- Email: [email protected]
Comparison of cost accounting methods from different DRG systems and their effect on health care quality
Health Policy 74 (2005) 46–55
Comparison of cost accounting methods from different DRG systems and their effect on health care quality Jan Eric Leister, J¨urgen Stausberg∗ Medical Faculty, University of Duisburg-Essen, Institute for Medical Informatics, Biometry and Epidemiology, Hufelandstr. 55, 45122 Essen, Germany
Abstract Background: Diagnosis related groups (DRGs) are a well-established provider payment system. Because of their immanent potential of cost reduction, they have been widely introduced. In addition to cost cutting, several social objectives – e.g., improving overall health care quality – feed into the DRG system. Objectives: The WHO compared different provider payment systems with regard to the following objectives: prevention of further health problems, providing services and solving health problems, and responsiveness to people’s legitimate expectations. However, no study has been published which takes the impact of different cost accounting systems across the DRG systems into account. Methods: We compared the impact of different cost accounting methods within DRG-like systems by developing six criteria: integration of patients’ health risk into pricing practice, incentives for quality improvement and innovation, availability of high class evidence based therapy, prohibition of economically founded exclusions, reduction of fragmentation incentives, and improvement of patient oriented treatment. Results: We set up a first overview of potential and actual impacts of the pricing practices within Yale-DRGs, AR-DRGs, G-DRGs, Swiss AP-DRGs adoption and Swiss MIPP. It could be demonstrated that DRGs are not only a ‘homogenous’ group of similar provider payment systems but quite different by fulfilling major health care objectives connected with the used cost accounting methods. Conclusions: If not only the possible cost reduction is used to put in a good word for DRG-based provider payment systems, maximum accurateness concerning the method of cost accounting should prevail when implementing a new DRG-based provider payment system. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Costs and cost analysis; Diagnosis related groups; Hospitals; Quality of health care; World Health Organisation
1. Introduction ∗
Corresponding author. Tel.: +49 201 723 4512; fax: +49 201 723 5933. E-mail address: [email protected] (J. Stausberg).
The World Health Organisation (WHO) has pointed to the enormous relevancy of health care quality concerning the economical and social progress of a coun-
0168-8510/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.healthpol.2004.12.001
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
47
Table 1 Provider payment mechanisms and provider behaviour [7] Provider behaviour mechanisms
Prevent health problems
Respond to legitimate expectations
Deliver services
Contain costs
Line item budget Global budget Capitation (with competition) Diagnosis related payment Fee-for-service
+/− ++ +++ +/− +/−
+/− +/− ++ ++ +++
−− −− −− ++ +++
+++ +++ +++ ++ −−−
Key: (+++) very positive effects; (++) some positive effects; (+/−) little or no effect; (−−) some negative effects; (−−−) very negative effects.
try within its ‘Health for all by the year 2000’ program. However, most approaches to reform the provider payment system are still focused on cost reduction. Although the opposite is intuitively appealing, it has been suggested that increasing service quality levels reduce health care costs [1]. Nevertheless improving health care quality remains second on most provider payment systems reform agendas within developed countries. An unexpectedly high proportion of health systems does not take social objectives (e.g. improving health or responsiveness to the expectations of the population) into account. While implementing new provider payment systems many countries accept a careless reduction in health care quality [2–4]. Within several studies [4–6] there was little doubt that almost any of the currently used provider payment systems causes unmeant provider behaviour and rewards providers who act against major quality aims. The same studies show considerable evidence that better quality could lead to cost savings. The challenge is to set purchasing goals that allow providers all necessary discretionary power in doing there business while giving the purchaser the capacity to influence the overall quality of services. Diagnosis related groups (DRGs) are generally thought to control health costs very well. Moreover, DRGs seem to support other health objectives like responsibility to legitimate expectations or good service delivery [7]. It is for this reason that several governments are implementing DRGs or comparable payment systems [8]. Diagnostic related payment is based on a fixed payment per patient related to the patient’s main diagnosis. The concept of diagnostic related payment has been outlined by a group at Yale University in 1975 originally [9,10]. Within their original concept the authors named three different objectives: (1) reducing
costs; (2) increasing benefit and (3) increasing quality of health care systems. During the next few years, solely the first aspect of this concept was been taken into account [11]. There is strong evidence [3,12] that the provider payment methodology provides high incentives for opportunistic provider behaviour [13]. An improper setup of the payment mechanisms can reduce the quality of services. In order to discourage opportunistic behaviour payment mechanisms should be based on medical evidence. Moreover, one needs to construct a cost accounting system that incorporates the lessons learned from activity based costing. Prior research only compared DRGs to other provider payment methods (see Table 1), without giving deeper insights to the cost accounting methodology. This work attempts to present an in-depth review of different cost accounting methods used in DRG systems with respect to health care quality. Based on a review of pricing practices (which are used to define the several cost weights at last) we analyzed implications for health care quality by using one of the several different cost accounting systems. Analyzing several pricing practices allows us to show the other objectives (apart from cost savings), which are integrated into each cost accounting model.
2. Methods Within its ‘World Health Report 2000’ the WHO designed an evaluation scheme modelling the impact of provider payment systems (see Table 1) [3,7]. Since most DRG systems have been designed to reduce health care costs, this aspect has been spared out within this review. The remaining three aspects had been refined
48
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
as follows, because they are too superficial to evaluate the consequences of different pricing practices:
3. Criteria set 3.1. Prevention of further health problems
• Prevention of further health problems: o integration of patients’ health risk into pricing practice; o incentives for quality improvement and innovation. • Providing services and solving health problems: o availability of high class evidence based therapy; o prohibition of economically founded exclusions. • Responsiveness to people’s legitimate expectations: o reduction of fragmentation incentives; o improvement of patient oriented treatment. There is no evidence to assume a disjunction of all chosen parameters. Furthermore strong positive or negative correlations between the constructed major health care objectives might exist. Especially the objective ‘Integration of patients’ health risk into pricing practice’ might conflict with the objective ‘Prohibition of economically founded exclusions’. In addition there is some indication that an evidence based calculation system has a negative impact on innovation and medical improvements [5]. Nevertheless, the criteria system should enable us to provide a first survey of major health objectives integrated into pricing practice because of this research’s exploratory nature. With the designed criteria framework we evaluated the cost accounting methods of five DRG systems worldwide: Yale-DRG, AR-DRG, G-DRG, Swiss AP-DRG adaptation and Swiss MIPP. The criteria are used to rate the cost accounting methods of the selected DRG systems concerning their quality aspects. To handle these parameters more comfortably a universal scale with the following dimension will be used: (−2)–(−1)–(0)–(+1)–(+2). Minus two (−2) indicates a high negative impact on the aspect in consideration while plus two (+2) indicates a strong positive influence. While evaluating the DRG systems only cost accounting aspects are considered. Other factors will be neglected consciously. The criteria set will be introduced in detail in the following section. The assessment of the five DRG systems is then presented in Section 4. Some appear comparable, but by considering the six criteria it is possible to show some important differences.
