Introduction

This concept describes a number of approaches to hospital costing that were used in MCHP research during the 1990's. NOTE: For an overview of general costing methods (sources of data, types of costs, and approaches) and the methods of costing specific health services (hospitals, physicians, home care, personal care homes and pharmaceutical/prescription drugs) that have been used in MCHP research over time, please see the Costing Methods: An Overview of Costing Health Services in Manitoba concept.

Background

Many methods exist for hospital costing. They can be broadly separated into two categories: 'bottom-up' costing and 'top-down' costing.


• Bottom-up , or micro-costing, consists of identifying and costing the resources that are used by a specific patient. This approach cannot necessarily be generalized to other patients, and is difficult to estimate, particularly in a system like ours, in which hospitals are funded by a global budget rather than for individual services.
  • For a discussion on when it might be appropriate to use micro-costing (herein referred to as the cost list), see Shanahan et al.(1999)
  
  
• Top-down costing is the approach used by MCHP for costing. This method starts at the top with total expenditures and then divides these by a measure of total output.

Two different methods of 'top-down' costing have been used by MCHP: "per diems" and "case-mix" costing. Hospital acute per diem costs are calculated by dividing the total acute inpatient costs of providing care by the number of patient days to give an "average" cost per patient per day. One limitation of this method is that no adjustments are made for differences in patient characteristics that are likely to affect resource utilization (i.e. case-mix) or other factors explaining between-hospital.
• See Wall et al. (1994) for an example of per diem costing.

The other 'top-down' costing approach is case-mix costing . This method has been used in various reports by MCHP (Shanahan et al. (1999), Shanahan et al., 1994; Shanahan, 1996; Shanahan et al. 1997; Mustard & Derksen, 1997 ). This method goes further than the per diem costing by dividing patients into more homogeneous groups - groups that are clinically meaningful of patients who can be expected to use similar amounts of hospital resources. Then case-mix measures are used to assign a higher weight to the patients who are expected to consume more resources. Weights can be adjusted depending on whether the case is typical or atypical (explained more below). The case-mix costing uses the total hospital costs for the numerator (as for the per diem calculations), but rather than dividing by the number of patient days to find an "average" cost per day, it is divided by the sum of these weights to find a cost per weighted case (CWC). Thus the cost for one type of case can be compared to costs of other types of cases.

RDRG and CMG Case Mix Systems

Two separate classification systems are available at MCHP to group inpatient cases into clinically meaningful resource use groups: see the Diagnosis Related Groups (RDRGs®) - Overview and the Case Mix Groups (CMGs™) - Overview concepts. Mustard & Derksen (1997) used CMGs™; Shanahan et al. (1994), Shanahan (1996); Shanahan et al. (1997) and Shanahan et al. (1999) used RDRGs®. At the time, the main difference between these two classification systems is that the RDRGs® allow for differing levels of severity based on complications and co-morbidities within similar diagnostic groupings. (A CMG™ overlay is available, starting with the 1998/99 data, to account for severity). There are also many similarities between these two systems (CMGs™ and RDRGs® were both developed directly from the original DRG™ system).

Despite these two different groupings, the costing methods remain somewhat similar. Each of the methods used involves resource intensity weights. For the RDRGs®, relative case weights (RCWs) were developed based on charge data from Maryland and Canadian LOS. The Maryland data was from the 1991 and 1992 Maryland Health Services Cost Review Commission data set. This data set includes the charges (which reflect the actual costs) for acute care patients in all Maryland general hospitals. For CMGs™, resource intensity weights (RIWs)™ are available from Canadian Institute for Health Information (CIHI) with the CMG™ grouper (also developed using the Maryland data). Average Length Of Stay (ALOS) and trim point for typical patients were also developed for Manitoba patients for each RDRG®. For typical CMG™ classes an average length of stay and a trim point are defined by CIHI.

For outpatient care, Day Procedure Groups (DPGs) and weights are available from CIHI.

