ante/Post Natal CNDMS-predicted hours 2518 54,896.57 NPR-predicted hours 2518 42,734.00 Actual inpatient hours 716 16,235.35 Adult Critical Care CNDMS-predicted hours 716 14,489.62 NPR-predicted hours 716 14,497.02 Actual inpatient hours 763 43,486.98 Non-Adult Critical Care CNDMS-predicted hours 763 42,340.25 NPR-predicted hours 763 27,246.00 Standard Patient Type Category Mean Deviation Actual inpatient hours 20.5319 16.10417 Medical/Surgical CNDMS-predicted hours 21.7437 17.85130 NPR-predicted hours 21.9580 17.85618 Actual inpatient hours 26.5462 14.09407 Pediatric CNDMS-predicted hours 26.9912 16.01600 NPR-predicted hours 22.1459 13.25130 Actual inpatient hours 22.0461 21.74175 Ante/Post Natal CNDMS-predicted hours 21.8017 22.75990 NPR-predicted hours 16.9714 18.76285 Actual inpatient hours 22.6751 14.58144 Adult Critical Care CNDMS-predicted hours 20.2369 14.50061 NPR-predicted hours 20.2472 13.36028 Actual inpatient hours 56.9947 17.72631 Non-Adult Critical Care CNDMS-predicted hours 55.4918 20.14025 NPR-predicted hours 35.7090 28.34835 Total Patient Type Category Cost Actual inpatient hours $3,303,999 Medical/Surgical CNDMS-predicted hours $3,498,999 NPR-predicted hours $3,533,480 Actual inpatient hours $2,568,609 Pediatric CNDMS-predicted hours $2,611,667 NPR-predicted hours $2,142,840 Actual inpatient hours $2,220,485 Ante/Post Natal CNDMS-predicted hours $2,195,863 NPR-predicted hours $1,709,360 Actual inpatient hours $649,414 Adult Critical Care CNDMS-predicted hours $579,585 NPR-predicted hours $579,881 Actual inpatient hours $1,739,479 Non-Adult Critical Care CNDMS-predicted hours $1,693,610 NPR-predicted hours $1,089,840 Table 6. Cost Analysis by Scenario (All Cases, All Australian Public Hospitals) Indicator Indicator for CNDMS for NPR Larger Larger than than Actual Actual N Mean CNDMS NPR CNDMS Variance 34,833 4.0273 predicted predicted (Actual--Predicted) less than less than NPR Variance 34,833 5.4519 actual actual (Actual--Predicted) ABS (NPR Variance)-- 34,833 1.4245 ABS (CNDMS Variance) NPR CNDMS Variance 18,525 2.3136 predicted (Actual--Predicted) at least NPR Variance 18,525 -4.2592 as much as (Actual--Predicted) actual ABS (NPR Variance)-- 18,525 1.9456 ABS (CNDMS Variance) CNDMS Variance 16,185 -1.9203 CNDMS NPR (Actual--Predicted) predicted predicted NPR Variance 16,185 3.4942 at least less than (Actual--Predicted) as much actual ABS (NPR Variance)-- 16,185 1.5738 as actual ABS (CNDMS Variance) NPR CNDMS Variance 33,726 -3.5644 predicted (Actual--Predicted) at least NPR Variance 33,726 -6.0663 as much as (Actual--Predicted) actual ABS (NPR Variance)-- 33,726 2.5018 ABS (CNDMS Variance) Indicator Indicator for CNDMS for NPR Larger Larger Percent than than Standard of Actual Actual Deviation Actual CNDMS NPR CNDMS Variance 4.38686 27.8% predicted predicted (Actual--Predicted) less than less than NPR Variance 6.13494 35.0% actual actual (Actual--Predicted) ABS (NPR Variance)-- 5.10348 7.2% ABS (CNDMS Variance) NPR CNDMS Variance 2.83058 17.2% predicted (Actual--Predicted) at least NPR Variance 5.48736 -34.3% as much as (Actual--Predicted) actual ABS (NPR Variance)-- 5.58337 17.1% ABS (CNDMS Variance) CNDMS Variance 2.66141 -10.8% CNDMS NPR (Actual--Predicted) predicted predicted NPR Variance 5.46970 20.9% at least less than (Actual--Predicted) as much actual ABS (NPR Variance)-- 5.36311 10.1% as actual ABS (CNDMS Variance) NPR CNDMS Variance 4.26754 -35.2% predicted (Actual--Predicted) at least NPR Variance 6.93401 -59.5% as much as (Actual--Predicted) actual ABS (NPR Variance)-- 5.97656 24.4% ABS (CNDMS Variance) Table 7. Cost Analysis by Scenario (Medical/Surgical Patient Type Category, Australian Public Hospitals) Indicator Indicator for CNDMS for NPR Larger than Larger than Actual Actual N Mean CNDMS NPR CNDMS Variance predicted predicted (Actual--Predicted) 18,506 3.5870 less than less than actual actual NPR Variance 18,506 3.9782 (Actual--Predicted) ABS (NPR Variance)-- 18,506 0.3912 ABS (CNDMS Variance) NPR CNDMS Variance predicted at (Actual--Predicted) 11,958 2.0582 least as much as NPR Variance 11,958 -3.5956 actual (Actual--Predicted) ABS (NPR Variance)-- 11,958 1.5374 ABS (CNDMS Variance) CNDMS NPR CNDMS Variance 9,764 -1.4741 predicted at predicted (Actual--Predicted) least as less than much as actual NPR Variance 9,764 2.1787 actual (Actual--Predicted) ABS (NPR Variance)-- 9,764 0.7046 ABS (CNDMS Variance) NPR CNDMS Variance 23,135 -3.4559 predicted at (Actual--Predicted) least as much as NPR Variance 23,135 -5.7192 actual (Actual--Predicted) ABS (NPR Variance)-- 23,135 2.2633 ABS (CNDMS Variance) Indicator Indicator for CNDMS for NPR Percent Larger than Larger than Standard of Actual Actual Deviation Actual CNDMS NPR CNDMS Variance 4.26279 24.9% predicted predicted (Actual--Predicted) less than less than actual actual NPR Variance 4.14125 29.5% (Actual--Predicted) ABS (NPR Variance)-- 3.42594 4.6% ABS (CNDMS Variance) NPR CNDMS Variance 2.44436 15.1% predicted at (Actual--Predicted) least as much as NPR Variance 3.98693 -29.3% actual (Actual--Predicted) ABS (NPR Variance)-- 4.40200 14.2% ABS (CNDMS Variance) CNDMS NPR CNDMS Variance 2.30324 -9.4% predicted at predicted (Actual--Predicted) least as less than much as actual NPR Variance 2.65084 18.6% actual (Actual--Predicted) ABS (NPR Variance)-- 2.92624 9.2% ABS (CNDMS Variance) NPR CNDMS Variance 4.14781 -27.7% predicted at (Actual--Predicted) least as much as NPR Variance 6.05151 -52.7% actual (Actual--Predicted) ABS (NPR Variance)-- 5.12234 25.0% ABS (CNDMS Variance)
|Printer friendly Cite/link Email Feedback|
|Author:||Heslop, Liza; Plummer, Virginia|
|Date:||Nov 1, 2012|
|Previous Article:||Are staffing changes in California's hospitals sensitive to individual hospital characteristics?|
|Next Article:||Changing tides: increasing evidence to embrace a patient classification system.|