The first criterion is the integration of patients’ health risks into pricing practice [2,3,5]. Providers have to be paid for low consultation frequencies or effective prevention and have to be penalized for making patients to re-consult repeatedly. In an optimized system health care providers are paid for the healthy patients and not for their patients diseases [14]. Concerning this aspect we will estimate all pricing practices which integrate the patients’ health risks into calculation methodology as positive, otherwise a negative or neutral mark is applied. In addition there is a positive impact on preventing further health problems if providers receive a reward for improving process quality and using medical innovations. In case of the absence of rewards for an innovative behaviour, providers will rest on the current level of medical knowledge or even use obsolete methods due to cost effectiveness [5]. To reach this aim it is necessary to install a balanced recalculation routine concerning pricing practice. In addition to this it might be quite useful to allow providers some kind of monopoly profit to avoid cutthroat competition and stagnation of quality improvement. Following this perception we estimate all pricing practices, which included such options as positive, while negative or neutral marks are given for the remaining. 3.2. Providing services and solving health problems The first criterion is the availability of high class evidence based therapy. Using an applicable pricing practice is shown to have a deep impact on standardization of medical treatment and the alignment to evidence based practices [1,15]. Taking the evidence base into consideration reduces inferior individual treatment agreements between the (sometimes uninformed) patient and the opportunistic provider. Concerning this aspect we will estimate all pricing practices as positive, which take the availability of high class evidence based therapy into consideration, while otherwise negative or neutral marks are given. In addition there is a negative impact on service providing if pricing practice causes economical risks for
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
some patient groups. Accepting economical risk groups would lead to rationed services or at least to the exclusion of risk group members from needed treatment [5,16]. These risk groups could be formed by clinical or expenses pooling inhomogeneities, but can also be a consequence of political processes [17]. Following this perception we evaluate all pricing practices, which prohibit risk group discrimination as positive, while otherwise negative or neutral marks are assigned. 3.3. Responsiveness to people’s legitimate expectations The first criterion is the reduction of fragmentation incentives. DRGs lead to a replacement of inpatient treatment by outpatient therapy. The so-called ‘bloody hospital release’ is often said to be the greatest risk of DRG based provider payment [15]. This kind of behaviour might reduce treatment quality and could neglect patients’ expectations for good services. We estimate all pricing practices as positive, which reduce fragmentation incentives by using a suitable calculation methodology, otherwise as negative or neutral. In addition an improvement of patient oriented treatment can have a positive impact on the enlargement of responsiveness to patient’s legitimate expectations. Using a diagnostic based provider payment system might reduce some services, patients are strongly interested in. Decreasing of nursing services or increasing waiting times in case of elective treatment might be examples for reduced patient orientation [15,18]. Following this view we estimate all pricing practices that avoid inferior service attributes as positive, otherwise as negative or neutral.
4. Results 4.1. Yale-DRGs—the origin of DRG movement At the end of the 1960th a group around Robert Fetter started to develop a casemix system that aimed at better cost and effectiveness control within the large American university hospitals [10,11]. Later on the resulting DRG system was adopted nationally to allow cost savings within the national health system. The State Department of Health and Health Services started its testing in New Jersey in the early 1980th and estab-
49
lished the system for Medicare and Medicaid in 1983. Since then the Health Care Financing Administration (HCFA) was charged with revising cost weights and definitions every year [19]. The pricing practice is based on a quite inexact costing system, which is referred to as top-down approach in the following. It is focused to the DRG as a unit of output. Patients are not taken into consideration individually. Cost accounting follows universal service weights, even for many direct expenses. These service weights are based on national cost studies, which are conducted by the HCFA periodically. Meanwhile service weights, length of stay or other inexact cost allocation bases [20] had been increasingly replaced by more accurate and use-oriented bases. Nevertheless too many cost groups were accounted by disregarding direct costing methods. • Integration of patients’ health risk into pricing practice: Using a top-down approach does not support an integration of patients’ health risk into accounting systems properly. Aggregated service weights will not allow detailed consideration with regard to special patient risk factors [19]. This causes a very negative (−2) rating for this aspect. • Incentives for quality improvement and innovation: The effects on quality incentives for improvement and innovation are negative, too. If there is no chance to observe the connection between costs and treatment for specific patients, there are no incentives to improve medical methods. Even complicationoriented splits within patient diagnostic groups will not be account for easily. As a consequence this also leads to a very negative (−2) rating. • Availability of high class evidence based therapy: As the Yale system is based on actual costs instead of budget costs there is no evidence based calculation approach visible. Again, this leads to a very negative (−2) rating. • Prohibition of economically founded exclusions: Since the top-down approach does not take into consideration any special patient characteristics, the prohibition of economically founded exclusions is out of reach at the same token [21]. Using universal service weights and cost accounting on a DRG level will include some patients with extremely high costs within the ‘homogeneous’ diagnosis group. Excluding these risky patients is economically useful for
50
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
Table 2 Final results for the different cost accounting methods AR-DRG TD
AR-DRG BU
G-DRG
Swiss AP-DRG
MIPP
Prevention of further health problems Integration of patients’ health risk into pricing practice −2 Incentives for quality improvement and innovation −2
Yale-DRG
−2 −1
+1 0
+1 +1
−1 0
+2 −2
Providing services and solving health problems Availability of high class evidence based therapy Prohibition of economically founded exclusions
−2 −2
−2 −2
−2 +1
−2 +2
−2 +1
+1 +2
Responsiveness to people’s legitimate expectations Reduction of fragmentation incentives Improvement of patient oriented treatment
0 −2
0 −2
0 +2
+1 +1
0 +2
+2 +2
−10
−9
+2
+4
0
+7
Sum TD: top-down, BU: bottom-up.
every service provider. This leads to a very negative (−2) rating again. • Reduction of fragmentation incentives: Combining an actual costing system with periodic service weight studies is an effective approach to reduce fragmentation incentives within a few revising periods [19]. On the other hand, several progress reports from the United States (US) frame a different reality. Within 15 years of using Yale DRGs in the US there is evidence for fragmentational and opportunistic behaviour between inpatient and outpatient treatment on the one and different clinical departments on the other hand [18,21]. This causes a merely neutral (0) rating. • Improvement of patient oriented treatment: The topdown approach does not consider any individual patient demands. Moreover, it has negative effects on the improvement of patient oriented treatment [18,19,21]. Without specific patient oriented cost weight calculation only a very negative (−2) rating is adequate. The final results for the different DRG systems with regard to their cost accounting methods are given in Table 2. 4.2. AR-DRGs—Australian’s derivative In 1984 the Australian government begun developing its own refined DRG system based on the American All-Patient-DRGs (AP-DRGs). The developed system was used as a provider payment method for inpatients in Victoria in 1993 for the first time. The cost weights
are based on iterated calculations using actual cost data from hospitals, which provide data voluntarily. To allow as many hospitals as possible to participate within the national cost weight studies, two costing systems have been in use: (1) a top-down and (2) a bottom-up approach [22,23]. The top-down approach is very similar to the original Yale-DRG system. Pricing practice is also based on an inexact costing system focusing on the DRG as unit of output. Once again specific patients are not considered. Cost accounting follows universal service weights, even for many direct expenses. These service weights have been taken from Maryland (US) initially. Later they have been revised using national cost studies to better reflect new clinical evidences [13]. However, most of the service weights use inexact bases like length of stay and neglect direct costing methods. In addition, hospitals are not forced to use national service weights within their calculation. Hence hospitals can use their own cost allocation base [24]. In consequence, estimations of the top-down approach are nearly equal to the Yale-DRGs. • Incentives for quality improvement and innovation: Only with regard to this aspect a slightly better assessment (−1) is acceptable taking into account that an individual calculation of service weights within every hospital allows to adapt to new treatment methods and possibly higher costs faster. Alternatively a bottom-up system called patient level clinical costing is offered for cost allocation. This costing system is more precise. Still more effort is needed. Based on an accurate clinical cost record-
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
51
Fig. 1. Comparison between AR-DRG top-down and bottom-up approach. The black line shows the overall estimation for the top-down approach, the grey line for the bottom-up approach. The light grey area between both lines represents the superiority of the bottom-up approach.