Notes:


• The cost assigned to a case is for a complete course of treatment, and dividing the cost by the length of stay will not accurately reflect the costs of any particular day.
  
  
• The cost for a particular case is an estimate of the average cost for a particular type of case, and may not accurately reflect the cost of a particular case.
  
  
• The cost for a particular type of case is calculated as a value relative to all other types of cases - it does not reflect the actual costs of providing care to a particular type of case or particular person.
  
  
• Weights, either from RDRG® or CMG™, were not developed solely from Manitoba cost data. Maryland data has been used to calibrate CIHI (CMG™) weights and Manitoba (RDRG®) weights and costs.
  
  
• The source of dollars is variable and should be noted along with exclusions. See notes below on where dollar amounts came from for the 4 deliverables discussed below.
  
  
• The time period and RDRG®/CMG™ version must be valid for costing or weight data. (That is, if weights have been developed for version 7 RDRG®, the appropriateness of applying the weights to version 9 RDRG® should be examined).
  
  
• Even though some methods don't require checking for atypical cases (e.g. deaths, transfers, long-term care, and very long or very short stays), these cases should always be examined to see if there are significant outliers.

At this point, it will be easier to talk about each of the approaches taken in the major MCHP reports separately; the differences will be summarized in a table in a later section . Currently Management Information System (MIS)/CMG™ costing work is being done, and will probably be available as an alternative sometime in 1999.

Details of Costing Methods Used in MCHP Deliverables in the 1990's

This section describes four different costing methods used in MCHP projects in the 1990's.

1 - Case-Mix Costing Using RDRGs

From: Shanahan et al. (1994); Lloyd et al. (1995) and Shanahan et al. (1996)

The goal of these reports was to develop a method for assessing the cost efficiency of Manitoba Hospitals. Using the 1991/92 (and subsequently the 1993/94) inpatient hospital data, all inpatient (in-year) hospital days were classified into RDRGs® (version 5 RDRGs® for 1991/92, version 7 RDRGs® for 1993/94); each was weighted using Manitoba RCWs. Adjustments were made to the weights for atypical cases (cases involving non-acute days, outliers (LOS > Trim), transfers or deaths). For example, for cases with LOS > Trim, a marginal case weight was added to the RCW for every day that the LOS was past the average LOS (that is, case weight = RCW + Marginal case weight * (LOS - ALOS)). Marginal case weights were developed for each RDRG®.

For each hospital these case weights were summed. Hospital-specific average case weights were calculated by summing all the case weights in each hospital and dividing by the total number of hospital cases. Hospital average cost per weighted case (CWC) was calculated by dividing the total inpatient dollars by the total hospital case weights. See the next section for how the hospital inpatient dollars were identified. The hospital average CWC was the focus of the report. The CWCs for each hospital represent average cost per case adjusted for the types of patients treated in the hospitals. The report went onto compare CWCs across the hospitals. These hospital-specific CWC (developed for both the 1991/92 and 1993/94 inpatient hospital data) are used by other studies to determine hospital costs (see Pie Project section below).

CWC (per hospital) = (Total $ per hospital) / (Sum of all RCW (per hospital))

To find the cost of a particular case:

Cost of a case = CWC * (RCW for that case)

For more information on costing methods using the RDRG case mix system, please see Appendix A: Overview of Methodology for Calculating Cost Per Weighted Case in the Hospital Case Mix Costing Project 1991/92: Methodological Appendix deliverable by Lloyd et al. (1995).

2 - Pie Project

From: Shanahan et al. (1997) )

This project looked at how heath care dollars were spent for residents of different regions. It included costs associated with hospitals, physicians, PCHs, long-term care hospitals, mental health hospitals and home care. Only the hospital portion will be discussed here. Using the methodology of the case-mix costing reports, specific case costs were estimated by multiplying a given case weight by the CWC in the hospital where the care was provided. (i.e. estimated inpatient cost for a case = RCW * CWC for that hospital). As in the case-mix costing reports, the RCWs were adjusted for cases that were classified as non-acute, long-stay outliers, deaths or transfers. The costs per weighted case (CWC) for each hospital was developed as for case-mix costing above.