ing system the expenses for any single patient should be generated. Since the majority of costs can be estimated as direct costs without accounting for overhead expenses, this cost accounting system guarantees that costs are accounted on the bases of actual cases. This is opposed to accounting based on DRGs. This way few indirect costs are accounted which circumvents the methodical shortcomings with regard to overhead expenses. Similar to the top-down approach each hospital can use individual service weights for some or all cost calculations. Thus there might be a great chance of calculating costs for any specific patient, but also the bottom-up system can only be another expensive top-down approach [24,25]. • Integration of patients’ health risk into pricing practice: Using a bottom-up approach does more or less allows a good integration of patients’ health risk into pricing practice [13]. This causes a positive (+1) rating for this evaluation criterion. • Incentives for quality improvement and innovation: As a consequence of the former aspect the effects on incentives for quality improvement and innovation
are positive, too. Nevertheless considering the fact that political processes within hospitals might allow a transcription of actual cost reductions to other burden centres [26], a neutral (0) assessment has been reached. • Availability of high class evidence based therapy: As the AR-DRGs are based on actual costs instead of budget costs, a very negative (−2) appraisal is given. • Prohibition of economically founded exclusions: Since the bottom-up approach considers individual patients [24,26], a positive rating is in range. Merely the fact that the used costing system is suspended to political processes gives it a slight negative back taste. Therefore, a moderate positive (+1) rating is adequate. • Reduction of fragmentation incentives: Combining an actual costing system with periodically rerun cost weight studies is an effective approach to reduce fragmentation incentives within a few revising periods [19]. Moreover, direct cost calculations for individual patients reduce such attempts. In the opposite political processes and the allowance of service weights within the bottom-up approach can lead to
52
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
Fig. 2. Comparison between AR-DRG bottom-up approach and G-DRGs. The black line shows the overall estimation for the Australian ARDRG bottom-up approach, the grey line for the G-DRG system. The light grey area between both lines represents the superiority of the G-DRG system, and the dark grey area of the Australian bottom-up approach.
some fragmentation activities [13,26]. This leads to a merely neutral (0) rating. • Improvement of patient oriented treatment: The bottom-up approach can take into account patients’ individual wishes. Without the clear definition of the cost calculation methods in place the integration of these interests into the cost accounting system could be negotiated [13]. As a consequence a very positive (+2) rating seems to be adequate. Concerning the fact that a chain is only as robust as its weakest link, the complete AR-DRG system has to be finally given the more negative results within every single aspect. Fig. 1 gives a comparison between both Australian approaches. 4.3. G-DRGs—Germany uses an adaptation of the Australian system Within a far reaching health system reformation in 2000 the Federal Republic of Germany decided to implement a refined DRG system based on the bottom-up
part of the AR-DRGs [27,28]. Although each hospital determinates its exact pricing practice within the patient level clinical costing approach, exact cost allocation bases are given within the calculation guidelines. This causes a high degree of decisiveness. Thus many hospitals were forced to use the given allocation bases. In addition, many costs were calculated on the level of direct expenses, so that inexact service weights could be avoided in many cases. For some departments like radiology or laboratory an explicit cost allocation system is mandatory. In contrast to the Australian bottom-up approach the G-DRG system (see Fig. 2) tries to reduce political processes within hospitals by giving concrete guidelines. On the one hand, these changes do not impact the integration of patients’ health risk into pricing practice (+1) and the availability of high-class evidence based therapy (−2). Eliminating political processes within hospitals allows a slightly better appraisal concerning the aspects of incentives for quality improvement and innovation (+1), prohibition of economically founded exclusions (+2) and reduction of fragmentation incen-
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
53
Fig. 3. Comparison between G-DRGs and MIPP. The black line shows the overall estimation for the German G-DRG approach, the grey line for the MIPP system. The dark grey area between both lines represents the superiority of the G-DRG system, and the light grey area of the MIPP approach.
tives (+1). On the other hand (as pointed out earlier) this has a negative impact on the improvement of patient oriented treatment (+1), because patients’ interests might be lost within a more ‘mathematical’ oriented calculation process. 4.4. Swiss AP-DRG adaptation—another ‘refined’ system? Since 1998 Switzerland is also experimenting with a DRG based provider payment system [29–34]. The Swiss AP-DRG adaptation follows the same approach as the G-DRG system, using an established DRG system and to recalculate the cost weights. Withal – similar to the G-DRGs – very concrete cost accounting and calculation guidelines are applied. Nevertheless there is an important difference to the German approach: because of the pricing practice, the medical definitions of DRGs have never been changed. With exception of the integration of patients’ health risk into pricing practice, which would be more difficult without changing the medical DRG definitions (−1), the same estimations as for the AR-DRGs were given.
4.5. MIPP—a path-oriented approach Although the MIPP (in German ‘Medizinisches System integrierter Patientenpfade’, which means medical system of integrated patient paths) has never been used for calculations within a provider payment system so far, it founds a remarkable approach [35–37]. By using patient management guidelines the MIPP approach uses evidence-based cost weights. In contrary to all other evaluated systems MIPP is consequently based on budgets instead of actual costing. By defining a complete patient episode from admission to discharge, a detailed definition of quality requirements and costs (including costs of optional treatments by probability) is given. Using a budget costing system means nearly perfect (+2) integration of patients’ health risk into pricing practice [38]. In contradiction to that the impact on incentives for quality improvement and innovation is very negative (−2), since the re-calculation is associated with considerable costs. Hence this re-calculation can only be staged infrequently. Moreover, real life cases are seldom observable. However using patient
54
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
management guidelines will increase (+1) the availability of high-class evidence-based therapy [39]. The impact on all other criteria will be very positive (+2), because all risks can be sawn and integrated within the calculation process. Every behaviour in contradiction to the defined standards might have consequences and patient deputies explicitly take place within calculation teams. See Fig. 3 for a comparison between G-DRG and MIPP.
5. Discussion It has been reasoned (refer to Figs. 1–3) that the cost accounting methods of the contemplated DRG systems have their strengths and weaknesses concerning the quality requirements, which have been established by the WHO. Every single DRG system’s cost accounting method has an impact on providers’ behaviour and the quality of the services provided. The presented criteria set can be used to compare the DRG costing models and make a distinct choice. The framework that has been presented within this paper explains which DRG system’s cost accounting system fulfils its objective best instead of merely cutting costs. To allow an impartial decision, a tool like the benefit-value-analysis might be helpful. Calculating a simple sum for every cost accounting method shows advantages for the MIPP system (see Table 2). It should not be neglected that the implementation and maintenance of the MIPP system results in considerable cost survey expenses. These costs render a mere 20% of DRGs feasible. Considering the ABC-analysis approximately 80% of the cases would fall into this category. What is more, the MIPP system is a very recent attempt with limited theoretical basis and no practical experience. By using the developed evaluation framework, the pricing practices impact to health care quality can be estimated, but there is little doubt, that the presented ratings could not be much more then a ‘first offer’. An unerring rating has to be based on empirical data and sophisticated economical models. Enhancing knowledge concerning these needs might allow governments to design better payment systems combining optimal cost effectiveness and high quality improvements. Furthermore, next to the six criteria used within this paper other objectives and requirements for an
optimal reward based health system are relevant for this discussion. Merely considering the given criteria could not be far reaching enough and could neglect risks and chances at the same time. Moreover, other methods and approaches apart from the cost calculations, which could provide incentives for reasonable provider behaviour, are imaginable. They could be used to ‘optimise’ the calculation results actively. Unfortunately, only few attempts exist and have been followed half-heartedly. Government have been cautioned that an unsolicited and inadequate method of calculation could impact provider behaviour badly and result in economic losses.
References [1] Glaeske G, Lauterbach KW, R¨urup B, et al. Weichenstellung f¨ur die Zukunft—Elemente einer neuen Gesundheitspolitik (Langfassung). Berlin: Friederich-Ebert-Stiftung; 2001. [2] Sachverst¨andigenrat f¨ur die konzertierte Aktion im Gesundheitswesen. Finanzierung, Nutzerorientierung und Qualit¨at. Berlin: Bundestagdrucksache; 2003. [3] Barnum H, Kutzin J, Saxenian H. Incentives and provider payment methods. International Journal of Health Planning and Management 1995;10:23–45. [4] Braun B, M¨uller R. Auswirkungen von Verg¨utungsformen auf die Qualit¨at der station¨aren Versorgung. Schw¨abisch Gm¨und: Edition GEK; 2003. [5] Institute of Medicine, Committee on Quality Health Care in America. Crossing the quality chasm: a new health system for the 21st century. Washington, DC: National Academic Press; 2001. [6] Institute of Medicine, Committee on Quality Health Care in America. To err is human: building a safer health system. Washington, DC: National Academic Press; 1999. [7] World Health Organisation. The world health report 2000—health systems: improving performance. Geneva: World Health Organisation; 2000. [8] Danish Ministry of Health. Hospital funding and casemix. Copenhagen: The Danish Ministry of Health; 1999. [9] Fetter RB, Thompson JD, Mills RE. A system for cost and reimbursement control in hospitals. The Yale Journal of Biologiy and Medicine 1976;49:123–36. [10] Fetter RB, Mills RE, Riedel DC, et al. The application of diagnostic specific cost profiles to cost and reimbursement control in hospitals. Journal of Medical Systems 1977;1:137– 49. [11] Fetter RB. Diagnosis related groups—understanding hospital performance. Interfaces 1991;21:6–26. [12] Hillman AL. Managing the physician: rules versus incentives. Health Affairs 1991;10:138–46. [13] Phelan PD, Tate R, Webster F, et al. DRG cost weights-getting it right. MJA 1998;169:36–8.
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55 [14] Sell S. Einf¨uhrung eines durchg¨angig (fall)pauschalierenden Verg¨utungssystems f¨ur Krankenhausleistungen auf DRG-Basis. Sozialer Fortschritt 2000;5:102–15. [15] Fischer W. Auswirkungen der Einf¨uhrung von DRGPauschalen in den USA. Wolfertswil: Zentrum f¨ur Medizinische Informatik; 2001. [16] Roos NP, Freeman JL. Potential for inpatient-outpatient substitution with diagnosis-related groups. Health Care Financing Review 1989;10:31–8. [17] Fischer W. Zur Einf¨uhrung und Nutzung von Patientenklassifikationssystemen in Krankenh¨ausern. Wolfertswil: Zentrum f¨ur Medizinische Informatik; 1997. [18] Lewis MA, Leake B, Leal-Sotelo M, et al. The initial effects of the prospective payment system on nursing home patients. AJPH 1987;77:819–21. [19] Coffey RM. Casemix information in the united states: 15 years of management and clinical experience. Casemix Quaterly 1999;1:13–26. [20] Stausberg J, Kretzschmar F, Weimar C, et al. Beschreibung von Problempatienten im Krankenhaus aus administrativen Daten mit Hilfe der Fuzzy-Logik und Vergleich der Ergebnisse mit DRG-Parametern und klinischen Skalen am Beispiel akuter cerebrovaskul¨arer Erkrankungen. Informatik. Biometrie und Epidemiologie in Medizin und Biologie 2002;1:27– 35. [21] Ellis RP. Hospital payment in the united states: an overview and discussion of current policy issues. In: Paper Prepared for International Conference on Setting Prices for Disease: Lessons from Foreign Experience, Paris; 2001. [22] Commonwealth Department of Health and Aged Care. Hospital reference manual round 5, 2000–2001. Canberra: Commonwealth of Australia; 2001. [23] Jackson T. Using computerised patient-level costing data for setting DRG weights—the Victorian (Australia) cost weight studies. Health Policy 2001;56:149–63. [24] Watts J, Richardson J, Segal L. Comparing national public hospital cost data collections for use in performance reporting. West Heidelberg: Commonwealth Department of Health and Aged Care; 2000. [25] Duckett SJ. Casemix funding for acute hospital inpatient services in Australia. MJA 1998;169:17–21.
55
[26] Jackson T, Watts J, Lane L, et al. Data compatibility in patient level clinical costing. Casemix Quaterly 1999;1:1–12. [27] Deutsche, Krankenhausgesellschaft. Kalkulation von Fallkosten-Handbuch zur Anwendung in Krankenh¨ausern, Version 2.0. Berlin: DKVG; 2002. [28] Roeder N. Anpassung des G-DRG-Systems an das deutsche Leistungsgeschehen. M¨unster: DRG-Research Group; 2003. [29] H+-Die Spit¨aler der Schweiz. Handbuch Kosten- und Leistungsrechnung. Bern: H+-Die Spit¨aler der Schweiz; 2001. [30] H+-Die Spit¨aler der Schweiz. Bern: Kostenstellenrechnung im Spital. H+-Die Spit¨aler der Schweiz; 2001. [31] H+-Die Spit¨aler der Schweiz. Handbuch Kostentr¨agerrechnung: Systematische Erarbeitung einer Kostentr¨agerrechnung im Spital. Bern: H+-Die Spit¨aler der Schweiz; 2001. [32] ISE. Finanzierung der Akutspit¨aler-Abteilungsbezogene Fallpauschalen im Kanton Wallis-Evaluation und Perspektiven. Prilly: ISE; 1999. [33] Guillain H, Kr¨uger R, Rey J. Handbuch APDRG Schweiz, Version 3.0. Prilly: ISE; 2001. [34] Kr¨uger R. Dritte Version der APDRG-Referenzwerte f¨ur die Schweiz (Kostengewichte): Allgemeiner Kontext und methodische Aspekte. Prilly: ISE; 2001. [35] Aargauischer Krankenkassenverband. Neues Spitalfinanzierungsmodell auf Basis von Behandlungsstandards. Aarau: Gesch¨aftsstelle mipp; 2001. [36] Kantonsspital Aarau, Gesch¨aftsstelle MIPP. Neues Spitalfinanzierungsmodell auf Basis von Behandlungsstandard— Evaluation der Phase 1.7.2000 bis 30.6.2001 des Pilotprojekts: Fallpreispauschalen nach dem Modell integrierter Patientenpfade >mipp<. Aarau: Gesch¨aftstelle MIPP Kantonspital Aarau; 2001. [37] Rieben E, M¨uller H, Holler T, et al. Pfadkostenrechnung als Kostentr¨agerrechnung-Kalkulation und Anwendung von Patientenpfaden. Landsberg/Lech: Ecomed; 2003. [38] G¨ussow J, Greulich A, Ott R. Beurteilung und Einsatz der Prozesskostenrechnung als Antwort der Krankenh¨auser auf die Einf¨uhrung der DRGs. Kostenrechnungspraxis 2002;46:179–89. [39] L¨ungen M, Lauterbach KW. Die Messung der Fallschwere unter einem DRG-basierten Verg¨utungssystem. Medizinische Klinik 2001;97:70–6.