Day care surgery costs were estimated using the DPG to classify cases and apply appropriate weights. The DPG weight was then multiplied by the CWC (as for inpatient costs) for the hospital providing the care to obtain an estimated cost per case.

Estimated day surgery cost for a case = DPG weight * CWC for that hospital.

For each area, the inpatient and day procedure costs were combined. Per capita expenditures were then looked at.

This report then looks at the difference between using hospital-specific CWC (described above) and using a provincial average CWC (sum all provincial inpatient dollars and divide by total provincial case weights). Note that this is the method discussed in the cost list approach (See Cost List ).

Two different methods for calculating outpatient expenditures are also discussed (emergency departments, outpatient clinics). The first method involved allocating outpatient dollars for each hospital based on inpatient discharges from the hospital (inpatient proxy). E.g. if a population in a RHA had 60% of a given hospital's inpatient cases, that RHA was allocated 60% of the hospital's outpatient costs. The second method (the combined method) used all available outpatient data. For the 33 hospitals with outpatient claims data, outpatient expenditures were distributed proportionately based on claims for service (this included the teaching hospitals). For the 5 Winnipeg community hospitals, data from a previous study (Emergency Room Use in Winnipeg Hospitals (1991/92) was used to distribute the dollars. For the remaining 38 hospitals, the inpatient proxy method was used. This combined method was the preferred method in the report. Per capita outpatient hospital expenditures were then looked at, as well as total inpatient plus outpatient hospital expenditures per capita.

3 - Funding Methodology

From: Mustard & Derksen (1997)

The goal of this report was to develop a funding methodology to allocate health care resources to Regional Health Authorities equitably in relation to the need for health care services in their populations. The first step in this process required precise age- and sex-specific estimates of provincial per capita use of health care resources, valued in dollars. A portion of this first step involved hospital inpatient, day surgery and other outpatient services. Hospital separations that occurred in FY 1994/95 were included, except for separations from extended care beds at 6 specified hospitals, and all separations from Deer Lodge and Riverview (these were all transferred to the Institutional Long Term Care Pool).

The CMG™ classification system was used to group the inpatient cases. The typical resource intensity weight (RIW™) was assigned to inpatient cases, except for very short cases (LOS < 0.5 * ALOS and the ALOS was > 6 days.) which did not end in death. These very short stay cases were assigned a weight = LOS * the routine auxiliary weight for daily care (per diem weight, assigned by the CMG™ grouper). No adjustments were made for cases that were very long, that were transferred or which ended in death.

Day surgery and other outpatient services were assigned the DPG weight, if they could be classified by the DPG grouper (cases which grouped into DPG 62 (hemodialysis), DPG 63 (transfusions), DPG 65 (chemotherapy) and DPG 99 (ungroupable) were excluded).

These weights (inpatient, day surgery and outpatient) were then summed within provincial age and sex categories. The distributions of resource intensity weights across the age and sex categories were used to distribute the FY 96/97-service pool budget.

To calculate dollars of utilization, the sum of the weights for each age/sex cell were divided by the provincial sum of the weights, and multiplied by the total dollars of utilization. These total dollars of utilization were for inpatient and outpatient combined, for 1996/97 (rather than for 1994/95, which is the year of hospital data that this was run on, due to the goal of this project).

It would also be correct to use this method on a case by case basis, rather than for cells of age/sex. Since this is the method that the other approaches take, it will be briefly summarized here:

The CWC can be found by dividing total dollars of utilization by the provincial sum of the weights.

CWC (provincial) = (Total provincial $) / (Sum of all RIW™ and DPG weights)

The dollars of utilization for each case could then be found by multiplying the CWC by the weight assigned to that particular case.