Comparison of cost accounting methods from different DRG systems and their effect on health care quality Jan Eric Leister, J¨urgen Stausberg∗ Medical Faculty, University of Duisburg-Essen, Institute for Medical Informatics, Biometry and Epidemiology, Hufelandstr. 55, 45122 Essen, Germany
Abstract Background: Diagnosis related groups (DRGs) are a well-established provider payment system. Because of their immanent potential of cost reduction, they have been widely introduced. In addition to cost cutting, several social objectives – e.g., improving overall health care quality – feed into the DRG system. Objectives: The WHO compared different provider payment systems with regard to the following objectives: prevention of further health problems, providing services and solving health problems, and responsiveness to people’s legitimate expectations. However, no study has been published which takes the impact of different cost accounting systems across the DRG systems into account. Methods: We compared the impact of different cost accounting methods within DRG-like systems by developing six criteria: integration of patients’ health risk into pricing practice, incentives for quality improvement and innovation, availability of high class evidence based therapy, prohibition of economically founded exclusions, reduction of fragmentation incentives, and improvement of patient oriented treatment. Results: We set up a first overview of potential and actual impacts of the pricing practices within Yale-DRGs, AR-DRGs, G-DRGs, Swiss AP-DRGs adoption and Swiss MIPP. It could be demonstrated that DRGs are not only a ‘homogenous’ group of similar provider payment systems but quite different by fulfilling major health care objectives connected with the used cost accounting methods. Conclusions: If not only the possible cost reduction is used to put in a good word for DRG-based provider payment systems, maximum accurateness concerning the method of cost accounting should prevail when implementing a new DRG-based provider payment system. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Costs and cost analysis; Diagnosis related groups; Hospitals; Quality of health care; World Health Organisation
1. Introduction ∗
Corresponding author. Tel.: +49 201 723 4512; fax: +49 201 723 5933. E-mail address: [email protected] (J. Stausberg).
The World Health Organisation (WHO) has pointed to the enormous relevancy of health care quality concerning the economical and social progress of a coun-
0168-8510/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.healthpol.2004.12.001
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
47
Table 1 Provider payment mechanisms and provider behaviour [7] Provider behaviour mechanisms
Prevent health problems
Respond to legitimate expectations
Deliver services
Contain costs
Line item budget Global budget Capitation (with competition) Diagnosis related payment Fee-for-service
+/− ++ +++ +/− +/−
+/− +/− ++ ++ +++
−− −− −− ++ +++
+++ +++ +++ ++ −−−
Key: (+++) very positive effects; (++) some positive effects; (+/−) little or no effect; (−−) some negative effects; (−−−) very negative effects.
try within its ‘Health for all by the year 2000’ program. However, most approaches to reform the provider payment system are still focused on cost reduction. Although the opposite is intuitively appealing, it has been suggested that increasing service quality levels reduce health care costs [1]. Nevertheless improving health care quality remains second on most provider payment systems reform agendas within developed countries. An unexpectedly high proportion of health systems does not take social objectives (e.g. improving health or responsiveness to the expectations of the population) into account. While implementing new provider payment systems many countries accept a careless reduction in health care quality [2–4]. Within several studies [4–6] there was little doubt that almost any of the currently used provider payment systems causes unmeant provider behaviour and rewards providers who act against major quality aims. The same studies show considerable evidence that better quality could lead to cost savings. The challenge is to set purchasing goals that allow providers all necessary discretionary power in doing there business while giving the purchaser the capacity to influence the overall quality of services. Diagnosis related groups (DRGs) are generally thought to control health costs very well. Moreover, DRGs seem to support other health objectives like responsibility to legitimate expectations or good service delivery [7]. It is for this reason that several governments are implementing DRGs or comparable payment systems [8]. Diagnostic related payment is based on a fixed payment per patient related to the patient’s main diagnosis. The concept of diagnostic related payment has been outlined by a group at Yale University in 1975 originally [9,10]. Within their original concept the authors named three different objectives: (1) reducing
costs; (2) increasing benefit and (3) increasing quality of health care systems. During the next few years, solely the first aspect of this concept was been taken into account [11]. There is strong evidence [3,12] that the provider payment methodology provides high incentives for opportunistic provider behaviour [13]. An improper setup of the payment mechanisms can reduce the quality of services. In order to discourage opportunistic behaviour payment mechanisms should be based on medical evidence. Moreover, one needs to construct a cost accounting system that incorporates the lessons learned from activity based costing. Prior research only compared DRGs to other provider payment methods (see Table 1), without giving deeper insights to the cost accounting methodology. This work attempts to present an in-depth review of different cost accounting methods used in DRG systems with respect to health care quality. Based on a review of pricing practices (which are used to define the several cost weights at last) we analyzed implications for health care quality by using one of the several different cost accounting systems. Analyzing several pricing practices allows us to show the other objectives (apart from cost savings), which are integrated into each cost accounting model.
2. Methods Within its ‘World Health Report 2000’ the WHO designed an evaluation scheme modelling the impact of provider payment systems (see Table 1) [3,7]. Since most DRG systems have been designed to reduce health care costs, this aspect has been spared out within this review. The remaining three aspects had been refined
48
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
as follows, because they are too superficial to evaluate the consequences of different pricing practices:
3. Criteria set 3.1. Prevention of further health problems
• Prevention of further health problems: o integration of patients’ health risk into pricing practice; o incentives for quality improvement and innovation. • Providing services and solving health problems: o availability of high class evidence based therapy; o prohibition of economically founded exclusions. • Responsiveness to people’s legitimate expectations: o reduction of fragmentation incentives; o improvement of patient oriented treatment. There is no evidence to assume a disjunction of all chosen parameters. Furthermore strong positive or negative correlations between the constructed major health care objectives might exist. Especially the objective ‘Integration of patients’ health risk into pricing practice’ might conflict with the objective ‘Prohibition of economically founded exclusions’. In addition there is some indication that an evidence based calculation system has a negative impact on innovation and medical improvements [5]. Nevertheless, the criteria system should enable us to provide a first survey of major health objectives integrated into pricing practice because of this research’s exploratory nature. With the designed criteria framework we evaluated the cost accounting methods of five DRG systems worldwide: Yale-DRG, AR-DRG, G-DRG, Swiss AP-DRG adaptation and Swiss MIPP. The criteria are used to rate the cost accounting methods of the selected DRG systems concerning their quality aspects. To handle these parameters more comfortably a universal scale with the following dimension will be used: (−2)–(−1)–(0)–(+1)–(+2). Minus two (−2) indicates a high negative impact on the aspect in consideration while plus two (+2) indicates a strong positive influence. While evaluating the DRG systems only cost accounting aspects are considered. Other factors will be neglected consciously. The criteria set will be introduced in detail in the following section. The assessment of the five DRG systems is then presented in Section 4. Some appear comparable, but by considering the six criteria it is possible to show some important differences.
The first criterion is the integration of patients’ health risks into pricing practice [2,3,5]. Providers have to be paid for low consultation frequencies or effective prevention and have to be penalized for making patients to re-consult repeatedly. In an optimized system health care providers are paid for the healthy patients and not for their patients diseases [14]. Concerning this aspect we will estimate all pricing practices which integrate the patients’ health risks into calculation methodology as positive, otherwise a negative or neutral mark is applied. In addition there is a positive impact on preventing further health problems if providers receive a reward for improving process quality and using medical innovations. In case of the absence of rewards for an innovative behaviour, providers will rest on the current level of medical knowledge or even use obsolete methods due to cost effectiveness [5]. To reach this aim it is necessary to install a balanced recalculation routine concerning pricing practice. In addition to this it might be quite useful to allow providers some kind of monopoly profit to avoid cutthroat competition and stagnation of quality improvement. Following this perception we estimate all pricing practices, which included such options as positive, while negative or neutral marks are given for the remaining. 3.2. Providing services and solving health problems The first criterion is the availability of high class evidence based therapy. Using an applicable pricing practice is shown to have a deep impact on standardization of medical treatment and the alignment to evidence based practices [1,15]. Taking the evidence base into consideration reduces inferior individual treatment agreements between the (sometimes uninformed) patient and the opportunistic provider. Concerning this aspect we will estimate all pricing practices as positive, which take the availability of high class evidence based therapy into consideration, while otherwise negative or neutral marks are given. In addition there is a negative impact on service providing if pricing practice causes economical risks for
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
some patient groups. Accepting economical risk groups would lead to rationed services or at least to the exclusion of risk group members from needed treatment [5,16]. These risk groups could be formed by clinical or expenses pooling inhomogeneities, but can also be a consequence of political processes [17]. Following this perception we evaluate all pricing practices, which prohibit risk group discrimination as positive, while otherwise negative or neutral marks are assigned. 3.3. Responsiveness to people’s legitimate expectations The first criterion is the reduction of fragmentation incentives. DRGs lead to a replacement of inpatient treatment by outpatient therapy. The so-called ‘bloody hospital release’ is often said to be the greatest risk of DRG based provider payment [15]. This kind of behaviour might reduce treatment quality and could neglect patients’ expectations for good services. We estimate all pricing practices as positive, which reduce fragmentation incentives by using a suitable calculation methodology, otherwise as negative or neutral. In addition an improvement of patient oriented treatment can have a positive impact on the enlargement of responsiveness to patient’s legitimate expectations. Using a diagnostic based provider payment system might reduce some services, patients are strongly interested in. Decreasing of nursing services or increasing waiting times in case of elective treatment might be examples for reduced patient orientation [15,18]. Following this view we estimate all pricing practices that avoid inferior service attributes as positive, otherwise as negative or neutral.
4. Results 4.1. Yale-DRGs—the origin of DRG movement At the end of the 1960th a group around Robert Fetter started to develop a casemix system that aimed at better cost and effectiveness control within the large American university hospitals [10,11]. Later on the resulting DRG system was adopted nationally to allow cost savings within the national health system. The State Department of Health and Health Services started its testing in New Jersey in the early 1980th and estab-
49
lished the system for Medicare and Medicaid in 1983. Since then the Health Care Financing Administration (HCFA) was charged with revising cost weights and definitions every year [19]. The pricing practice is based on a quite inexact costing system, which is referred to as top-down approach in the following. It is focused to the DRG as a unit of output. Patients are not taken into consideration individually. Cost accounting follows universal service weights, even for many direct expenses. These service weights are based on national cost studies, which are conducted by the HCFA periodically. Meanwhile service weights, length of stay or other inexact cost allocation bases [20] had been increasingly replaced by more accurate and use-oriented bases. Nevertheless too many cost groups were accounted by disregarding direct costing methods. • Integration of patients’ health risk into pricing practice: Using a top-down approach does not support an integration of patients’ health risk into accounting systems properly. Aggregated service weights will not allow detailed consideration with regard to special patient risk factors [19]. This causes a very negative (−2) rating for this aspect. • Incentives for quality improvement and innovation: The effects on quality incentives for improvement and innovation are negative, too. If there is no chance to observe the connection between costs and treatment for specific patients, there are no incentives to improve medical methods. Even complicationoriented splits within patient diagnostic groups will not be account for easily. As a consequence this also leads to a very negative (−2) rating. • Availability of high class evidence based therapy: As the Yale system is based on actual costs instead of budget costs there is no evidence based calculation approach visible. Again, this leads to a very negative (−2) rating. • Prohibition of economically founded exclusions: Since the top-down approach does not take into consideration any special patient characteristics, the prohibition of economically founded exclusions is out of reach at the same token [21]. Using universal service weights and cost accounting on a DRG level will include some patients with extremely high costs within the ‘homogeneous’ diagnosis group. Excluding these risky patients is economically useful for
50
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
Table 2 Final results for the different cost accounting methods AR-DRG TD
AR-DRG BU
G-DRG
Swiss AP-DRG
MIPP
Prevention of further health problems Integration of patients’ health risk into pricing practice −2 Incentives for quality improvement and innovation −2
Yale-DRG
−2 −1
+1 0
+1 +1
−1 0
+2 −2
Providing services and solving health problems Availability of high class evidence based therapy Prohibition of economically founded exclusions
−2 −2
−2 −2
−2 +1
−2 +2
−2 +1
+1 +2
Responsiveness to people’s legitimate expectations Reduction of fragmentation incentives Improvement of patient oriented treatment
0 −2
0 −2
0 +2
+1 +1
0 +2
+2 +2
−10
−9
+2
+4
0
+7
Sum TD: top-down, BU: bottom-up.
every service provider. This leads to a very negative (−2) rating again. • Reduction of fragmentation incentives: Combining an actual costing system with periodic service weight studies is an effective approach to reduce fragmentation incentives within a few revising periods [19]. On the other hand, several progress reports from the United States (US) frame a different reality. Within 15 years of using Yale DRGs in the US there is evidence for fragmentational and opportunistic behaviour between inpatient and outpatient treatment on the one and different clinical departments on the other hand [18,21]. This causes a merely neutral (0) rating. • Improvement of patient oriented treatment: The topdown approach does not consider any individual patient demands. Moreover, it has negative effects on the improvement of patient oriented treatment [18,19,21]. Without specific patient oriented cost weight calculation only a very negative (−2) rating is adequate. The final results for the different DRG systems with regard to their cost accounting methods are given in Table 2. 4.2. AR-DRGs—Australian’s derivative In 1984 the Australian government begun developing its own refined DRG system based on the American All-Patient-DRGs (AP-DRGs). The developed system was used as a provider payment method for inpatients in Victoria in 1993 for the first time. The cost weights
are based on iterated calculations using actual cost data from hospitals, which provide data voluntarily. To allow as many hospitals as possible to participate within the national cost weight studies, two costing systems have been in use: (1) a top-down and (2) a bottom-up approach [22,23]. The top-down approach is very similar to the original Yale-DRG system. Pricing practice is also based on an inexact costing system focusing on the DRG as unit of output. Once again specific patients are not considered. Cost accounting follows universal service weights, even for many direct expenses. These service weights have been taken from Maryland (US) initially. Later they have been revised using national cost studies to better reflect new clinical evidences [13]. However, most of the service weights use inexact bases like length of stay and neglect direct costing methods. In addition, hospitals are not forced to use national service weights within their calculation. Hence hospitals can use their own cost allocation base [24]. In consequence, estimations of the top-down approach are nearly equal to the Yale-DRGs. • Incentives for quality improvement and innovation: Only with regard to this aspect a slightly better assessment (−1) is acceptable taking into account that an individual calculation of service weights within every hospital allows to adapt to new treatment methods and possibly higher costs faster. Alternatively a bottom-up system called patient level clinical costing is offered for cost allocation. This costing system is more precise. Still more effort is needed. Based on an accurate clinical cost record-
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
51
Fig. 1. Comparison between AR-DRG top-down and bottom-up approach. The black line shows the overall estimation for the top-down approach, the grey line for the bottom-up approach. The light grey area between both lines represents the superiority of the bottom-up approach.