Cost of a case = CWC * (RIW™ or DPG weight for that case)


• Details regarding the case by case application of this method as well as the estimated annual cost per weighted case values to be used with the CMG™/RIW™/DPG data can be found in the MCHP concept titled Calculating Hospital Costs Using Cost Per Weighted Case (CPWC) / Cost of a Standard Hospital Stay (CSHS) Values.

4 - Cost List

From: Jacobs et al. (1999)

This report is somewhat different from the ones above in that its goal is not to look at distributions of costs as a final product, but rather to recommend approaches to finding costs. Among other costing methods (Pharmaceuticals, Home Care, etc.), approaches to hospital costing are recommended.

The recommended approach in the cost list for inpatient hospital cases uses the RDRG® classification, rather than the CMG™ classification, because the RDRG® classification further subdivides the patients into levels of severity as defined by complications or co-morbidities which would be expected to affect the amount of hospital resources used. The provincial cost of a typical case for each RDRG® is reported in the paper. These costs are calculated by the following steps (note similarity and differences) to the case-mix costing method described above. Using the 1993/94 inpatient hospital data, all inpatient hospital days were classified into RDRGs® (version7); each was weighted using Manitoba RCWs. Adjustments were made to the weights for atypical cases (cases involving non-acute days, outliers (LOS > Trim), transfers or deaths).

Whereas for the case-mix costing study, hospital average cost per weighted case (CWC) were calculated by dividing the total inpatient dollars by the total hospital case weights - the cost list found the provincial inpatient cost per average weighted case (CWC) by dividing the provincial inpatient hospital expenditure by the sum of all the inpatient RDRG® weights.

CWC (provincial) = (Total provincial inpatient $) / (Sum of all RCWs)

An estimate of the cost of a typical case in each RDRG® was found by multiplying the inpatient cost per average weighted cases (CWC) by the RDRG® typical case weight.

Cost of a typical case for a RDRG® = CWC * RCW for the RDRG®

Appendix 1 of the report contains, for each RDRG®, an estimate of the cost of a 'typical' case. Also reported in the appendix for each RDRG® are the average LOS, the TRIM point, the average cost per day and the marginal cost per day. When using this data, keep in mind that they were developed for version 7 RDRG®, and they are estimates for typical cases only.

The cost list report recommends how to use the inpatient hospital cost list for 4 possible circumstances, as described in the following table:

Circumstances Related to Measure	Recommended Cost Measure
1. The intervention under consideration affects the number of hospitalizations but not the treatment patterns (including LOS) within an RDRG®.	Use the cost of a typical case, and apply it to the number of cases within each RDRG®.
2. The intervention will only affect the LOS, resulting in a longer or shorter stay.	Use the marginal cost per day.
3. The intervention is expected to affect both the number of cases and the LOS (i.e. the average LOS is not expected to equal the typical average length of stay), OR if one wants to cost a group of cases.	Use the average cost per case OR the average cost per case + (the marginal cost * the number of days the stay is beyond or before the mean).
4. The treatment will affect how resources are used within an RDRG®.	One should consider using micro-cost analysis.

The circumstance that compares most directly to the 3 reports listed above is the third circumstance, the costing of a group of existing cases. The recommended approach in this situation is:

If the LOS is equal to the typical average LOS within the RDRG®,

Cost = average cost per case

If the LOS is not equal to the typical average LOS within the RDRG®,

Cost = average cost per case + (marginal cost * (LOS - ALOS))

For day procedure costs the recommended approach is to group the cases by DPG.

Appendix 2 of the report gives the costs for each DPG (which were estimated by multiplying the inpatient (provincial) cost per weighted case by the DPG weight, that is, cost for a DPG = CWC * DPG weight.

The provincial cost for cases that are reported in the appendices are likely to be on average higher or lower, depending on the type of hospital in which the care was received. Thus hospital type adjustment factors are reported in the paper, which can be applied to both inpatient costs and day procedure costs.

For calculating emergency department outpatient costs, the paper reports that it is possible to calculate the average nursing cost of an average emergency department visit (see the report for details).

Source of Global Budgets

1. The case-mix costing reports (1991/92 and 1993/94), the Pie project, and the cost list report all used as their primary source of financial data the HS-1 forms. This information was supplemented with data from various other sources, including:
   
   
   • Hospital Statistics Part 1 (HS-1): All hospitals annually file HS-1 forms with Statistics Canada. The HS-1 consists of hospital costs and statistics in an aggregate form.
     
   • Financial Information Systems (FIS): used to provide audited and inventory-adjusted cost data for drugs and medical and surgical supplies for the rural hospitals.
     
   • Laboratory and Imaging Services (LIS) - provides diagnostic services for many rural hospitals.
     
   • Community Therapy Services (CTS) and South Central Therapy Services (SCTS) - occupational therapy and physiotherapy provided by outside agencies.
     
     NOTE: Some costs were excluded, such as medical reimbursements, interns' and residents' salaries, capital costs and depreciation, and costs not directly related to patient care, such as education and research programs.
   
   
2. The Funding Methodology report used as its global budget the expected service pool costs as provided by Manitoba Health. Exact costs were not necessary since this paper is a methodological paper and not a cost paper.
   
   
3. A future source of costing data will come from the Management Information System (MIS) data work that is currently underway at MCHP. This will probably be available as a longer term (more stable) alternative sometime in 1999.

Table Summarizing the Similarities/Differences in the Approaches

Note that the goals of each of the projects are all different, which explains in part why different methodologies were used.

Note: For the Cost list, two columns are really needed to summarize what the approach is: one column for the development of the CWC, and one for how it should be applied. Due to limited space, only the column on how it should be applied will be shown. The development of the CWC for the cost list follows that outlined for the Case-Mix project, except that the CWC was found for all hospitals combined, not per hospital.

Table 1: Hospital Expenditure Methods



Variable	Case-Mix Costing Shanahan et al. (1994); Shanahan (1996)	Pie Project Shanahan et al. (1997)	Funding Methodology Mustard & Derksen (1997)	Cost List Jacobs et al. (1999)
Output/Goal	CWC (hospital specific cost per weighted case)	Per Capita Expenditures for inpatient and day procedure cases for residents of different regions; also outpatient	Distribution of RIW™ across age/sex cells (used to distribute the FY 1996/97 service pool budget). (Inpatient, day procedure and outpatient included).	Methods of how to apply hospital costing. To achieve this: Appendix 1: cost of typical case for each RDRG®. Appendix 2: DPG costs
Adjusted for Case-Mix	Yes	Yes	Yes	Yes
Classification System used for inpatient cases	RDRGs®	RDRGs®	CMGs™	RDRGs®
Weights applied to inpatient cases	Case Weights	Case Weights	Typical Weights	N/A - using Appendix 1 the cost is applied directly by RDRG®
Hospital Specific Costs	Yes	Yes	No	No (although adjustments can be made for hospital type)
Day Procedure	N/A	DPG weights	DPG weights	DPG cost
Outpatient Hospital Expenditure allocation	N/A (inpatient only)	Mixture of inpatient/ outpatient utilization data	Inpatient Utilization data	Nursing costs
Short stay	Adjustment to weight only if transfer or death	Adjustment to weight only if transfer or death	Adjustment to weight if LOS < 0.05 * ALOS for that CMG™ and ALOS > 6 and not a death.	Adjustment to cost if LOS < ALOS.
Long stay	Adjustment to weight if LOS > TRIM	Adjustment to weight if LOS > TRIM	No adjustment	Adjustment to cost if LOS > ALOS.
Other adjustments	Adjustment to weight if death, transfer or non-acute days	Adjustment to weight if death, transfer or non-acute days	None	None