ing system the expenses for any single patient should be generated. Since the majority of costs can be estimated as direct costs without accounting for overhead expenses, this cost accounting system guarantees that costs are accounted on the bases of actual cases. This is opposed to accounting based on DRGs. This way few indirect costs are accounted which circumvents the methodical shortcomings with regard to overhead expenses. Similar to the top-down approach each hospital can use individual service weights for some or all cost calculations. Thus there might be a great chance of calculating costs for any specific patient, but also the bottom-up system can only be another expensive top-down approach [24,25]. • Integration of patients’ health risk into pricing practice: Using a bottom-up approach does more or less allows a good integration of patients’ health risk into pricing practice [13]. This causes a positive (+1) rating for this evaluation criterion. • Incentives for quality improvement and innovation: As a consequence of the former aspect the effects on incentives for quality improvement and innovation
are positive, too. Nevertheless considering the fact that political processes within hospitals might allow a transcription of actual cost reductions to other burden centres [26], a neutral (0) assessment has been reached. • Availability of high class evidence based therapy: As the AR-DRGs are based on actual costs instead of budget costs, a very negative (−2) appraisal is given. • Prohibition of economically founded exclusions: Since the bottom-up approach considers individual patients [24,26], a positive rating is in range. Merely the fact that the used costing system is suspended to political processes gives it a slight negative back taste. Therefore, a moderate positive (+1) rating is adequate. • Reduction of fragmentation incentives: Combining an actual costing system with periodically rerun cost weight studies is an effective approach to reduce fragmentation incentives within a few revising periods [19]. Moreover, direct cost calculations for individual patients reduce such attempts. In the opposite political processes and the allowance of service weights within the bottom-up approach can lead to
52
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
Fig. 2. Comparison between AR-DRG bottom-up approach and G-DRGs. The black line shows the overall estimation for the Australian ARDRG bottom-up approach, the grey line for the G-DRG system. The light grey area between both lines represents the superiority of the G-DRG system, and the dark grey area of the Australian bottom-up approach.
some fragmentation activities [13,26]. This leads to a merely neutral (0) rating. • Improvement of patient oriented treatment: The bottom-up approach can take into account patients’ individual wishes. Without the clear definition of the cost calculation methods in place the integration of these interests into the cost accounting system could be negotiated [13]. As a consequence a very positive (+2) rating seems to be adequate. Concerning the fact that a chain is only as robust as its weakest link, the complete AR-DRG system has to be finally given the more negative results within every single aspect. Fig. 1 gives a comparison between both Australian approaches. 4.3. G-DRGs—Germany uses an adaptation of the Australian system Within a far reaching health system reformation in 2000 the Federal Republic of Germany decided to implement a refined DRG system based on the bottom-up
part of the AR-DRGs [27,28]. Although each hospital determinates its exact pricing practice within the patient level clinical costing approach, exact cost allocation bases are given within the calculation guidelines. This causes a high degree of decisiveness. Thus many hospitals were forced to use the given allocation bases. In addition, many costs were calculated on the level of direct expenses, so that inexact service weights could be avoided in many cases. For some departments like radiology or laboratory an explicit cost allocation system is mandatory. In contrast to the Australian bottom-up approach the G-DRG system (see Fig. 2) tries to reduce political processes within hospitals by giving concrete guidelines. On the one hand, these changes do not impact the integration of patients’ health risk into pricing practice (+1) and the availability of high-class evidence based therapy (−2). Eliminating political processes within hospitals allows a slightly better appraisal concerning the aspects of incentives for quality improvement and innovation (+1), prohibition of economically founded exclusions (+2) and reduction of fragmentation incen-
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
53
Fig. 3. Comparison between G-DRGs and MIPP. The black line shows the overall estimation for the German G-DRG approach, the grey line for the MIPP system. The dark grey area between both lines represents the superiority of the G-DRG system, and the light grey area of the MIPP approach.
tives (+1). On the other hand (as pointed out earlier) this has a negative impact on the improvement of patient oriented treatment (+1), because patients’ interests might be lost within a more ‘mathematical’ oriented calculation process. 4.4. Swiss AP-DRG adaptation—another ‘refined’ system? Since 1998 Switzerland is also experimenting with a DRG based provider payment system [29–34]. The Swiss AP-DRG adaptation follows the same approach as the G-DRG system, using an established DRG system and to recalculate the cost weights. Withal – similar to the G-DRGs – very concrete cost accounting and calculation guidelines are applied. Nevertheless there is an important difference to the German approach: because of the pricing practice, the medical definitions of DRGs have never been changed. With exception of the integration of patients’ health risk into pricing practice, which would be more difficult without changing the medical DRG definitions (−1), the same estimations as for the AR-DRGs were given.
4.5. MIPP—a path-oriented approach Although the MIPP (in German ‘Medizinisches System integrierter Patientenpfade’, which means medical system of integrated patient paths) has never been used for calculations within a provider payment system so far, it founds a remarkable approach [35–37]. By using patient management guidelines the MIPP approach uses evidence-based cost weights. In contrary to all other evaluated systems MIPP is consequently based on budgets instead of actual costing. By defining a complete patient episode from admission to discharge, a detailed definition of quality requirements and costs (including costs of optional treatments by probability) is given. Using a budget costing system means nearly perfect (+2) integration of patients’ health risk into pricing practice [38]. In contradiction to that the impact on incentives for quality improvement and innovation is very negative (−2), since the re-calculation is associated with considerable costs. Hence this re-calculation can only be staged infrequently. Moreover, real life cases are seldom observable. However using patient
54
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55
management guidelines will increase (+1) the availability of high-class evidence-based therapy [39]. The impact on all other criteria will be very positive (+2), because all risks can be sawn and integrated within the calculation process. Every behaviour in contradiction to the defined standards might have consequences and patient deputies explicitly take place within calculation teams. See Fig. 3 for a comparison between G-DRG and MIPP.
5. Discussion It has been reasoned (refer to Figs. 1–3) that the cost accounting methods of the contemplated DRG systems have their strengths and weaknesses concerning the quality requirements, which have been established by the WHO. Every single DRG system’s cost accounting method has an impact on providers’ behaviour and the quality of the services provided. The presented criteria set can be used to compare the DRG costing models and make a distinct choice. The framework that has been presented within this paper explains which DRG system’s cost accounting system fulfils its objective best instead of merely cutting costs. To allow an impartial decision, a tool like the benefit-value-analysis might be helpful. Calculating a simple sum for every cost accounting method shows advantages for the MIPP system (see Table 2). It should not be neglected that the implementation and maintenance of the MIPP system results in considerable cost survey expenses. These costs render a mere 20% of DRGs feasible. Considering the ABC-analysis approximately 80% of the cases would fall into this category. What is more, the MIPP system is a very recent attempt with limited theoretical basis and no practical experience. By using the developed evaluation framework, the pricing practices impact to health care quality can be estimated, but there is little doubt, that the presented ratings could not be much more then a ‘first offer’. An unerring rating has to be based on empirical data and sophisticated economical models. Enhancing knowledge concerning these needs might allow governments to design better payment systems combining optimal cost effectiveness and high quality improvements. Furthermore, next to the six criteria used within this paper other objectives and requirements for an
optimal reward based health system are relevant for this discussion. Merely considering the given criteria could not be far reaching enough and could neglect risks and chances at the same time. Moreover, other methods and approaches apart from the cost calculations, which could provide incentives for reasonable provider behaviour, are imaginable. They could be used to ‘optimise’ the calculation results actively. Unfortunately, only few attempts exist and have been followed half-heartedly. Government have been cautioned that an unsolicited and inadequate method of calculation could impact provider behaviour badly and result in economic losses.
References [1] Glaeske G, Lauterbach KW, R¨urup B, et al. Weichenstellung f¨ur die Zukunft—Elemente einer neuen Gesundheitspolitik (Langfassung). Berlin: Friederich-Ebert-Stiftung; 2001. [2] Sachverst¨andigenrat f¨ur die konzertierte Aktion im Gesundheitswesen. Finanzierung, Nutzerorientierung und Qualit¨at. Berlin: Bundestagdrucksache; 2003. [3] Barnum H, Kutzin J, Saxenian H. Incentives and provider payment methods. International Journal of Health Planning and Management 1995;10:23–45. [4] Braun B, M¨uller R. Auswirkungen von Verg¨utungsformen auf die Qualit¨at der station¨aren Versorgung. Schw¨abisch Gm¨und: Edition GEK; 2003. [5] Institute of Medicine, Committee on Quality Health Care in America. Crossing the quality chasm: a new health system for the 21st century. Washington, DC: National Academic Press; 2001. [6] Institute of Medicine, Committee on Quality Health Care in America. To err is human: building a safer health system. Washington, DC: National Academic Press; 1999. [7] World Health Organisation. The world health report 2000—health systems: improving performance. Geneva: World Health Organisation; 2000. [8] Danish Ministry of Health. Hospital funding and casemix. Copenhagen: The Danish Ministry of Health; 1999. [9] Fetter RB, Thompson JD, Mills RE. A system for cost and reimbursement control in hospitals. The Yale Journal of Biologiy and Medicine 1976;49:123–36. [10] Fetter RB, Mills RE, Riedel DC, et al. The application of diagnostic specific cost profiles to cost and reimbursement control in hospitals. Journal of Medical Systems 1977;1:137– 49. [11] Fetter RB. Diagnosis related groups—understanding hospital performance. Interfaces 1991;21:6–26. [12] Hillman AL. Managing the physician: rules versus incentives. Health Affairs 1991;10:138–46. [13] Phelan PD, Tate R, Webster F, et al. DRG cost weights-getting it right. MJA 1998;169:36–8.
J.E. Leister, J. Stausberg / Health Policy 74 (2005) 46–55 [14] Sell S. Einf¨uhrung eines durchg¨angig (fall)pauschalierenden Verg¨utungssystems f¨ur Krankenhausleistungen auf DRG-Basis. Sozialer Fortschritt 2000;5:102–15. [15] Fischer W. Auswirkungen der Einf¨uhrung von DRGPauschalen in den USA. Wolfertswil: Zentrum f¨ur Medizinische Informatik; 2001. [16] Roos NP, Freeman JL. Potential for inpatient-outpatient substitution with diagnosis-related groups. Health Care Financing Review 1989;10:31–8. [17] Fischer W. Zur Einf¨uhrung und Nutzung von Patientenklassifikationssystemen in Krankenh¨ausern. Wolfertswil: Zentrum f¨ur Medizinische Informatik; 1997. [18] Lewis MA, Leake B, Leal-Sotelo M, et al. The initial effects of the prospective payment system on nursing home patients. AJPH 1987;77:819–21. [19] Coffey RM. Casemix information in the united states: 15 years of management and clinical experience. Casemix Quaterly 1999;1:13–26. [20] Stausberg J, Kretzschmar F, Weimar C, et al. Beschreibung von Problempatienten im Krankenhaus aus administrativen Daten mit Hilfe der Fuzzy-Logik und Vergleich der Ergebnisse mit DRG-Parametern und klinischen Skalen am Beispiel akuter cerebrovaskul¨arer Erkrankungen. Informatik. Biometrie und Epidemiologie in Medizin und Biologie 2002;1:27– 35. [21] Ellis RP. Hospital payment in the united states: an overview and discussion of current policy issues. In: Paper Prepared for International Conference on Setting Prices for Disease: Lessons from Foreign Experience, Paris; 2001. [22] Commonwealth Department of Health and Aged Care. Hospital reference manual round 5, 2000–2001. Canberra: Commonwealth of Australia; 2001. [23] Jackson T. Using computerised patient-level costing data for setting DRG weights—the Victorian (Australia) cost weight studies. Health Policy 2001;56:149–63. [24] Watts J, Richardson J, Segal L. Comparing national public hospital cost data collections for use in performance reporting. West Heidelberg: Commonwealth Department of Health and Aged Care; 2000. [25] Duckett SJ. Casemix funding for acute hospital inpatient services in Australia. MJA 1998;169:17–21.
55
[26] Jackson T, Watts J, Lane L, et al. Data compatibility in patient level clinical costing. Casemix Quaterly 1999;1:1–12. [27] Deutsche, Krankenhausgesellschaft. Kalkulation von Fallkosten-Handbuch zur Anwendung in Krankenh¨ausern, Version 2.0. Berlin: DKVG; 2002. [28] Roeder N. Anpassung des G-DRG-Systems an das deutsche Leistungsgeschehen. M¨unster: DRG-Research Group; 2003. [29] H+-Die Spit¨aler der Schweiz. Handbuch Kosten- und Leistungsrechnung. Bern: H+-Die Spit¨aler der Schweiz; 2001. [30] H+-Die Spit¨aler der Schweiz. Bern: Kostenstellenrechnung im Spital. H+-Die Spit¨aler der Schweiz; 2001. [31] H+-Die Spit¨aler der Schweiz. Handbuch Kostentr¨agerrechnung: Systematische Erarbeitung einer Kostentr¨agerrechnung im Spital. Bern: H+-Die Spit¨aler der Schweiz; 2001. [32] ISE. Finanzierung der Akutspit¨aler-Abteilungsbezogene Fallpauschalen im Kanton Wallis-Evaluation und Perspektiven. Prilly: ISE; 1999. [33] Guillain H, Kr¨uger R, Rey J. Handbuch APDRG Schweiz, Version 3.0. Prilly: ISE; 2001. [34] Kr¨uger R. Dritte Version der APDRG-Referenzwerte f¨ur die Schweiz (Kostengewichte): Allgemeiner Kontext und methodische Aspekte. Prilly: ISE; 2001. [35] Aargauischer Krankenkassenverband. Neues Spitalfinanzierungsmodell auf Basis von Behandlungsstandards. Aarau: Gesch¨aftsstelle mipp; 2001. [36] Kantonsspital Aarau, Gesch¨aftsstelle MIPP. Neues Spitalfinanzierungsmodell auf Basis von Behandlungsstandard— Evaluation der Phase 1.7.2000 bis 30.6.2001 des Pilotprojekts: Fallpreispauschalen nach dem Modell integrierter Patientenpfade >mipp<. Aarau: Gesch¨aftstelle MIPP Kantonspital Aarau; 2001. [37] Rieben E, M¨uller H, Holler T, et al. Pfadkostenrechnung als Kostentr¨agerrechnung-Kalkulation und Anwendung von Patientenpfaden. Landsberg/Lech: Ecomed; 2003. [38] G¨ussow J, Greulich A, Ott R. Beurteilung und Einsatz der Prozesskostenrechnung als Antwort der Krankenh¨auser auf die Einf¨uhrung der DRGs. Kostenrechnungspraxis 2002;46:179–89. [39] L¨ungen M, Lauterbach KW. Die Messung der Fallschwere unter einem DRG-basierten Verg¨utungssystem. Medizinische Klinik 2001;97:70–6.